|USFDC Home | USF Electronic Theses and Dissertations||| RSS|
This item is only available as the following downloads:
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 2200373Ka 4500
controlfield tag 001 002317429
007 cr bnu|||uuuuu
008 100901s2009 flua ob 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0003043
Falls in bone marrow transplant patients
h [electronic resource] :
b a retrospective study /
by Lura Henderson.
[Tampa, Fla.] :
University of South Florida,
Title from PDF of title page.
Document formatted into pages; contains 50 pages.
Thesis (M.S.)--University of South Florida, 2009.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
ABSTRACT: Falls are a contributing factor to increased morbidity in the elderly and chronically ill populations and can affect overall quality of life. The literature indicates that oncology patients are a particularly vulnerable population who are further at risk for falls due to increased age, treatment related fatigue, side effects of medications, co-morbidities, decreased muscle tone, altered mental status, and anemia. Although patients with cancer are at a high risk for falls, this is not a well-documented patient problem in the nursing literature. This study examined the validity of the use of the Morse Fall Assessment Tool for use with Bone Marrow Transplant patients and explored other variables that might influence fall outcomes. This study was a retrospective chart review. The sample consisted of a total of 59 patients, which included 29 fallers and 30 non-fallers on a bone marrow transplant unit.There were 22 males and 37 females, ranging in age from 20 to 70 with a mean age of 53.9 (SD= 12.2).The results of this study indicate that there is a significant difference between fallers' (M= 43.8) and non-fallers' (M= 26.8) scores on the Morse Fall Scale (p= 0.000). Significant differences between groups were found with history of falls (p= 0.042), secondary diagnosis (p= 0.015), and muscle weakness (p= 0.025). Laboratory results from fallers and non-fallers revealed significant differences in platelet count (p= 0.003), BUN (p= 0.032), glucose (p= 0.009), and phosphorous (p= 0.001). This is the first study to document falls in the bone marrow transplant population. This study should be a stimulus for future studies conducted in the oncology and/or bone marrow transplant population. Studying falls in these patients is essential to understanding the physiological risk factors that may contribute to patient falls.Findings lay the foundation for studying falls in the bone marrow transplant population. It is crucial to study falls in this population in order to make appropriate assessments and interventions to keep this population free from injury.
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
Advisor: Susan C. McMillan, Ph.D., A.R.N.P.
Peripheral stem cell transplant
Morse Fall Scale
Bone Marrow Transplantation.
t USF Electronic Theses and Dissertations.
Falls in Bone Marrow Transplant Patients: A Retrospective Study by Lura Henderson, R.N., B.S.N. A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science College of Nursing University of South Florida Major Professor: Susan C. McMillan, Ph.D., A.R.N.P Cindy Tofthagen, Ph.D., A.R.N.P. Versie Johnson-Mallard, Ph.D., A.R.N.P. Date of Approval: July 13, 2009 Keywords: peripheral stem cell transplant, oncology, falls, r isk factors, Morse Fall Scale Copyright 2009, Lura Henderson
Dedication This is dedicated to my amazing family. To my fath er, Robert, thank you for always being there for me, to listen and lend advic e, and financially aiding in my graduate school endeavors. To my mother, Reena, tha nk you for believing in me, encouraging me to aim high, and for introducing me to civic opportunities at a young age. To my wonderful fianc, Kelly, I am so blessed to h ave you in my life and thank you for your love, support, advice, and for being my rock w hile I am in graduate school. To my cousin, Steven Jones, who in his short life, inspir ed me to become a Bone Marrow Transplant Nurse in his honor.
Acknowledgements I would like to express my deepest gratitude to Dr Susan McMillan for taking the time to mentor a young oncology nurse and guiding m e throughout my endeavors in this Oncology Masters Program. To my preceptors Kimberly Tope, ARNP and Laura Besaw, ARNP my sincerest appreciation for receiving the va lues of your teaching, nursing wisdom, and especially to Laura for the extra clini cal assignments and differential diagnosis case studies. Dr. Cindy Tofthagen and Dr. Versie Johnson-Mallard, thank you for kindly agreeing to be on my thesis committee an d for all of your insight on oncology nursing and patient falls.
i Table of Contents List of Tables iii Abstract iv Chapter One Introduction 1 Problem Statement 2 Research Objectives 2 Definition of Terms 3 Significance to Nursing 3 Chapter Two Review of the Literature 5 Risk Factors and Falls in the General Population 5 Risk Factors and Falls in Oncology Patients 10 Falls Assessment Tools 14 Synthesis of Literature 17 Chapter Three Methods 20 Sample 20 Instruments 21 Procedures 22 Data Analysis 22 Chapter Four Results, Discussion, and Conclusions 24 Results 24 Demographic Data 2 4 Morse Fall Scale 26 Temperature 27 Laboratory Values 27 Inter-Rater Reliability 29 Discussion 29 Demographic Data 30 Morse Fall Scale 31 Other Contributing Factors 32 Temperature 32 Laboratory Values 33 Vital Signs and Medications 34 Inter-Rater Reliability 35
ii Implications for Nursing 36 Conclusions 37 Recommendations for Future Research 37 References 39 Appendices 41 Appendix A: Chart Audit for Fallers and Non-Faller s 42 Appendix B: Morse Fall Scale 45 Appendix C: Letter of Approval from Florida Hospit al 47 Appendix D: IRB Approval from Florida Hospital 48 Appendix E: USF IRB Approval 49
iii List of Tables Table 1. Mean and Percent of Demographic Variables for Fallers and 24 Non-Fallers Table 2. Frequency and Percent of Falls among Falle rs 25 Table 3. Frequency and Percent of Cancer Diagnoses 25 Table 4. Chi-square comparison of Morse Fall Scale Variables between Fallers 27 and Non-Fallers Table 5. Temperature in Bone Marrow Transplant Pati ents 27 Table 6. Means, t -scores, and Significance of Laboratory Values of F allers and 28 Non-Fallers
iv Falls in Bone Marrow Transplant Patients: A Retrospective Study. Lura Henderson ABSTRACT Falls are a contributing factor to increased morbi dity in the elderly and chronically ill populations and can affect overall quality of life. The literature indicates that oncology patients are a particularly vulnerabl e population who are further at risk for falls due to increased age, treatment related fatig ue, side effects of medications, comorbidities, decreased muscle tone, altered mental status, and anemia. Although patients with cancer are at a high risk for falls, this is n ot a well-documented patient problem in the nursing literature. This study examined the val idity of the use of the Morse Fall Assessment Tool for use with Bone Marrow Transplant patients and explored other variables that might influence fall outcomes. This study was a retrospective chart review. The s ample consisted of a total of 59 patients, which included 29 fallers and 30 non-fall ers on a bone marrow transplant unit. There were 22 males and 37 females, ranging in age from 20 to 70 with a mean age of 53.9 (SD= 12.2).The results of this study in dicate that there is a significant difference between fallers (M= 43.8) and non-falle rs (M= 26.8) scores on the Morse Fall Scale (p= 0.000). Significant differences betw een groups were found with history of
v falls (p= 0.042), secondary diagnosis (p= 0.015), a nd muscle weakness (p= 0.025). Laboratory results from fallers and non-fallers rev ealed significant differences in platelet count (p= 0.003), BUN (p= 0.032), glucose (p= 0.009 ), and phosphorous (p= 0.001). This is the first study to document falls in the b one marrow transplant population. This study should be a stimulus for future studies conducted in the oncology and/or bone marrow transplant population. Studying falls in the se patients is essential to understanding the physiological risk factors that m ay contribute to patient falls. Findings lay the foundation for studying falls in the bone m arrow transplant population. It is crucial to study falls in this population in order to make appropriate assessments and interventions to keep this population free from inj ury.
