USF Libraries
USF Digital Collections

Back massage

MISSING IMAGE

Material Information

Title:
Back massage long term effects and dosage determination for persons with pre-hypertension and hypertension
Physical Description:
Book
Language:
English
Creator:
Olney, Christine M
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:

Subjects

Subjects / Keywords:
Hypertension   ( mesh )
Massage   ( mesh )
Relaxation Therapy   ( mesh )
Blood pressure
Body mass index
Effleurage
Randomized
Relaxation response
Dissertations, Academic -- Nursing -- Doctoral -- USF   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Summary:
ABSTRACT: Significance: Complementary and alternative therapies (CAM) are widely used however the efficacy of many CAM therapies for specific diseases has yet to be verified. Massage therapy, specifically back massage, used to assist in the management of elevated blood pressure is one such unverified therapy. A pilot study completed in 2002 resulted in significant changes in blood pressure using a repeated application of the 10 minute back massage. Research Aims: This study, evolving from a psychophysiology framework, aimed to determine the long term efficacy of a back massage treatment and possible dosage needed to effectively assist in the management of elevated blood pressure. Primary Research Hypotheses: After adjusting for covariates: A. Systolic blood pressure (SBP) and or diastolic blood pressure (DBP) would decrease significantly over time using a back massage treatment in subjects with pre-hypertension or controlled hypertension.B.There would be a significant difference in the SBP (and or DBP) changes over time using 10 applications of back massage versus five applications of back massage in the subjects with pre-hypertension and controlled hypertension.Methods: A priori power analysis determined the three groups by four time points (repeated measures) design required a sample of 45 participants. The sample of men and women, 18-75 years of age, were recruited from a university setting. Outcome Variables: Systolic Blood Pressure, Diastolic Blood Pressure Potential Covariates: Age, BMI, Medications, Years of Hypertension, Salivary cortisol, and State and Trait Personality Indicators (anger, anxiety, depression). Intervention: Group 1: Ten 10-minute back massages given three times a week for 3.5 weeks. Group 2: Five 10-minute back massages given three times a week for 1.5 weeks. Control (group 3): Ten 10-minute relaxation sessions using learned techniques for 3.5 weeks.Findings: For participants with elevated body mass index (>̲27.85) in the 10-massage group, systolic and diastolic blood pressure changed significantly over time. The dosage analysis did not clearly reveal the direction of the trends, therefore further exploration is warranted.
Thesis:
Dissertation (Ph.D.)--University of South Florida, 2007.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Christine M. Olney.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 164 pages.
General Note:
Includes vita.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001928609
oclc - 212009636
usfldc doi - E14-SFE0001923
usfldc handle - e14.1923
System ID:
SFS0026241:00001


This item is only available as the following downloads:


Full Text
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam 2200397Ka 4500
controlfield tag 001 001928609
005 20091216104838.0
007 cr mnu uuuuu
008 080227s2007 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001923
035
(OCoLC)212009636
040
FHM
c FHM
049
FHMM
090
RT41 (ONLINE)
1 100
Olney, Christine M.
0 245
Back massage :
b long term effects and dosage determination for persons with pre-hypertension and hypertension
h [electronic resource] /
by Christine M. Olney.
260
[Tampa, Fla] :
University of South Florida,
2007.
520
ABSTRACT: Significance: Complementary and alternative therapies (CAM) are widely used however the efficacy of many CAM therapies for specific diseases has yet to be verified. Massage therapy, specifically back massage, used to assist in the management of elevated blood pressure is one such unverified therapy. A pilot study completed in 2002 resulted in significant changes in blood pressure using a repeated application of the 10 minute back massage. Research Aims: This study, evolving from a psychophysiology framework, aimed to determine the long term efficacy of a back massage treatment and possible dosage needed to effectively assist in the management of elevated blood pressure. Primary Research Hypotheses: After adjusting for covariates: A. Systolic blood pressure (SBP) and or diastolic blood pressure (DBP) would decrease significantly over time using a back massage treatment in subjects with pre-hypertension or controlled hypertension.B.There would be a significant difference in the SBP (and or DBP) changes over time using 10 applications of back massage versus five applications of back massage in the subjects with pre-hypertension and controlled hypertension.Methods: A priori power analysis determined the three groups by four time points (repeated measures) design required a sample of 45 participants. The sample of men and women, 18-75 years of age, were recruited from a university setting. Outcome Variables: Systolic Blood Pressure, Diastolic Blood Pressure Potential Covariates: Age, BMI, Medications, Years of Hypertension, Salivary cortisol, and State and Trait Personality Indicators (anger, anxiety, depression). Intervention: Group 1: Ten 10-minute back massages given three times a week for 3.5 weeks. Group 2: Five 10-minute back massages given three times a week for 1.5 weeks. Control (group 3): Ten 10-minute relaxation sessions using learned techniques for 3.5 weeks.Findings: For participants with elevated body mass index (>27.85) in the 10-massage group, systolic and diastolic blood pressure changed significantly over time. The dosage analysis did not clearly reveal the direction of the trends, therefore further exploration is warranted.
502
Dissertation (Ph.D.)--University of South Florida, 2007.
504
Includes bibliographical references.
516
Text (Electronic dissertation) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
500
Title from PDF of title page.
Document formatted into pages; contains 164 pages.
Includes vita.
590
Adviser: Mary Webb, Ph.D.
2 650
Hypertension.
Massage.
Relaxation Therapy.
653
Blood pressure.
Body mass index.
Effleurage.
Randomized.
Relaxation response.
690
Dissertations, Academic
z USF
x Nursing
Doctoral.
773
t USF Electronic Theses and Dissertations.
949
FTS
SFERS
ETD
RT41 (ONLINE)
4 856
u http://digital.lib.usf.edu/?e14.1923



PAGE 1

Back Massage: Long Term Effects and Dosage Determination for Persons With Pre-Hypertension and Hypertension By Christine M. Olney A dissertation submitted in partial fulfillment of the requirement s for the degree of Doctor of Philosophy College fo Nursing University of South Florida Major Professor: Mary Webb, Ph. D. Jason Beckstead, Ph.D. Theresa Beckie, Ph.D. Doug Schocken, M.D. Maureen Groer, Ph.D. Date of Approval: March 26, 2007 Keywords: blood pressure, body mass index effleurage, randomized, relaxation response Copyright: 2007, Christine M. Olney

PAGE 2

Dedication This dissertation is dedicated to my husband, Ronald Olney, who has been my best friend and a constant source of encour agement for more than 28 years. It is also dedicated to our children, Hannah and Elizabeth; they ar e the world’s best cheering squad a mom could ever desire. It is my wish to always be an inspiration to them. Lastly, this dissertat ion is dedicated to my parents, Paul and Alice Neppel. Two statements (among t heir many sage words of wisdom) were of significant inspiration fo r me as a child. I believe these often repeated words laid the foundation for me to become a researcher: “because” and “you can do anything you choose to do”, the former stat ement being the most unsatisfactory answer to my question of “why” and the latter statement instilling in me the confidence to go forth.

PAGE 3

Acknowledgements I would like to express my gratitude to my dissertation committee for their guidance and support throughout the dissertat ion process. I would also like to acknowledge Dean Patricia Burns for prov iding a research space at the College of Nursing.

PAGE 4

i Table of Contents List of Tables vi List of Figures vi Abstract viii Chapter One: Introduction 1 Background 2 Significance of the Study 4 Purpose and Aims 6 Research Hypotheses 8 Chapter Summary 9 Chapter Two: Review of the Literature 10 Blood Pressure and Hypertension 11 Physiology of Blood Pressure 11 Hypertension Pathophysiology 14 Massage 20 History of Massage 20 Types of Massage 24 Benefits of Massage 24 Summarizing the Connection of Massage and Nursing 27 Review of Massage Literature 28 Meta-Analysis and Systematic Reviews 28 Selected Research Studies 31 Theoretical Framework 44 Chapter Summary 46 Chapter Three: Method 47 Study Design 47 Setting 47 Sample 48 Sampling Method 51 Protection of Human Subjects 52 Randomization 52 In strumentation 54 Blood Pressure 56 Salivary Cortisol 57

PAGE 5

ii State Trai t Personality Inventory 58 Expectations Rating Scale 59 Touch Perception Rating Scale 59 Perceived Rapport Scale 59 Subject Information form 60 Procedure 60 The Massage Therapist 61 Massage Routine 61 10-Massage Group and 5-Massage Group Procedure 62 Control Group Procedure 63 Post In tervention Procedures 63 Data Management 63 Security 63 Data Analysis Plan 64 Chapter Summary 64 Chapter Four: Overview of Findings 66 Analytical Strategy 66 Preliminary Analyses 67 Sample 67 Assumptions 75 Missing data 75 Normality and Outliers 77 Homogeneity of Variance 81 Sphericity 81 Homogeneity of Regression Slopes 81 Correlations 82 Hypotheses Testing 92 Hypothesis 1 92 Hypothesis 2 95 Hypothesis 3 99 Hypothesis 4 100 Hypothesis 5 101 Systolic lying & sitting blood pressure changes 102 over time Diastolic lying & sitting blood pressure changes 103 over time Additional Analysis of Interest 107 Expectations 107 Chapter Five: Discussion 108 A dditional Discussion 113 St udy Limitations 116 Implications for Practice 117 Future Research Recommendations 118

PAGE 6

iii References 120 Appendices 130 Appendix 1: CD list 131 Appendix 2: Recruitment Flyer 132 Appendix 3: Lette r from Instituti on Review Board 133 Appendix 4: Approved Consent form 134 Appendix 5: Approved HIPAA form 136 Appendix 6: Subject Information form 140 Appendix 7: Mini Mental Exam 142 Appendix 8: St ate Trait Personality Inventory 144 Appendix 9: Pe rmission from Dr Spielberger 148 Appendix 10: Su bject Touch Perception Scale 149 Appendi x 11: Subject Perceived Rapport Scale 150 Appendix 12: Schedule for Participants 151 Appendix 13: Medica tions Used by Study Participants 152 Appendix 14: Means and Standard Deviation Bars for 153 Trait Anxiety, Anger and Depression Appendix 15: Corre lations of pertinent variables 155 Appendix 16: Scatter Plot s Correlations of pertinent variables 159 Appendix 17: Correlations of Antihypertensive Medication, 161 BMI and Baroreceptor Response About the Author End Page

PAGE 7

iv List of Tables Table 1 JNC VII Cla ssification of Blood Pressure 16 Table 2 Types of Massage 24 Table 3 Benefits of Massage 25 Table 4 Summary of Massage Studies in which Blood Pressure Measured 32 Table 5 Reasons Inte rested Persons did not Join Study 50 Table 6 Reasons Cons ented Participants Dr opped out of Study 54 Table 7 Variables Measured and Instruments used to Measure each Variable 55 Table 8 C haracteristics of Groups 68 Table 9 Antihypertens ive Medications Used by Study 70 Participants Table 10 Means (Standard Deviations) and Significance of 71 State Anxiety, Anger and Depression Table 11 Means (Standard De viations) and Signi ficance of 72 Trait An xiety, Anger and Depression Table 12 Cortisol Means (Standard Deviations) and 73 Significance Table 13 The Subject Touch Perception Scale Means 74 (Standard Deviation) Table 14 Subject Perceived Rapport Scale Means 75 (Standard Deviation) Table 15 Demographics of D ependent Variable Means 76 (Standard Deviations)

PAGE 8

v Table 16 Heterogeneity of Variance of BMI and Systolic 82 Blood Pressure Table 17 Correlations of Anti hypertensive Medication, BMI and 87 Blood Pressure Difference Scores: 10-Massage Group Table 18 Correlations of An tihypertensive Medication, BMI and 88 Blood Pressure Difference Scores: 5-Massage Group. Table 19 Correlations of Antihyper tensive Medication, BMI and 89 Blood Pressure Difference Scores: Control Group. Table 20 Analysis of Covari ance of Massage Effects on Systolic 93 Blood Pressure: A Report of Simple Main E ffects after Controlling for BMI Table 21 Analysis of Covariance of Massage Effects on Diastolic 96 Blood Pressure: A Report of Simple Main Effects after Controlling for BMI Table 22 Laying and Sitti ng Blood Pressure Group Means 101 (Standard Deviations) Table 23 Lying to Sitting Ch ange in Systolic Blood Pressure: 106 Simple Main Effects of Position by Time within Group Table 24 Lying to Sitting C hange in Diastolic Blood Pressure: 107 Simple Main Effects of Position within Group

PAGE 9

vi List of Figures Figure 1 Psychophysiol ogy and Blood Pressure Model 4 Figure 2 Ba roreceptor Pathway 13 Figure 3 Model of Massage Influence on Blood Pressure 45 within the Psychophysiology Framework Figure 4 Participant Eligibility and Attrition 53 Figure 5 State Anxiety Means and Standard Deviation Bars 78 Figure 6 State Anger Means an d Standard Deviation Bars 78 Figure 7 State Depression Means and Standard Deviation 79 Bars Figure 8 Group Means of Cortisol Levels over Time 79 Figure 9 Group Means of Systolic Blood Pressures over Time 80 Figure 10 Group Means of Diastolic Blood Pressures over Time 80 Figure 11 Systolic Blood Pressu re Difference and BMI for 90 10-Massage Group Figure 12 Systolic Blood Pressu re Difference and BMI for 90 5-Massage Group Figure 13 Systolic Blood Pressu re Difference and BMI for 91 Control Group Figure 14 Systolic Blood Pr essures over Time, Total Group 94 and Split by BMI: 10-Massage Group Figure 15 Systolic Blood Pressu res over Time, Total Group 95 and Split by BMI with Trend lines:10-Massage group

PAGE 10

vii Figure 16 Diastolic Blood Pressu res over Time, Total Group 98 and Split by BMI: 10-Massage group Figure 17 Diastolic Blood Pressu res over Time, Total Group 98 and Split by BMI with Trend lines:10-Massage group Figure 18 10-Massage Group Lying & Sitting Systolic Blood 102 Pressure Relationship Figure 19: 5-Massage Group Lying & Sitting Systolic Blood 102 Pressure Relationship Figure 20: Control Group Lying & Sitting Systolic Blood 103 Pressure Relationship Figure 21: 10-Massage Group Lying & Sitting Diastolic Blood 103 Pressure Relationship Figure 22: 5-Massage Group Lying & Sitting Diastolic Blood 104 Pressure Relationship Figure 23: Control Group Lying & Sitting Diastolic Blood 104 Pressure Relationship

PAGE 11

viii Back Massage: Long Term Effects and Dosage Determination for Persons With Pre-Hypertension and Hypertension Christine M. Olney ABSTRACT Significance: Complementary and alternative therapies (CAM) are widely used however the efficacy of many CAM therapies for specific diseases has yet to be verified. Massage therapy, specifically back massage, used to assist in the management of elevated blood pressure is one such unverified therapy. A pilot study completed in 2002 resulted in signi ficant changes in blood pressure using a repeated application of t he 10 minute back massage. Research Aims: This study, evolving from a psychophysiology fr amework, aimed to determine the long term efficacy of a back massage tr eatment and possible dosage needed to effectively assist in the managem ent of elevated blood pressure. Primary Research Hypotheses: After adjusting for covariates: A. Systolic blood pressure (SBP) and or diastolic blood pressure (DBP) would decrease significantly over time using a back massage treatment in subjects with pre-hypertension or controlled hypertension. B. There would be a significant di fference in the SBP (and or DBP) changes over time using 10 applications of back massage versus five applications of back massage in the subjects with pre-hypertension and controlled hypertension. Methods: A priori power analysis determined the th ree groups by four time points (repeated measures) design required a sample of 45 participants. The sample of men and women, 18-75 years of age, were recruited from a university setting. Outcome Variables: Systolic Blood Pressure, Diast olic Blood Pressure

PAGE 12

ix Potential Covariates: Age, BMI, Medicati ons, Years of Hypertension, Salivary cortisol, and State and Trait Personality I ndicators (anger, anx iety, depression). Intervention: Group 1: Ten 10-minute ba ck massages given three times a week for 3.5 weeks. Group 2: Five 10-minute back massages given three times a week for 1.5 weeks. Control (group 3): T en 10-minute relaxation sessions using learned techniques for 3.5 weeks. Findings: For participants with elevated body mass index (> 27.85) in the 10massage group, systolic and diastolic blood pressure changed significantly over time. The dosage analysis did not clearly reveal the direction of the trends, therefore further explor ation is warranted.

PAGE 13

1 Chapter One Introduction Massage has long been thought to have an effect on blood pressure levels (e. g. Longworth, 1982; Hernandez -Rief, Field, Krasnegor et al., 2000; Moyer, Rounds & Hannum, 2004). Howeve r, specific effects of massage on blood pressure levels remain controve rsial due to the scarcity of rigorous randomized clinical trials (Moyer et al ., 2004). Recently an organized effort by the scientific community has emerged to understand the efficacy of massage. In 2004, Moyer et al. published a meta-analysis of 37 experimental clinical trials of massage therapy. This work calculated the effect sizes for nine dependent variables; five studies measured systo lic and diastolic blood pressures as dependent variables. The meta-analysis results provided scientific evidence and a better understanding of the efficacy of massage in the management of blood pressure. Further the meta-analysis ident ified specific questions in need of exploration regarding t he effects of massage. The intent of this study was to ex plore specifically identified areas in question regarding the efficacy of massage effects on blood pressure. An indepth review of literature about mass age and blood pressure, along with the findings from the PI’s pilot study (O lney, 2005), guided the proposal for this study. This study sought to determi ne the long-term effects of back massage

PAGE 14

2 and to compare two dosages of back massage on persons with elevated blood pressure. Further, this study exami ned a specific physiological mechanism, baroreceptor response, which could possibl y assist in the understanding of the effects of back massage on blood pressure. Background Massage therapy, in particular, therapeutic back massage, has long been of interest to nursing. Historically back massage has been an integral part of nursing care, usually administered when preparing the patient for nighttime sleep. As the American health care syst em has changed, the workload of the hospital staff nurse intensified an d back massage fell from the working vernacular of nursing care (Meintz, 1995). By the 1990’s, complementary and alternative medicine became embedded in heal thcare, reinvigorating interest in massage therapy and back massage. Curre ntly, back massage, along with other forms of massage are now considered co mplementary therapies. Under the auspices of complementary therapy, ba ck massage is commonly referred to as therapeutic back massage. Elevated blood pressure is a serious health problem throughout the world. Unmitigated elevated blood pressure, be it pre-hypertensi on or hypertension often results in dire health consequences. Hypertension is a major contributor to cardiovascular disease, the primary cause of death in the United States (American Heart Association, 2006). Researchers, internationally, have historically identified multiple causes of hypertension, yet with more than 90% of cases of hypertension, causation remain s elusive (Kaplan, 2002). Scientists

PAGE 15

3 have studied and verified many phenomena t hat add to the understanding of the development of hypertension. One o ften addressed phenomenon for elevation of blood pressure is unrelenting stress. Psychophysiology, the study of how psychological factors influence physiology, resulting in health or illness, provided a conceptual framework for this study. Psychophysiology considers t he influence of stress on homeostasis, adaptation and allostasis to help explai n the relationships between man and his internal and external environment. Stress has long been recognized as a contributing factor to the development of elevated blood pressure (Benson, 2000). Movement from a Cartesian approach toward a holistic model of psychophysiology has allowed scientists to begin to explain the integrat ive interactivity of the human element within one’s endogenous environment. Sp ecifically, environmental chronic physical (external) stressors such as obes ity, nicotine, and caffeine are known to contribute to the genesis of hypertension. Perceived stress (internal) is also recognized a as contributory factor in tr ansitory and chronic hypertension. These external and internal stressors influence the autonomic nervous system, hormonal-endocrine system and the i mmune system (Segerstrom & Miller, 2004). The result is that blood pressure changes in response to the arousal or lack of arousal (Figure 1). The psychoph ysiology theory is developed further in Chapter two. The relationship between stressors and their effects on blood pressure has been well studied by the scientific community. The knowledge gained

PAGE 16

4 Figure 1: Psychophysiology and Blood Pressure Model regarding stress and blood pressure has lead to the development of interventions to reduce the external and internal stresso rs. Stress-relieving methods such as controlled breathing, yoga, progressive muscle relaxation and transcendental meditation have been advocated as forms of non-pharmacological methods useful in decreasing blood pressure (B enson, Beary & Carol, 1974; Grossman, Grossman, Schein, Zimlichman, & Gavi sh, 2001; Jacobson, 1939). The recent public interest in non-drug stress-relie ving practices demonstrates the need to determine their efficacy for preventing and assisting in the management of hypertension. Significance of the Study Complementary and alternative medi cine was defined nearly 15 years ago as “interventions neither taught widely in medical schools nor generally available in US hospitals” (Eisenberg, Kessler, Fost er, Norlock, Calkins, and Delbanco, External Stressors Arousal of: Autonomic Nervous System HormonalEndocrine System Blood Pressure Changes Internal Stressors

PAGE 17

5 1993, page 246). The interest in comp lementary and alternative therapies, including massage therapy, has evolved r apidly. Reports of the usage of complementary and alternative therapi es among Americans has varied from 28.9% (Ni, Simile, and Hardy, 2002) to 62% (Barnes, Powell-Griner, McFann, & Nahin, 2004) of the populatio n. In a recent study measuring trends of complementary and alternative therapy use over the past 50 years by adults in the United States, researchers found the use of massage therapy experienced rapid growth in the 1980’s and 1990’s (Kes sler, Davis, Foster et al., 2001). Tindle, Davis, Phillips, and Eisenberg (2005) further evaluated the trends of complementary and alternative therapi es between 1997 and 2002 and found the trend in massage has stayed approximately the same as previous years. Barnes, Powell-Griner, McFann, and Nahi n (2004), in a search for types of complementary and alternative therapi es usage among American adults, agreed with Kessler et al. (2001) that massage therapy was us ed by approximately five percent of the population for many pur poses. Yeh, Davis and Phillips (2006) reported that 36% of patients with card iovascular disease use complementary and alternative therapies. Further, 5% specifically use complementary and alternative therapies to tr eat their hypertension. Massage is used for a multitude of pur poses including muscle relaxation, circulatory stimulation, myofascial release, and lymphatic drainage. Back massage is one type of massage used by nurses even though the efficacy of back massage is not well established. Therefore, research on back massage is important.

PAGE 18

6 The media has led the general populat ion to believe that massage can indeed influence their blood pressure traj ectories (Prevention Magazine, 2006; Kurashova-Wine, 1999). Though there ar e some findings in the literature towards this end, the research is not rigorous enough to stat e that there are indeed any long-term effects. Therefore, long-term effects of therapeutic back massage are of relevant interest and rigorous studies are urgently needed to protect the public from possible harm. Further, the literature does not address the dosage needed to effect change. Is one back massage as effective as are several? There is little information ava ilable to advise a c linician for patient teaching. Another reason for further t herapeutic back massage research is to clarify the underlying mechanism(s) influenc ing the changes in blood pressure with back massage. If indeed, there is a long-term effect of back massage on blood pressure, what mechanism(s) are involved? Finally, offsetting the onset of hype rtension has potential health care cost savings, including a delay in drug treatment intervention, a possible decrease in dosage needed, and the delay in overall health de-conditioni ng sequela that occurs with the onset of uncontrolled el evated blood pressure. This study evaluated a therapy that could easily be tau ght to family mem bers for utilization, thus alleviating the need for cost ly massage therapy clinicians. Purposes and Aims The primary purpose of this study was to determine if therapeutic back massage could provide long-term effect s to a person with elevated blood pressure. This study tested the findings of a pilot study in which an experimental

PAGE 19

7 design compared back massage to relaxati on (Olney, 2005). A second purpose of this study was to compare two dosages of therapeutic back massage on persons with elevated blood pressure. Fina lly, this study tested for changes in baroreceptor activity as a possible ex planation of effects of therapeutic back massage. The specific aims of this study were to: 1. Determine the difference in systolic blood pressure over time in a treatment versus control group of prehypertensive and hypertensive subjects. 2. Determine the difference in diastolic blood pressure over time in a treatment versus the control group of pre-hypertensive and hypertensive subjects. 3. Compare the systolic blood pressu re changes over time of two dosages of back massage in prehypertensive and hypertensive subjects. 4. Compare the diastolic blood pressure changes over time of two dosages of back massage in prehypertensive and hypertensive subjects. 5. Determine the difference in lying versus sitting blood pressure change over time in the treatment versus control groups of pre-hypertensive and hypertensive subjects.

PAGE 20

8 Research Hypotheses The study tested the long-term ef fects and dosage of therapeutic back massage on persons with elevated blood pr essure. Furthe rmore, the study aimed to examine a possible physiologic al mechanism, baroreceptor response, explaining part of the change which may oc cur with TBM. The following research hypotheses were tested: 1. After adjusting for covariates, systolic blood pressure will decrease significantly over time using back massage treatment in subjects with prehypertension and cont rolled hypertension. 2. After adjusting for covariates, di astolic blood pressure will decrease significantly over time using back massage treatment in subjects with prehypertension and cont rolled hypertension. 3. After adjusting for covariates, there will be a significant difference in systolic blood pressure changes usi ng 10 applications of back massage versus five applications of back mass age in subjects with prehypertension and controlled hypertension. 4. After adjusting for covariates, there will be a significant difference in diastolic blood pressure changes us ing 10 applications of back massage versus five applications of back mass age in subjects with prehypertension and controlled hypertension. 5. There will be a significant change over time in reclining blood pressures compared to sitting blood pressures in the subjects with prehypertension

PAGE 21

9 and controlled hypertension in the back massage treatment group of 10 applications when compared to the control group. Chapter Summary Chapter One presented a brief over view of what is known about the effects of massage on blood pressure and a statement of the purpose of this study to further understand that relati onship. A short discussion about how the practice of the ba ck rub with bedtime preparation has been lost only to have back rubs re-emerge under the auspi ces of complementary therapies. The seriousness of elevated blood pr essure was briefly discussed. The psychoneuroimmunology theory provi des a framework within which understanding of how elevated blood pr essure develops and how a stress reduction method such as massage may assist in managing a healthy blood pressure level. The primary purpose of this study was presented: to test for long term effects of back massage. Furt her a second purpose was to test two dosages of massage for effects. Finally the study tested for a mechanism, change baroreceptor function. The signi ficance of the study discussed the increase in complementar y therapy usage by the general population and the need to ensure that using massage for bl ood pressure management is indeed a safe practice. The financial impact of offsetting the onset of hypertension through the use of a complementar y therapy such as massage was discussed. Finally the five study hypotheses were clearly stated.

PAGE 22

10 Chapter Two Review of the Literature Chapter Two presents a review of t he relevant literature regarding the state of the science on hypertension and massage therapy. A brief overview of blood pressure physiology is presented as well as the pathophysiology of hypertension. A full review of massage lit erature as it relates to this study follows. The framework for the study is discussed, clarifying the theoretical underpinnings for the study. Initial computerized searches on Medline, Cochrane Database of Systematic Reviews, CINAHL and OVID-F ull Text, using the headings of blood pressure, elevated blood pressure, hypert ension, baro-receptor s, or high blood pressure yielded more than 280,000 articl es. Searches for massage literature under headings of massage, massage therapy, back massage, back rub, therapeutic massage, yield ed more than 8,000 articl es. In Pub Med alone, combining search terms using the B oolean “AND”, for example, massage AND blood pressure, yielded a much-reduced result of 413 hits. Limiting the articles to the English language, adults and human st udies further reduced the number of articles to 120. Perusal of the referenc e lists from recovered articles led to additional relevant research discovery. In addition, computer searches and

PAGE 23

11 expert opinion were sought regardi ng the phenomenon of baroreceptors, relaxation response and their relationshi p with blood pressure maintenance. Blood Pressure and Hypertension Physiology of Blood Pressure This discussion will review the definition and key components of blood pressure maintenance. Several specific mechanisms of blood pressure which are of particular interest in this study will also be discussed. Blood pressure is the force exerted by blood against the arterial walls within the circulatory system (Sherwood, 2007). Although blood pressure is regulated by mean arterial pressure, we normally assess arterial blood pressure by measuring systolic and diastolic blood pressure. Specifically, systolic blood pressure is the maximum pressure within the arteries when the left ventricle of the heart is in systole (contracted) (Sherw ood). Diastolic blood pressure occurs during ventricular relaxation and is t he minimum measured pressure exerted against the arterial walls (Sherwood). The physiology of blood pressure maintenance is complex. The circulatory system, the heart and blood vessels, is a closed system that serves to transport oxygen, nutrients, waste and hormones throughout the body. Mean arterial pressure within that closed system is regulated by two variables: cardiac output and total peripheral resistance. Ca rdiac output is a function of stroke volume and heart rate. Total peripheral resistance is determined by arteriolar radius and blood viscosity (Kaplan, 2002). There are many influences on cardiac output and peripheral resistance which in tu rn influence mean arterial pressure.

PAGE 24

12 Influences such as autonomic activity, ske letal muscle activity, myogenic activity of the arterioles, nitric oxide, endot helin, histamine, oxygen and carbon dioxide levels and hormones such as insulin, vasopressin and renin-angiotensin II are constant factors in blood pressure maintena nce. Other factors such as acid-base balance, potassium levels, osmolarity, and adenosine and prostaglandin release also affect local arteriolar radius thereby affecting peripheral resistance (Sherwood, 2007). Blood pressure maintenance is complex when considering the multi-factorial influences. Nearly every organ, including the hear t, is stimulated by the sympathetic and parasympathetic branches of the aut onomic nervous system. There are some exceptions, such as the blood vessels, sweat and salivary glands which are not stimulated by both branches. In t he case of blood vessels, arteries and capillaries are stimulated by t he sympathetic fibers only, causing vasoconstriction. For the most part, there is no parasympat hetic innervation at the blood vessel site to counter the sympathetic norepinephr ine stimulation (Sherwood, 2007). Thus, arteries can be stimulated to constrict but not stimulated to dilate by the sympathetic ner vous system. Clarifying, it is the lack of sympathetic stimulation that contributes to the relaxati on or dilation of arteries. Baroreceptors, mechanisms within the autonomic nervous system, are of particular interest to this discuss ion about blood pressure maintenance. Baroreceptors, also known as mechanoreceptors, are pressure sensors located in the carotid sinus and the aortic ar ch and are sensitive to changes in mean arterial pressure. Baroreceptors are also sensitive to pulse pressure fluctuations.

PAGE 25

13 Pulse pressure, the difference between systolic and diastolic pressures, can be altered by slight changes in systolic and diastolic pressures without changing the mean arterial pressure (Sherwood, 2007). When mean arterial pressure or pulse pressure changes, the baroreceptors sens e the change. An example would be when a healthy supine person sits up. The action of sitting up initially lowers blood pressure (through grav ity) thereby decreasing stroke volume and cardiac output. The sensed lower blood pressure c auses the baroreceptors to fire at a slower rate, sending a message via the nint h and tenth afferent cranial nerves to the autonomic center in the brain (nucl eus of the solitary tract) (Sherwood). Figure 2: Baroreceptor Pathway Baroreceptors in Aortic Arch and Carotids Nucleus of Solitary Tract Cardiac Vagal Motorneurons Spinal cord: Sympathetic Preganglionic Motorneurons Heart Rate Blood Pressure Medulla

PAGE 26

14 A responding sympathetic activity is init iated in an attempt to restore blood pressure to normal by increasing t he cardiac output and constricting blood vessels (Figure 2), (Purves et al., 2001). Alt hough debated whether baroreceptors play a role in long-term m ean arterial pressure maintenance, the latest publications appear to suppor t this theory (Thrasher, 2005). There are other responses, or reflex es, besides the baroreceptor reflex that influence blood pressure maintenance. For instance volume receptors respond to salt and water balance, chemorec eptors sense oxygen or acid levels, the hypothalamus controls cutaneous arte rioles for temperature maintenance, and cardiovascular responses are associat ed with exercise (Sherwood, 2007). One other response of interest is the cardiovascular response to emotions or certain behaviors, such as the “fight or flight” response, which is mediated through the cerebral cortex-hypothalamic pathway to the sympathetic nervous system (Izzo, 2000). This last respons e and its influence on blood pressure are of particular interest in this study and will be discussed further within the psychoneuroimmunology model. Hypertension Pathophysiology The statistics regarding hypertension ar e stunning and clarify the impact on our society. In 2003, approximately 65 million Amer icans, ages 20 years and older, had elevated blood pressure (Ameri can Heart Association, 2006). This infers that nearly one of every 3 adults has hypertens ion. Of those with the disease, approximately 63% are awar e they have hypertension but only 29% have their blood pressure controlled at a cceptable levels. Although awareness of

PAGE 27

15 hypertension has improved over the past 10 years, the death rate due to elevated blood pressure has increased by almost 30%. In 2003, of those with hypertension, the overall death rate was 18%. Of all deaths in 2003, hypertension was either t he primary cause or contri buted to 12% of the deaths. Almost $64 billion is spent annually on direct and indirect cost due to hypertension (American Heart Association, 2006). Clearly the fi nancial impact on our society is extreme. Nearly 90% of hypertension ca ses are categorized as primary hypertension (American heart Associati on, 2006). Secondary hypertension, which has known renal, endocrine and neurogenic causes, comprises the remaining 10% of hypertension cases and will not be discussed. Primary hypertension is of concern to health care providers because it alters normal function of the heart and blood vessels. With hypertension development, there is an abnormal incr ease in peripheral resistance and/or cardiac output thus increasing the heart workload and exposing blood vessels to excessive internal pressure (Kaplan, 2002). Unmitigated, chronic elevated blood pressure leads to target organ damage re sulting in congestive heart failure, stroke, renal failure, and retinal dam age. The majority of persons who experience their first heart attack, first stroke or have congestive heart failure have elevated blood pressure (American Heart Association, 2006). The World Health Organization and the International Society of Hypertension have defined hypertension as a blood pressure measuring above 140/90 mmHg (World Health Organization, International Societ y of Hypertension

PAGE 28

16 Writing Group, 2003; Chalmers, 1999.) The Seventh Report from the Joint National Committee on Prevention, detec tion, Evaluation and Treatment of High Blood Pressure (JNC7) amended the def inition by adding an additional blood pressure classification called prehy pertension for adults age 18 and older (Chobanian et. al., 2003). Prehypertens ion is defined as a systolic blood pressure of 120 -139 mmHg or higher and/or a diastolic blood pressure of 80 89 mmHg or higher (Table 1). Table 1 JNC VII Classification of Blood Pressure ________________________ __________________ __________________ B/P Classification Systolic (mmHg) Diasto lic (mmHg) ________________________ _____________________ _______________ Normal < 120 and < 80 Prehypertensive 120-139 or 80-89 Stage 1 Hypertension 140-159 or 90-99 Stage 2 Hypertension >160 or > 100 __________________ __________________ _________________________ A person whose blood pressure is in the prehypertensiv e category is at greater risk for developing hypertension than one whose blood pressure is normal (Vasan et al., 2001). Lifestyle modifi cation is the treatment of choice for this classification (Chobanian et al ., 2003) and includes recommendations of dietary changes such as a reduction in sodium and increase in calcium intake (Oparil, 2000), reduction in weight (Kaplan, 2002; McCowen, Chan, & Blackburn,

PAGE 29

17 2000), cessation and avoidance of tobacco (O nken, White, Cooney et al., 2001), decrease in alcohol consumption (Cushman, 2000), increase in exercise (Herrera & Lowenthal, 2000), and stress reduction (L ight, 2000). Results from stress reduction studies (Patel, 1997) have shown significant changes in blood pressure and support a strong argument for further research (Webb, 2002; Webb & Beckstead, 2002, Yucha, et. al., 2001). Ma ssage therapy is one such stress reduction method researched for its impac t on cardiovascular, immunological, and psychological outcomes (Groer, 1994; Ironson & Field, 1996; Moyer et al., 2004). Any one or all of the highly intera ctive neural, hormonal, chemical, and mechanical systems may contri bute to the development of high blood pressure (Kaplan, 2002). Further, these interact ions are uniquely expressed and where the exact maladaptation occurs leading to the development of chronically elevated blood pressure is unique. Therefore using a reductionist approach to explain hypertension pathophysiology has hi storically proven difficult (Delgado & Weder, 2000). Research has also identified contribut ory factors in t he development of hypertension such as genetics, age, race/et hnicity, obesity, smoking, diet, lack of exercise, inflammation, salt sensitivity as well as psychological factors of anxiety, anger, hostility and depression (Charmandari, Kino, Souvatzoglou, & Chrousos, 2003; Chobanian et al., 2003; He & MacG regor, 2003; Li & Chen, 2005; Li, 2006; Rutledge & Hogan, 2002; & Yan et al., 2003).

PAGE 30

18 Baroreceptors appear to have a role in hypertension. These mechanoreceptors “reset” thei r firing threshold to a higher level in the presence of chronically elevated blood pressure When resetting occurs, a higher threshold is required to fire an action potential, allowing for a higher level of pressure against the vessel walls befor e the impulse signal is generated and responded to through the autonomic nervous system. Presently, there is great debate in the literature regar ding the resetting of bar oreceptors and hypertension (DiBona, 2004; Dickerson, 2004; Slei ght, 2004; Thrasher, 2004). This dissertation study will not be resolving the debate, but does examine baroreflex sensitivity. The association of psychological factors with hyper tension has been studied for many years, although me thodological issues have provided inconsistent findings (Rutledge & Hogan, 2002) For example, in their review of the literature relating anger and anxiety to essential hypertension, Spielberg and colleagues concluded that t he evidence was controversial. It was clear that suppressed anger and hypertension had an a ssociation, but causation was not verified (Spielberger et al., 1991). A la ter review by Rutledge and Hogan (2002) assessed the associations between hypert ension and psychological factors such as anger, anxiety and depression, arguing that the effect sizes of the 15 studies were not attributable to methodological characteristics. They concluded the effect sizes for the psychological vari ables of anger, anxiety and depression were significant. Scalco, Scalco, Azul and Lotufo (2005) reviewed the relationship between hypertension and depression. T he evidence demonstrated a complex

PAGE 31

19 relationship between depression and hyper tension with suggestions that depression increases the risk of developi ng hypertension. Additionally, studies reviewed by Light (2001) clearly demonstr ate the relationship between perceived stress and hypertension. This dissertati on study measured psychological factors of anger, anxiety, and depression as cova riates to the blood pressure outcome variables. Persons initially diagnosed with prehypertension and hypertension should be advised to strictly adhere to lifestyle m odifications such as weight reduction, following the Dietary Approaches to Stop Hypertension (DASH) diet and decreasing in dietary sodium cons umption (Chobanian et al., 2003). Recommended modifications also included are increase in exercise, moderation of alcohol consumption, and cessation of smoking. Usual pharmacological treatments of hypertension include thiazide-type diuret ics as an initial therapy alone, or in combination with one of the following clas ses of drugs: angiotensin converting enzyme inhibitors, angiotensin-receptor blockers, calcium channel blockers, or beta-blockers (Chobanian et al., 2003). Other medications are available to assist in gaining a desir able blood pressure level. The right combination of medication is often infl uenced by other disease processes the person may be experiencing (Chobanian et al., 2003). Beyond the lifestyle modifications re commended by the Seventh Report of the Joint National Committee (Chobanian et al., 2003), some practitioners recommend non-pharmacological interventions to assist in the management of hypertension. Some of these interv entions include biofeedback, relaxation

PAGE 32

20 training, massage therapy, transcendental meditation, breathing exercises, cognitive therapy and herbal supplements (Light, 2001; Weil, 2007). Massage History of Massage Massage has a long and rich history. Nearly every early culture in the world has recorded some form of manual m anipulation of soft ti ssue. From the first known civilization, clay tablet writings dat ed 2100 BC, described rubbing and friction for a remedy of a diseased part (B enjamin, 2005). Ancient recordings of massage from China, Japan, India, Gr eece, Rome, and Turkey have provided the underlying foundation for modern massage practice (Salvo, 1999; Benjamin, 2005). The knowledge and use of massage, just as many other health practices, traveled with mankind as they traded and warred. Hippocrates of Cos, ( 460-375 BC) the father of modern western medicine, and Galen, a Roman physician, promoted t he use of rubbing and friction. The Greek’s had gymnasiums, where their m en exercised, and received massages regularly. The Roman’s borrowed the Gr eek ideas to develop Roman baths, where massage, along with the bath, was utilized by both men and women (Benjamin, 2005). During the Middle Ages, after the Ro man Empire deteriorated, massage was mostly lost as a health practice, al though there is some evidence that folk healers and mid-wives used massage. Fr om that same period, the Arabic Empire offers writings from a phy sician/philosopher, Avicenna (980-1037 BC) who promoted the use of massage, exer cise and bathing (Beck, 1999).

PAGE 33

21 The modern western massage evolved in the 19th and 20th centuries. Pehr Henrik Ling (1776-1839) and Johann Mezger (1838-1909) are given credit for giving a scientific basis to massage. Ling, a Swedish fencing master and educator, believed that movement of t he body was the key to good health. His writings on educational gymnastics influenc ed physical education in the western world. Ling’s system of passive and acti ve movements was used to treat medical conditions. The passive movements included shaking, pressing, stroking pinching, squeezing, kneading, clapping, vibr ations and rolling (Benjamin, 2005). It was Mezger, a physician from Amster dam who categorized the soft tissue movements (massage) into four categories, as we know them today: effleurage (stroking), petrissage (kneading), friction (r ubbing), and tapotement (tapping). The United States was introduced to Swedish massage techniques in the mid to late 19th century. It was through the pr actice and writings of people such as Drs. George and Charle s Taylor, orthopedic practi tioners, and Hartvig Nissen who published Swedish Movement and Massage in several medical journals, that medical practitioners in the United Stat es became interested in the benefits of massage (Salvo, 1999). These renowned m edical practitioners offered legitimate credence to massage health benefits. Massage was even more popularized in the late 19th century by John Kellogg (of Battle Creek, Michigan) who opened a natural healing resort which advocat ed healthy living through vegetarianism, sunshine, exercise, colonics, and massage. Swedish massage held importance in health and healing early in the 20th century. Schools of massage taught st udents how to apply Swedish movement,

PAGE 34

22 hydrotherapy, heat lamps, diathermy and co lonic irrigation. After both World Wars I and II, massage therapists were em ployed for rehabilitation of soldiers with injuries. However, there was a decli ne in the use of massage as a treatment for illness (Goldstone, 2000). Post-war m odern western medicine gave way to pharmaceuticals and new technology for disease treatment, leaving the hands-on therapies to specialties of physical ther apy and rehabilitation. While at Harvard Medical School (1930-1950’s), Mary Mc Millan, a renowned physical therapy educator wrote Massage and Therapeutic Exercise. This influential text touted massage and its benefits for physical therapi sts, which further separated medical practice from massage. During the mid 20th century the reputation of massage was harmed when massage parlors became fronts for prosti tution. It has taken the massage profession several decades of worki ng with law enforcement and state legislatures to seriously prosecute prostitution fronts and resurrect their reputation as healthcare prov iders. Most states now have strict licensure through their health departments to maintain le gitimate massage businesses. The profession has a very strict code of et hics, similar to nursing and medicine. Although the stigma of prostitution cont inues to linger most people today frequent massage professionals for stress reduc tion as well as deep tissue work with injuries. Massage therapy, was revitalized in the United States during the 196070’s cultural revolution. A holistic m odel evolved, espousing the concepts of wellness and prevention. During the following 20 years the profession of

PAGE 35

23 massage therapy gained acceptance. The American Massage Therapy Association Foundation was established in the early 1990’s. A major mission of the American Massage Therapy Foundation has been to fund research and establish a data base for massage’s efficacy. The University of Miami, College of Medicine, created the Touch Research Inst itute in the early 1990’s with a goal to understand the role of t ouch in human health and development. The International Symposium on the Science of Touch will hold its third conference in 2008 with an effort to share the latest research and knowledge gained. The National Institutes of Health established an Office of Alte rnative Medicine in the early 1990’s which in 1998 became the National Center for Complementary and Alternative Medicine. Millions of research dolla rs have been granted to the study of massage therapy‘s effectiveness through National Center for Complementary and Alternative Medicine. The American Nurses Association acknowledges massage therapy as a sub-specialty for nursing. An active National Association of Nurse Massage Therapists organization has evolved from this recognition. The National Association of Nurse Massage Therapist s promotes the use of touch with patients, the intention to heal and the us e of current research findings (National Association of Nurse Massage Therapists, 2006). With this national support, entrepreneurial nurses have been able to own and manage private businesses with touch and massage being the holistic intervention used for healing (MitzelWilkinson, 2000).

PAGE 36

24 Types of Massage Massage therapy treatments are marketed in many forms (Table 2). The Table 2: Types of Massage Influence Types/Names ________________________ __________________ ______________________ Cultural influences Ayuvedic (India) Lomi-LomiHawaii Tuina (Chinese) Swedish Focused massage Infant Sports CranialSacral Lymphatic Drainage Myofascial Neuromuscular Specific Developer Ro lfing, Feldenkrais, and Esalen ________________________ _____________________ ___________________ variation is due to the influence or focus during the development of that particular therapy. Some types of massage are co mbinations of several influences. Tappan suggests this availability for cross cu ltural sharing is due to the internet and access to international travel (Benj amin, 2005). Yet, each massage has a specific purpose that separates it from another form. Benefits of Massage A single good massage will leave a healthy person with a sense of wellbeing. Regular massage treatments ar e known to have many benefits (Beck,

PAGE 37

25 1999). Table 3 is a listing of benefits often experienced with regular massage treatments, complied by expert and aut hor Mark Beck (Beck, 1999). Table 3 Benefits of Massage ________________________ _____________________ ___________________ Listed Benefits of Massage* Relief of stress and tensions. Mental and physical fatigue relief. Pain in the shoulders, neck, and back is relieved. Muscles and joints become suppler, soreness and stiffness are relieved. Muscle soreness from overexertion is relieved. Circulation is improved. Digestion, assimilation and elimination improved. Facial massage tomes the skin and softens fine lines. Headache and eyestrain are often relieved. Deep relaxation, relief of insomnia. Muscular spasm relief. Obesity and flabby muscles improved w hen combined with exercise and diet programs. Pain in joints, sprains and poor circulation are relieved. Increased circulation of nourishing bl ood to skin and other parts of the body to promote healing. Mental strain reduced resu lting in better productivity. Mildly high blood pressure is temporarily reduced. Renewed sense of confidence and control is experienced. Constrictions and adhesions can be reduced and prevented as traumatized muscle tissue heals. Joint mobility can be increased. Beck (1999) does not denote which of these listed benefits are scientifically based.

PAGE 38

26 For critically ill persons there are some benefits with gentle massage such as improved mobility, help in contro lling pain and discomfort, reduction in disorientation and confusion, reduction in isolation and fear, ease of emotional distress, and the development of a more positive attitude about condition. With all of these benefits listed, few are actually based upon sci entific research. Most of these benefits are bas ed upon observation of massage performed over the millenniums of massage practice. The benefit of temporary reduction of mild high blood pressure along with several other listed benefits to include relief of stress, tension, mental strain, and increase in circulation and gai n in a sense of relaxation are being addressed by a growing body of literature. This growing literature is building the psychophysiology framework for under standing the relationship between massage and blood pressure. The purpose of the massage used in th is study was to invoke relaxation by the use of a regularly applied back massage. Swedish massage uses five classic strokes of effleurage, petrissage, tapotement, friction and vibration. The effleurage stroke, when applied with moderat e pressure, slowly and smoothly on the back, may stimulate the parasympathetic nervous system and evoke the relaxation response (Benjamin, 2005). Nu rses are taught to use rhythmic effleurage strokes to provide comfort, relaxation, induce sleep, and reduce anxiety in the ill patient (Potter & Pe rry, 2001). Cambron, Dexheimer and Coe (2006) examined the six different types of massage. Interestingly, the researchers found that some baseline charac teristics, i.e. younger age and taller

PAGE 39

27 stature, were associated significantly with greater changes in systolic blood pressure. Although no significant rela tionships were found with gender, race, medication, or weight, there were trends of decreased pressure in males, heavier persons and Hispanics (Cambron, Dexh eimer & Coe, 2006). Duration and pressure of the six different massage types were associated with change in blood pressure. Cambron and colleagues found t hat although not significant, Swedish massage had the greatest effect on bl ood pressure reduction (Cambron, Dexheimer & Coe, 2006). The author s noted that several therapy types increased blood pressure; therefore they encouraged the use of Swedish massage on persons with hypertension. In vi ew of the review of literature, the specific technique used in this research project was Swedish massage which utilizes the effleurage stroke on the back. Summarizing the Connection of Massage and Nursing Massage has been used in nursing practice for years, particularly the back massage. Massage is recognized as part of nursing practice. A professional organization of nurses, active National Association of Nurse Massage Therapists promotes the use of massage. Holistic nursing, which embraces the practice of nursing through a bio-psycho-social-spiri tual model, has energized and promoted the use of touch to heal (Dossey, Keegan & Guzzetta, 2000). Many of the research studies establishing the validit y of massage have come from nursing, a number of them evolving from the holistic framework.

PAGE 40

28 Review of Massage Literature Massage literature is rich in many ways. There are hundreds of studies that have been completed. Some val uable information gleaned from these studies has help to validate the practice of massage. For example, research has validated that depression and anxiety are influenced by multiple doses of massage (Moyer, Rounds, & Hannum, 2004). Outcomes from massage studies have in fluenced nursing practice. For example, research about massaging ov er boney prominences resulted in a change in nursing practice for prevention of pressure ulcers (Dyson, 1978; Buss, Halfens, & Abu-Saad, 1997). Prior to this research, nurses were encouraged to massage vigorously the boney prominences of a bedridden patient. Nursing presently espouses not to massage bony pr ominences, particularly if they are inflamed. The story is not finished, though, with yet another review of the literature recently publis hed arguing there is not st rong evidence either way regarding the massaging of bony prominences (Duimel-Peeters, Halfens, Berger, & Snoeckx, 2005). This is a good example demonstrating that even though there have been studies completed and the findings have changed practice, the findings may not necessarily be str ong enough to confirm the practice. Meta-Analysis and Systematic Reviews Strength of evidence is very important when practice is being challenged. Meta-analysis and systematic reviews of ex perimental studies offer the strongest evidence in research (Burns & Grove, 2005). The literature search on massage yielded one meta-analysis of the effe cts of massage on ni ne dependent variables

PAGE 41

29 (Moyer, Rounds, & Hannum, 2004). Mo yer et al. (2004) analyzed five experimental studies that measured blood pressure. One of the five studies measured hypertensive partici pants (Hernandez-Reif et al., 2000), one measured patients undergoing cardiac catheterizati on (Okvat, et al., 2002) and the other three studies measured blood pressures of healthy participants (Delaney, et al., 2002; Mueller Hinze, 1988; & Wendler, 1999). The effect sizes (Hedge’s g ) for systolic and diastolic blood pressures were combined in this analysis, resulting in a medium effect ( g = 0.25, p < .05). Hedge’s g is essentially the same as a Cohen’s d except in the case of a very sma ll sample size, as both are the within group pooled standard deviation (UCLA Depa rtment of Educ ation, 2007) A second meta-analysis specifically focused on the effects of the effleurage back massage on components of re laxation included calculation of systolic and diastolic blood pressures ( Labyak & Metzger, 1997). Nine studies were examined, of which most were wit h-in group designs. Although the lack of randomization weakened the re sults, the authors found decreases in systolic and diastolic blood pressures with a signific ant gender effect. Specifically, blood pressure in female subjects rises during the first 3 to 5 minutes of the massage session, then blood pressure declines in the following 10 minut e rest period. Contrasting, male subjects experience a decline in blood pressure during and 10 minutes following the massage (Labyak & Metzger, 1997). Systematic reviews, the second hi ghest level of evidence, have expanded what is known and not known about mass age. PubMed lists over 600 reviews on massage. Many of the reviews focus ed on massage effects on neck pain, low

PAGE 42

30 back pain, cancer pain, perineal care in the birthing mother, infant or neonatal care, nausea and vomiting, sports injuries and lymphedema. These reviews, for the most part, did not discuss blood pr essure changes due to massage. Some reviews did include blood pressure as a variable to measure and discuss. Richards and colleagues (2000) included 22 articles in t heir review of the effects of massage in acute and critical care patients (Richards, Gibson, & OvertonMcCoy, 2000). The framework of the revi ew was relaxation and sleep promotion. Nine of the 22 articles included systolic and diastolic blood pressure changes as variables reflecting level of relaxation. Seven of the 10 studies that reported physiological measures, showed that massage produced physiological relaxation. At least one or more of the physiological indicators (heart rate, blood pressure respiratory rate, EMG activity cortisol level, skin temperature and galvanic skin response) had significant chan ges. This review (Richards et al., 2000) included all levels of inquiry; theref ore the findings of the systematic review are not as strong as meta-analyses wh ich include only randomized studies. Two methodological critiques that evaluate massage research have been published (i.e. Cawley, 1997; Hobbs & Davi es, 1998). Methodological variations acknowledged within these crit iques included type of desi gn, number of variables evaluated, number of treat ments (massages given), length of massage, and timing of measurement. Furt her, samples varied greatly from very ill patients in critical care units to elderly in nursi ng homes to healthy nurses. The authors noted that sample size had been inconsiste nt. These critiques clearly stated a need for better research design, use of r andomization and increase sample size.