1 Chapter I Introduction Falls are a contributing factor to increased morbid ity in the elderly and chronically ill populations and can affect overall quality of life. Preventing falls has been an area of concern and a focus of research for deca des. Multiple intrinsic factors can contribute to increased patient falls such as acute and chronic illnesses, medications, increased age, mental status, and muscle tone (Krau ss, Evanoff, Hitcho, Ngugi, Dunagan, Fischer, et al. 2005). Therefore, preventing patien t falls is a high priority and is included in the National Patient Safety Goals Hospital Program written by the Joint Commission on the Accreditation of Healthcare Organizations (J CAHO) as goal number nine to reduce the risk of patient harm resulting from fal ls (JCAHO, 2008). Shever and colleagues (2008), concluded that patien ts who received high nursing surveillance (patient care once every two hours) ex perienced 157 falls compared to 324 falls experienced by patients who received limited or no high nursing surveillance. This study suggested that the additional $191 high surve illance costs per hospitalization is associated with a cost savings of $17,483, an amoun t estimated to cover the medical expenses for a patient who has fallen (Shever, Titl er, Kerr, Qin, Kim, & Picone, 2008). This savings is of considerable importance for nurs ing and hospital administrators. As of October 1, 2008 if a patient experiences a fall whi le in the hospital, Medicare will no longer pay hospitals for the increased costs of car e related to injury. In addition, Medicare forbids hospitals to charge patients for the increa sed medical expenses associated with
2 falls and other hospital-acquired conditions (Cente rs for Medicare and Medicaid Services, 2008). Oncology patients are a particularly vulnerable pop ulation who are further at risk for falls due to increased age, treatment related f atigue, side effects of medications, comorbidities, decreased muscle tone, altered mental status, and anemia. Although patients with cancer are at a high risk for falls, this is n ot a well-documented patient problem in the nursing literature (Holley, 2002). Fall risk-as sessment tools utilized in nursing are not designed for cancer patients and may not accurately represent or predict falls in the oncology population (OConnell, Baker, Gaskin, & Ha wkins 2007). Problem Statement While the importance of identifying fall risk facto rs in oncology patients is high, empirical analysis has been limited.Oncology nurses need to be cognizant and have a detailed understanding of falls and associated risk factors in their patient populations (Holley, 2002).The purpose of this study was to evaluate the valid ity of the use of Morse Fall Assessment Tool for Bone Marrow Transplant (BM T) patients by known group comparisons and exploration of other variables that might influence fall outcomes. Research Objectives The following objectives guided this study: 1. To determine if there was a significant differe nce in the Morse Fall Assessment Score in BMT patients who fall and those who do not fall. 2. To determine if there were significant differen ces in age, gender, diagnosis, history of falls, confusion, muscle weakness, blood laboratory values, blood pressure, temperature, and medications taken within 24 hours between falle rs and non-fallers.
3 Definition of Terms Falls are defined as any sudden, unanticipated inc ident that causes a person to accidently land on any lower surface (Pearse, Nicho lson, & Bennett, 2004). It is also suggested that a fall is a sudden, unintentional c hange in position causing an individual to land at a lower level, not as a consequence of a sudden onset of paralysis, epileptic seizure, or force (Overcash, 2007, p. 342). The ge neral theme of falls definitions are that they are spontaneous and do not result from loss of consciousness. Falls can be further differentiated into anticipated physiological falls which includes ambulation difficulty and confusion, unanticipated physiological falls th at occur when a patient faints or seizes, and accidental falls such as tripping or slipping ( Morse, Black, Oberle, & Donahue, 1989). Significance to Nursing Nurses are committed to providing safe environment s for their patients. Assessment and prevention of patient falls are top priorities to nurses and risk management. In the view of risk management, failure to identify patients at risk for falls and appropriately providing interventions may lead to a decreased level of safety and the potential for malpractice lawsuits (Tommasini, Tala mini, Bidoli, Sicolo, & Palese, 2008). Patient falls may occur as a result of environmenta l factors and intrinsic patient characteristics such as incontinence, medications, muscle weakness, confusion, and anemia (Krauss, et al., 2004; Dharmarajan, et al., 2006). The hospital that is the target of this study is c urrently using the Morse Fall Assessment Tool without any validity data signifyin g that this tool is sensitive for use in Bone Marrow Transplant patients. Understanding onco logy falls risk factors may enable
4 nurses to appropriately and more accurately identif y patients at higher risk of falling. Assessment of patient falls and ensuring proactive interventions to prevent injury from falls are important issues in nursing research. Ide ntification of fall risk factors in the bone marrow transplant population should be explored to determine the validity of the Morse Fall Assessment Tool. Exploration of other variable s that might influence fall outcomes identifies potential oncology risk factors while ad ding to the current nursing literature.