PAGE 43

31 Selected Research Studies Twenty eight studies are presented in Table 4. All included studies measured blood pressure. Seventeen of the 28 studies used a control group and randomized participants to groups. T hese 17 studies offer higher level of evidence and are depicted by an “L= 1” in the last column. Within subject designs (not randomized) do not offer as much strength theref ore are depicted by “L = 2”. The primary purpose of four of the studies was to measure the effects of massage on blood pressure (Aourell, Skoog, & Carleson, 2005; Combron, Dexheimer, & Coe, 2006; Hernandez-Reif et al., 2000; Olney, 2005). Two of the four studies evaluated participants wit h elevated blood pressure (HernandezReif, 2000; Olney, 2005) while the other studies evaluated normotensive subjects. The evidence regarding the effect of massage on blood pressure is murky in that there is little consistency am ong the studies regarding the number of applications of massage. The massage interventions vary from a single application (Bauer& Dracup, 1987; Cady & Jones, 1997; Corley, Ferriter, Zeh, & Gifford, 1995; Delaney, Leong, Watkins, & Brodie, 2002; Felhandler & Lisander, 1999; Goodfellow, 2003; Longw orth, 1982; McNamara, Bu rnham, Smith, & Carroll, 2003; McRee, Noble, & Pasvogel 2003; Okvat, Oz, Ting, & Namerow, 2002; Stevenson, 1994) to multiple app lications such as Hernandez-Reif and colleagues (2000) study in which they applied a 30 minute body massage two

PAGE 44

32 Table 4 Summary of Massage Studies in wh ich Blood Pressure was Measured Author (year) Journal Purpose Sample Intervention Measures Outcomes Level of EvidenceAhles, TA, Tope, DM, Pinkson, B, et al. (1999). Journal of Pain and Symptom Management. 18( 3), 157-163. To test the efficacy of massage therapy in reducing symptoms of physical and psychological distress with bone marrow transplant. N = 34 (all bone marrow transplant) Massage = 16 Control = 18 Female = 26 Male = 8 Mean Age= 41 years 20 minutes of Swedish massage to face, scalp, shoulders, neck, spine, plus acupressure in upper back shoulders neck and face. 11 pts received 8-9 massages three times per week, 5 pts received 4-7 massages. Control= standard care Pre – Post* Sig. DBP decrease in tx group L = 1 Aourell, M, Skoog, M, & Carleson, J (2005). Complementary Therapies in Clinical Practice, 11 243-246. To evaluate the effects on blood pressure in healthy young males. N = 15 Single Group Males Median Age = 32 years 30 minute massage to back, neck, and chest 2 times per week for 4 weeks 4 week washout period. 30 minute massage to legs, arms, and face 2 times per week for 4 weeks Pre – Post* Sig SBP decrease each tx. in first 4 wk period both grps. Sig SBP decrease 2nd period back neck chest massage. Sig. DBP decrease in tx. 1 to tx. 4 back neck and chest tx. L= 2

PAGE 45

33Author (year) Journal Purpose Sample Intervention Measures Outcomes Level of EvidenceBarr, JS, & Taslitz, N (1970). Physical Therapy, 50 (12), 1679-1691. To test the influence of back massage on autonomic functions. N =10 Single group Females Age range= 19 to 21 years. 20 minute back massage. Administered 3 times over several days PrePost** No Sig. changes in BP. L=2 Bauer, WC, & Dracup, KA (1987). Focus on Critical Care, 14 (6), 42-46. To determine the physiologic effects of back massage in patients with acute myocardial infarction. N = 25 Single group Female = 7 Male =18 Mean Age = 55.6 years 6 minutes of back massage Administered once. PrePost*** No Sig. changes in BP. L = 2 Beeken, JE, Parks, D, Cory, J, & Montopoli, G (1998). Clinical Nursing Research, 7 (3), 309-317. To determine the effectiveness of neuromuscular release massage therapy in five individuals with chronic obstructive lung disease. N = 5 Single group Female = 1 Male = 4 Age range = 57-74 years 24 diaphragmatic release massage. Administered once a week for 24 weeks. PrePost*** Sig. SBP decrease interaction L = 2 Bost, N & Wallis, N (2006). Australian Journal of Advanced Nursing, 23 (4), 28-33. To determine if massage therapy decreases physical and psychological indicators of stress in nurses. N = 58 Massage = 30 Control = 28 Sex: not reported Median Age = 42 years 15 minute back massage Once a week for 5 weeks PrePost No Sig. changes in BP L = 1

PAGE 46

34Author (year) Journal Purpose Sample Intervention Measures Outcomes Level of Evidence Cady, SH & Jones, G E (1997). Perceptual and Motor Skills, 84 157-158. To determine if chair massage results in significant decrease in systolic and diastolic blood pressure. N = 52 Single group Female = 12 Male = 40 Mean Age = 40 years 15 minute chair massage Administered once. PrePost* Sig. SBP & DBP decrease L = 2 Corley, MC, Ferriter, J, Zeh, J, & Gifford, C (1995). Applied Nursing Research, 8 (1), 39-43. To determine if the physiological and psychological responses are different between a 3-minute back rub and a 3minute control period of undisturbed rest in the institutionalized elderly. N = 19 Massage = 12 Control = 7 Female = 11 Male = 8 Mean Age= 78 years 3-minute back rub Administered once. Pre-Post* SIg. SBP in males vs. females in tx. group L = 1 Delaney, JP, Leong, KS, Watkins, A, & Brodie, D (2002). Journal of Advanced Nursing, 37 (4), 364-371. To investigate the effects of myofascial trigger-point massage therapy to the head, neck, and shoulder areas on cardiac autonomic tone and psychological outcomes. N = 30 Massage = 15 Control = 15 Female = 16 Mean age = 31 Male = 14 Mean Age = 34 20 minutes of myofascial trigger point therapy. Controls sat for same period of time. Administered once. PrePost* Sig. SBP & DBP decrease in tx. group only. L = 1

PAGE 47

35Author (year) Journal Purpose Sample Intervention Measures Outcomes Level of EvidenceDunn, C, Sleep, J, & Collett, D (1995). Journal of Advanced Nursing 21, 34-40. To determine the effectiveness of aromatherapy and massage used in the nursing care of patients in an ICU. N = 122 Massage = 43 Aromatherapy = 41 Control = 38 Female = 53 Male = 69 Mean Age = 60 years 15-30 minutes Massage to back outer legs or scalp. Aromatherapy: same with essential oil of lavender. Control: 30+ min undisturbed rest. One to 3 treatments within 5 days. PrePost* No significant values. L=1 Fakouri, C, & Jones, P (1987). Journal of Gerontological Nursing, 13 (2), 32-35. To assess the effectiveness of the slow stroke back rub in promoting relaxation. N = 18 Single group Female = 14 Male = 4 Mean Age 73.7 years. 3-minute slow stroke back massage applied at bed time. For 3 nights. PrePost Sig. SBP change each tx. Sig. DBP on third tx. only. L = 2 Felhandler, D & Lisander, B (1999). Complementary Therapies in Medicine 7, 231-234. To compare two noninvasive methods (with a control) to stimulate acu-points on the cardiovascular system N = 24 Accupoint = 8 Stroking = 8 Control = 8 Pressure using dental tool on accupoints. Stroking along the meridians. Same amount of time for each group. One time application PrePost* Pressure: Sig. SBP & DBP decrease (compared to control) Stroking: Sig. DBP (compared to stroking). L =1

PAGE 48

36Author (year) Journal Purpose Sample Intervention Measures Outcomes Level of EvidenceFerrell-Torry, AT & Glick, OJ (1993). Cancer Nursing, 16 (2), 93-101. To examine the effects of therapeutic massage on pain perception anxiety, and relaxation levels of hospitalized patients with cancer pain. N = 9 (day 1) N = 7 (day 2) Single group All men Mean Age = 56.6 years 30 min. massage of effleurage & petrissage to feet back neck and shoulders plus 6 trigger points in upper middle and lower trapezium muscles. 2 consecutive nightly tx. Pre-Post*** Day 1: Sig. SBP & DBP decrease @ T1 to T2 Day 2: Sig. SBP decrease @ T1 to T3 L = 2 Fraser, J. & Kerr, J. R. (1993). Journal of Advanced Nursing, 18, 238-245. To measure the effects of back massage on anxiety and perceived relaxation levels of elderly. N = 21 Massage=NR Control 1=NR Control 2=NR Females = 17 Males = 4 Mean Age = NR Group 1: 5 minute Back massage and conversation. Group 2: Conversation only. Group 3: No intervention. Four consecutive days at bedtime. PrePost** Group 1: Within subject SBP & DBP decrease L = 1 Goodfellow, L. (2003). Nursing Research, 52 (5), 318-328. To determine back massage influence on psychosocial, physiologic and immune function of spouses of patients with cancer. N = 42 Massage = 21 Control = 21 Females = NR Males = NR Mean Age = 52 years 20 minute therapeutic back massage Administered once PrePost* No sig. changes in BPs L = 1

PAGE 49

37Author (year) Journal Purpose Sample Intervention Measures Outcomes Level of EvidenceHayes, J & Cox, C (2000). Complementary Therapies in Nursing & Midwifery, 6, 9-13. To describe research findings from a complementary therapy intervention designed to reduce anxiety in the critical care environment. N = 25 Single group Female = 12 Male =13 Mean Age = 53.9 years 5 minute foot massage Mean number of foot massages = 2.7 (range 1-10 sessions) Pre-Post** Sig. MAP time effect from pre to during measurement. L = 2 Hernandez-Reif, Field, T, Krasnegor,J, Theakston,H, Hossain,Z & Burman, I (2000). Journal of Bodywork and Movement Therapies, 4 (1), 31-38. To assess massage therapy versus progressive muscle relaxation effects on adults with hypertension. N = 30 adults: Massage = 15 Control = 15 Females = 21 Males = 9 Mean Age: 51.6 years 30 minute massage to head, neck arms torso, legs, and back 2 times per week X 5 wks Control: 30 min progressive muscle relaxation 2 times per wk X 5 wks at home. Sitting: Pre-Post* Reclining: PrePost* No Sig. SBP changes. Sig. DBP decrease: Sitting: after first and last day tx. Reclining: first to last day pre tx. L = 1 Holland, B, & Pokorny (2001). Rehabilitation Nursing, 26 (5), 182-186. To determine the physiological and psychological effects of three consecutive days of SSBM on adult patients in a rehabilitation setting. N = 24 adults Single Group Females = 21 Males = 3 Mean Age 71.8 years 3 minute SSBM. Three consecutive nights. PrePost* Sig. SBP & DBP decrease time effect L = 2

PAGE 50

38Author (year) Journal Purpose Sample Intervention Measures Outcomes Level of EvidenceLongworth, JCD (1982). Advances in Nursing Science, 4 (4), 44-61. To examine the psychophysiological effects of the slow-stroke back massage (SSBM) on normal individuals who were free of disease conditions that might influence the results. N = 32 Single group All female Mean Age = 31.5 years 3 and 6 minutes of SSBM. Administered once. PrePost** Sig. SBP decrease in mean scores over time. No Sig. DBP changes. L = 2 McNamara, ME, Burnham, DC, Smith, C, & Carroll, DL (2003). Alternative Therapies in Health and Medicine, 9 (1), 50-57. To measure the effects of a 20 minute back massage on the physiological and psychological human responses in patients admitted for diagnostic cardiac catheterization N = 46; Massage = 23 Control = 23 Female=12 Male=34 Mean Age = 64.9 years 20 minutes back massage. Administered once. PrePost*** Sig. SBP & DBP decrease for within subjects main effect of time. Sig. SBP decrease between group. Sig. SBP decrease for time by group L= 1 Meek, SS (1993). IMAGE: Journal of Nursing Scholarship, 25 (1), 17-21. To examine the effects of SSBM on SBP, DBP HR and Skin temperature in hospice patients. N = 30 Single Group Female =14 Male =16 Mean Age = 70.8 years. 3 minute SSBM 2 consecutive days PrePost*** Sig. SBP & DBP decrease main effect due to time. Tx effect persisted for five min. L = 2

PAGE 51

39Author (year) Journal Purpose Sample Intervention Measures Outcomes Level of Evidence McRee, LD Noble, S, & Pasvogel(2003). AORN Journal, 78 (3), 433-447. Determine effects of preoperative massage and music therapy on pt’s pre-op, intra-op and post-op experiences. N = 52 Grp1 = 13 Grp2 = 13 Grp3 = 13 Control = 13 Female=33 Male=19 Mean age= 43 years. Grp 1: 30 minute massage (neck, back arms and lower limbs). Grp 2: 30 min. massage and music. Grp 3: 30 min piano music. Control: 30 min. sitting in waiting room. Administered once. Pre-Post** Sig. SBP decrease pre to intra-op Sig. DBP decrease intra-op to pre-op in massage and music Sig. DBP decrease intra-op to post-op in control L = 1. Mok, E. & Woo, CP (2004). Complementary Therapies in Nursing & Midwifery, 10 209-216. To determine effects of SSBM on anxiety, BP HR pain perception immediately and three days after SSBM on stroke pts with shoulder pain. N=102 Massage = 51 Control = 51 Female = 51 Male = 51 Mean age 73.2 10 minute SSBM at bedtime Administered for 7consecutive days Pre-Post*** Sig. SBP & DBP decrease pre to post and post 2. L = 1 Okvat, HA, Oz, MC, Ting. W, & Namerow, P.B. (2002). Alternative Therapies, 8 (3), 2002. To determine feasibility and efficacy of massage for anxiety reduction in cardiac catheterization patients. N = 78 Massage = 43 Control = 35 Female = 19 Male = 59 Mean age = 60.1 years 10 minute massage to hand, wrist arm shoulder, upper back scapular region and neck and scalp. Control: 10 minutes quiet time. Administered once. PrePost** No Sig. findings on any measures. L = 1

PAGE 52

40Author (year) Journal Purpose Sample Intervention Measures Outcomes Level of Evidence Olney, CM (2005). Biological Research for Nursing, 7 (2), 98-105. To determine feasibility and effectiveness of TBM on BP, HR and anxiety in hypertensive persons. N = 14 Massage = 8 Control = 7 Female = 10 Male = 5 Mean Age = 54.5 years 10 ten minute back massages. Three times per week. Control: relaxation (same schedule) One therapist. PrePost Sig. SBP and DBP decrease (group by time) L = 1 Post-White, J, et al., (2003). Integrative Cancer Therapies, 2 (4), 332-344. To test the effects of Healing Touch (HT) and Massage Therapy (MT) or caring presence (P) vs control (own) to inducing relaxation or other cancer tx related symptoms. N = 164 Massage = 63 Healing Touch= 56 Caring Pres. = 45 Female = 142 Male = 22 Mean age 54.7 years 45 minute sessions of massage, healing touch or caring presence Once per week for 4 weeks. Pre-Post Short term effects: MT & HT: Sig. SBP & DBP decrease. MT & HT vs P: Sig SBP decrease. L = 1 Stevenson, CJ (1994). Complementary Therapies in Medicine, 2 27-35. To assess the effects of aromatherapy foot massage, plain vegetable oil foot massage and verbal contact only vs control group in cardiac patients. N = 100 Four groups Group sizes = NR Gender= NR Mean Age = NR 20 minute foot massage with: 1) plain foot massage 2) aroma therapy or 3) chatting or 4) no intervention Administered once. Pre-Post *** No Sig. BP changes. L = 1

PAGE 53

41Author (year) Journal Purpose Sample Intervention Measures Outcomes Level of EvidenceTaylor, AG et al. (2003). The Journal of Alternative and Complementary Medicine, 9 (1), 77-89. To examine the effects on postoperative massage and vibration therapy on short-term post surgical pain, negative affect and physiologic stress reactivity. N = 105 Massage = 35 Vibration = 35 Usual Care = 36 All female. Mean age = 56.2 years 45 minute Swedish massage to upper and lower body plus usual care; 20 minute vibration therapy plus usual care; Usual care (UC). Administered for 3 consecutive evenings Pre-Post No Sig. differences found. L = 1 N = sample size; Sig. = significan t; BP = blood pressure; SBP= systolic blood pre ssure; DBP = diastolic blood pressure; tx. = treatment group; L = 1 is level of evidence : experimental with control group; L = 2 is level of evidence: with in group design; Pre-Post = BP measurements taken before and after each treatment. PrePost ** = BP measur ements before, during and after the treatme nt; PrePost *** = several BP measurements be fore and/ or after each treatment; NR = not reported; SSBM = Slow Stroke Back massage

PAGE 54

42 times a week for five weeks (H ernandez-Reif et al., 2000; see also Ahles, Tope, Pinkson et al., 1999; Aourell, Skoog & Carleson, 2005; Barr & Taslitz, 1970; Beeken, Parks, Cory, & Montopoli, 1998; Bost & Wallis, 2006; Dunn, Sleep & Collett, 1995; Fakouri & Jones, 1987; Ferrell-To rry & Glick, 1993; Fraser & Kerr, 1993; Hayes & Cox, 2000; Holland & Pokorny, 2001; Meek1993; Mok & Woo, 2004; Olney, 2005; Post-White, Kinney, Sa vik, Gau, Wilcox, & Lerner, 2003; Taylor, Galper, Taylor, Rice, Andersen, Irvin, Wang, & Ha rrell, 2003). Moyer and colleagues (2004) determined that single dose massage did elicit a moderate effect size for blood pressure (Moyer, Rounds & Hannum, 2004). They did not analyze multi dose effect on blood pressure. Further, the variation in the length of massage application time, such as a 3-minute back massage (Holland & Pokorny, 2001) versus a 20-minute back massage (Goodfellow, 2003) has made it diff icult to determine the most effective dosage. Moyer and colleagues (2004) report ed in their meta-analysis that length of massage session did not reach statisti cal significance, but felt a higher powered study may prove differently. T hey encouraged further studies to include two levels of application time (M oyer, Rounds & Hannum, 2004). Another issue of differentiation among t he research is the type of massage applied. Some researchers utilized the slow stroke back massage which is a systematic effleurage applied to the ba ck. The predominant massage utilized effleurage to the back and sometimes ot her parts of the body. When the effleurage is combined with other ther apies, for instance acupressure point,

PAGE 55

43 vibration, foot massage and or aromatherapy, determination of which stroke is effective becomes problematic. The relationship of massage and psychological factors has been studied, particularly depression, anxiety, and anger which are known to also be associated with hypertension (Bost & Wa llis, 2006; Field et al., 1996; Field, Quintino, Henteleff, Wells-Keife, & Delve cchio-Feinberg,1997; Groer et al., 1994; Hernadez-Reif, Dieter, Field, Swedlow & Diego, 1998; Hernandez-Reif, Field Ironson, Beutler, & Vera, 2005; Ironson et al., 1996; McRee, Noble & Pasvogel, 2003; Mok & Woo, 2004; Olney, 2005; Post -White et al., 2003; Rexilius, Mundt, Megel, & Agrawal, 2002; Shulman & J ones, 1996; Smith, Kemp, Hemphill & Vojir, 2002) Depression, anxiety and anger have been shown to change significantly with application of massage. A meta-analysis (Moyer, Rounds & Hannum, 2004) found significant effects of massage on factors state anxiety, trait anxiety and depression (Hedge’s) g = .37, g = .75 and g = .62, p< .01, respectively. Neuroendocrine measures of the ph ysiological stress response have been shown to be affected by massage therapy. Specifically, cortisol, a common measure of the hypothalamic pituitar y adrenal axis, is often measured in massage studies. Cortisol is known to increase with perceived stress. Massage therapy research has mixed results regarding its effect on cortisol levels. In the Moyer, Rounds, and Hannum (2004) meta -analysis, massage did not produce a significant mean effect size in cortisol Field and colleagues argue that cortisol

PAGE 56

44 decreases on average 31% in reviewed studies using massage therapy (Field, Hernandez-Reif, Diego, Schanberg a & Kuhn, 2005). Theoretical Framework Psychophysiology provides the framework for this research. In this study, the goal was to determine characteristi cs (long-term effects and dosage) of the relationship between a known stress reducti on therapy (massage) and a state of health (hypertension). In the review of literature, evidence was presented demonstrating the associations betw een psychological factors (anxiety, depression and anger), neuroendocrine function (cortisol), and health (levels of blood pressure) (Light, 2001; Lovello, 2005; Rutledge & Hogan, 2002 ). Figure 3 depicts this relationship. The intervention of massage has a relationship with blood pressure, as discussed in the review of literature. Because of the integrated relationship between the environmental, psychological and physiological factors within hypertension, each must be measured when blood pressure is measured. Blood pressure is, in essence, evaluated in the presence of the psychophysiologic status. In other words, t he internal and external factors are covariates to the blood pressure at the moment it is measur ed. It is only in this context, evaluating blood pressure in the presence of the co variates, that the influence of back massage can be understood. No one piece ex ists in isolation of the other. Further, this very simplified model, Fi gure 3, does not int end to exemplify the intricately complex full model of massage, blood pressure and psychophysiology. Lovallo (2005) has built numerous elaborat e models to depict the psychological

PAGE 57

45 Figure 3: Model of Massage Influence on Blood Pre ssure within the Psychop hysiology Framework Massage Intervention Systolic Blood Pressure Diastolic Blood Pressure Baroreceptor Res p onse Systolic Blood Pressure Diastolic Blood Pressure Baroreceptor Res p onse Psychological Factors Physiological Factors Environmental Factors Age Alcohol BMI Ethnicity/Race Exercise Gender Medications Smoking S-T Anger S-T Anxiety S-T Depression Salivary Cortisol Potential Covariates:

PAGE 58

46 and neuroendocrine relationship with health status. Figure 3 is intended to demonstrate that massage is known to hav e an influence on blood pressure as it exists within the model of psychophysiol ogy. The environmental, psychological, and physiological factors shown in this m odel are operationalized by the potential covariates: age, alcohol, body mass inde x (BMI), ethnicity/ race, exercise, gender, medications, smoking, anxiety, anger depression, and salivary cortisol and are not exclusive of other influenc es that may be included in a full psychophysiology model. Chapter Summary Chapter Two has reviewed the physiology of blood pressure and the pathophysiology of hypertens ion conveying the comple xity and the seriousness of the disease process. The histor y of massage was discussed developing a background of why effleurage stroke is the type of massage recommended for eliciting the relaxation response. The re view of massage liter ature clarified the troubling lack of consistency in massage resear ch thus far, specifically the lack of systematic approach to type of massage, length of massage, and number of dosages needed to elicit particular res ponses. The psychophysiology framework has provided a solid theoret ical basis upon which to explore the effects of massage on persons with hypertension. Th is framework is a holistic model in which the relationship of the covariat es in the model can be explored.