5 Chapter II Review of the Literature This chapter reviews current nursing literature re levant to patient risk factors for falling. The first part of the review discusses pat ient risk factors and falls in the general population. This is followed by a review of current studies of risk factors and falls in the oncology population. Current falls assessment tools are then discussed. This chapter ends with a synthesis of the literature and illustrates that falls research in the bone marrow transplant population is crucial. Risk Factors and Falls in General Population Morse and colleagues are well known in the nursing literature for their research on patient falls and for encouraging the use of the Morse Fall Scale in hospital settings. Their study, published in 1989, was the first study of this widely used scale. In a prospective study by Morse and colleagues (1989), t he Morse Fall Scale was utilized to predict patient daily risk factors for falling in a cute medical and surgical, long term geriatric, and rehabilitative units with 2689 patie nts over a four month period to determine the feasibility of using this scale in nu rsing practice. Measures used to predict patient falls were: history of falling, presence of secondary diagnosis, use of ambulatory aids, current intravenous administration, gait char acteristics, and mental status. Patients were identified as low risk of falling if the score was less than 20, medium risk for falling if scores ranged from 25-40, and were considered hi gh risk if given a score greater than 45. Nurses scored patients risk of falling daily, documented fall prevention interventions,
6 and if a fall occurred, documented the type of fall and attributing factors. Analysis of patient falls revealed that 61.9% were physiologica l anticipated falls, 13.6% were unanticipated falls, and 24.5% were accidental fall s. Further, 76.9% of patients who fell were considered to be high risk for falls. This sca le was found to be a convenient and effective predictor of patient falls and guided nur sing staff in implementing fall prevention strategies. However, oncology patients w ere not identified as specifically included. Stevenson, Mills, Welin, and Beal (1998) conducted a retrospective, descriptive, comparative study designed to compare 301 hospitali zed fallers and 301 hospitalized non-fallers matched as well as possible by age and medical diagnosis at discharge. The purpose of this study was to increase nursing knowl edge beyond established risk factors of age and medical diagnosis, by comparing groups o f fallers and non-fallers in an acutecare setting. Of 602 patients, 54% were female and 46% were male with a median age of 61.8 years. Cardiac patients represented the majori ty of the group at 23.2%. The remaining diagnoses were psychiatric (14%), rehabil itation (9.8%), gastrointestinal (8.8%), cancer (8.1%), orthopedic (6.9%), pulmonary (6.2%), neurologic (3.9%), and various others (19%). The authors designed a datacollection form associated with fall risk factors identified in the literature and conte nt validity was evaluated by the hospitals nursing research group. Data for this study was ret rospectively collected from medical records and incident reports of patients who had fa llen. Logistic regression was their method of data analysis and aided in estimating the odds ratios to interpret their results. Collinearity was not found to be present among the independent variables. The primary risk factors associated with falls and increased od ds ratio of falling in this study were
7 incontinence (11.3), length of stay > 18 days (9.9) dependent for ambulation (6.0), independent for hygiene (2.5), and lack of regular exercise (2.0). Stevenson and colleagues concluded that no two studies have found the same set of fall risk factors. Therefore, it was suggested that risk factors for f alls are patient population specific and each hospital should evaluate and modify fall risk factors on a continuing basis (Stevenson, Mills, Welin, & Beal, 1998). In a case control study by Krauss and colleagues (2 005), patient risk factors that were significantly correlated with an increased ris k of falling included: gait or balance deficits, confusion, activity level, use of sedativ es and/or hypnotics, antiarrhythmic, nonnarcotic analgesics, and diabetes medications. T he aim of this study was to identify possible risk factors of falling in the hospital an d describe the events leading up to the fall. Ninety-eight inpatient falls were matched to three hundred and eighteen controls that had the same length of hospitalization stay until t he index fall occurred. Environmental circumstances revealed that 82% of patients fell in their rooms, while 85% of those falls occurred without needed assistance, and the majorit y of patient activity prior to falling pertained to toileting needs. Further, it was noted that 42% of patients who used assistance devices at home were actually using one in the hospital compared to 53% of the controls. Care related risk factors indicated t he higher the number of patients a nurse had, the more likely one of their patients would fa ll. These investigators recommended that nursing efforts should be focused on making sc heduled rounds to offer ambulation or assistance for toileting, utilizing bed rails, impr oving nurse-patient ratios, and monitoring patients taking medications known to increase risk of falls.
8 OHagan and OConnell (2005) conducted a retrospect ive analysis to examine the relationship between patient blood pathology levels and patient falls in the acute-care setting. These researchers used a convenience sampl e of 220 patient charts. These charts were divided among patients that fell and were matc hed on variables of casemix type and length of hospitalization among those patients who did not fall. Included as variables were age, gender, presence of intravenous therapy, history of falls, confusion and continence status as these are documented in the li terature as known patient risk factors. There was no statistical difference found in the bl ood pathology levels for the variables of Na, K, Cl, bicarbonate, urea, creatinine, Ca, P, Mg bilirubin, liver enzymes (ALT, AST), GGT, total protein, albumin, globulin, Hb, RCC (red cell count), PCV, MCHC, MCV, RDW, platelets, ESR, leukocytes, mature neutrophils lymphocytes, monocytes, eosinophils, or basophils. The variable found to be of significance was alkaline phosphatase with p= 0.036. These elevated values ar e typically seen in patients with liver or bone disease and have been present in cancer or orthopedic patients. Chi-squared tests were conducted to determine the relationship of the above mentioned variables with those who fell and those who did not fall. In this study, the only variable of significance was confusion 24 hours prior to the fall with p=0.001. These authors questioned the reliability of fall risk assessment tools as they found the maj ority of the risk variables insignificant in their findings including gender, continence stat us, medications, intravenous therapy, and history of falls. Implications for nursing shou ld include continual and regular patient falls assessment as patients in the acute-care sett ing are prone to frequent changes in acuity (OHagan & OConnell, 2005).
9 In a prospective analysis by Hitcho and colleagues (2004), data on patient characteristics, fall environments, and injury were obtained through interviews with patients and/or nurses and a review of incident rep orts and medical records. Two hundred patient falls were studied with falls occurring in medical, cardiology, neurology, orthopedics, surgery, oncology, and women and infan t services. A total of 183 patients actually fell, 168 (92%) of these patients fell onc e during the investigation, 13 (7%) fell twice, and two (1%) fell three times. The differenc e between falling once and repeatedly falling was gender. In this study, the men tended t o fall more than the women (p=0.03). Confusion (44%), muscle weakness (81%), diabetes (3 9%), urinary frequency (36%), and lower extremity problems (38%) accounted for the ma jority of patient falls. It is likely that medications may have contributed to these fall s as well. Those who received central nervous system (58%) or vasoactive/blood pressure ( 56%) medications 24 hours prior to the fall were at risk for falling; whereas patients who received a sedative were least likely to fall at 12%. The majority of falls occurred at n ight (59%), in the patients room (85%), and with no assistance (79%). The medicine and neur ology units had the highest fall rates both at 6.12 falls per 1,000 patient days and had h igher nurse patient ratios at 6.5 and 5.3, respectively. The oncology unit had the third highe st rate of falls at 3.75 falls per 1,000 patient days with an average nurse-patient ratio of 4.6. Falls were more likely to occur between day 10 and 11 of the hospital stay on the o ncology floor with a range of (1 to 38) (Hitcho, et al., 2004). This is of particular impor tance for oncology nurses as these patients may have an increased risk of falls around this time due to chemotherapy and anti-emetic side effects, radiation, and treatment related fatigue.
10 Dharmarajan, Avula, and Norkus (2007) studied 362 a mbulatory older adults hospitalized from June 2001 through December 2004. The purpose of their study was to identify whether a relationship exists between anem ia laboratory values and the occurrence of falls in older individuals in long te rm care and community settings. The sample ranged in age 59 to 104 years old and includ ed 166 males and 196 females. A majority of these patients (210) lived in the commu nity and the remaining 152 were nursing home residents. Patient demographics includ ed history of prior falls, Hgb, Hct, serum iron, total iron binding capacity (TIBC), fer ritin, serum B12, and serum folate were collected (if available) from medical records. The authors used the World Health Organization (WHO)s definition of anemia defined a s Hgb < 12.0 in females and Hgb < 13.0 in males. These investigators did not find a s ignificant difference between gender distribution, Hgb level, proportion of anemia, leng th of hospital stay, and serum albumin levels in patients from nursing homes and the commu nity. Interestingly, Hispanic patients were reported to have significantly higher mean hem oglobin levels (p=.018) and significantly less anemia (p=.001) than African Ame rican, white, and Asian patients. It was also noted that anemia occurred more frequently in fall patients compared with controls (p=.001). Further, they found patients had a 22% decreased risk for falls for every 1.0g/dL increase in Hgb (p<0.001) and 1.9 fol d increased risk for falls in the presence of anemia (p=0.008) (Dharmarajan, et al., 2007). Risk Factors and Falls in Oncology Patients OConnell, Cockayne, Wellman, and Baker (2005), con ducted a prospective study to examine fall risk factors and the circumstances of patient falls in oncology and palliative care settings. Two hundred and twenty-se ven patients and twenty-four nurses
11 participated in this study. Of this sample group, 3 4 patients fell and 193 did not fall. The measures utilized in this study were the Eastern Co operative Oncology Group (ECOG) performance scale, total confusion score, orientati on to person, time, and place score, muscle strength test, history of prior falls, and s elf-rated fatigue level. The post-falls questionnaire was answered through patient intervie w and nurse interview or selfadministration. Results of this study indicated tha t fallers had a significantly higher mean age of 74.79 compared to 66.45 (p=0.000). The perce ntage of female fallers was 55.9% (n=19) and non-faller females were 50.3% (n=97). Ch i-square tests indicate there was no significant gender difference between fallers and n on-fallers. Chi square tests were used to determine significance of previous history of fa lls and it was determined that fallers and non-fallers were equally likely to have a fall in the preceding 12 months. It was determined that fallers had reduced physical functi oning and were more likely to have answered incorrectly on one of the questions on ori entation compared to non-fallers. The self-rated fatigue level of fallers was significant ly higher than those of non-fallers (p=0.01). These investigators questioned nurses on what could have caused the fall, and reasons they cited included poor condition of the p atient, patient lack of knowledge regarding use of equipment, lack of nursing assista nce, ambulating factors, and elimination issues. Length of hospitalization for t hese oncology patients ranged from 1 to 33 days with a mean length of stay at 11.38 days. T his study identified three fall risk factors not previously mentioned in the literature: ECOG performance scale, muscle strength test, and self-rated fatigue score. It was recommended by these authors that on admission, oncology and palliative care patients l evels of physical and cognitive function be assessed by the ECOG performance scale, the self-rated fatigue scale, muscle
12 strength test, and the orientation in person, place and time score. These items would be very helpful in a falls risk assessment tool for on cology and palliative care settings (OConnell, Cockayne, Wellman, & Baker, 2005). In 2007, nurses in Australia studied patient risk f actors associated with an increased incidence of falls in oncology and medica l settings (OConnell, Baker, Gaskin, & Hawkins, 2007). The purpose of their study was to analyze whether items on the current Falls Risk-Assessment Tool (FRAT), which me asures cognition and physical functioning levels, were truly indicative of distin guishing between fallers and nonfallers in oncology and medical settings. Numerous FRAT too ls utilized across nurse practice settings have limited application for specialty pop ulations, such as oncology. Oncology patients have occasionally been included in falls s tudies; however, they are not the focus of falls research interest. This study was retrospe ctive (reviewed falls within the past 12 months and whether they occurred in the community o r hospital) and the prospective portion studied current hospital patients and fall status. The retrospective group included 184 male and 193 female patients with ages ranging from 23 to 97 years (M = 73). The prospective portion contained 14 men and 20 females with ages ranging from 46 to 89 (M = 77). Twenty-nine patients of the prospective comp onent were medical patients and the remaining five were oncology patients. For this study, the nurses used a FRAT that includ ed sections on demographics, prior history of falls, continence issues, physical functioning, confusion, orientation (person, place, and time), muscle strength, and fat igue. Demographics were recorded from the patients chart and patients were asked if they had fallen within the past 12 months. Physical functioning was determined through the ECOG scale with scores
13 ranging from 0-4. Lower scores indicated greater ph ysical functioning which are as follows: 0 fully active, 1 somewhat restricted, 2 a mbulatory and able to perform activities of daily living (ADLs), restricted to bed or chair > 50% of hours awake, and 4 completely disabled and not able to perform ADLs or ambulate. The bedside confusion item was analyzed through the modified mini-mental state exa mination. Patients were asked to recite the months of the year in reverse order. Thi s was scored 0-4, with 0 being normal and 4, inability to perform. Orientation was measur ed by asking the patient to repeat his or her name, the year and month, and location. The correct responses were summed and scores could vary between 0 and 4, with the lower s cores indicating higher degrees of orientation. Muscle strength was evaluated as 1, 2, or 3 indicating firm, medium, or weak, respectively. Finally, fatigue was determined using a symptom distress scale. Patients were asked to quantify their fatigue from the past 24 hours on a 100 point visual analog scale, where 0 indicated feeling very tired and 1 00 represented feeling your best. The study was conducted in a private hospital duri ng a fourteen month period. Research assistants collected demographics from pat ient charts and conducted a 10 minute fall assessment of each patient using the de tailed FRAT described above within 48 hours of hospital admission. From this data, the researchers concluded that a prior history of falls within the past 12 months was not related to whether they had fallen during the current admission. Further, they conclud ed that muscle strength can differentiate between fallers and nonfallers in the hospital setting. They agreed that those who fell tended to be more confused and less orient ed than nonfallers. Yet, after applying the Bonferroni adjustments, the authors found no si gnificant difference between the FRAT items of fallers and nonfallers (OConnell, et al., 2007).