PAGE 59

47 Chapter Three Method Chapter three presents the methodology used in this study. Included are: the study design, setting, sample, sampli ng method, instrumentation, procedure and data analysis plan. Study design The study was a 3 x 4 experimental design. Participants were randomized to one of three groups. Treatment 1 received 10 back massages, treatment 2 received 5 back massages and the control group received ten sessions using self-selected relaxation. Four m easurement time poi nts included preintervention, 48-72 hours post-intervention, at week eight of the study (2 weeks post intervention for the 10-Massage group and Control group / 4 weeks post intervention for the 5-Massage group) and at week ten of the study. The study design attempted to blind the researcher to the data by having a research assistant do data collection. Setting The research study took place at a large metropolitan university. A suite provided by the College of Nursing had tw o rooms, the first being a sitting area where participants were able to sit in co mfortable chairs with a water fountain providing a continuous, relaxing soft sound and magazines to page through while

PAGE 60

48 waiting. The second room contained a t able at which the participant sat to complete written instruments, a comf ortable adjustable chair in which the participant sat to have blood pressures ta ken, and the massage table upon which the participant lie to receive the massage or to do relaxation. The room was divided by a folding screen behind whic h the massage table was positioned, offering participants a sense of privacy wh ile gaining the horizontal position onto the massage table. The room s, painted in a soft butter color, were approximately 12 x 20 feet, providing adequate space. The air temperature in the rooms averaged 72 degrees Fahrenheit. Soft re laxing music played continuously throughout the day from a st ereo in the second room. See Appendix 1 for a list of the musical CDs used in this study. Participants were granted permission by the university parking and transportation department to park in the par king lot adjacent to the College of Nursing during the period of data gathering. The entire data collection took 235 days beginning April 7, 2006 and endi ng November 28, 2006. Sample An a priori power analysis, based on the res earcher’s pilot study, (Olney, 2005) suggested that a sample size of seven per group for diastolic pressure ( = .05, = .80, effect size 1.56) or sample size of three per group for systolic pressure ( = .05, = .80, effect size 2.25) wa s needed to detect a difference. With no other literature suggesting signific ant findings using such a small sample, the researcher consulted experienced rese archers to gain a better estimate for sample size needed in this study. In a phone conversation with expert massage

PAGE 61

49 researcher, Dr. Maria Hernandez-Reif of the Touch Research Institute, University of Miami, Miami, FL, advised that sm aller samples are appropriate with massage because the effect sizes on outcome variabl es tend to be large. Therefore, it was determined that 15 subjects per group would be adequate to detect a significant change in systolic or diastolic blood pressure. The researcher proposed to recruit 18 subjects per group to allow for attrition due to dismissals and drop out. Therefore, the goal was to recruit and enroll 54 participants. Recruitment Given that nearly one in three adults has hypertension (American Heart Association, 2006), the researcher det ermined the local university population would be adequate from which to pull a sa mple of hypertensive participants. Therefore a convenience sample was recr uited from the University of South Florida faculty, staff, students and associ ates, which has a total population of approximately 50,000 with >43, 000 students (University of South Florida, 2005), and >7,000 faculty and staff, (University of South Florida, 2004). Therefore, an estimated 16,000 hypertensive persons, not considering age variance, were available to be recruited from the uni versity campus. Recruitment began with a mass mailing of an approved recruitment flyer via the University of South Florida Health Sciences Center’s e-mail system (Appendix 2). A hard copy of the flyer was also posted on approved bulletin boar ds through out the Health Science Center’s campus. Due to the nature of the study, a rolling enrollment had to occur; therefore the inter net flyer was posted on average every 3 weeks to recruit more participants. After several m onths of posting throughout the Health

PAGE 62

50 Sciences, recruitment numbers began to decli ne. Therefore, the flyer was also posted on other campus e-mail lists and hard copies were posted on approved boards throughout the university campus. All flyers were removed from the boards when recruitment was completed. Over the eight months of re cruitment, 220 people contacted the researcher via either phone or e-mail to i nquire about the study. Each interested person was screened by the researcher for in clusion into the study. Sixty-three of the 220 (28.6%) consented into the study Table 5 lists the reasons for the 157 inquirers to not join the study. Table 5 Reasons Interested Persons did not Join Study Reason Number of people (%) Did not return researcher’s return call or e-mail 46 (20.9%) Too far to drive 28 (12.7%) Schedule would not allow time required 26 (11.8%) Recently changed medications 18 (8.2%) Inquired for another person 11 (5.0%) No reason 11 (5.0%) Family Crisis 9 (4.1%) Insulin Dependent Diabetes Melitus 3 (1.4%) BP not elevated 3 (1.4%) Accident or Illness 2 (1%) On weight loss program 2 (1%) Money not offered 2 (1%) Total 157 (71.4%)

PAGE 63

51 Sampling Method The researcher initially screened each person in a phone conversation for inclusion criteria, described the study and burden to the potential subject. The inclusion and exclusion criteria based upon literature review. Inclusion criteria included: 1) Between ages 18 to 75 years of age, 2) Diagnosis of prehypertension as defined by the JNC VII or controlled hypertension as defined by participant’s health care provider, 3) Pe rsons with diagnosis of hypertension had to be under the care of a primary health care provider and have no changes in medications in the past 6 months, 4) Be able to speak, read, understand, and write English at the sevent h grade level, 5) Must be available for ten weeks. Exclusion criteria included: 1) Less than 18 years or over 75 years, 2) Persons with uncontrolled hypertens ion as defined by systolic blood consistently > 140 mmHg, 3) Persons wit h hypertension not under the care of a primary care provider, 4) Persons wit h recent (past 6 months) changes in antihypertensive medications includi ng dosage changes, 5) Non-English speaking, reading and writing persons, 6) Persons with inflammation or open sore or rash on the back, shoulders and or neck areas, acute infections, advanced osteoporosis, aneur ysm, hematoma or edema of the area being massaged, diabetes, active treatment of cancer, experiencing extreme fatigue, intoxication, psychosis, frailty or any other process that deemed contraindicated by their personal health care provi der, 7) Persons with insulin dependent diabetes, 8) Pregnant women beyond the second month, due to discomfort of lying on abdomen during the massage, 9) Unable to commit to entire length of

PAGE 64

52 study, and 10) Persons with hist ory of thoracotom y. This last criteria was added after an adverse event (See Appendix 3 for le tter from Institutio n Review Board). Of the 220 inquiries, 63 interested per sons met the study criteria. A 30 minute appointment was arr anged for the participant to co mplete the initial paper work, meet the researcher, and have t he study explained again to ensure the participant understood the commitment. Protection of Human Subjects The initial paper work in cluded the approved Univer sity of South Florida Human Subjects Institutional Review Board consent form (Appendix 4), the approved University of South Florida Research Authorization for Use and Disclosure of Protected Health Info rmation form (Appendix 5), and a Subject Information form (Appendix 6) If the participant was ol der than 65 years of age, a Mini-Mental State Exam was administer ed (Appendix 7). Once the researcher was assured the participant was committed to the study, that they understood the commitment and the consent and health in formation protection form was signed, a schedule was arranged with the parti cipant for the entire study. Randomization Fourteen of the consented 63 participant s dropped from the study prior to randomization. The reasons for leavi ng the study are listed in Table 6. Randomization to group occurred on the fi fth visit to the research lab (see procedure). At the study’s completi on, there were 42 participants who had

PAGE 65

53 Figure 4: Participant Eligibility and Attrition Eligible Participants (n=63) Randomized (n=49) Excluded from study (n = 14) 10-Massage (n = 17) 5-Massage (n = 17) Control: (n = 15) Received allocated intervention (n=16) Did not receive allocated intervention due to elevated blood pressure (n=1) Received allocated intervention (n=15) Did not receive allocated intervention due to elevated blood pressure (n=1), adverse event (n = 1) Received allocated intervention (n=14) Did not receive allocated intervention due to schedule change (n=1) Lost to follow up due to: medication change (n = 1), Moved (n = 1) Lost to follow up due to medication change (n = 1) Analysed (n = 15) Analysed (n = 13) Analysed (n = 14) Lost to follow up: (n = 0)

PAGE 66

54 Table 6 Reasons Consented Participants Dropped out of Study Reason Number Scheduling conflicts with work or family 7 Change in BP medication 2 Blood pressure too high* 2 Blood pressure too low 1 Refusal to comply with protocol 1 No reason 1 Total 14 Referred to primary physician. completed all 4 times points in this study Figure 4 depicts t he break out of how the sample of 42 participants occurred. Instrumentation The main outcome variables in this st udy were systolic and diastolic blood pressure. Unlike most stress intervent ion studies, this study measured several stress related variables as covariates as opposed to outcome variables to test the theoretical structure of this design. For example within the psychophysiology theory, it is known that blood pressure changes are possibly due to perceived stress. Understanding the levels of stress that the participant was experiencing at the time of each measurement helps with the interpretation of the blood pressure level. Levels of stress ar e measured many ways in science. This researcher chose to measure stress biologically and ps ychologically using salivary cortisol, and anxiety, depression, and anger, respectively. Table 7 lists and discusses the instruments used to measure the outco me variables and the covariates.

PAGE 67

55 Each participant was measured at four time points during the study: prerandomization (week two), post intervention (week four for the 5-massage group, week six for the 10-massage group and cont rol), at week eight and week ten (Appendix 12). Written inst ructions in preparation to the data gathering sessions were provided to the participants. The inst ructions asked that the participant not Table 7 Variables Measured and Instruments used to Measure each Variable Variable Instrumentation Systolic Blood Pressure Dinamap Pro Care 400 automated oscillometric monitor with digital screen and printout. Diastolic Blood Pressure Dinamap Pro Care 400 automated oscillometric monitor with digital screen and printout. Heart rate Dinamap Pro Care 400 automated oscillometric monitor with digital screen and printout. Baroreceptor Response Dinamap Pro Care 400 automated oscillometric monitor digital screen and printout. Salivary Cortisol DRG Diagnostics Cortisol ELISA, DRG International, Inc. USA Anxiety Spielberger’s revised State-Trait Personality Inventory Depression Spielberger’s revised State-Trait Personality Inventory Anger Spielberger’s revised State-Trait Personality Inventory Expectations Subjective Rating Scale Touch Perception Subject Touch Perception Scale (Semantic Rating Scale) Rapport with Therapist Subject Perceived Rapport Scale (Semantic Rating Scale) Age Alcohol BMI Ethnicity/Race Exercise Gender Medications Smoking Subject Information form: Self reported.

PAGE 68

56 eat or drink at least one hour prior to the measurements. The routine for data gathering went as follows: saliva sample (5-minutes), five blood pressures of three sitting, one lying and one just after sitt ing up (~ 17 minut es) and State-Trait Personality Inventory (~ 8 mi nutes). The routine was strictly followed for all cases except one in which the meas urement routine was reversed due to participant having only fasted for 45 minutes prior to co ming in for data gathering session. Blood Pressure This study used a Dinamap Pro Care oscillometric monitor, donated by General Electric Medical Systems Inform ation Technologies to measure blood pressure. Blood pressure has tradi tionally been measur ed by auscultatory methods using a mercury sphygmomanometer (Pickering et al., 2005). Other methods, such as the oscillometric monitor, are becoming more common and have several advantages over the mercur y method (Pickering et al., 2005). One detriment to using a oscillometric monito r is that the algorit hms used to measure blood pressure vary from one monitor to t he next. To avoid this problem in the study, the same monitor was used thr oughout the study. The Dinamap Pro Care 400 automated oscillometric monitor is a highly respected and commonly used instrument within health care systems (GE Healthcare, 2007). The instrument used in this study was calibrated and te sted repeatedly for reliability to measure blood pressure accurately. The instrum ent was tested for f unction and safety prior to the study’s use by General Elec tric technicians. During this study the researcher tested the in strument against a manual sphygmomanometer on a

PAGE 69

57 regular basis to ensure the Dinamap Pro Care 400 continued to measure accurately. Correct blood pressure cuff size is very important fo r accurate blood pressure measurement (Pickering et al., 2005). Each participant had their nondominant arm measured by the researcher to determine correct cuff size prior to the first blood pressure measurement according to the American Heart Association Guidelines fo r cuff sizes. Three systolic and diastolic blood pressures were measured and averaged at each of the four time points. The Dinamap Pro Care 400 monitor was also used to record the heart rate during each blood pressure measurement. The monitor stored the recordings and provided a paper print out of those blood pressure an d heart rate recordings at the end of each participant’s visits. To determine baroreceptor response, blood pressures sitting and lying are compared. Therefore the appropriate instrument to measure this variable was the Dinamap Pro Care 400. Salivary Cortisol Salivary cortisol was used as a physi ological marker of perceived stress (Smyth et. al., 1998). Salivary cortis ol kits obtained from DRG Diagnostics Cortisol ELISA, DRG International, Inc. USA were used by the researcher. Dr. Maureen Groer, University of South Florida, College of Nu rsing, provided the lab, freezer space, and guidance in co rrect laboratory methodology.

PAGE 70

58 State Trait Personality Inventory The revised State Trait Personality Inventory (STPI) is an 80-item inventory comprised of eight 10-item scale s (Spielberger et al., no date). The scales measure emotional states and personality traits of anxiety, anger, depression and curiosity. This self r eport questionnaire is arranged so that the 40 state items are listed on the front of a single page and the 40 trait items listed on the back of that page (Appendix 8). The psychometrics of the ST PI were reported in the Preliminary Manual for the State Trait Personality Inventory by Charles C. Spielberger and colleagues, unpublished, unless other wise indicated (Appendix 9: correspondence with Dr Spielberger about the manual). The 10-items for state and trait anxiety were selected as the bes t items from the 20item parent scales, the State and Trait Anxiety Inventory (f orm Y), correlating .95 or higher with college students. Alpha coefficients based on working adults were: > .88 for trait anxiety and > .91 for state anxiety (Spi elberger et al., unpublished). The state and trait anger scales are t he same as the original State-Trait Anger Expression Inventory (S TAXI). Correlations with original State and Trait Anger Inventory based on 198 Navy pers onnel and over 200 college students were high ( r > .94). The anger alpha coefficient s for working adults were: trait anger .88 or higher and for state anger .93 or higher (Spielberger et al., unpublished).

PAGE 71

59 The internal consistency (alpha coe fficients) for state and trait depression scales (and subscales) were .93 or hi gher for university students (81 males and 170 females) (Spielberger, Ritterband, Reheiser & Brunner, 2003). Further, there is good concurrent validity (r =. 72 to .85) with the Beck Depression Inventory (BDI), Center for Epidemiological Studies-Depression scale (CES-D) and the Zung Depression Rating Scale (Z DRS) for the trait depression scales (Spielberger, Ritterband, Reheiser & Brunner, 2003). Expectations Rating Scale The Expectations Rating Scale was dev eloped by the researcher. This subjective scale asked the participants a dichotomous question: “Do you expect the massage therapy (or relaxation) to lo wer your blood pressure?” The purpose of this question was to provide insight r egarding the influence of the participant’s expectation on the outcome variables. Touch Perception Scale The Subject Touch Perception Scale (Appendix 10) was developed by the researcher. This semantic scale asked the participant to respond to their perception of the back ma ssage and it was anticipated that this scale would provide insight regarding the influence of touch perception on the outcome variables. This scale was administered to only those participants that received a massage intervention and only at t he second measurement session. Perceived Rapport Scale The Subject Perceived Rapport Scale (Appendix 11) was developed by the researcher. This semantic scale asked participants who received a massage

PAGE 72

60 intervention to respond to their percepti on of the relationship with the massage therapist. This scale was administered at the second measur ement session only. It was anticipated that this scale would provide insight regar ding the influence of perceived rapport on the outcome variables. Subject Information form The Subject Information form is a self report form developed by the researcher. The body mass index (BMI) used in this study was calculated from the reported height and wei ght (705 X weight / height2). Procedure The participant’s first visit was a data gathering session taking a total of 30 minutes time. Consent, Protection of Health Information, and demographics information was obtained from the partici pant. A detailed schedule for meeting times was agreed upon by the participant and the researcher. The planned schedule asked the participant to come fo r visits three times per week, on a Monday, Wednesday, Friday schedule, for 30 minutes each (See Appendix 12). The purpose of the second, third and fourth meetings was to acclimate the participant to having their blood pressure measured in that environment. During these three visits every participant was t aught methods of relaxation to include controlled breathing, progressive mu scle relaxation, and imagery. Every participant was instructed to use these methods through out the study whenever blood pressure measurement was occurring. Baseline data was gathered during the fifth visit including saliva for cortis ol, three blood pressure s and the State Trait Personality Inventory. After baseline data was completed, the subject was

PAGE 73

61 randomized to one of three groups by dr awing an assignment from a canister. The canister had equal numbers of folded slips of paper with a designated group assignment to: Treatment 1 (10-Massage group), Treatment 2 (5-Massage group) or Control The specified group procedure was then explained to the participant. The sixth through fifteenth visits for the 10-Massage and Control groups and the sixth through tenth vi sits for the 5-Massage group was the intervention period. During the 30-minute intervention visi t the participant first sat quietly for 3-5 minutes, utilizing the lear ned methods of relaxation. Once the participant indicated he or she felt re laxed, one measurement of blood pressure was obtained. Then the individual treatment procedure was instituted. The Massage Therapist For fidelity purposes, this study utilized one massage therapist. The researcher is a nationally certified, state licens ed massage therapist. The researcher developed the back massage based on professional experience, the empirical evidence and advice from a fo rmer massage instruct or, Maggie Kelly, Licensed Massage Therapist. Former clients offered feedback during the development phase of the back massage, allowing refinement to the pressure used and the speed of the strokes. Massage Routine The massage routine used a sequence of three types of strokes for the duration of the 10-minute back massage. Using both hands the therapist applied long effleurage strokes that started from the base of the neck, along both sides of the spine, to the base of the sacrum, then returning to the neck base by a smooth

PAGE 74

62 pulling stroke to the lateral sides of t he back. The hands of the therapist never lost contact with the subject’s back. This effleurage stroke was rhythmic and repeated three times. The second stro ke, also repeated three times, was a circular stroke starting at the base of t he occipital bone, at ear level. The stroke was moved along the lateral neck and over the upper shoulders (deltoids) and returned to the ear area. The third stro ke, repeated three times, again started at the base of the neck, circling the perimete r of the scapula. Each set of three strokes took approximately five sec onds to complete. The sequence was repeated for the full 10-mi nute application. 10-Massage Group and 5-Massage Group Procedure The therapist stepped out of the room while the subject disrobed (behind a screen) from the waist up and then la y prone on a cleanly draped massage table with her/his face resting on the face support extension. The researcher reentered the room only after the participant indicated she/he was prepared. A towel was tucked into the participant’s waist band to protect the clothing and a pillow placed under the lower leg and ankle to support the feet. A rhythmic 10minute back massage was administered using a non-allergenic massage lotion. The back was dried with a clean towel. Be fore leaving the room, the therapist reminded the participant to sit at the edge of the table and breathe for several breaths prior to getting off the massage tabl e to redress. T he therapist left the room and returned once the participant in dicated she/he was dressed and sitting in the chair. The participant sat quietly fo r 2-5 minutes allowing the researcher to

PAGE 75

63 assess any adverse effects of the massage such as dizziness, and then a second blood pressure and heart rate was measured. Control Group Procedure The control group participants experi enced the same procedure as the treatment group participants ex cept they did not receiv e a massage. The control participants were directed to use the l earned relaxation techniques for the 10 minutes as they lie supine on the massage t able. They were offered pillows for support under the head and knees and a blanket for warmth. Post Intervention Procedures Three post-intervention visits took plac e in this study. For the 10-Massage and Control group participants the first post measurement was the sixteenth visit (~week 6). For the 5-Massage group par ticipants, the first post measurement was eleventh visit (~week 4). The thir d and fourth post measurements occurred during the eighth and tenth week afte r their enrollment (Appendix 12). Data Management Security All data was gathered on paper and filed in the respective participant’s file. The files were maintained in a locked cabi net at the study site throughout the duration of the study. Du ring the analysis phase the f iles were maintained in a locked cabinet in the researcher’s offi ce. The data were loaded into an EXCEL file using only coded identification numbers. A password was needed to access the data file.

PAGE 76

64 Data Analysis Plan The data were analyzed using the SPSS Graduate Pack 11.5. Alpha was set at .05, protecting against a Type I erro r. Prior to analysi s, all data were examined for accuracy in data entry, mi ssing values, and whether assumptions of each analysis were met. Basic analysi s of variance was used to determine differences in the mean group demographic characteristics. Hypotheses 1 through 4 required adjustm ent of covariates prior to the analysis comparing changes in group means over time. The MANOVA procedure provided insight and adjustments needed to meet the assumptions of analysis of covariance (ANCOVA). An analysis of covariance tested for mean group changes over time in systolic and diastolic blood pressure. Follow up orthogonal analysis was planned for the expected interactions. Hypothesis 5 was concerned with the c hange in baro-receptor function. This hypothesis stated that the change in the lying bl ood pressure group mean compared to the sitting blood pressure group mean would be significantly different for the participants who receiv ed 10 massages than the control group. This hypothesis was analyzed using an ANOVA procedure. Chapter Summary Chapter 3 described the method used in this study. The process of sample selection was described, includ ing protection of t heir human rights and their personal information. A figure (Figure 4) of participant eligibility depicted the number of interested individuals and how, once completely randomized into the study, the attrition rate prol onged the recruitment time per iod. The instruments to

PAGE 77

65 measure each variable were fully descri bed. The procedure for each group was different therefore full discloser was incl uded in this report. Data security throughout the study was ensured. T he analysis plan for each hypothesis was discussed and the results will be fully disclosed in Chapter four.

PAGE 78

66 Chapter Four Overview of Findings The purpose of this study was to evaluate the long term effects and dosage of back massage for persons with hy pertension or prehypertension. This randomized clinical trial tested five hypotheses addressing the effectiveness of back massage. The first two hypotheses were related to the long term effects of back massage. The third and fourth hypotheses compared dosage of back massage. The fifthl hypothesis addre ssed a possible causal physiological mechanism for change in blood pressure using back massage. Each participant was measured for all variables at four time points during the study: prerandomization (week two), post intervention (week four for the 5-massage group, week six for the 10-massage group), at w eek eight, and week ten (Appendix 12). The probability of making a Type I error was set at p<.05 for all the analyses. Analytical Strategy Unlike many stress intervention studi es that measure biological and psychological variables as outcome va riables, this study was designed to evaluate blood pressure changes within t he context of the psychological and biological status of the par ticipant at each measurement session. Therefore, this study used analysis of variance (ANOVA) and analysis of covariance (ANCOVA) to determine the effects of massage on blood pressure. The assumptions of ANOVA and ANCOVA were assessed for vi olations. If violations to the

PAGE 79

67 assumptions were found, the necessary adjustments were made. The purpose for using covariates in the analysis was to decrease error variance and increase power. Further, each covariate was analyzed for its relationship to the dependent variables prior to adding to t he model. Significant correlations regarding covariate to dependent variable and group by time interactions and patterns helped determine support for each hypothesis. Preliminary Analyses Sample The sample consisted of 42 partici pants (26 females, 16 males). The sample had a 67% completion rate for the study (see Figure 4). The data set from these 42 participant s’ self-reported characteristics had no missing data (Table 8). The age range was 26 to 70 years with a mean age of 48.67 (SD = 12.1) years. The racial and ethnic make up of the sample was mostly white (86 %). African Americans m ade up 9 % of the total and As ians and Unknown/ Other categories each made up 2% to the sample The three groups were compared to determine if there were significant differ ences among them on specific variables. The only significant difference found among the three groups was body mass index (BMI), a weight-height rati o (weight(kg) / [height (meter)]2): (F [2, 39] = 3.38, p<.045) (Table 8). Antihypertensiv e and diuretic medication usage (Table 9) did not differ between the three groups. N early half (53%) of the participants in the 10-Massage group were on medications, and nearly two thirds (64%) of the

PAGE 80

68 Table 8 Characteristics of Groups Group 10-Massage 5-Massage Control Total Group Significance Characteristic (n = 15) (n = 13) (n = 14) (n = 42) P-value (.05) Gender 2 (2) = 2.108, p =.349 Female 9 (60%) 10 (77%) 7 (50%) 26 (62%) Male 6 (40%) 3 (23%) 7 (50%) 16 (38%) Age* 48 (13.95) 50.15 (10.60) 48. 00 (12.06) 48.67 (12.1) F = .136, p = .873 Race/Ethnicity 2 (8) = 7.877, p =.446 AfricanAmerican 2(13%) 2 (15%) 0 4 (9%) Asian 0 0 1 (7%) 1 (2%) Caucasian 11(73%) 11 (85%) 11 (79%) 33 (86%) Hispanic 1 (7%) 0 2 (14%) 3 (7%) Other 1 (7%) 0 0 1 (2%)

PAGE 81

69 Table 8 (continued) Group 10-Massage (n = 15) 5-Massage (n = 13) Control (n = 14) Total Group (n = 42) Significance P-value (.05) Body Mass Index (kg/m2)* Self-reported. 27.85 (3.67) 30.33 (7.68) 34.22 (8.01) 30.74 (7.03) F = 3.35, p = .045 Smoker (Yes or No) 1 (7%) 2 (15%) 1 (7%) 4 (9.5%) 2 (2) = .752, p =.686 Drinks Alcohol (Yes or No) 10 (67%) 8 (62%) 6 (43%) 24 (57%) 2 (2) = 1.83, p =.402 Exercise (Mean Minutes per day)* 31 (24.78) 22 (20.40) 18 (23.51) 23.65 (23.44) F = 1.29, p = .286 Hypertension (Mean years diagnosed)* 2.9 (3.73) 9.5 (11.26) 6.4 (6.41) 6.10 (7.89) F = 2.65, p =.083 Takes at least one medication for Hypertension 8 (53%) 11 (85%) 10 (71%) 29 (69%) 2 (2) = 3.244, p =.197 Mean (Standard Deviation).

PAGE 82

70 Control group were on medications. See Appendix 13 for all medications used by study participants. Table 9 Antihypertensive Medications Used by Study Participants Medication 10-Massage Group # (% of group) 5-Massage Group # (% of group) Control # (% of group) Total # (% of total sample) ACE Inhibitors 3 (20%) 3 (23%) 2 (14%) 8 (19%) Angiotensin II Receptor Blockers 3 (20%) 3 (23%) 2 (14%) 8 (19%) Beta Blockers 0 3 (23%) 2 (14%) 5 (12%) Calcium Channel Blockers 1 (6%) 0 2 (14%) 3 (7%) Diuretics 4(27%) 4 (31%) 5 (36%) 14 (33%) # of Participants on HBP meds [# on 2 meds] 11 [3] (53%) 13 [2] (85%) 13 [3] (71%) 37 [8] (69%) The psychological variables were measured to reflect the level of perceived stress at the time of each m easurement session. The State Trait Personality Inventory (Spielberger et al., unpublished) was used to measure three psychological variables that are re lated to hypertension: anxiety, anger and depression. The reliabilities for the invent ories in this study were strong (i. e. State inventory = .88 and trait inventory = .94). The means and standard deviations for state anxiety, anger and depression are reported in Table 10. The means and standard deviations for trait anxiety, anger and depression are reported in Table 11. There were no between-group differences on any of these p sychological measures at any of the time points.