14 Overcash (2007), conducted a descriptive, prospecti ve, and quantitative study aimed at exploring the incidence of falls that occu r in community-dwelling older oncology patients and how these falls relate to sco res on a comprehensive geriatric assessment (CGA) consisting of depression, age, fun ctional status, and cognition screening instruments. The sample consisted of 165 oncology outpatients aged 70 years or older at a cancer center. The four measures that were utilized in this study are Activities of Daily Living (ADL), Instrumental Acti vities of Daily Living (IADL), Geriatric Depression Scale (GDS), and the Mini-Ment al State Examination (MMSE). Each patient was screened once using the CGA instru ment. The relationship among falls and each of the four scales was determined by using a point biserial correlation, and a multiple regression analysis aided in constructing a model to predict falls in this patient population. The most significant correlation was be tween IADL total scores and falls. These scores were found to be predictive of falls w hen controlling for age and gender using multiple logistic regression. For instance, a score of 9 of a possible score of 24 suggests an 81% risk of a fall compared to a score of 17 of 24 which suggests a 43% risk of a fall. The author concluded that there is limit ed research in the area of falls and oncology patients which made comparisons in the lit erature difficult. In the future, research needs to include an increased sample size and study more potential risk factors in oncology patients such as sensory deficits, anem ia, fatigue, and medications (Overcash, 2007). Falls Assessment Tools Kim, Mordiffi, Bee, Devi, and Evans (2007), conduct ed a prospective descriptive design in an acute care hospital in Singapore to ev aluate the sensitivity, specificity,
15 positive predictive value (PPV), and negative predi ctive values (NPV) of the Morse Fall Scale (MFS), St. Thomas Risk Assessment Tool in Fal ling Elderly Inpatients (STRATIFY), and the Hendrich II Fall Risk Model (HF RM). One hundred and forty-four patients were studied for the inter-rater reliabili ty study from the medical (38.9%), oncology (36.1%), and surgical (25%) units. These p atients were screened within 24 hours of admission during the week and on the next business day for Saturday, Sunday, or public holiday admissions. Sensitivity measures the actual number of patients with high-risk scores who fell divided by the total numb er of patient falls. Specificity measures the actual number of patients with low-ris k scores who did not fall divided by the total number of patients who did not experience a fall. Literature review reveals, the sensitivity of the MFS ranged between 72% and 83% w ith the specificity ranging between 29% and 83%, the sensitivity of STRATIFY ra nges from 54% to 93% depending on the cutoff score, and the HFRM sensiti vity and specificity were 74.9% and 73.9% respectively. Results of the validity study, which included 5489 patients, revealed that even though MFS and STRATIFY specificities wer e high, the low sensitivity renders these as ineffective tools to predict patients at h igh risk of falling. This study indicated that HFRM with a sensitivity of 70% and specificity of 61.8% appeared to be the more appropriate tool for predicting patients at risk fo r falling (Kim, et al., 2007). Nurse researchers in Taiwan conducted a quasi-exper imental study to determine the effectiveness of fall prevention among hospital ized patients based on a modified Fall Risk Factors Assessment Tool. Upon review of 108 pa tient fall cases between 1996 and 2001, these nurses discovered that falls actually o ccurred more often in patients with a fall risk factor of less than three points (56.5%) than the patients who actually received a
16 higher score of more than three points (43.5%). It is possible that the high risk patients did not fall because the nurses were more conscious of the higher score and may have adopted fall prevention interventions, causing the falls rate to decrease. These chart audits revealed that the current hospital assessment tool was not as accurate for high risk fall patients. The control group (n=43) contained patien ts who had fallen before the implementation of the new FRAT between March 1 thro ugh November 30, 2001. The same dates (March 1 November 30, 2002) were used a year later for the use of the experimental group (n=39) who had fallen after the implementation of the new FRAT (Hsu, et al. 2004). After reviewing the literature, the authors of thi s study added balance, lower limb muscle strength, and the will of patient getti ng off bed to their current FRAT which also assessed patients for consciousness level, wal king capability, self care level, history of falls, and medicine administered throughout hosp italization. Further, information was collected for the experimental group to determine a ctivity patterns prior to the fall and reasons why patients did not utilize call lights. B oth groups were more likely to fall while they were sober (as opposed to sleepy or lost ), weak/needing support, and while attempting to get out of bed. The majority of the f alls occurred during evening and night time shifts (3pm to 7am) even though patients had f amily or friends present. This study further supports research conducted by OConnell, e t al. (2007), by indicating prior history of falls is not a good predictor of patient s likelihood of falling during the current hospital admission. Fifty-one percent of the contro l group and sixty-seven percent of the experimental group indicated they had no history of falling in the past; however, they proceeded to fall during current hospitalization. T he control group had an unequal
17 representation of fallers with all scores which ind icated the old FRAT is not a good predictor of the likelihood of patient falls. Afte r the implementation of the new FRAT, the experimental group experienced less patient fal ls correlating with lower scores and an increase in falls as the FRAT scores increased. Pat ients walking ability and fall assessment scores signify significant differences ( p < .01). Interviews with these patients indicate that they did not want to bother their nur se or caregiver and tried to help themselves (Hsu, et al. 2004). Statistical data revealed that 71.4% of patients i n the experimental group had been appropriately identified by nurses as being at a hi gher risk for falls due to poor balance and weaker lower limb muscle strength. Further, the modified risk assessment tool identified high risk patients 74.4% of the time com pared to the control group and the use of the old assessment tool at 60.5%, with the avera ge assessment score increasing from 2.74 to 3.64. The correlation between falls and ass essment scoring showed statistically significant differences (p < .01), signifying that the modified Fall Risk Factor Assessment Tool is better than the original FRAT at correctly identifying high risk patients. While this study adds another perspective to patient fall risk factors, the authors agree that further research needs to be conducted in this area and the falls assessment tool can always be improved (Hsu, et al. 2004). Synthesis of Literature Numerous research studies indicate that risk facto rs for falling in the general population include: history of falling, increased a ge, co-morbidities, gender, altered mental status, gait characteristics, incontinence, length of hospitalization > 18 days, medications (sedatives/hypnotics, antiarrhythmics, nonnarcotic analgesics, antidiabetics),