PAGE 83

71 Table 10 Means (Standard Deviations) and Significance of State Anxiety, An ger and Depression Variable* 10-Massage Group 5-Massage Group Control TOTAL Significance Pre State Anxiety 15.53 (5.69) 13.23 (2.89) 15.36 (4.88) 14. 76 (4.71) F(2,39) = 1.002, p = .376 Post State Anxiety 15. 80 (6.96) 14.85 (7.05) 14.07 (6.44) 14.93 (6.69) F(2,39) = 0.234, p = .793 Post 1 State Anxiety 15.27 (4. 45) 17.38 (8.17) 15.43 (8.12) 15. 98 (6.94) F(2,39) = 0.378, p = .688 Post 2 State Anxiety 16.40 (6. 67) 14.00 (6.10) 14.64 (5.58) 15. 07 (6.09) F(2,39) = 0.581, p = .564 Pre State Anger 12.40 (5.29) 10.38 (0.96) 11.29 (2.20) 11.40 (3.47) F(2,39) = 1.197, p = .313 Post State Anger 11.47 (3.91) 12.92 (8.04) 11.00 (2.72) 11. 76 (2.72) F(2,39) = 0.485, p = .620 Post 1 State Anger 10.73 (1.28) 13.08 (5.65) 11.93 (5.36) 11. 86 (4.46) F(2,39) = 0.960, p = .392 Post 2 State Anger 11.60 (4.36) 10.31 (0.86) 11.00 (1.84) 11. 00 (2.84) F(2,39) = 0.712, p = .497 Pre State Depression 13.20 (2.91) 12.31 (1.70) 15.64 (5.49) 13. 74 (3.91) F(2,39) = 2.928, p = .065 Post State Depression 16.47 (5. 58) 15.08 (4.72) 15.86 (7.89) 15. 83 (6.10) F(2,39) = 0.174, p = .841 Post 1 State Depression 14.67 (3 .50) 16.31 (6.26) 16.50 (7.50) 15.79 (5.85) F(2,39) = 0.418, p = .661 Post 2 State Depression 15.67 (5 .12) 16.38 (6.46) 16.29 (7.45) 16.10 (6.23) F(2,39) = 0.053 p = .948 *Pre = measurement at week 2 Post = measurement at w eek 4 if in 5-Massage group or at week 6 if in 10Massage group or Control Post 1= measurement at week 8 Post 2 = measurement at week 10

PAGE 84

72 Table 11 Means (Standard Deviations) and Significance of Trait Anxiety, Anger and Depression Variable* 10-Massage Group 5-Massage Group Control TOTAL Significance Pre Trait Anxiety 18.53 (5.42) 16.85 (5.51) 18.50 (5.13) 18. 00 (5.28) F(2,39) = 0.438, p = .649 Post Trait Anxiety 17.80 (4.90) 15.69 (5.20) 17.43 (6.16) 17. 02 (5.39) F(2,39) = 0.581, p = .564 Post 1 Trait Anxiety 17.67 (4.97) 16.54 (4.98) 17.64 (6.34) 17.31 (5. 36) F(2,39) = 0.187, p = .830 Post 2 Trait Anxiety 17.87 (5.28) 17.23 (5.97) 17.64 (6.33) 17.60 (5. 72) F(2,39) = 0.042, p = .959 Pre Trait Anger 18.40 (5.54) 18.31 (5.09) 16.57 (3.78) 17.76 (4.83) F(2,39) = 0.628, p = .539 Post Trait Anger 17.27 (4.42) 16.31 (3.33) 16.21 (3.47) 16. 62 (3.74) F(2,39) = 0.342, p = .713 Post 1 Trait Anger 17.27 (4.18) 16.77 (4.83) 15.64 (4.11) 16.57 (4.32) F(2,39) = 0.519, p = .599 Post 2 Trait Anger 16.93 (4.33) 16.23 (5.09) 15.57 (4.24) 16.26 (4.47) F(2,39) = 0.325, p = .724 Pre Trait Depression 16.73 (5.09) 15.62 (4.79) 17.64 (5.42) 16. 69 (5.05) F(2,39) = 0.531, p = .592 Post Trait Depression 17.07 (5. 43) 15.23 (5.05) 17.93 (7.11) 16. 79 (5.90) F(2,39) = 0.721, p = .493 Post 1 Trait Depression 16.60 (5.12) 15.62 (5. 30) 16.57 (6.70) 16.29 (5. 62) F(2,39) = 0.128, p = .880 Post 2 Trait Depression 17.87 (5.72) 16.54 (7. 13) 17.43 (8.00) 17.31 (6. 83) F(2,39) = 0.129, p = .879 *Pre = measurement at week 2 Post = measurement at w eek 4 if in 5-Massage group or at week 6 if in 10Massage group or Control Post 1= measurement at week 8 Post 2 = measurement at week 10

PAGE 85

73 The cortisol means and standards deviations and ANOVA results are reported in Table 12. There were no group differences at any of the four time points. Table 12 Cortisol Means (Standard De viations) and Significance Variable* 10-Massage Group 5-Massage Group Control TOTAL Significance Pre Cortisol 12.54 (8.38) 14.14 (7.18) 10.65 (4.79) 12.45 (6.99) F(2,38) = 0.804, p = .455 Post Cortisol 15.37 (5.92) 16.33 (7.11) 14.64 (6.46) 15.44 (6.36) F(2,38) = 0.223, p = .801 Post1 Cortisol 14.35 (6.81) 15.47 (7.42) 12.83 (7.09) 14.22 (7.00) F(2,38) = 0.455, p = .638 Post2 Cortisol 13.44 (7.70) 14.73 (6.60) 11.28 (6.38) 13.16 (6.93) F(2,38) = 0.817, p = .450 *Pre = measurement at week 2 Post = measurement at week 4 if in 5-Massage group or at week 6 if in 10-Massage group or Control Post 1= measurement at week 8 Post 2 = measurement at week 10 The expectation question, a dichotomous variable, asked if the participant believed if the massage (or relaxation, if in the control group) would help lower their blood pressure. Over 86% of t he participants in the 10-Massage group and 93% of participants in the 5-Mass age group responded to the expectation question in the affirmative. Approximately 76% of t he Control group believed the relaxation would decrease their blood pr essure. The groups did not differ 2 (4) = 4.616, p = .329.

PAGE 86

74 The Subject Touch Perception Scale is a 6-item semantic differential scale developed by the researcher used to gain a sense of what the participant may be experiencing when receiving the back massage. The groups were similar in their responses to these items. Findings are r eported in Table 13. The reliability of this scale was poor (alpha = -.045). Table 13 The Subject Touch Perception Sc ale Means (Standard Deviation) Item 10-Massage Group (n = 15) 5-Massage Group (n = 13) Soothing / Irritating 1.13 (.352) 1.08 (.277) Unpleasant/ Pleasant* 1.07 (.258) 1.15 (.376) Pressure 2.53 (.640) 2.69 (.480) Rough/ Smooth* 1.20 (.414) 1.38 (.870) Rhythmic/ Jerky 1.13 (.516) 1.00 (.000) Relaxing/ Tensing 1.27 (.594) 1.08 ( .277) indicates items that were reverse scored. The Subject Perceived Rapport Scale is a 10-item Semantic Differential scale developed by the researcher used to gain a sense of the participant’s perceived relationship with the massage therapist/researcher. Findings are reported in Table 14. The reliability of this scale was good (alpha = .83). The means and standard deviations of t he dependent variables of systolic blood pressure and diastolic blood pressure are reported in Table 15. There were no significant differences between grou ps at any time points of the study.

PAGE 87

75 Table 14 Subject Perceived Rapport Scal e Means (Standard Deviation) Item 10-Massage Group (n = 15) 5-Massage Group (n = 13) Comfortable/ Uncomfortable 1.20 (.561) 1.00 (.000) Distrustful/ Trustful 1. 27 (1.03) 1.00 (.000) Less Anxious/ More Anxious 1.27 (.594) 1.31 (.630) Calm/Annoyed 1.20 (.561) 1.00 (.000) Worse/ Better* 1.33 (.617) 1.23 (.599) Physically Tense/ Physically Relaxed/ 1.53 (.915) 1.08 (.277) Willing to share feelings/ Unwilling to share feelings 1.40 (.737) 1.15 (.555) Unsafe/ Safe 1.00 (.000) 1.00 (.000) Respected/ Ignored 1.13 (.516) 1.00 (.000) Unpleasant/ Pleasant 1.13 (.516) 1.08 (.277) indicates items that were reverse scored Assumptions Missing data All data concerning participant charac teristics, and considered for use as covariates, were complete. All other va riables that were considered potential covariates in the analysis we re assessed for missing data. Missing data were present on several it ems in the State Trait Personality Inventory (STPI). The State Trait Personality Inventory c onsists of 80 items to which the participant responds. Each it em was responded to four times during the study. Of the 13,440 ( 80 items/ inventory x 4 times x 42 participants)

PAGE 88

76 Table 15 Demographics of Dependent Variabl e Means (Standard Deviations) Variable 10-Mass ageGroup 5-Massage Group Control (n = 15) (n = 13) (n = 14) Pre Systolic 121.04 (11.32) 118.95 (12.49) 124.19 (10.53) Post Systolic 119.38 (17.97) 120.79 (13.91) 119.83 (9.35) Post Systolic1 119.76 (13.09) 122.38 (14.80) 119.43 (8.12) Post Systolic2 120.73 (15.06) 115.82 (12.80) 118.31 (11.22) Pre Diastolic 76.91 (7.18) 75.87 (5.59) 75.69 (10.86) Post Diastolic 76.73 (10.81) 76.51 (6.27) 72.64 (9.41) Post Diastolic1 77.58 (8.97) 76.82 (8.46) 76.14 (7.18) Post Diastolic2 76.69 (10.60) 75.31 (7.34) 76.62 (9.07) *Pre = measurement at week 2 Post = meas urement at week 4 if in 5-Massage group or at week 6 if in 10-Massage group or Control Post 1= m easurement at week 8 Post 2 = m easurement at week 10 Note: The groups did not differ; all p values were > .05

PAGE 89

77 possible responses in the State Trait Pe rsonality Inventory, 10 participants failed to respond to 17 items. The provi ded instructions for managing missing data were followed for all missing data (Spielber ger, unpublished). Briefly, a prorated score was computed by determining a mean score for the subscale from which the item was missing. The value wa s then multiplied by 10 and rounded to the next highest whole number. Heart rate data were missing on two par ticipants in the control group at pre-measurement. Three heart rate post-measurements were missing on one participant. These cases were dropped in an y analyses that involved heart rates. Salivary cortisol was measured at the four established ti me points. One case in the control group had missing data at each of the four time points. The case was dropped from any analyses that involved cortisol as a covariate. There were no missing data from the S ubject Touch Perception Scale and the Subject Perceived Rapport Scale data set. The Expectations question, which asked for a “yes” or “no” response, had one missing case in each of the massage groups. These cases were left out of any subsequent analyses. There were no missing data on the dependent variables, systo lic and diastolic blood pressures. Normality and Outliers There was significant positive skewne ss in the potential covariates of anxiety, anger, and depression. Figures 5, 6 and 7 depict the means with standard deviations of state anxiety, anger and depression, respectively.

PAGE 90

78 6 11 16 21 26 246810 WEEKSTATE ANXIETY SCOR E 10M 5M CNT Figure 5: State Anxiety M eans and Standard Deviation Bars Figure 6: State Anger Means and Standard Deviation Bars Appendix 14 shows a similar depi ction of the means with standard deviation bars for trait anxiety, anger and depression. Of particular interest, preliminary analysis resulted in no differ ence (effect sizes did not change, error did not decrease) using both the ske wed data and the transformed data. Therefore for ease of interpretation, the un-transformed data of anxiety, anger and depression were used in the analysis. 4 6 8 10 12 14 16 18 20 2 4 6 8 10 WEE K STATE ANGER SCORE 10M 5M CNT

PAGE 91

79 Figure 7: State Depression M eans and Standard Deviation Bars Cortisol data were normally distri buted. Figure 7 depicts the means with standard deviation bars for the th ree groups over time. Figure 8: Group Means of Cortisol Levels over Time Systolic blood pressure and diastolic blood pressure data were normally distributed. See Figures 8 and 9 fo r group means with standard deviation bars depicting the overlap of the distribution of systolic and diastolic blood pressures. 7 9 11 13 15 17 19 21 23 25 2 4 6 8 10 WEE K STATE DEPRESSION SCORE 10M 5M CNT 3 8 13 18 23 28 2 4 6 8 10 WEE K CORTISOL LEVELS 10M 5M CNT

PAGE 92

80 Figure 9 Group Means of Systolic Blood Pressures over Time Figure 10 Group Means of Diastolic Blood Pressure over Time The between group values did not differ at any of the four time points. Specifically, systolic blood pressure at week 2: F(2, 39) = 721, p = .493; systolic blood pressure at week 10: F(2, 39) = 484, p = .620; diastolic blood pressure at week 4 & 6: F(2, 39) = .734, p = .487. 100 105 110 115 120 125 130 135 140 2 46 8 10 WEEKSYSTOLIC BP 10M 5M CNT 62 67 72 77 82 87 92 2 46 8 10 WEEKDIASTOLIC BP 10M 5M CNT

PAGE 93

81 Homogeneity of Variance The dependent variables were homogeneou s in their variances. Systolic blood pressure: Box’s M = 28.59, F(20, 5295) = 1.213, p = .232; diastolic blood pressure: Box’s M = 34.12, F = (20, 5295) = 1.448, p = .089. Sphericity The assumption of sphericity was te sted on the dependent variables. The test for sphericity was found to be not si gnificant, thus protecting from Type 1 error. Homogeneity of Regression Slopes One of the assumptions using ANCO VA is that there is homogeneity of the regression slopes i.e. the dependent variable-covari ate slopes are the same for the three groups. All covariates we re tested with both dependent variables for homogeneity of regression slopes. T here was significant heterogeneity of regression slopes with the covariate BMI and the systolic blood pressure but not diastolic blood pressure. Table 16 depict s the significant relationship of BMI within the 10-Massage group, F (1, 36) = 5. 14, p = .029. The BMI within the 5-Massage group and the BMI within the Control group appear to have flat slopes. All other covariates were tested with the dependent variables and appear to have homogeneity of regressi on slopes. This means that the covariate-dependent variable slopes ar e the same for th e three groups.

PAGE 94

82 Table 16 Heterogeneity of Variance of BMI and Systolic Blood Pressure Source of Variance SS df MS F P Value Group 3107.0 2 1553.5 3.38 .045 TBMI within 10Massage Group 2362.82 1 2362.82 5.14 .029 TBMI within 5Massage Group 734.35 1 734.35 1.60 .214 TBMI within Control Group 33.31 1 33.31 .07 .789 Note: TBMI is the variable representing body mass index Correlations The framework of this study requir ed that the relationship between the dependent variables and the covariates be assessed within the context of time. Therefore the correlation matrices (A ppendix 15) were designed to evaluate relationships of dependent and covariates at each of the four measuring time points: pre-randomization (week 2) postintervention (week 4 for the 5-Massage group or week 6 for the 10-Mass age group and Control group), post 1intervention (week 8) and post 2-inte rvention (week 10). The following discussion will point out interesting significa nt relationships at each time point. Correlations at pre-randomization: pre-systolic and pre-diastolic blood pressures had significant negative correlati ons with antihypertensive medications (r = -.438 and r = -.348) respectively. This most likely indicates that the more medications are used, the lower the blood pressure. Antihypertensive medications had a positive correlation with number of years with hypertension

PAGE 95

83 (r = .326) which can be interpreted as the more years diagnosed with hypertension, the more apt to be on medication. The negative relationship between medications and minutes per day of exercise (r = -.415) infer that the more one exercises the fewe r medications one is on. The negative correlation between age and minutes per day of exer cise (r = -.443) infers the older participants exercised fewer minutes per day. State anger was negatively correlated with age (r = -.319) and positiv ely correlated with state anxiety (r = .703). State depressi on had positive relations wit h BMI (r = .354), state anxiety (r = .636) and state anger (r = .393). Trait anxiety related positively to BMI (r = .398), state anxiety (r = .504), state anger (r = .366) and state depression (r = .620). Trai t anger was positively related to state anger (r = .362) and trait anxiety (r = .327). Trait depre ssion was positively related to BMI (r = .407), state anxiety (r = .476) state anger(r = .390) state depression (r = .622) and trait anxiety (r = .872). Finally, the biological marker for perceived stress, cortisol, was positively related to trait depression. The second measurements were done at week 4 for the 5-Massage group and week 6 for the 10-Massage group and t he Control group. The correlations changed with any variables that were time s ensitive. Systolic and diastolic blood pressures had negative correlations with antihypertensive medications (r = -.422 and r = -.391) respectively. Unlike the pre-randomiz ation there was a positive correlation with systolic and diastolic blood pressures and state anxiety (r = .443 and r = .401) respectively. St ate anger was positively related to years on medication (r = .367) and state anxie ty (r = .779). State depression

PAGE 96

84 relationship at this second measure bore about the same positive relations as the pre-randomization with BMI (r = .374), state anxiety (r = .712) and state anger (r = .429), but not state depr ession. Trait anxiety compared to pre-randomization was similar with positive relations to BMI (r = .450), state anxiety (r = .420), and state depression (r = .661). Trait anger was positively related to state depression (r = .661).The biological marker for perceiv ed stress is positively related to years on medication which is different from the pre randomization. The third measurements were done at week 8. Systolic and diastolic blood pressures continued to have negative correlations with antihypertensive medications. State anxiety continued to be positively related to systolic and diastolic blood pressures. State anger had a positive relationship with at this measurement diastolic blood pressure (r = .347), which is different from the previous times. State anger and state anx iety continued to be positively related (r = .839). State depression related to BMI about the same as week 2/4 and related to positively to state anxiety (r = .849) and state anger (r = .833). Trait anxiety continued to be positively related to BMI. Trait anxiety also was positively correlated with all three state anxiety anger and depression. Trait and state depression were positively related (r = .370). Trait depression had positive correlations with state anxiety(r = .417), state anger(r = .415), state depression (r = .681) and trait anxie ty(r = .870) and trait depression(r = .347). There were no correlations with cortisol at week 8. Th is is interesting in view of the high correlations with depression and the other psychological markers.

PAGE 97

85 The fourth measurements were comp leted at week 10 of the study. Systolic and diastolic blood pressures continued to be negatively related to the number of medications and positively relate d to state anxiety. Diastolic blood pressure showed a positive correlation with state anger (r = .406) similar to week eight. State anger and state anxiety conti nued to be positively related. State depression continued to be positively rela ted to BMI, state anxiety and state anger. Trait anxiety was positively correlated to BMI (r = .46) as well as with state anxiety and state depression but not state anger. Trait anger was different from the previous measurements with positive correlations wi th BMI, state anxiety, state anger, state depression and trait anxiety. Trai t depression correlations were similar to week 8 except there was no significant relationship with state anger. Cortisol had a positive correlation with state anxiety (r = .334) and depression (r = .315). Between-group variances were assessed on the sample characteristics that were considered potent ial covariates. As noted earlier, body mass index was the only potential covariate that wa s found to be significantly different between the groups, rendering adjustment necessary. Two covariates in particular came in to focus as areas of concern. Antihypertensive medications relationship was of concern because of the strong relationship with blood pressures at each ti me point. Also BMI was of concern because of the great variance between grou ps. A variable was created to help evaluate the blood pressure change over ti me and the relationships of these two covariates. A new variable was creat ed and represented the difference scores

PAGE 98

86 between the first and the last blood pre ssure measurements. Correlations between the two covariates of concern (BMI and antihypertens ive medications) and systolic and diastolic pressures c hange scores, by group, are shown in Tables 17, 18 and 19. In the 10-Mass age group there are si gnificant negative correlations between all BMI and the pre ssure change scores except one (Post 2 diastolic blood pressure – Pre-diastolic blood pressure). Antihypertensive medications did not have a significant correlation with any of the change scores in the 10-Massage group. In the 5-Ma ssage group, antihypertensive medications had a negative correlation with BMI and all but the SYSDIFF (Post Systolic blood pressure – Pre-systolic blood pressure ) change scores, none being significant. Further, the BMI correlation with t he medications and change scores appear benign. The Control group has signifi cant positive correlations between antihypertensive medications and BM I, antihypertensive medications and DYSDIFF1 (Post1 Diastolic blood pressu re – Pre-diastolic blood pressure), and BMI and DYSDIFF1. Perhaps the rela tionship could be summed up as the data appear to say there is a negative rela tionship between body mass index and blood pressure in the 10-Massage group, unlike the other two groups. The antihypertensive medication correlation with blood pressure difference scores over the four time points revealed little evidence of a strong systematic relationship. Scatter plots of the three groups’ rela tionships of BMI and blood pressure difference scores helps to clarify the natur e of the relationship. Figures 11, 12

PAGE 99

87 Table 17 Correlations of Antihypertensive Medication, BMI and Blood Pressure Difference Scores: 10-Massage Group. AHBP MED BMI SYS DIFF SYS DIFF1 SYS DIFF2 DYS DIFF DYS DIFF1DYS DIFF2 AHBP MED .318 -.360 -.414 -.326 -.266 -.336 -.178 BMI -.643** -.639* -.636* -.572* -.588* -.433 SYS DIFF .568* .784** .816** .370 .568* SYS DIFF1 .576* .378 .829** .228 SYS DIFF2 .767** .399 .681** DYS DIFF .403 .823** DYS DIFF1 .290 DYS DIFF2 Pearson’s Correlation signific ant at the .05 level (2-tailed). ** Pearson’s Correlation significant at the .01 level (2-tailed). A-HPN = Antihypertensive Medication (angiotensin conver ting enzyme inhibitors, angiotensin-receptor blockers, calcium channel blockers, bet a-blockers, and diuretics) BMI = Body Mass Index (705 Weight/ Height2) SYSDIFF = Post Systolic blood pre ssure – Pre-systolic blood pressure SYSDIFF1 = Post1 Systolic blood pre ssure – Pre-systolic blood pressure SYSDIFF2 = Post2 Systolic blood pre ssure – Pre-systolic blood pressure DYSDIFF = Post Diastolic blood pressure – Pre-diastolic blood pressure DYSDIFF1 = Post1 Diastolic blood pre ssure – Pre-diastolic blood pressure DYSDIFF2 = Post2 Diastolic blood pre ssure – Pre-diastolic blood pressure

PAGE 100

88 Table 18 Correlations of Antihypertensive Medication, BMI and Bl ood Pressure Difference Scores: 5-Massage Group. AHBP MED BMI SYS DIFF SYS DIFF1 SYS DIFF2 DYS DIFF DYS DIFF1 DYS DIFF2 AHBP MED -.444 .138 -.094 -.181 -.196 -.333 -.451 BMI .004 .163 -.037 .309 .190 .065 SYS DIFF .080 .489 .634* .309 .373 SYS DIFF1 .057 .051 .711** .367 SYS DIFF2 .205 .018 .609* DYS DIFF .502 .644* DYS DIFF1 .626* DYS DIFF2 Pearson’s Correlation signific ant at the .05 level (2-tailed). ** Pearson’s Correlation significant at the .01 level (2-tailed). A-HPN = Antihypertensive Medication (angiotensin conver ting enzyme inhibitors, angiotensin-receptor blockers, calcium channel blockers, bet a-blockers, and diuretics) BMI = Body Mass Index (705 Weight/ Height2) SYSDIFF = Post Systolic blood pre ssure – Pre-systolic blood pressure SYSDIFF1 = Post1 Systolic blood pre ssure – Pre-systolic blood pressure SYSDIFF2 = Post2 Systolic blood pre ssure – Pre-systolic blood pressure DYSDIFF = Post Diastolic blood pressure – Pre-diastolic blood pressure DYSDIFF1 = Post1 Diastolic blood pre ssure – Pre-diastolic blood pressure DYSDIFF2 = Post2 Diastolic blood pre ssure – Pre-diastolic blood pressure

PAGE 101

89 Table 19 Correlations of Antihypertensive Medication, BMI and Blood Pressure Difference Scores: Control Group. AHBP MED BMI SYS DIFF SYS DIFF1 SYS DIFF2 DYS DIFF DYS DIFF1 DYS DIFF2 AHBP MED .689** .074 .400 -.067 .278 .603* .259 BMI .299 .463 .136 .225 .583* .477 SYS DIFF .660* .676** .568* .394 .418 SYS DIFF1 .579* .641* .703** .735** SYS DIFF2 .320 .109 .319 DYS DIFF .763** .753** DYS DIFF1 .755** DYS DIFF2 Pearson’s Correlation signific ant at the .05 level (2-tailed). ** Pearson’s Correlation significant at the .01 level (2-tailed). A-HPN = Antihypertensive Medication (angiotensin conver ting enzyme inhibitors, angiotensin-receptor blockers, calcium channel blockers, bet a-blockers, and diuretics) BMI = Body Mass Index (705 Weight/ Height2) SYSDIFF = Post Systolic blood pre ssure – Pre-systolic blood pressure SYSDIFF1 = Post1 Systolic blood pre ssure – Pre-systolic blood pressure SYSDIFF2 = Post2 Systolic blood pre ssure – Pre-systolic blood pressure DYSDIFF = Post Diastolic blood pressure – Pre-diastolic blood pressure DYSDIFF1 = Post1 Diastolic blood pre ssure – Pre-diastolic blood pressure DYSDIFF2 = Post2 Diastolic blood pr essure – Pre-diastolic blood pressure

PAGE 102

90 and 13 show the systolic blood pressure difference scores plotted against the BMI. Appendix 16 shows the same plots for diastolic blood pressure change GROUP: 1BMI36 34 32 30 28 26 24 22 20SYSDIFF220 10 0 -10 -20 Figure 11: Systolic Blood Pressure Difference and BMI for 10-Massage Group GROUP: 2BMI50 40 30 20SYSDIFF230 20 10 0 -10 -20 Figure 12: Systolic Blood Pressure Difference and BMI for 5-Massage Group

PAGE 103

91 GROUP: 3BMI50 40 30 20SYSDIFF220 10 0 -10 -20 Figure 13: Systolic Blood Pressure Difference and BMI for Control Group scores and BMI. The scatter plots dem onstrate a negative slope for systolic blood pressure in the 10-Massage group (G roup 1), but not t he 5-Masage group (Group 2) and Control group (Group 3). It appears that BMI may be obscuring the relationship of massage and systolic blood pressure. The interpretation might be that for those participants that have a higher BMI the massage routine tended to lower their blood pressure score s and for those persons who were of lower BMI, the massage routine tended to increase their blood pressure. But further analysis is necessary before this or other interpretations can be made. Therefore in Hypothesis 1, analyses were directed to determining the relationship in the context of BMI. With the dias tolic blood pressure, the slopes (Appendix 16) in the scatter plots were similar to t he systolic blood pressure scatter plots. Appropriate analyses follow ed in Hypothesis 2.