18 and abnormal lab values (Morse, et al., 1989; Steve nson, et al., 1998; Krauss, et al., 2005; OHagan & OConnell, 2005; Hitcho, et al., 2004; Dh armarajan, et al., 2007). Tommasini, Talamini, Bidoli, Sicolo, and Palese (20 08) noted that 21% of elderly patients were diagnosed with fevers prior to fallin g in the hospital. Research by Hitcho, et al. (2004) noted that the oncology unit had the thi rd highest rate of falls at 3.75 falls per 1,000 patient days. Falls were more likely to occur between day 10 and 11 of the hospital stay on the oncology unit with a range of (1 to 38) (Hitcho, et al., 2004). This is of particular importance for oncology nurses as this m ay mean that these patients have an increased risk of falls around this time due to che motherapy, radiotherapy, treatment related fatigue, and anti-emetic side effects. OCo nnell, Cockayne, Wellman, and Baker (2005) noted that hospitalization for oncology pati ents also ranged from 1 to 33 days with a mean length of stay at 11.38 days. Their study id entified three fall risk factors not previously mentioned in the literature: performance status as measured by the ECOG score, muscle strength test, and self-rated fatigue score. These items might be very helpful in a falls risk assessment tool for oncolog y and palliative care settings (OConnell, Cockayne, Wellman, & Baker, 2005). Furt her, OConnell, Baker, Gaskin, and Hawkins (2007), concluded that muscle strength can differentiate between fallers and nonfallers in hospital oncology and medical patient s. Overcash (2007) found a significant correlation between Instrumental Activities of Dail y Living (IADL) total scores and falls. Research indicates that the Morse Fall Assessment Tool has a sensitivity of 72%83% with specificity ranging between 29%-83%. Resul ts of a validity study conducted by Kim, et al. (2007), which included 5489 patients, r evealed that even though the Morse Fall Scale (MFS) specificities were high, the low s ensitivity renders this as an ineffective
19 tool to predict patients at high risk of falling. I n addition, the Morse Fall Scale measures patients history of falling. Multiple authors ques tion the reliability of fall risk assessment tools as they have found two risk variables insigni ficant in their findings such as gender and history of falls (OHagan & OConnell, 2005; O Connell, Cockayne, Wellman, & Baker, 2005; OConnell, et al., 2007; Hsu, et al., 2004). While nursing studies add additional perspectives to patient fall risk factor s, the authors agree that further research needs to be conducted in this area and that falls a ssessment tools can always be improved (Hsu, et al. 2004). Further, Stevenson and colleagu es concluded that no two studies have found the same set of fall risk factors. Therefore, it is suggested that risk factors for falls are patient population specific and each hospital s hould evaluate and modify fall risk factors on a continuing basis (Stevenson, Mills, We lin, & Beal, 1998). Currently no studies on risk factors and falls hav e been reported in the bone marrow transplant population. The majority of these patients receive high-doses of chemotherapy or high-dose chemotherapy with total b ody irradiation prior to transplantation. Bone marrow transplant patients ar e among the most acutely ill oncology patients and require excellent clinical care and mo nitoring. Nurses need to have a better understanding of patient risk factors and ways to p revent patient falls. A retrospective chart review may make a contribution to oncology fa lls research and improve the quality of life in these patients.
20 Chapter III Methods The purpose of this study was to evaluate the valid ity of the use of Morse Fall Assessment Tool for Bone Marrow Transplant (BMT) pa tients by known group comparisons and exploration of other variables that might influence fall outcomes. This chapter outlines the research methods. First the sa mple and setting are described. Variables included in the chart audit are then disc ussed. Research procedures and approval by the University of South Florida and Flo rida Hospitals Institutional Review Boards are then discussed. This section concludes w ith data analysis and methods to answer the objectives of this study Sample This study was conducted on a Bone Marrow Transpla nt Unit (BMTU) at a large, public, metropolitan, hospital in Florida. Total en umeration was utilized to include the most recent 30 patients on the BMTU who had fallen and a matching group of 30 of these patients, randomly selected from BMT non-fallers wi thin the same time frame. According to Hitcho, et al. (2005), the mean number of days in the hospital prior to falls in oncology patients is 10.5. For comparison purpos es, the Morse Fall Score of the nonfallers were collected within 24 hours of day 11 of admission. Patients under the age of 18 and falls associated with physical therapy sessi ons were excluded.
21 Instruments A chart audit form (Appendix A) was developed for this study based on the work of Morse, et al. (1989), Krauss, et al. (2005), Hit cho, et al. (2005), OConnell, et al. (2005), Dharmarajan, et al. (2007), OConnell, et a l. (2007). Demographic variables audited included age, gender, diagnosis, history of falls, confusion, and muscle weakness. Blood laboratory variables included white blood cou nt, platelets, hemoglobin (Hgb), hematocrit (Hct), sodium (Na), chloride (Cl), pota ssium (K), glucose, bicarbonate, carbon dioxide, albumin, calcium, blood urea nitrog en (BUN), creatinine, and alanine transaminase (ALT), aspartate aminotransferase (AS T), alkaline phosphatase (alk. phosphatase), bilirubin (liver enzymes), blood pres sure, temperature, medications taken within 24 hours of fall, and the Morse Fall score ( Appendix B). Inter-rater reliability was utilized to evaluate th e consistency of the chart audit form. An additional hospital employee used the same chart audit form to collect information from five patient charts from each grou p. These forms were then compared to the principal investigators forms to determine inter-rater reliability. This guided whether any changes needed to be made to clarify it ems on the chart audit. Validity was initially built in by the process of c arefully basing the audit form on variables found in the literature. In addition, com parison of known groups provided evidence of validity of the Morse Fall Scale in bon e marrow transplant patients. The scores of fallers and non-fallers, who were expecte d to have contrasting scores on this scale, were compared (Burns & Grove, 2005).