PAGE 104

92 Hypothesis Testing Hypothesis 1 The first hypothesis stated that after adjusting for covariates, systolic blood pressure would decrease signific antly over time using back massage treatment in subjects wit h prehypertension and contro lled hypertension. When the data were analyzed using an ANO VA without consider ing any of the covariates, the group x time interactions we re not significant F(6, 117) = 1.38, p = .248. Each covariate with a significant correlation, (p = 05), was placed into the analysis, hypothesizing that the cova riates would decrease error variance and increase power of the analysis. The ANCOVA analysis with BMI as a co variate, assuming homogeneity of regression slopes, yielded the same non-signi ficant F-test. E rror variance was the same for both ANOVA and ANCOVA test s (MS within + residual = 51.88). The heterogeneity of regression slopes method was then used to determine if adjusting for BMI would decrease the erro r variance. The test yielded a nonsignificant group x time interacti on, F(6,108) = 2.08, p = .062. In the context of the hypothesis in wh ich the study design is expecting to find change over time within the 10-Massage group and 5-Massage group and not the Control group and cons idering that the interact ion was nearly significant, a test for simple main effects was employed. The simple main effect of time for 10-Massage group was significant F(3, 108) = 3.50, p = .018. Error variance did decrease (MS within + residual = 49.85). No other significant simple main effects

PAGE 105

93 for time were detected (Table 20). The e ffect size and power of the simple main effects analysis increased from a basic ANOVA with a partial 2 = .005, power = .083 to partial 2 = .089, power = .77 when using the heterogeneity slopes method in ANCOVA. Table 20 Analysis of Covariance of Massage Effect s on Systolic Blood Pressure: A Report of Simple Main Effects after Controlling for BMI Source of Variance SS (adjusted)* df MS F P Value Time within 10Massage Group 532.70 3 174.57 3.50 .018 Time within 5Massage Group 12.04 3 4.01 .08 .970 Time with Control Group 145.73 3 48.58 .97 .408 Error Within Groups 5383.39 108 49.85 Note: SME = simple main effects of ti me. These tests were based on ANCOVA using heterogeneous slopes adjustment of the covariate BMI. Further analysis was needed to understand the significant finding in the 10-Massage group, particularly in view of the group means plotted in Figure 9. The scatter plots, Figures 11, 12, and 13, help to conceptualize the significant finding in the 10-Massage group, as t he negative slope of t he 10-Massage group is visually obvious. The results appear to be obscured by BMI. Therefore a graph was constructed depicting the 10-ma sssage group data when split at the median of BMI (Figure 14). The heavy lin e depicts the entire group (n = 15), and appears to have little, if any, slope over ti me. But when the group is split by the

PAGE 106

94 BMI (Median BMI = 27.48), a variable called “Size”, with “Light” being those with less than a BMI = 27.48 and “Heavy” being t hose with a BMI equal to or greater than 27.48, the two trends appear very diffe rent. Indeed, analysis of BMI x time within the 10-Massage group was significant, F( 3,39) = 5.88, p = .0 02. Further, a trend analysis was conducted to determine if the slopes were significantly different between the Total and the Light and the Heavy groups. The Total group linear trend was not significant, F(1, 13) = .02, p = .888 (see heavy line in Figure 14). The size x time trend analysis found significant linear trends F(1, 13) = 14.08, p = .002. Refer to the upper and lower line on Figure 14. The size x time analysis also yielded nearly significant quadratic effects, F(1, 13) = 4.37, p = .0 57. No significant cubic trends were found. 105 115 125 135 Week 2Week 6 Week 8 Week 10Systolic BLood Pressure Total Group Light Heavy Figure 14: Systolic Blood Pressures over Time, Total Group and Split by BMI: The 10-Massage group To determine the slope of the Light and Heavy trend lines, further trend analyses were conducted. The linear e ffect of time in the Light group was

PAGE 107

95 significant F(1, 13) = 7.12, p = .019, (slope coefficient = 4.567) but the quadratic effects was not significant, (F(1, 13) = .93, p = .353 (s lope coefficient = -2.261). The linear effect of time in the Heavy group was significant F(1, 13) = 6.97, p = .020, (slope coefficient = -4.229) and the quadratic effect was nearly significant, (F(1, 13) = .4.12, p = .069, (slope coefficient = 4.458). There were no significant cubic effects. Figure 15 depicts the trend lines. 105 115 125 135 Week 2Week 6 Week 8 Week 10Systolic BLood Pressure Total Group Light Heavy Figure 15: Systolic Blood Pressures over Time, Total Group and Split by BMI with Trend Lines: The 10-Massage group This hypothesis was partially suppor ted using the heterogeneity of slopes method in ANCOVA (Table 19). The 10-Massage group changed significantly over time whereas the 5-Massage group and the Control group did not change significantly over time in the c ontext of the covariate BMI. Hypothesis 2 The second hypothesis stated that afte r adjusting for covariates, diastolic blood pressure would decrease signific antly over time using back massage

PAGE 108

96 treatment in subjects wit h prehypertension and contro lled hypertension. This analysis was conducted in the same sequenc e as the first hypothesis. The data were analyzed using an ANOVA, without cons idering any of the covariates. The group x time interaction was not signific ant F(6, 117) = .70, p = .648. An ANCOVA analysis with BMI as the covariat e was not significant. Error variance was the same for both tests (MS within + residual = 20.67). The heterogeneity of regression slopes method was then used to determine if adjusting for BMI would decr ease the error variance. The test yielded a non-significant group x time intera ction, F(6,108) = 2. 02, p = .069. In the context of the hypothes is in which the study design is expecting to find change over time with in the 10-Massage group and 5-Massage and not the Control group and considering that the inte raction was nearly significant, a test Table 21 Analysis of Covariance of Massage E ffects on Diastolic Blood Pressure: A Report of Simple Main Effe cts after Controlling for BMI Source of Variance SS (adjusted)* df MS F P Value Time within 10Massage Group 155.61 3 51.87 2.64 .053 Time within 5Massage Group 23.62 3 7.87 .40 .753 Time with Control Group 85.34 3 28.45 1.45 .233 Error Within Groups 2120.57 108 19.63 Note: SME = simple main effects of ti me. These tests were based on ANCOVA using heterogeneous slopes adjustment of the covariate BMI.

PAGE 109

97 for simple main effects was employed. T he simple main effect of time for the 10-Massage group was nearly significant F( 3, 108) = 2.64, p = .053. Error variance did decrease (MS within + residual = 19.63). No other significant simple main effects for time were detected (Table 21). T he effect size and power of the simple main effects analysis increased from a basic ANOVA with a partial 2 = .003, power = .072 to partial 2 = .068, power = .63 when using the heterogeneity of regression slo pes method in ANCOVA. In view of the scatter plots in Table 16, the nearly significant simple main effects, further analysis was conducted to determine if there was a similar obscuring of effects by BMI with diastoli c blood pressure. As in Hypothesis 1, a graph was constructed depicting the 10-ma sssage group data when split at the median of BMI (Figure 16). The heavy lin e depicts the entire group (n = 15), and appears to have little, if any, slope over ti me. But when the group is split by the BMI (Median BMI = 27.48) with “Light” bei ng those with less than a BMI = 27.48 and “Heavy” being those with a BMI equal to or greater than 27. 48, the two lines appear very different. Indeed, analysis of BMI x Time within the 10-Massage group interaction was significant, F(3, 39) = 4.32, p = .010. Further, a trend analysis was conducted to determine if t he slopes were significantly different between the Total and the Light and the Heavy groups. The Total group linear trend was not significant, F(1, 13) = .05, p = .824 (see heavy line in Figure 16). The size x time trend analysis found signifi cant linear trends F(1, 13) = 6.80, p = .022. Refer to the upper and lower line on Figure 16. The size x time analysis

PAGE 110

98 also yielded significant quadrat ic effects, F(1, 13) = 7.01 p = .020. No significant cubic trends were found. 65 70 75 80 85 Week 2Week 6 Week 8 Week 10 WeekDiastolic Blood Pressure Total Group Light Heavy Figure 16: Diastolic Blood Pr essures over Time, Total Group and Split by BMI: 10-Massage group To determine further the slope of t he Light and Heavy trend lines, further trend analyses was conducted. The linea r effects of the Light group were not significant F(1, 13) = 3.77, p = .0749, (slope coefficient = 2.108) but the quadratic 65 70 75 80 85 Week 2Week 6 Week 8 Week 10 WeekDiastolic Blood Pressure Total Group Light Heavy Figure 17: Diastolic Blood Pr essures over Time, Total Group and Split by BMI with Trend lines: 10-Massage group

PAGE 111

99 effects were significant, (F(1, 13) = 4.88, p = .046 (slope coe fficient = -2.857). The linear effects of the Heavy group were not significant F(1, 13) = 3.03, p = .105, (slope coefficient = -1.770 ) and the quadratic effects were not significant. There were no significant c ubic effects. Figure 17 depicts the trend lines. This hypothesis was not supported using the heterogeneity of slopes method in ANCOVA (Table 20). The 10-Ma ssage group did show trends towards effectiveness in that the participants of the Hea vy group had a significant decreasing trend in their diastolic blood pressure over time. Hypothesis 3 Hypothesis 3 stated that after adjusti ng for covariates, there would be a significant difference in systolic blood pr essure changes using 10 applications of back massage versus five applications of back massage in subjects with prehypertension and controlled hypertension. This hypothesis requires a test of interaction between group and time. Contrast statements were designed to compare the effectiveness of the 10-Massage group to the 5-Massage group over time. This group by time interacti on was not significant, F (6, 117) = 1.33, p = .248, partial 2 = .064, power = .505. A detailed trend analysis was conduc ted to detect if there were differences in the linear and quadratic effects over time between the two massage groups. The group x time linear effects was not significant, F(2, 39) = 1.43, p= .252. The quadratic group by time effects were significant, F(2,39) = 3.27, p = .048. Further, linear effects contrasting the 5-Massge group

PAGE 112

100 to the 10-Massage group were not signific ant F(1, 39) = .47, p = .498, but the quadratic effect for this contrast was sign ificant F(1, 39) = 4.82, p = .034. This indicates that the shape of the tre nd lines for the 10-Massage group and the 5-Massage group are not the same. Hypothesis 4 Hypothesis 4 stated that after adjusti ng for covariates, there would be a significant difference in diastolic blood pr essure changes using 10 applications of back massage versus five applications of back massage in subjects with prehypertension and controlled hypertension. This hypothesis requires a test of interaction between group and time. Contrast statements were designed to compare the effectiveness of the 10-Massage group to the 5-Massage group over time. The group x time interaction was not significant, F (6, 117) = .70, p = .648, partial 2 = .0354, power = .270. A detailed trend analysis was perfo rmed to detect if there were differences in the linear and quadratic effects over time between the two massage groups. The group by time linear effect was not significant, F(2, 39) = .51, p = .603. The quadratic group by time e ffect was not significant, F(2,39) = 1.45, p = .246. Hypothesis 5 Hypothesis 5 stated that t here would be a significant change over time in reclining blood pressures compared to sitti ng blood pressures in the subjects with prehypertension and controlled hypertens ion in the back massage treatment

PAGE 113

101 Table 22 Laying and Sitting Blood Pressure Group Means (Standard Deviations) Variable at 4 Time Points* Group1 (n = 15) Group2 (n = 13) Control (n = 14) Total Group (n = 42) Group differences p-values 1 Lying SBP 118.60 (10.97) 122.08 (12.81) 121.79 (10. 92) 120.74 (11.38) .672 1 Sitting SBP 120.47 (10.90) 119.54 (13. 48) 127.79 (10.16) 122.62 (11.84) .947 2 Lying SBP 117.27 (16.02) 118.62 (14.83) 119.00 (13. 43) 118.26 (14.49) .935 2 Sitting SBP 120.27 (18.38) 121.92 (11. 42) 121.57 (9.70) 121.21 (13.56) .916 3 Lying SBP 118.67 (12.00) 120.23 (18.90) 118.21 (13. 57) 119.00 (14.59) .132 3 Sitting SBP 120.93 (11.25) 126.15 (17. 71) 121.57 (11.37) 122.76 (13.47) .945 4 Lying SBP 118.40 (14.19) 118.23 (13.06) 120.21 (14.06) 118.95 (13.5) .557 4 Sitting SBP 122.33 (14.59) 118.38 (11. 92) 122.14 (14.71) 121.05 (13.65) .708 1 Lying DBP 72.33 (6.31) 72.85 (6. 39) 70.93 (10.66) 72.02 (7.89) .812 1 Sitting DBP 78.87 (6.97) 75.62 (6.08) 77.50 (11.74) 77.40 (8.54) .688 2 Lying DBP 71.73 (9.44) 72.38 (5.04) 69.86 (8.25) 71.31 (7.79) .469 2 Sitting DBP 75.73 (9.05) 76.77 (6.77) 76.29 (9.44) 76.24 (8.36) .974 3 Lying DBP 73.80 (8.32) 71.46 (11. 18) 69.79 (6.29) 71.74 (8.70) .614 3 Sitting DBP 78.40 (9.82) 74.15 (10.79) 76.86 (7.78) 76.57 (9.45) .950 4 Lying DBP 71.60 (8.13) 71.23 (7.38) 70.93 (8.19) 71.26 (7.7) .502 4 Sitting DBP 75.47 (8.70) 75.08 (9.49) 76.21 (8.27) 75.60 (8.61) .943 *1 = pre randomization 2 = first post measure (at w eek 4 for the 5-Massage group and Week 6 for the 10-Massage group) 3= second post measure at week 8 4 = third post measure at week 10.

PAGE 114

102 Systolic lying & sitting blood pressure changes over time Figure 18: 10-Massage Group Lying & Sitting Systolic Blood Pressure Relationship Figure 19: 5-Massage Group Lying & Si tting Systolic Blood Pressure Relationship 90 100 110 120 130 140 150 2 4 6 8 10 WEEK10-MASSAGE GROUP SBP LYING SITTING 100 105 110 115 120 125 130 135 140 145 150 2 4 6 810 WEEK5-MASSAGE GROUP SBP LYING SITTING

PAGE 115

103 Figure 20: Control Group Lying & Sitting S ystolic Blood Pressure Relationship Diastolic lying & sitting blood pressure changes over time. Figure 21: 10-Massage Group Lying & Si tting Diastolic Blood Pressure Relationship 100 105 110 115 120 125 130 135 140 2 4 6 8 10 WEEKCONTROL GROUP SBP LYING SITTING 60 65 70 75 80 85 90 2 46 8 10 TIME10-MASSAGE GROUP DBP LYING SITTING

PAGE 116

104 Figure 22: 5-Massage Group Lying & Sitting Diastolic Blood Pressure Relationship Figure 23: Control Group Lying & Sitting Di astolic Blood Pressure Relationship 60 65 70 75 80 85 90 24 6 8 10 TIME5-MASSAGE GROUP DBP LYING SITTING 50 55 60 65 70 75 80 85 90 95 2 4 68 10 TIMECONTROL GROUP DBP LYING SITTING

PAGE 117

105 group of 10 applications when compared to the control group. The group means for systolic and diastolic blood pressures, lying and sitting, are shown in Table 20. There were no significant differenc es between groups at any time point. Further, a visual depiction of each group’ s systolic and diastolic blood pressures, sitting and lying, with standard deviation bar s are further depicted in Figures 18 through 23. Correlation matrices were calc ulated to view the strength of the relationship between the lying to sitting chan ges and two covariates in particular. The two covariates, considered for s eparate reasons. BMI was considered because of its previously established relationship in the previous analysis. Antihypertensive medications (including di uretics) were considered because of their effect on blood pressure. Medi cations can often cause orthostatic hypotension. Variables depicting the mean change in blood pressure at each time point were established. The matrices are depicted in Appendix 17. As seen in the matrices, when considering gr oups together, there is a significant correlation between diastolic blood pressure at time 1 (BARO5) and antihypertensive medications (r = .358) Further, the 5-Massage group was the group that reflected this significant corre lation with diastolic blood pressure lying to sitting change with medications (r = .598). The analysis therefore was run using the covariate antihypertensive medications. Analysis of variance was conducted to determine the effect of time on lying to sitting differences in systolic bl ood pressure for each group. There were no significant findings over time. A nalysis of variance was conducted to

PAGE 118

106 determine the effect of position by time on lying to sitting differences in systolic blood pressure for each group. Positi on X time appears to make a difference with a significant interaction in the 5 -Massage group, F (3,117) = 2.77, p = .045, partial 2 = .066, power = .657 (Table 23). Fi gure 19 depicts the interaction, thus the significant change, in lying to si tting blood pressure over time for the 5-Massage group. No other significant effects were found in systolic blood pressure lying to sitting differences. Analyses of covariance using BMI and antihypertensive medications as covariates were conducted. Neither medications nor BMI changed the error vari ance of the analysis. Table 23 Lying to Sitting Change in Systolic Bl ood Pressure: Simple Main Effects of Position by Time within Group Source of Variance SS df MS F P Value Position X Time within 10-Massage Group 18.57 3 6.19 .19 .900 Position X Time within 5-Massage Group 265.03 3 88.34 2.77 .045 Position X Time within 10-Massage Group 67.18 3 22.39 .70 .552 Error Within Group 3727.35 117 31.86 Analysis of variance was conducted to determine the lying to sitting differences in diastolic blood pressure fo r each group. A signi ficant effect for position was found in all three groups (T able 24). Figures 21, 22, and 23 depict that significant gap between the lying and sitting positions at each time point,

PAGE 119

107 visually supporting this finding. No signif icant effects for group or group by time by position were found with the diastolic blood pressure comparisons. Analyses of covariance using BMI and antihypertensive medications as covariates were conducted. Neither covariates changed the error variance of the analysis. Table 24 Lying to Sitting Change in Diastolic Bl ood Pressure: Simple Main Effects of Position within Group Source of Variance SS df MS F P Value Position within 10Massage Group 676.88 1 676.88 14.67 .000 Position within 5Massage Group 304.65 1 304.65 6.06 .014 Position within Control Group 1125.22 1 1125.22 24.39 .000 Error Within Groups 1799.50 39 46.13 Overall, the analyses do not suppor t the hypothesis of significant differences between the 10-massage group and the control group for baroreceptor response changes over time. Additional Analyse s of Interest Expectations The expectation question, a dichotomous response of “yes” or “no”, did not appear to have a relationship with t he change in systolic blood pressure changes, t (36) = .988, p = .330, or diastolic blood pressure changes t (36) = -.336, p = .717.

PAGE 120

108 Chapter Five Discussion Nursing has a long history using ma ssage, particularly back massage, as a comfort measure for thei r patients. The knowl edge generated regarding the efficacy of massage comes mostly from t he nursing literature. However, today massage is usually considered an alternat ive or complementary therapy rather than a conventional nursing therapy. T he general population ha s embraced it as one of the many alternative non-pharmace utical treatments for a number of disease processes, including heart dis ease and hypertension. The efficacy of massage for treatment of persons with pr e-hypertension and hypertension is not yet established. Therefore, with the safe ty of the public in mind, the current study proposed to determine if there are long term effects of massage. The study also compared two dosages of massage for their impact on systolic and diastolic blood pressure and tested fo r change in baroreceptor response in persons receiving massage treatments. The study used a 3-group randomized clin ical trial design. The sample consisted of 42 adults, (62% female, 86% Caucasian) predominantly from a southwest central United States universit y setting. The mean age for the total sample was 48.7 years (range 26 to 70 years) and the mean length of time reported to be living with hy pertension was 6.1years. The participants in this study exercised, on average, 24 minutes per day. Exercise included walking,

PAGE 121

109 jogging, weight lifting, and aerobic sports The study sample was obese with a BMI greater than 30, on average. More than half (57%) of the study sample drank alcohol at least occasionally. Most (90.5%) did not smok e cigarettes at the time of the study. Two subjects, older than 70 years of age, were screened for mental capacity (MMSE), each scoring the maximum (30) on the exam. A majority of the participants believed that participating in this study would impact their blood pressure. Considering the positive expectation, the voluntary nature and willingness of parti cipation, adherence to the study protocol remained tasking for the participants. The study protocol requested participants to visit the research office for 30 mi nutes, three times per week totaling 13 or 18 visits, depending on the group to which the partici pant was randomized. Although only twelve of the 355 treatment (massage and relaxation) sessions were missed and not rescheduled, compliance to the schedule was a challenge for most participants. The average time between vi sits was greater than the planned 2.3 days. Even though it was difficult for par ticipants to adhere to the regularity of three times per week visits, most were abl e to stick to the daily allotted 30 minute time slot for their treatments. For ex ample if participant #1 was scheduled for 3 pm, she came on time at 3 pm each visit. Rarely did the treatment times deviate from the schedule. To achieve a sense of long-term effe cts of changes in blood pressure due to the back massage, the study was se t up so that blood pressures were measured four times during the study: pr ior to randomization to group (week 2), at the completion of the treatment se ssions (week 4 for the 5-Massage group and

PAGE 122

110 week 6 for the 10-Massage group and the C ontrol group), again at week 8 of the study and again at week 10 of the study. The second measurement was planned to take place 48 hours after the last treatme nt. In reality, this appointment in which the initial effects of massage on bl ood pressure were to be measured was more than 48 hours but less than 72 hours on average after the last treatment. In respect to timeliness, however, all participants were able to be present within 30 minutes of the scheduled time fo r the four data gather ing appointments. This was important for the fi delity of the study in that the variables with circadian cycles could be comparable over time. Although a power analysis was done a priori based on effect sizes of a previous study, (Olney, 2005) the analysis showed no significant main effects. However, after adjusting for the BMI heterogeneity of regression slope, the analysis yielded significant simple main effects in the 10-massage group over time. This analysis suggested that givi ng regular ten-minute back massages over approximately 3.5 weeks, is related to a significant decrease in systolic blood pressure, but is limited to persons with a BMI greater than 27.48 kg/m2. The long term effect of the back massages rout ine was four weeks after the back massages were discontinued. The diastolic blood pressure analysis i ndicated no significant main effects. Because of the influence of body mass i ndex on systolic blood pressure in the 10-Massage group, a follow up trend analys is was conducted and did reveal significant diastolic blood pressure sl ope differences between the heavier (BMI

PAGE 123

111 equal to or greater than 27.48 kg/m2) and the lighter (BMI less than 27.48 kg/m2) participants within the 10-Massage group. This is the first study known to t he principle investigator designed to analyze long term effects of back massage. The liter ature has suggested that one back massage can elicit short term (usually measured within 15 minutes after end of massage) decreases in both systolic and diastolic blood pressure (e. g. Cady, & Jones, 1997; McNamara, Bu rnham, Smith & Carroll 2003). In the majority of massage studies, the blood pres sure is measured just prior to the massage and just after the massage. In other reported studies, when the massage is repeated for several days the significant effects were calculated with each massage session (Aourell, Skoog, & Ca rleson, 2005; Ferrell-Torry & Glick, 1993; Holland & Pokorny, 2001). He rnandez-Reif and colleagues (2000) have applied massage for 30 minutes, three times a week and this resulted in diastolic blood pressure changes over time. Mok and Woo (2004) found significant systolic and diastolic changes, applying a 10-minute back massage for seven evenings and measuring pre and post interv ention at day 1, day 7 and three days after stopping the application. Although it is unclear which blood pressure measures (pre-intervention or post intervent ion or all) they used in their repeated measures analysis, this is the best compar ison for the current study, supporting changes over time. Unfortunately, Mok and Woo’s study did not discuss BMI. Mok and Woo did measure for effects of age and gender, finding none. The Moyer, Rounds and Hannum (2004) meta-anal ysis suggested that testing of the long term effects as one of the necessary questions to further the science.

PAGE 124

112 Summarizing, the results of the current st udy imply that there may be beneficial effects on systolic and diastolic blood pressu re in persons of higher BMI (greater than or equal to 27.48 kg/m2). The randomization process appears to have equalized all of the potential covariates other than BMI. This study sample was small (n=42). Perhaps a larger, randomized sample would equalize the BMI thus allo wing for a better sense of the weight issue. A larger samp le would also decrease the possibility of a Type I error. Dosage of massage is another recomm ended area of research suggested by numerous studies including the Moye r and colleague meta-analysis (Moyer, Rounds, & Hannum, 2004). The current st udy evaluated if giving 5 back massages could be as effective as 10 back massages for lowering systolic and diastolic blood pressure. However, in t he current study main effect of group by time interaction for dosage was not si gnificant thus recommendations for an appropriate prescription can not be made at this time. Baroreceptor response can be m easured by comparing the lying blood pressure to the sitting upright blood pressu re. In this study, systolic and diastolic blood pressure were measured in the ly ing position and then again immediately upon sitting upright. The current study attempted to establish baroreceptor response as the mechanism that c ould explain why back massage may have influenced the long term blood pressure. It was hypothesized that by inducing a therapeutic level of relaxation (repeated experiences of lo wered blood pressure) the baroreceptor firing

PAGE 125

113 threshold would be re-set to a healthier leve l. This resetting to a healthier level would be manifested by a narrower gap between lying and sitting blood pressures. When viewing the position dependent systolic pressure changes over time for each group in Figures 18, 19 and 20, the group that most closely resembles the expectation is the contro l group (Figure 20). When viewing the position dependent diastolic pressure c hanges over time for each group in Figures 21, 22, and 23, the gr oup that most closely rese mbles the expectation is the 10-massage group (Figure 21). The analysis determined the changes were not significant. Additional Discussion The intimacy of massage demands that researchers begin to measure the relationship between the massage therapist and the one receiving the massage. This study did measure two concepts to further the understanding of touch perception and perceived rapport for those experiencing the massage. However, the therapist was constant across all participants in this study therefore the findings are not generalizable to other therapists-patient relationships. To gain a sense of what the parti cipant may be experiencing when receiving the back massage, a Subject T ouch Perception Scale was developed. This semantic differential-like scale listed opposing pairs of each characteristic of the massage. The reliability test on the six items used to describe the perception of touch was very poor. Removal of a ll items but the ‘sm ooth’ and ‘relaxing’ items from the reli abilities test yielded a higher alpha = .5722. Thus it appears that this instrument needs further develop ment. Keeping the reliabilities issue in

PAGE 126

114 mind, for the purposes of beginning the discussion on perception of touch the following summary of the data is present ed. Overall, the massage was perceived as soothing, pleasant, smooth, rhythmic and relaxing. All but the pressure items were scaled such that a score of 1 was perceived as a more positive experience. Mean group scores ranged from 1.00 (examp le: rhythmic) to 1.27 (example: relaxing) indicating that the perception of the massage touch on these five items was a positive experience for the participants. The pres sure-item on this scale interpreted differently from the other five items. Pressure was scaled from too soft (1.00) to too hard (5.00) Thus the resulting score of 3.00 may be interpreted by the participant as “just right”. Both groups felt the pressure of the massage was a bit on the soft side with the mean scores for the 10-massage group (2.53) and 5-massage group (2.69). These data are supported by the verbal requests by several participants to use more pressure. Due to the nature of the study, the researcher explained to the participant the same pressure must be used on all subjects. The implication here may be that response is determined by pressure. For instance, maybe some participants needed a deeper pressure to achieve a “relaxation response”. This is clearly an area for future research. Beyond the touch perception results, anecdotally, there were several comments made by participant s about the sound of the massage. Usually there was little sound elicited by the massage st rokes using oil. However, on several participants, the massage strokes elicit ed a loud friction sound when the back was covered by body hair or a suction sound when the back was hair free. In

PAGE 127

115 these particular cases the participants stat ed that the sound pr evented them from truly relaxing. Touch perception is a phenomen on which has been measured and reported in populations of abused persons. However, measuring the perception of touch in massage research is relatively new. This researcher did not find a developed instrument that measured touch perception from a non-abusive framework. Therefore, t he development of an instrum ent would be helpful to massage researchers when attempting to determine the effects of massage. The experienced relationship with the ma ssage therapist is of interest to researchers who evaluate the effects of massage (Moyer, Hanum, & Rounds, 2004). To gain a sense of the partici pant’s perceived relationship with the massage therapist a perceived rapport scale was developed. This semantic differential-like scale listed 10 opposing ch aracteristics of what the participant may have been experiencing. A score of 1 related to a more positive experience whereas a score of 5 related to a more negative experience. Scores for the two groups ranged from 1.00 (example: safe) to 1.53 (example: physically relaxed). The two groups were not different in t heir perception and the participants felt positive about their relationship with the massage therapist. This information is not as useful as the researcher had hypothesized. The variance was insignificant to determine if there is any influence on the variables being measured in this study. A larger scale, with 10 rather than 5 levels of response may have generated a greater c ontrast among responses. Content and concurrent validity would add to its in terpretability. Also measuring the

PAGE 128

116 control group for their interpretation of the relationship may have added to this analysis. Theoretically, development a va lid instrument w ould be beneficial in future massage research, as it would help explain more of the variance found in the research analysis. Study Limitations This study was underpowered for the analyses of diastolic pressure and baroreceptor. The original power analysis was based upon the researcher’s pilot work (Olney, 2005). A larger sample wo uld increase the power and better reflect dosage comparison. This study was conducted on a universit y campus which is not necessarily reflective of the general hypertensive populat ion. Therefore caution is advised to not generalize these findings beyond like populations. The participants’ characteristics were se lf-reported. Items such as weight and height, minutes per day of exercise years on diagnosed with hypertension and years on medications were subject to the participants ability to recall correctly and report accurately the request ed data. Therefore it is possible that the self-reported data lack full and accurate disclosure. Inaccurate reporting of the characteristics such as height and wei ght, which formulate the BMI covariate, could lead to analysis that does not reflect reality. Another important limitation in this study is that the research assistant left the study during the second month. This cr eated several issues for study fidelity. First, when the researcher took over the measurement sessions, she became unblinded to the data. It is possible that exposure to the data caused bias in the

PAGE 129

117 interactions of the researcher towards the participants. A nother possible issue with the research assistant leaving the study was that 14 participants had their pre-randomization data gathered by the re search assistant. Only 9 of those participants had their first post measur e and none had third and forth measures conducted by the research assistant. Th is inconsistency of data gatherer across time may have impacted the expectation of some participants, thus outcomes. Implications for Practice There are implications for practice resulting from this study. The significant findings of 10-back massage treat ment imply that t here may be lasting treatment effects for up to 4 weeks for systolic and diastolic blood pressure for persons with elevated body mass index. The linear and quadrat ic appearance of the analysis indicates there maybe an immediate treatment effect and a sustained effect out to a 4 week period of time. This study did not indicate the duration of effects beyond 4 weeks post applic ation. It is most likely that the effectiveness weakens the further away from the last tr eatment, based upon the psychophysiology theory. The data are not clear enough to make implications regarding dosage. Further re search is needed before definitive recommendations can be made as to exac t dosage and length of effects. This study does indicate that systolic and diasto lic blood pressure levels did decrease with a regularly applied 10-minute back massage for persons who had a higher body mass index. Back massage is a safe treatment for most persons. Recommending back massage as an adjunct t herapy to drug therapy for persons with pre-hypertension or hy pertension would certainly cause no harm in those

PAGE 130

118 with body mass index above 27.43 kg/m2 provided the person is otherwise healthy. But recommending the 10minut e back massage for those with lower body mass index should be done with caution. It would not be safe to recommend ba ck massage as a sole treatment for persons with pre-hypertension or hypert ension. Treatment of elevated blood pressure needs to be multifaceted, to in clude lifestyle modifications including weight maintenance at a healthy level, exercise, diet management, and stress reduction. Once the diagnosis of hypert ension is established, anti-hypertension medications are necessary to help lowe r and maintain a healthy blood pressure level. At this point, the evidence only supports back massage to be used as an adjunct therapy. Future Research Recommendations This study was based on psychoph ysiology theory and the design built upon previously published research and recommendations. The aim of this study, to determine long-term effe cts and dosage for person with prehypertension or hypertension, was clearly the next step in knowledge building. This study design needs to be expanded to a larger more heterogeneous population. This study did not analyze gender effects. However comparing gender effects would be interesting par ticularly because there are studies reporting differences among gender Also, race effects must be tested, in view of the disparate onset of hyper tension and heart disease in minority populations. It is possible due to cultural differences that massage would have a different effect.