22 Procedures First, approval was sought from the Nursing Scienc e Review Board at Florida Hospital. Following approval, it was reviewed by th e Institutional Review Board (IRB) at Florida Hospital. The letter of approval from Flori da Hospital was appended to the application for the IRB at the University of South Florida (USF). The IRB granted exempt status so this study could be exempt for the requirement for signed consent. This is the usual procedure for retrospective chart revi ews. After approval by the USF IRB, the Risk Manager wa s contacted to identify the most recent 30 bone marrow transplant patients who had fallen and had Morse Fall Scores recorded within 24 hours before the fall. Pa tient unique identifier numbers and pertinent information from the incident report were recorded and taken to medical records for further demographic information. Further, demog raphics and Morse Fall Scores were collected on 30 randomly selected bone marrow trans plant patients who had not fallen. For comparison purposes, the Morse Fall Score of th e non-fallers was collected within 24 hours of day 11 of admission based on research by H itcho et al. (2005). Data Analysis Demographic data was analyzed using frequencies, p ercentages, means, and standard deviations. To answer objective one, if th ere was a significant difference in the Morse Fall Assessment Score in BMT patients who fal l and those who do not fall, independent t-tests were utilized to determine if t here was a significant difference in the Morse Fall Assessment Score in BMT patients who fal l and those who do not fall. To answer objective two, if there were significant differences in age, gender, diagnosis, history of falls, confusion, muscle weak ness, blood laboratory values, blood
23 pressure, temperature, and medications taken within 24 hours between fallers and nonfallers, independent t-tests were utilized to deter mine if there were significant difference in age, blood laboratory values, and temperature be tween fallers and non-fallers. Differences in gender significance, diagnoses, hist ory of falls, secondary diagnosis, ambulatory aids, intravenous administration, confus ion, and muscle weakness were determined with chi-square tests. The level of stat istical significance for this study was set at p < 0.05.
24 Chapter IV Results, Discussion, and Conclusions This chapter presents the findings of the retrospe ctive study. The results, discussion of the results, limitations of the study implications for nursing and research, and conclusions are presented. Results Demographic Data The sample consisted of a total of 59 patients, wh ich included 29 fallers and 30 non-fallers, 22 males and 37 females, ranging in ag e from 20 to 70 years with a mean age of 53.9 (SD= 12.2) (Table 1). An independent t-test was utilized to determine if there was a significant difference in age among fallers and n on-fallers. No significant difference was found. Of the patients who fell (n=29), three p atients had repeated falls (Table 2). The most common diagnoses in this sample were: acut e myeloid leukemia (AML), multiple myeloma (MM), and non-hodgkins lymphoma (N HL) (Table 3). Table 1. Mean and Percent of Demographic Variables for Fallers and Non-Fallers. Mean Age % of Females % of Males Fallers 55.6 55 45 Non-Fallers 52.1 70 30
25 The majority of falls occurred during day shift ho urs of 7am to 7pm (55%) and the night shift experienced 45% of those falls. The mean number of days in the hospital prior to a patient experiencing a fall on the bone marrow transplant unit was 14.5. Table 2. Frequency and Percent of Falls among Falle rs Number of Falls Frequency Percent One Fall 26 89.7 Two Falls 2 6.9 Three Falls 0 0 Four Falls 1 3.4 Table 3. Frequency and Percent of Cancer Diagnoses a Type of Cancer Frequency Percent Acute Myeloid Leukemia 15 26.3 Multiple Myeloma 12 21.1 Non-Hodgkins Lymphoma 1 0 17.5 Hodgkins Disease 2 3.5 Chronic Myeloid Leukemia 3 5.3 Acute promyelocytic leukemia 2 3.5 Aplastic Anemia 1 1.8 Chronic Lymphocytic Leukemia 1 1.8 T-Cell Leukemia 3 5.3 Acute Lymphocytic Leukemia 7 12.3 Myelodysplastic Syndrome 1 1.8 a Diagnoses (n=57, missing data on 2 patients) There were 22 males and 37 females in this study. T he fallers consisted of 16 females and 13 males with the non-faller group cons isting of 21 females and 9 males.
26 Chi-square comparison of gender revealed that there is no statistically significant difference between fallers and non-fallers. Morse Fall Scale The Morse Fall Scale measured the following six va riables: history of falling, secondary diagnosis, use of ambulatory aids, IV the rapy, gait characteristics, and mental status. Scores for the MFS can range from 0 through 125. If patients have a score of 0-24 they are at low risk for falling, 25-49 gives patie nts a moderate risk, and > 50 puts patients at a higher risk of falling. In this retro spective study, fallers (n=29) had a mean MFS of 43.79 (SD= 20.644) and the non-fallers (n=30 ) had a mean MFS of 26.83 (SD= 12.421). Independent t-tests were utilized to deter mine if there was a significant difference in the Morse Fall Assessment Score in BM T patients who fell and those who did not fall. There was a statistically significant difference between patients who fell and who did not fall (t=3.839; p=0.000). The individual variables on the MFS that showed si gnificant differences between groups in this study were: history of falls (p=0.04 2), secondary diagnosis (p=0.015), and muscle weakness (p=0.025) (Table 4). Difference in groups according to use of intravenous therapy, use of ambulatory aids, and co nfusion were not significant.