PAGE 131

119 Regarding methodological recommendations first, blinding the researcher to the data to prevent possi ble bias is necessary. Biological responses to massage need to be measured in future studies. Biological data in particular can add to the knowledge about t he physiological response to such interventions. Psychological measurement should continue to be a part of future studies both as in this study as covariates, or as out come variables. The relationship with the therapist needs to be further explored. Establishment of validity for a touch perception instrument and rapport perception instrument would enhance measurement of these variables.

PAGE 132

120 References Ahles, T. A., Tope, D. M., Pinkson, B., Walch, S., H ann, D., Whedon, M., et al. (1999). Massage therapy for pati ents undergoing autologous bone marrow transplantation. Journal of Pain and Symptom Management. 18( 3), 157-163. Altshul, S. (2006, August). Massage your mood: a friendly touch eases depression and anxiety. Prevention Magazine ,100. American Heart Association (2006). High Blood Pressure Statistics. Retrieved November 26, 2006, from http://www.americanheart.org/pre senter.jhtml?identifier=4621 Aourell, M., Skoog, M., & Carleson, J. (2005). Effects of Swedish massage on blood pressure. Complementary Therapies in Clinical Practice, 11 243246. Barnes, P. M., Powell-Griner, E., MeFann, K., & Nahin, R. L. (May, 2004). Complementary and alternative medi cine use among adults: United States, 2002. Advance Data from Vital & Health Statistics, 343 1-19. Retrieved January 3, 2007 from http://www.cdc.gov/nchs/ products/pubs/pubd/ad/ad.htm Barr, J. S., & Taslitz, N. (1970). The influence of back massage on autonomic functions. Physical Therapy, 50 (12), 1679-1691. Bauer, W. C., & Dracup, K. A. (1987). Physiological e ffects of back massage in patients with acute myocardial infarction. Focus on Critical Care, 14 (6), 4246. Beck, M. (1999). Milady’s Theory and practice of therapeutic massage Albany, NY: Milady. Beeken, J. E., Parks, D., Cory, J., & M ontopoli, G. (1998). The effectiveness of neuromuscular release massage therapy in five individuals with chronic obstructive lung disease. Clinical Nursing Research, 7 (3), 309-317. Benjamin, P. J., and T appan, F. M. (2005). Tappan’s handbook of healing massage techniques (4th ed.). Upper Saddle River, NJ: Pearson Prentice Hall. Benson, (2000). The Relaxation Response. New York: First WholeCare.

PAGE 133

121 Benson, H., Beary, J. F., & Carol, M. (1974). The relaxation response. Psychiatry, 37 37-46. Bost, N. & Wallis, N. (2006). The effe ctiveness of a 15 minute weekly massage in reducing physical and psyc hological stress in nurses. Australian Journal of Advanced Nursing, 23 (4), 28-33. Burns, N., & Groves, S. K. (2005). The practice of nursing research: Conduct, critique, and utilization (5th ed.). St. Louis, MO: Elsevier Saunders. Buss, I. C., Halfens, R. J., & Abu-Saad, H. H. (1997). The effectiveness of massage in preventing pressure sores: A literature review. Rehabilitative Nursing, 22 (5), 229-234. Cady, S. H., & Jones G. E. (1997). Mass age therapy as a workplace intervention for reduction of stress. Perceptual and Motor Skills, 84 157-158. Cawley, N. (1997). A criti que of the methodology of research studies evaluating massage. European Journal of Cancer Care, 6 23-31. Cambron, J. A., Dexheimer, J., & Coe, P. (2006). Changes in blood pressure after various forms of therapeutic massage: A preliminary study. The Journal of Alternative and Complementary Medicine, 12 (1), 65-70. Chalmers, (1999). WHO-ISH Guidelines for the Management of Hypertension. Medical Journal of Australia, 171 (9), 458-459. Charmandari, E., Kino, T., Souvatzoglou, E., & Chrousos G. P. (2003). Pediatric stress: Hormonal mediators and human development. Hormone Research 59 (4), 161-79. Chobanian, A. V., Bakris, G. L., Black, H. R., Cushman, W. C., Green, L. A., Izzo, J. L., et al. (2003). The seventh r eport of the joint national committee on prevention, detection evaluation, and tr eatment of high blood pressure: The JNC 7 report. Journal of the American Medical Association, 289 (19), 25602572. Corley, M. C., Ferriter, J., Zeh, J., & Gifford, C. (1995). Physiological and psychological effects of back rubs. Applied Nursing Research, 8 (1), 39-43. Cushman, W. C. (2000). Alcohol consumption and the management of hypertension. In S. O paril & M. Weber (Eds.), Hypertension: A companion to Brenner and Rector's the kidney (pp. 466-470). New York: W. B. Saunders Company.

PAGE 134

122 Delaney, J. P., Leong, K.S., Watkins, A., & Brodie, D. (2002). The short-term effects of myofascial trigger point massage therapy on cardiac autonomic tone in healthy subjects. Journal of Advanced Nursing, 37 (4), 364-371. Delgado, C., & Weder, A. (2000). Pathophysiology of hy pertension. In S. Oparil & M. A. Weber (Eds.), Hypertension: A companion to Brenner and Rector’s the kidney (pp. 21-28). New York: W. B. Saunders Company. DiBona, G. F. (2004). The sympathetic nervous system and hypertension: Recent developments. Hypertension, 43 147-150. Dickerson, C. J. (2004). Commentary on Hot Topic Review: Baroreceptors and the long-term control of blood pressure. Experimental Physiology, 89 (4), 335-337. Dossey, B. M., Keegan, G., Guzzetta, C. E. (1999). Holistic Nursing: A handbook for practice (3rd ed.). Gaithersburg, MD: Aspen Publishers. Duimel-Peeters, I. G., Halfens, R. J., Be rger, M. P., Snoeckx, L. H. (2005). The effects of massage as a method to prevent pressure ulcers: A review of the literature. Ostomy Wound Management, 51 (4), 70-80. Dunn, C., Sleep, J., & Collett, D. (1995). Sensing an improvement: An experimental study to evaluate the use of aromatherapy, massage and periods of rest in an intensive care unit. Journal of Advanced Nursing 21, 34-40. Dyson, R. (1978). Bed sore s the injuries hospital st aff inflict on patients. Nursing Mirror 146 (42), 30-32. Eisenberg, D. M., Kessler, R. C., Foster, C., Norlack, F. E., Calkins, D. R., & Delbanco, T. L. (1993). Unconventional medicine in the United States. Prevelance, costs and patterns of use. The New England Journal of Medicine, 326, 246-252. Fakouri, C., & Jones, P. (1987). Rela xation Rx: Slow stroke back rub. Journal of Gerontological Nursing, 13 (2), 32-35. Felhandler, D., & Lisander, B. (1999). Ef fects of non-invasive stimulation of acupoints on the cardiovascular system. Complementary Therapies in Medicine 7 231-234. Ferrell-Torry, A. T. & Glick, O. J. (1993) The use of therapeutic massage as a nursing intervention to modify anxiety and the perception of cancer pain. Cancer Nursing, 16 (2), 93-101.

PAGE 135

123 Field, T. Grizzle, N., Scafidi, & Schanber g, S. (1996). Massage and relaxation therapies’ effects on depre ssed adolescent mothers. Adolescence, 31 (124), 903-911. Field, T., Quintino, O., H enteleff, T., Wells-Keife, L. Delvecchio-Feinberg, G. (1997). Job stress reduction therapies. Alternative Therapies in Health and Medicine, 3 (4), 54-56. Field, T., Hernandez-Reif, M., Diego, M. Schanberg, S., & Kuhn, C. (2005). Cortisol deceases and serotonin and dopamine increase following massage therapy. International Journal of Neuroscience, 115 (10), 1397-1413. Fraser, J. & Kerr, J. R. (1993). Psycho physiological effects of back massage on elderly institutionalized patients. Journal of Advanced Nursing, 18 238-245. GE Healthcare (2007). Pro 400: Gold standard BP, Temp, SpO2 and Pulse. Retrieved January 5, 2007 from http://www.gehealthcare.com/euen/ patient_monitoring/products/immmonitoring/dinamapp/pro400.html Goldstone, L. A. (2000). Massage as an orthodox medical treatment past and future. Complementary Therapies in Nursing & Midwifery, 6 169-175. Goodfellow, L. (2003). The effe cts of therapeutic back massage on psychophysiologic variables and immune function in spouses of patients with cancer. Nursing Research, 52 (5), 318-328. Groer, M., Mozingo, J., Droppleman, P., Davi s, M., Jolly, M. L., Boynton, M., et al. (1994). Measures of salivary se cretory immunoglobulin A and state anxiety after a nursing back rub. Applied Nursing Research, 7(1), 2-6. Grossman, E., Grossman A., Schein, M.H., Zimlichman, R., & Gavish, B. (2001). Breathing-control lowers blood pressure. Journal of Human Hypertension, 15 263-269 Hayes, J. & Cox, C. (2000). Immediat e effects of a five-minute foot massage on patients in critical care. Complementary Therapies in Nursing & Midwifery, 6, 9-13. He, F. J., & MacGregor, G. A. (2003). Potassium: More beneficial effects. Climacteric, 6 (Supplement 3), 36-48. Hernandez-Reif, M., Dieter, J., Field, T ., Swerdlow, B., & Diego, M. (1998). Migraine headaches are reduced by massage therapy. International Journal of Neuroscience 96, 1-11.

PAGE 136

124 Hernandez-Reif, M., Field, T., Krasnegor J., Theakston, H., Hossain, Z., Burman, I. (2000). High blood pressu re and associated symptoms were reduced by massage therapy. Journal of Bodywork and Movement Therapies, 4 (1), 31-38. Hernandez-Reif, M., Ironson, G., Field, T., Hurley, J., Ka tz, G., Diego, M., et al. (2004). Breast cancer patients have improved immune and neuroendocrine functions following massage therapy. Journal of Psychosomatic Research, 57 (1), 45-52. Herrera, A. H., & Lowenthal, D. L. (2000). Exercise and hypertension. In S. Oparil & M. Weber (Eds.), Hypertension: A companion to Brenner and Rector's the kidney (pp. 470-478). New York: W. B. Saunders Company. Hobbs, S., and Davies P. D. (1998). Critical review of how nurses research massage therapy: Are they using the best methods? Complementary Therapies in Nursing & Midwifery, 4 35-40. Holland, B., & Pokorny, M. E. (2001). Slow stroke back mass age: Its effect on patients in a rehabilitation setting. Rehabilitation Nursing, 26 (5), 182-186. Ironson, G., Field, T., Scafidi, F., Hashim oto, M., Kumar, A., Price, A., et al. (1996). Massage therapy is associat ed with enhancement of the immune system’s cytotoxic capacity. International Journal of Neuroscience, 84 (1-4), 205-217. Izzo, J. L. (2000). Sympathetic nerv ous system in acute and chronic blood pressure elevation. In S. Oparil & M. A. Weber (Eds.), Hypertension: A companion to Brenner and Rector ’s the kidney, (pp. 42-58). New York: W. B. Saunders Company. Jacobson (1938). Progressive relaxation Chicago: Chicago University Press. Kaplan, N. M. (2002). Kaplan’s clinical hypertension (8th ed.). Philadelphia: Lippincott Williams & Wilkins. Kessler, R. C., Davis, R. B ., Foster, D. F., Van Rompay M. I., Walters, E. E., Wilkey, et al. (2001). Long-term trends in the use of complementary and alternative medical therapi es in the United States. Annuls of Internal Medicine, 135 (4), 262-268. Kurashova-Wine, Z. (1999). Massage for hypertension. Massage Magazine, 67, 67-68. Labyak, S. E., & Metzger, B. L. (1997). The effects of effleurage back rub on the physiological components of re laxation: a meta-analysis. Nursing Research, 46 (1), 59-62.

PAGE 137

125 Li, J. J., & Chen, J. L. (2005). Inflammation may be a bridge connecting hypertension and atherosclerosis Medical Hypotheses, 64 (5), 925-929. Light, K. C. (2000). Environmental and p sychosocial stress in hypertension onset and progression. In S. Opar il & M. A. Weber (Eds.), Hypertension: A companion to Brenner and Rector’s the kidney, (pp. 59–70). New York: W. B. Saunders Company. Longworth, J. C. D. (1982). Psychophysiological effect s of slow stroke back massage in normotensive females. Advances in Nursing Science, 4 (4), 4461. Lovello, W. R. (2005). Stress & health: Biological an d psychological interactions (2nd ed.). London: Sage Publications. McCowen, K. C., Chan, S., & Blackburn, G. L. (2000). Diet-calo ries. In S. Oparil & M. Weber (Eds.), Hypertension: A companion to Brenner and Rector's the kidney (pp. 460-465). New York: W. B. Saunders Company. McNamara, M. E., Burnham, D. C., Smith, C. & Carroll, D. L. ( 2003). The effects of back massage before diagnostic cardiac catheterization. Alternative Therapies in Health and Medicine, 9 (1), 50-57. McRee, L. D., Noble, S., & Pasvogel, A. (2003). Using massage and music therapy to improve postoperative outcomes. AORN Journal, 78 (3), 433447. Meek, S. S. (1993). Effects of slow stro ke back massage on relaxation in hospice clients. IMAGE: Journal of Nursing Scholarship, 25 (1), 17-21. Meintz, S. L. (1995). Whatev er became of the backrub? RN, (April, 1995), 49-56 Mitzel-Wilkinson, A. (2000). Mass age therapy as a nursing practice. Journal of Holistic Nursing 14 (2), 48-56. Mok, E. & Woo, C. P. (2004). The effects of slow-stroke back massage on anxiety and shoulder pain in elderly stroke patients. Complementary Therapies in Nursing & Midwifery, 10 209-216. Moyer, C. A., Rounds, J., & Hannum, J. W. (2004). A meta-analysis of massage therapy research. Psychological Bulletin, 130 (1), 3-18. Mueller Hinze, M. L. (1988) The effects of therapeut ic touch and acupressure on experimentally-induced pain (Doctoral di ssertation, University of Texas at Austin, 1988). Dissertation Abstracts International, 49 4755.

PAGE 138

126 Munro, B. H. (2001). Statistical methods for health care research (4th ed.). New York: Lippincott. National Association of Nurse Massage Therapists (2007). NANMT philosophy Retrieved December 2, 2007, from http://www.nanmt.org/philosophy.html Ni, H., Simile, C., & Hardy, A. M. (2002). Utilization of complementary and alternative medicine by United States adults: results form the 1999 national health interview survey. Medical Care, 40 (4), 353-358. Okvat, H. A., Oz, M. C., Ting. W., & Namerow, P. B. (2002). Massage therapy for patients undergoing cardiac catheterization. Alternative Therapies, 8 (3), 68-75. Olney, C. M. (2005). The effect of therapeutic ba ck massage in hypertensive persons: A preliminary study. Biological Research for Nursing, 7 (2), 98-105. Onken, D. A., White, W. B., Cooney, J. L., et al. (2001). Impact of Smoking cessation on ambulatory blood pressu re and heart rate in postmenopausal women. American Journal of Hypertension, 14, 942-949. Pickering, T. G., Hall, J. E., Appel, L. J., Falkner, B. E., Graves, J., Hill, M. N., et al., (2005). Recommendations for blood pressure measurement in humans and experimental animals. Part 1: blood pressure measurement in humans. A statement for professi onals from the subcommi ttee of professional and public education of the American Heart Association Council on High Blood Pressure Research. Circulation, 111 697-716. Post-White, J., Kinney, M. E., Savik, K., Gau, J. B., Wilcox C., & Lerner, I. (2003). Therapeutic massage and heali ng touch improve symptoms in cancer Integrative Cancer Therapies, 2 (4), 332-344. Potter, P. A., & Perry, A. G. (2001). Fundamentals of Nursing (5th ed.). ST. Louis: Mosby. Purves, D., Augustine, G. J., Fitzpatr ick, D., Katz, L. C., LaMantia, A., McNamara, J. O., et al. (Eds.). (2001). Neuroscience (2nd ed.). Sunderland, Massachusetts: Sinauer Associates, Inc. Rexilius, S. J., Mundt C., Erickson Megel, M. & Agra wal, S. (2002). Therapeutic effects of massage etherapy and handling touch on caregivers of patients undergoing autologous hem atopoietic stem cell transplant. Oncology Nursing Forum, 29 (3), E35-44. Richards, K. C., Gibson, R., & Overton-McCo y, A. L. (2000). Effects of massage in acute and critical care. AACN Clinical Issues, 11 (1), 77-96.

PAGE 139

127 Rutledge, T. R., & Hogan, B. E. (2002). A quantitative review of prospective evidence linking psychological fact ors with hypertension development. Psychosomatic Medicine, 64, 758-766. Salvo, S.G. (1999). Massage therapy: Principles and practice Philadelphia: Saunders. Segerstrom, S. C., & Miller, G. E. ( 2004). Psychological stress and the human immune system: A meta-analytic study of 30 years of inquiry. Psychological Bulliten, 130 (4), 601-630 Scalco, A. Z., Scalco, M. Z., Azul, J. B. S., a nd Lotufo, N. F. (2005). Hypertension and Depression. Clinics, 60 (3), 241-250. Shulman, K. R., & Jones, G. E. (1996). The effect iveness of massage therapy intervention on reducing anxiety in the workplace. The Journal of Applied Behavioral Science, 32 (2), 160-173. Sherwood, L. (2007). Human physiology: From cells to systems (6th ed.). Belmont, CA: Thomson. Sleight, P. (2004). Commentar y on Hot Topic Review: Arterial baroreflexes can determine long-term blood pressure. Experimental Physiology, 89 (4), 337341. Smith, M. C., Kemp, J., He mphill, L., & Vojir, C. P. (2002). Outcomes of therapeutic massage for hospita lized cancer patients. Journal of Nursing Scholarship, 34 (3), 257-262. Smyth, J., Ockenfels, M. C., Porter, L., Kirschbaum, C., Hellhammer, D. H., & Stone, A. A. (1998). Stressors and mood measured on a momentary basis are associated with salivary cortisol secretion. Psychoneuroendocrinology, 23 (4), 353-370. Spielberger, C. D., Crane, R. S., Kearns, W. D., Pellegr in, K. L., Rickman, R. L., & Johnson, E. H. (1991). Anger and anxiety in ess ential hypertension. In C. D. Spielberger, I. G. Sarason, Z. Kulcs, & G. L. Van Heck (Eds.), Stress and emotion (Vol. 14, pp. 265-283). New York: Hemisphere/Taylor & Francis. Speilberger, C. D., Ritter band, L. M., Reheiser, E. C., & Brunner, T. M. (2003). The nature and measurement of depression. Revista Internacional de Psicologia Clinica de la Salud/ Internat ional Journal of Cli nical and Health Psychology 3(2), 209-234.

PAGE 140

128 Speilberger, C. D., Jacobs, C., Crane, R., Russell, S., Westbe rry, L., Barker, L., et al. (n.d.). Preliminary manual for the Stat e-Trait Personality Inventory (STPI) University of South Florida: unpublished. Stevenson, C. J. (1994). The psychoph ysiological effects of aromatherapy massage following cardiac surgery. Complementary Therapi es in Medicine, 2 27-35. Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate statistics (4th ed .). Boston: Allyn and Bacon. Taylor, A. G. Galper, D. I., Taylor, P., Rice, L. W., Ander sen, W., Irvin, W., et al. (2003). Effects of adjunctive Swed ish massage and vibration therapy on short-term postoperative outcomes: A randomized, controlled trial. The Journal of Alternative and Complementary Medicine, 9 (1), 77-89. Tindal, H. A., Davis, R. B ., Phillips, R. S.,& Eisenber g, D. M. (2005). Trends in use of complementary and alternativ e medicine by US adults: 1997-2002. Alternative Therapies in Health Medicine, 11 (1), 42-49. Thrasher, T. N. (2005). Baroreceptors, baroreceptor unloading, and the long-term control of blood pressure. American Journal of Ph ysiology: Regulatory, Interrogative and Compar ative Physiology, 288 (4), R8 19-27. UCLA Department of Education (2007). Introduction to research design and statistics: Effect size. Retrieved February 2, 2007, from http://www.gseis.ucla.edu/c ourses/ed230a2/notes/effect.html University of South Florida (2004). Gender and ethnicity distribution and employee headcount: Fall 2003 Retrieved March 10, 2004, from http://isis.fastma il.usf.edu/bpa/factbook/fb2004/default.htm University of South Florida (2006). 2005 – 2006 academic year, Tampa physical campus, intended and declared majors in all colleges, all departments, all majors: Annual unduplicated student headcount. Retrieved December 29, 2006, from http://usfweb2.usf.edu/in fomart/infomartapps/Major10.aspx Vasan, R. S., Larson, M. G. Leip, E. P., Evans, J. C .. O'Donnell, C. J., Kannel, W. B., et al. (2001). Impact of HighNormal Blood Pressure on the Risk of Cardiovascular Disease. New England Journal of Medicine, 345 (18), 12911297. Webb, M. S. (2002). Stress management strategies in hypertension control. Ethnicity & Disease, 12 (Supplemental 3), S3-95-S3-100.

PAGE 141

129 Webb, M. S., & Beckstead. J. W. (2002). Stress-related influences on blood pressure in African American women. Research in Nursing & Health, 25 383-393. Wendler, M. C. (1999). An investigation of selected outcomes of Tellington touch in healthy soldiers (Doctoral dissertati on, University of Colorado, 1999). Dissertation Abstracts International, 60 5439. Wendler, M. C., (2003). Effects of Tellin gton touch in healthy adults awaiting venipuncture. Research in Nursing & Health, 26, 40-52. Weil, A. (2007). Hypertension Retrieved January 27, 2007, from http://www.alternativemedicine.co m/common/adam/DisplayMonograph.asps toreID=02AD61F001A74B5887D3BD1 1F6C28169&name=ConsConditions_ Hypertensioncc#MassagePT World Health Organization, International Society of Hypertens ion Writing Group (2003). 2003 World Health Organization (W HO)/ International Society of Hypertension (ISH) statement on management of hypertension. Journal of Hypertension, 21 (11), 1983-1992. Yan, L. L., Liu. K., Matthews, K. A., Daviglus, M. L., Fe rguson, F., & Kiefe, C. I. (2003). Psychosocial factors and risk of hypertension: The coronary artery risk development in young adults (CARDIA) study. Journal of American Medical Association, 290 (16), 2138-2148. Yeh, G. Y., Davis, R. B., & Phillips, R.S. (2006). Use of complementary therapies in patients with cardiovascular disease. American Journal of Cardiology, 98 (5), 673-680. Yucha, C. B., Clark, L., Smith, M., Uris P., LaFleur, B., & Du val, S. (2001). The effect of biofeedback in hypertension. Applied Nursing Research, 14 (1), 2935.

PAGE 142

130 Appendices

PAGE 143

131 Appendix 1 List of Music CDs Benghiat, M. (2002). At Peace: Flow [CD]. Riverside, CT: At Peace Media. Benghiat, M. (2003). At Peace: Spirit [CD]. Riverside, CT: At Peace Media. Frantzich, P. (2005). Lifescapes: The Best of Massage & Meditation [ 2 CD set]. Montreal:Compass Productions. Gibson, D. (1989). Solitudes: Harmony [CD]. Toronto: Solitudes, Ltd.

PAGE 144

132 Appendix 2

PAGE 145

133 Appendix 3

PAGE 146

134 Appendix 4

PAGE 147

135 Appendix 4 (continued)

PAGE 148

136 Appendix 5

PAGE 149

137 Appendix 5 (continued)

PAGE 150

138 Appendix 5 (continued)

PAGE 151

139 Appendix 5 (continued)

PAGE 152

140 Appendix 6 Subject Information No.________________ Name___________________________________________________________________ Address_________________________________________________________________ Contact Phone Number(s)___________________________________________________ Date of Birth______________________ Gender: Female ________ Male________ Height___________________________ Weight_____________________________ Pregnant: Yes________ No________ If yes, how many months?_________________ Smoke: Yes______ No_______ Quit (year) _________ Packs per day____________ Alcohol: Yes______ No _______ Quit (year) ________ Drinks per day___________ Daily Exercise: Yes ______ No_____ Minutes per day_________________________ Ethnicity: Indian/Alaskan Native _____ Asian_______ Black/African American_____ Hispanic/Latino_____ Native Hawaiian/Pacific Islander _____ White/Caucasian_____ Other/Unknown _____ Primary Physician Name____________________________________________________ Contact Phone for Physician_________________________________________________ 1. Are you being seen by a physical or mental health care provider at this time? Yes________ No_________ 1a. If so, for what are you being treated? ________________________________________________________________________ ________________________________________________________________________ ________________________________________________________________________ ________________________________________________________________________ ________________________________________________________________________ ________________________________________________________________________ 1c. Have you ever been told you have hi gh blood pressure or hypertension? Yes_________ No_________

PAGE 153

141 Appendix 6 (continued) 1d. If you answered “yes”, for how many years? _______________________________ 1e. Do you take medication for hi gh blood pressure or hypertension? Yes_________ No_________ 1f. If you answered “yes”, what is/are th e name(s) of the medication(s) and how long have you been taking the medication(s)? ________________________________________________________________________ ________________________________________________________________________ ________________________________________________________________________ 2. What other medications are you taking at this time? Include prescriptions, over the counter and illegal medications, herbals, vitami ns or any other treatments that you either take orally, rectally, inject or put on your skin. Also include dosage and how often you take the medication. ________________________________________________________________________ ________________________________________________________________________ ________________________________________________________________________ ________________________________________________________________________ ________________________________________________________________________ ________________________________________________________________________ 3. List any allergies you have to medi cations, food, animals or household items (or other). ________________________________________________________________________ ________________________________________________________________________ ________________________________________________________________________ 4. Check any therapies you are receiving: _________ Massage therapy _________ Physical Therapy _________ Acupuncture _________ Occupational Therapy _________ Herbal therapy _________ Other 3a. If you check any of the above, for what are you being treated? ________________________________________________________________________ ________________________________________________________________________ ________________________________________________________________________

PAGE 154

142 Appendix 7

PAGE 155

143 Appendix 7 (continued)

PAGE 156

144 Appendix 8

PAGE 157

145 Appendix 8 (continued)

PAGE 158

146 Appendix 8 (continued)

PAGE 159

147 Appendix 8 (continued)

PAGE 160

148 Appendix 9

PAGE 161

149 Appendix 10 Subject Touch Perception Scale: A back massage feels different to every person that gets one. The researcher would like to understand how the back massages felt to you. Below is a list of words that will help you tell the researcher what the back massages felt like to you. Directions: Please circle the “X” that most correctly describes the feeling that you experienced when you got your back massages. Example: On the soothing/irritating word pair if you felt the back massages were soothing, you would circle the “X” closer to the word “soothing” but if the back massages were irritating to you, you would ci rcle the last “X”. The “X” that you circle is the right response for you. To me the back massages felt: --------------------very------less----neutral----less ---very----------------------1. Soothing X X X X X Irritating 2. Unpleasant X X X X X Pleasant 3. Too soft X X X X X Too hard 4. Rough X X X X X Smooth 5. Rhythmic X X X X X Jerky 6. Relaxing X X X X X Tensing

PAGE 162

150 Appendix 11 Subject Perceived Rapport Scale: When we interact with another person we ex perience many feelings. This list of words will help define your feelings about your interaction you have had with your back massage therapist. Directions: Please circle the “X” that most a ccurately describes the feelings that you experienced with your back massage therapist. Example: On the comfortable/ uncomfortable word pair, if you felt quite uncomfortable with the back massage therapis t, you would circle the “X” closer to the word “uncomfortable” but if you were absolutely co mfortable you would circle the first “X”. The “X” that you ci rcle is the right response for you. My back massage therapist made me feel: --------------------very------less----neutral----less ---very----------------------1. Comfortable X X X X X Uncomfortable 2. Distrustful X X X X X Trustful 3. Less Anxious X X X X X More Anxious 4. Calm X X X X X Annoyed 5. Worse X X X X X Better 6. Physically X X X X X Physically tense relaxed 7. Willing to X X X X X Unwilling to share feelings share feelings 8. Unsafe X X X X X Safe 9. Respected X X X X X Ignored 10. Unpleasant X X X X X Pleasant

PAGE 163

151 Appendix 12: Schedule for Participants If you are in the group that receives 10 back massages OR 10 relaxation sessions: Week 1 1 1 2 2 2 3 3 3 4 4 4 5 5 5 6 8 10 Visit: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Consent Habituation T1* Treatments: Vi sits 6-15 T2 T3 T4 If you are in the group that receives 5 back massages: Week 1 1 1 2 2 2 3 3 3 4 4 8 10 Visit: 1 2 3 4 5 6 7 8 9 10 11 12 13 Consent Habituat ion T1* Treatments: Visits 6-10 T2 T3 T4 Legend Consent: the first visit you will read and sign cons ent and HIPAA forms and fill out a demographics form. Habituation: At these three visit y ou will have your blood pressure taken 3 times and you will be adjusting to the research environment. T1, T2, T3, & T4: These are the days t he investigator will take your blood pr essure for the actual data base, collect your saliva and have you fill out a personality inventory. *At the end of the first visit (T1) you will be randomized, in other words, you will draw the group you will be in, for the study. Treatments: You will either be receiv ing 10 ten-minute back massages, 5 ten-minute back massages or 10 ten-minute relaxation sessions.

PAGE 164

152 Appendix 13: Medications Used by Study Participants Medication 10-Massage Group # (% of group) 5-Massage Group # (% of group) Control # (% of group) Total # (% of total sample) ACE Inhibitors 3 (20%) 3 (23%) 2 (14%) 8 (19%) Angiotensin II Receptor Blockers 3 (20%) 3 (23%) 2 (14%) 8 (19%) Beta Blockers 0 3 (23%) 2 (14%) 5 (12%) Calcium Channel Blockers 1 (6%) 0 2 (14%) 3 (7%) Diuretics 4(27%) 4 (31%) 5 (36%) 14 (33%) # of Participants on HBP meds [# on 2 meds] 11 [3] (53%) 13 [2] (85%) 13 [3] (71%) 37 [8] (69%) Anti-Lipids 5 (33%) 4 (31%) 2 (14%) 11 (26%) Anti-Anginal 0 0 1 (7%) 1 (2%) Hormones 2 (13%) 4 (31%) 1 (7%) 7 (17%) Bone Reabsorbtion Inhibitor 1 (6%) 2 (15%) 0 3 (7%) Anti-Histamines 2 (13%) 3 (23%) 2 (14%) 7 (17%) Nasal Glucocorticoids 1 (6%) 2 (15%) 2 (14%) 5 (12%) Bronchodilators 0 3 (23%) 0 3 (7%) Anti-Ulcer/ Appetite Suppressor 3(20%) 2 (15%) 2 (14%) 7 (17%) Anti-Platelet 0 0 2 (14%) 2 (5%) NSAID 5 (33%) 6 (46%) 8 (57%) 19 (45%) Opioid 0 0 2 (14%) 2 (5%) Anti-Depressant 5 (33%) 4 (31%) 4 (28%) 13 (31%) Anti-Anxiety 1 (6%) 2 (15%) 1 (7%) 4 (10%) Anti-Convulsant 1 (6%) 2 (15%) 0 3 (7%) Anti-Psychotic 0 0 1 (7%) 1 (2%) Muscarine Receptor Agonist 0 0 1 (7%) 1 (2%) Muscle Relaxants 0 1 (7%) 2 (14%) 3 (7%) Oral Diabetic 1 (6%) 1 (7%) 2 (14%) 4 (10%) Vitamins 10 (67%) 11 (84%) 6 (42%) 27 (64%)

PAGE 165

153 Appendix 14: Means and Standard Deviation Bars for Trait Anxiety, Anger and Depression 10 12 14 16 18 20 22 24 246810 WEEKTRAIT ANXIETY SCOR E 10M 5M CNT 10 12 14 16 18 20 22 24 26 2 4 6 8 10 WEE K TRAIT ANGER SCORE 10M 5M CNT

PAGE 166

154 Appendix 14: Means and Standard Deviation Ba rs for Trait Anxiety, Anger and Depression (Continued) 7 9 11 13 15 17 19 21 23 25 27 246810 WEEKTRAIT DEPRESSION SCORE 10M 5M CNT

PAGE 167

155 Appendix 15: Correlations of pertinent variables Correlation of pertinent variables at pr e-randomization meas urement (Week 2) PRE SYS PRE DIAS AGE BMI YRS HBP MPD EXC A-HPN SANX1 SANG1 SDEP1 TANX1 TANG1 TDEP1 CORT 1 PRESYS .490 -.131 .053 -.053 .116 -.438** .165 .039 .072 .048 .027 .005 .079 PRE DIAS -.220 -.232-.033 .187 -.348* .235 .117 .208 .081 .130 -.051 .040 AGE -.068 .236 -.443* .304 -.260 -.319* -.091 -.075 -.216 -.093 .033 BMI .140 -.268 .251 .207 .133 .354* .398* .087 .407* -.255 YRSHBP -.126 .326* -.168 -.037 -.037 -.197 .034 -.048 .193 MPDEXC -.415** -.158 .013 -.216 -.160 .158 -.211 -.015 A-HPN -.023 .184 .021 -.049 .150 -.042 -.108 S-ANX1 .703* .636* .504* .102 .476* .028 S-ANG1 .393* .366* .362* .390* -.121 S-DEP1 .620* -.015 .622* -.008 T-ANX1 .327* .872* .047 T-ANG1 .258 .019 T-DEP1 .322* CORT1 Pearson’s Correlation significant at the .05 level (2-tailed). ** Pearson’s Correlation significant at the .01 level (2-tailed). A-HPN = Antihypertensive Meds (angiotensin converting enzyme in hibitors, angiotensin-receptor bl ockers, calcium channel blocker s, beta-blockers, and diuretics) PRESYS = Systolic blood pressure measured prior to randomization PRE DIAS = Diastolic blood pressure measured prior to randomization BMI = Body mass index YRSHBP = Years the participant has been diagnosed with hypertension MPDEXC = Average number of minutes the participant exercises every day S-ANX1 = State anxiety measured prior to randomization S-ANG1 = State anger measured prior to randomization T-ANX1 = Trait anxiety measured prior to randomization T-ANG1 = Trait anger measured prior to randomization T-DEP1 = Trait depression measured prior to randomization CORT1 = Salivary cortisol measur ed prior to randomization

PAGE 168

156 Appendix 15 (continued) Correlations of pertinent variabl es at first post intervention measurement (w eek 4 for the 5-massage group or week 6 for the 10-Massage group and Control group), POST SYS POST DIAS AGE BMI YRS HBP MPD EXCS A-HPN SANX2 SANG2 SDEP2 TANX2 TANG2 TDEP2 CORT 2 POSTSYS .635* -.013 -.056.155 .127 -.422** .443* .256 .265 .101 .034 .036 .049 POSTDIA -.123 -.236.122 .238 -.391** .401* .208 .183 .070 .034 -.034 .184 AGE -.068.236 -.443* .304 .008 .084 .012 -.065 -.180 -.011 -.039 BMI .140 -.268 .251 .128 -.096 .374* .450* .261 .493* .045 YRSHBP -.126 .326* .235 .367* .280 -.080 .084 .100 .308* MPDEXC -.415** .065 .085 -.084 -.104 .019 -.200 .000 A-HPN -.140 -.021 -.010 -.142 .280 -.069 .122 S-ANX2 .779* .712* .420* .300 .411* .231 S-ANG2 .429* .054 .267 .132 .074 S-DEP2 .661* .436* .738* .235 T-ANX2 .268 .857* .172 T-ANG2 .375* -.054 T-DEP2 .097 CORT2 Pearson’s Correlation significant at the .05 level (2-tailed). ** Pearson’s Correlation significant at the .01 level (2-tailed). A-HPN = Antihypertensive Meds (angiotensin converting enzyme in hibitors, angiotensin-receptor bl ockers, calcium channel blocker s, beta-blockers, and diuretics) PRESYS = Systolic blood pressure measured prior to randomization PRE DIAS = Diastolic blood pressure measured prior to randomization BMI = Body mass index YRSHBP = Years the participant has been diagnosed with hypertension MPDEXCS = Average number of minutes the participant exercises every day S-ANX2 = State anxiety measured prior to randomization S-ANG2 = State anger measured prior to randomization S-DEP2 = State depression m easured prior to randomization T-ANX2 = Trait anxiety measured prior to randomization T-ANG2 = Trait anger measured prior to randomization T-DEP2 = Trait depression measured prior to randomization CORT2 = Salivary cortisol measur ed prior to randomization

PAGE 169

157 Appendix 15 (continued) Correlations of pertinent variables at se cond post intervention measurement (Week 8) POST SYS1 POST DIAS1 AGE BMI YRS HBP MPD EXCS A-HPN SANX3 SANG3 SDEP3 TANX3 TANG3 TDEP3 CORT 3 POSTSYS1 .586* .063 .061 .015 .068 -.441** .367* .201 .299 .142 .213 .189 -.136 POSTDIA1 -.121 -.072 .003 .297 -.367* 430* .347* .278 .217 .276 .155 -.064 AGE -.068 .236 -.443* .304 -. 107 .062 -.022 .043 -.085 .005 -.014 BMI .140 -.268 .251 .127 172 .367* .485* .254 .406* -.005 YRSHBP -.126 .326* .066 .118 .096 .010 .067 .096 .256 MPDEXCS -.415** 076 -.148 -.151 -.150 .056 -.179 .075 A-HPN -.115 .107 -.034 .020 .198 -.012 .279 S-ANX3 .839* .849* .454* .256 .417* .177 S-ANG3 .833* .409* .289 .415* .135 S-DEP3 .689* .290 .681* .037 T-ANX3 .370* .870* .007 T-ANG3 .347* -.023 T-DEP3 .008 CORT3 Pearson’s Correlation significant at the .05 level (2-tailed). ** Pearson’s Correlation significant at the .01 level (2-tailed). A-HPN = Antihypertensive Meds (angiotensin converting enzyme in hibitors, angiotensin-receptor bl ockers, calcium channel blocker s, beta-blockers, and diuretics) PRESYS = Systolic blood pressure measured prior to randomization PRE DIAS = Diastolic blood pressure measured prior to randomization BMI = Body mass index YRSHBP = Years the participant has been diagnosed with hypertension MPDEXCS = Average number of minutes the participant exercises every day S-ANX3 = State anxiety measured prior to randomization S-ANG3 = State anger measured prior to randomization S-DEP3 = State depression m easured prior to randomization T-ANX3 = Trait anxiety measured prior to randomization T-ANG3 = Trait anger measured prior to randomization T-DEP3 = Trait depression measured prior to randomization CORT3 = Salivary cortisol meas ured prior to randomization

PAGE 170

158 Appendix 15 (continued) Correlations of pertinent variables at se cond post intervention measurement (Week 10) POST SYS2 POST DIAS 2 AGE BMI YRS HBP MPD EXCS A-HPN SANX4 SANG4 SDEP4 TANX4 TANG4 TDEP4 CORT 4 POSTSYS2 .604* -.061 -.088 -.009 .059 -.556** .361* .166 .080 -.015 -.036 .060 .186 POSTDIA2 -.130 -.115 .195 .127 -.397** 386* .406* .283 .191 .097 .208 .139 AGE -.068 .236 -.443* .304 -.057 -.026 .022 -.005 -.080 .011 -.042 BMI .140 -.268 .251 .147 .072 .365* .460* .313* .338* .047 YRSHBP -.126 .326* .096 .181 .165 -.016 .155 .067 .205 MPDEXCS -.415** 089 -.014 -.150 -.221 -.013 -.124 .004 A-HPN .008 .165 .052 .043 .296 -.084 .064 S-ANX4 .740* .725* .460* .545* .419* .334* S-ANG4 .477* .248 .504* .189 .085 S-DEP4 .826* .552* .816* .315* T-ANX4 .441* .874* .166 T-ANG4 .419* .101 T-DEP4 .127 CORT4 Pearson’s Correlation significant at the .05 level (2-tailed). ** Pearson’s Correlation significant at the .01 level (2-tailed). A-HPN = Antihypertensive Meds (angiotensin converting enzyme in hibitors, angiotensin-receptor bl ockers, calcium channel blocker s, beta-blockers, and diuretics) PRESYS = Systolic blood pressure measured prior to randomization PRE DIAS = Diastolic blood pressure measured prior to randomization BMI = Body mass index YRSHBP = Years the participant has been diagnosed with hypertension MPDEXCS = Average number of minutes the participant exercises every day S-ANX4 = State anxiety measured prior to randomization S-ANG4 = State anger measured prior to randomization S-DEP4 = State depression m easured prior to randomization T-ANX4 = Trait anxiety measured prior to randomization T-ANG4 = Trait anger measured prior to randomization T-DEP4 = Trait depression measured prior to randomization CORT4 = Salivary cortisol meas ured prior to randomization

PAGE 171

159 Appendix 16: Scatter Plots for Dia stolic difference scores and BMI GROUP: 1BMI36 34 32 30 28 26 24 22 20DYSDIFF220 10 0 -10 -20 GROUP: 2BMI50 40 30 20DYSDIFF210 0 -10 -20

PAGE 172

160 Appendix 16: Scatter Plots for Diasto lic difference scores and BMI (continued) GROUP: 3BMI50 40 30 20DYSDIFF220 10 0 -10

PAGE 173

Appendix 17 ALL Groups: Correlations of Antihypertensive M edication, BMI and Baroreceptor Response A-HPN BMI BAR01 BAR02 BAR03 BAR04 BAR05 BAR06 BAR07 BAR08 A-HPN .251 .098 .103 .017 .130 .358* -.128 .036 .175 BMI -.074 .114 -.066 -.107 -.130 .025 .003 -.078 BAR01 .280 .536** .406** .680** .385* .519** .614** BAR02 .382* .493** .265 .231 .392* .059 BAR03 .441** .360* .324* .226 .394** BAR04 .365* .170 .432** .270 BAR05 .462** .483** .582** BAR06 .497** .560** BAR07 .313** BAR08 Pearson’s Correlation signific ant at the .05 level (2-tailed). ** Pearson’s Correlation significant at the .01 level (2-tailed). A-HPN = Antihypertensive Medication (angiotensin converti ng enzyme inhibitors, angiotensin-receptor blockers, calcium channel blockers, bet a-blockers, and diuretics) BMI = Body Mass Index (705 Weight/ Height2) BARO1 = Lying Systolic Blood Pressure Sitting Systolic Blood Pressure at Pre randomization BARO2 = Lying Systolic Blood Pressure Sitti ng Systolic Blood Pressure at Post measure BARO3 = Lying Systolic Blood Pressure Sitti ng Systolic Blood Pressure at Post 1 measure BARO4 = Lying Systolic Blood Pressure Sitti ng Systolic Blood Pressure at Post 2 measure BARO5 = Lying Diastolic Blood Pressure Sitti ng Diastolic Blood Pressure at Pre randomization BARO6 = Lying Diastolic Blood Pressure Sitt ing Diastolic Blood Pressure at Post Measure BARO7 = Lying Diastolic Blood Pressure Sitti ng Diastolic Blood Pressure at Post 1 Measure BARO8 = Lying Diastolic Blood Pressure Sitti ng Diastolic Blood Pressure at Post 2 Measure

PAGE 174

162 Appendix 17 (continued) 10-Massage Group: Correlations of Antihypertensiv e Medication, BMI and Baroreceptor Response A-HPN BMI BAR01 BAR02 BAR03 BAR04 BAR05 BAR06 BAR07 BAR08 A-HPN .318 .051 .235 .122 .086 .165 -.477 -.168 .238 BMI -.350 .351 -.162 .109 -.146 -.007 .115 -.294 BAR01 .170 .435 .248 .815** .420 .492 .824** BAR02 .156 .219 .208 .047 .220 .045 BAR03 .040 .367 .307 -.077 .339 BAR04 .103 .467 .270 .532* BAR05 .316 .653** .597* BAR06 .523* .423 BAR07 .308 BAR08 Pearson’s Correlation signific ant at the .05 level (2-tailed). ** Pearson’s Correlation significant at the .01 level (2-tailed). A-HPN = Antihypertensive Medication (angiotensin converti ng enzyme inhibitors, angiotensin-receptor blockers, calcium channel blockers, bet a-blockers, and diuretics) BMI = Body Mass Index (705 Weight/ Height2) BARO1 = Lying Systolic Blood Pressure Sitting Systolic Blood Pressure at Pre randomization BARO2 = Lying Systolic Blood Pressure Sitti ng Systolic Blood Pressure at Post measure BARO3 = Lying Systolic Blood Pressure Sitti ng Systolic Blood Pressure at Post 1 measure BARO4 = Lying Systolic Blood Pressure Sitti ng Systolic Blood Pressure at Post 2 measure BARO5 = Lying Diastolic Blood Pressure Sitti ng Diastolic Blood Pressure at Pre randomization BARO6 = Lying Diastolic Blood Pressure Sitt ing Diastolic Blood Pressure at Post Measure BARO7 = Lying Diastolic Blood Pressure Sitti ng Diastolic Blood Pressure at Post 1 Measure BARO8 = Lying Diastolic Blood Pressure Sitti ng Diastolic Blood Pressure at Post 2 Measure

PAGE 175

163 Appendix 17 (continued) 5-Massage Group: Correlations of Antihypertensiv e Medication, BMI and Baroreceptor Response A-HPN BMI BAR01 BAR02 BAR03 BAR04 BAR05 BAR06 BAR07 BAR08 A-HPN -.444 -.109 .093 .057 .397 .596* -.153 .104 .192 BMI .201 .111 .245 -.116 -.437 .029 -.001 .205 BAR01 .477 .605* .535 .433 .189 .559* .257 BAR02 .556* .495 .353 .492 .677* .014 BAR03 .619* .360 .350 .417 .349 BAR04 .703** .227 .667* .103 BAR05 .342 .415 .384 BAR06 .415 .499 BAR07 .150 BAR08 Pearson’s Correlation signific ant at the .05 level (2-tailed). ** Pearson’s Correlation significant at the .01 level (2-tailed). A-HPN = Antihypertensive Medication (angiotensin conver ting enzyme inhibitors, angiotensin-receptor blockers, calcium channel blockers, bet a-blockers, and diuretics) BMI = Body Mass Index (705 Weight/ Height2) BARO1 = Lying Systolic Blood Pressure Sitting Systolic Blood Pressure at Pre randomization BARO2 = Lying Systolic Blood Pressure Sitti ng Systolic Blood Pressure at Post measure BARO3 = Lying Systolic Blood Pressure Sitti ng Systolic Blood Pressure at Post 1 measure BARO4 = Lying Systolic Blood Pressure Sitti ng Systolic Blood Pressure at Post 2 measure BARO5 = Lying Diastolic Blood Pressure Sitti ng Diastolic Blood Pressure at Pre randomization BARO6 = Lying Diastolic Blood Pressure Sitt ing Diastolic Blood Pressure at Post Measure BARO7 = Lying Diastolic Blood Pressure Sitti ng Diastolic Blood Pressure at Post 1 Measure BARO8 = Lying Diastolic Blood Pressure Sitti ng Diastolic Blood Pressure at Post 2 Measure

PAGE 176

164 Appendix 17 (continued) Control Group: Correlations of Antihypertensive Medication, BMI and Baroreceptor Response A-HPN BMI BAR01 BAR02 BAR03 BAR04 BAR05 BAR06 BAR07 BAR08 A-HPN .689** .270 -.005 -.006 -.114 .285 .230 .191 .121 BMI -.040 .022 -.396 -.261 .134 .230 .193 -.103 BAR01 .245 .781** .427 .842** .516 .308 .764** BAR02 .337 .696** .268 .163 .199 .131 BAR03 .548* .507 .353 .246 .581* BAR04 .162 -.005 .285 .246 BAR05 .673** .419 .804** BAR06 .640* .728** BAR07 BAR08 Pearson’s Correlation signific ant at the .05 level (2-tailed). ** Pearson’s Correlation significant at the .01 level (2-tailed). A-HPN = Antihypertensive Medication (angiotensin converti ng enzyme inhibitors, angiotensin-receptor blockers, calcium channel blockers, bet a-blockers, and diuretics) BMI = Body Mass Index (705 Weight/ Height2) BARO1 = Lying Systolic Blood Pressure Sitting Systolic Blood Pressure at Pre randomization BARO2 = Lying Systolic Blood Pressure Sitti ng Systolic Blood Pressure at Post measure BARO3 = Lying Systolic Blood Pressure Sitti ng Systolic Blood Pressure at Post 1 measure BARO4 = Lying Systolic Blood Pressure Sitti ng Systolic Blood Pressure at Post 2 measure BARO5 = Lying Diastolic Blood Pressure Sitti ng Diastolic Blood Pressure at Pre randomization BARO6 = Lying Diastolic Blood Pressure Sitt ing Diastolic Blood Pressure at Post Measure BARO7 = Lying Diastolic Blood Pressure Sitti ng Diastolic Blood Pressure at Post 1 Measure BARO8 = Lying Diastolic Blood Pressure Sitti ng Diastolic Blood Pressure at Post 2 Measure

PAGE 177

About the Author Christine M. Olney received her Bachelor’s Degree in Nursing from Montana State University, in 1978. Mrs. Olney’s Master’s Degree in Nursing was received in May of 2002 from South Dakota State University. During her doctoral education program at the University of South Florida, Tampa, Mrs. Olney developed a role as a nursing research resident in a local hospital. In the role as a research resident she initiated and maintained a nursing research program. The role led to several nursing research projects and publications. Mrs. Olney has accepted a fellowship at the James A. Haley Veteran’s Administration Patient Safety Center. She plans to continue pursuing the impact of complementary therapies.