27 Table 4. Chi-square comparisons of Morse Fall Scale Variables between Fallers and NonFallers. Faller Non-Faller Chi-Square p History of Falls 8 2 4.6 0.042 Presence of Secondary 15 6 6.5 0.015 Diagnosis Ambulatory Aids 3 0 3.3 0.112 IV Therapy 29 30 (NS)a (NS)a Confusion 5 1 3.1 0.103 Muscle Weakness 13 5 5.5 0.025 a Not Significant, no variation in these scores, ever y patient had an IV Temperature The majority of bone marrow transplant patients we re afebrile during this chart review. The mean temperature maximum for both falle rs and non-fallers was 99.4 (SD= 1.6) and 99.5 (SD= 1.3), respectively. Temperatures greater than 100.5 occurred in 12 patients (n=59). The faller group had five fevers p resent 24 hours prior to falling (17.2%) and seven non-fallers had fevers present within 24 hours of day 11 of the chart review (23.3%). Table 5. Temperature in Bone Marrow Transplant Pati ents. Tmax Mean Tmax SD Temp > 100.5 (Frequency) Perce nt Fallers 99.4 1.6 5 17.2 Non-Fallers 99.5 1.3 7 23.3 Tmax = Maximum Temperature within 24 hours Laboratory Values For this study, laboratory results from complete b lood counts (CBC), complete metabolic profile (CMP), magnesium, and phosphorous were collected from both fallers
28 and non-fallers. The platelet count (p=0.003), BUN (p=0.032), glucose (p=0.009), and phosphorous (p=0.001) were found to be significantl y different between fallers and nonfallers (Table 6). Table 6. Means, t -Scores, and Significance of Laboratory Values of F allers and NonFallers. Fallers Non-Fallers Mean Mean t p Hemoglobin 9.86 9.38 1.54 0.129 Hematocrit 28.89 27.36 1.64 0.108 White Blood Count 9.16 0.91 1.6 0.115 Platelets 90.11 31.83 3.1 0.003 Sodium 136.74 137.73 -1.01 0.316 Potassium 3.97 3.84 1.11 0.272 Chloride 104.89 106.17 -1.04 0.303 Carbon Dioxide 25.19 25.07 0.15 0.885 Blood Urea Nitrogen 15.48 10.37 2.2 0.032 Creatinine 0.87 0.82 0.25 0.807 Glucose 135.67 11 0.67 2.69 0.009 ALT 38.65 35.40 0.42 0.674 AST 30.96 31.13 -0.03 0.974 Alk. Phosphatase 85.27 77.60 0.76 0.449 Bilirubin 0.91 0.94 -0.25 0. 802 Albumin 2.7 2.85 -1.27 0.210 Calcium 8.22 8.44 -1.39 0.17 1 Magnesium 1.78 1.84 -1.14 0.258 Phosphorous 2.54 3.85 -3.71 0.001
29 Inter-Rater Reliability Inter-rater reliability was conducted to compare t he results of the chart audits between the principal investigator and the lead res earch scientist at this hospital. A total of ten chart audits were compared for inter-rater r eliability, with five chart audits from each group. Chemotherapy, adjunct medication, and d iabetic medication were excluded from the total list of items since these proved dif ficult to obtain during the study. The principal investigator included 18 items from the c hart audit for purposes of inter-rater reliability. The following items counted as one poi nt on the chart audit: category of patient (faller versus non-faller), days from admis sion, date and time of fall, age, gender, diagnosis, blood pressure, temperature, Morse Fall Scale score, primary hematology labs (hemoglobin, hematocrit, white blood count, and pla telets), part of the CMP laboratory levels (discussed in the order they are abbreviated from top to bottom sodium, potassium, chloride, carbon dioxide, BUN, creatinine, and gluc ose), ALT, AST, ALP, bilirubin, albumin, calcium, and magnesium. The scores of the five fallers charts were average d together to receive an interrater reliability score of 11.8 out of a possible s core of 18. The scores of the five nonfallers charts were also averaged together to recei ve an inter-rater reliability score of 4.6 out of the possible score of 18. Discussion This study may serve to raise awareness of numerou s patient falls in the oncology setting and particularly in the BMT setting. Falls are a continued concern for nurses in all settings especially as it is the focus of the Joint Commissions National Patient Safety
30 Goals Hospital Program (2008). Early fall identification and prevention should be the goals of all nurses in their practice setting. Demographic Data A list of patients who fell on the Bone Marrow Tra nsplant Unit was obtained from risk management from January 2007 through January 2 009. A comparison group of nonfallers was randomly selected for this same time pe riod. The main limitation of this retrospective study is the relatively small size of the Bone Marrow Transplant Unit. This was not a problem in this study, as the principal investigator had 59 patients from the desired sampl e size of 60. However, 59 is a relatively small sample size. Limitations can exist with the majority of these patients residing in the same geographical area. Further res trictions exist if previous patient falls were not documented in the hospital-wide Incident R eports. Reporting incidents is strongly encouraged; however, it is not always comp leted. This can result with some falls going unexamined during this time frame. An additio nal limitation may have occurred due to the design of the study. Non-fallers were as sessed on day 11, based on results of previous research but this study found that the mea n number of days when patients fell was 14.5, which is around the time they reached nad ir. The sample largely consisted of females (62.7%) wi th the three most prominent diagnoses seen were acute myeloid leukemia (AML), m ultiple myeloma (MM), and nonhodgkins disease (NHL). Consistent with studies con ducted previously, (Hsu, et al., 2004; OConnell, et al., 2005; OHagan, et al., 200 5; Overcash, 2007) there is no significant gender difference between fallers and n on-fallers.
31 There was no distinctive pattern or events likely linked to falls during the day shift. Patients appeared to fall inadvertently and sporadically throughout the 12 hour shift. However, 77% percent (n=10) of falls on the night s hift occurred between 1:30am and 6:20am. This may be the result of patients waking i n the middle of the night for toileting needs or carrying out activities of daily living wh ile unable to sleep in the hospital environment. Morse Fall Scale The Morse Fall Scale is indeed reliable and valid i n the bone marrow transplant population at this hospital. This was supported by evidence that there was a significant difference in Morse Fall Scale scores between falle rs and non-fallers. Further study of the validity of the MFS for use in BMT and whether addi tional variables might be needed is recommended. For other hospitals and patient popula tions, it is recommended to review the literature on several fall assessment scales an d review the circumstances relating to each units fall characteristics. When individual items were evaluated, history of fa lls was found significant in this bone marrow transplant population. In this stu dy, three patients had repeated falls. The first patient was initially identified on the M orse Fall Scale as being at moderate risk prior to falling. On this patients subsequent admission he was correctly identified as being at high risk for falls due to a past histor y of falling. Two other patients were initially identified as high risk patients for fa lling, which also supports validity of this scale in this population. However, on subsequent ad missions these patients were not entered on the MFS as having a history of falling. Yet, history of falls was found
32 insignificant in earlier studies (OConnell, Cockay ne, et al., 2005; OConnell, et al., 2007). Secondary diagnosis was found to be significant in this retrospective chart review. This is consistent with a case control study by Kra uss and colleagues (2005) who found patient risk factors that were significantly correl ated with an increased risk of falling included the use of sedatives and/or hypnotics, ant iarrhythmic, nonnarcotic analgesics, and diabetes medications. Muscle weakness was also found to be significant i n the bone marrow transplant population. This is consistent with analysis by Hit cho and colleagues (2004), who found muscle weakness contributed to 81% of patient falls In addition, Hsu and colleagues (2004), found that 71.4% of their patients in an ex perimental group had been appropriately identified by nurses as being at a hi gher risk for falls due to poor balance and weaker lower limb muscle strength. Other contributing factors Temperatures Temperatures were assessed to determine if there w as a significant difference between fallers and non-fallers. In this study, no significant difference was found with elevated temperatures; however, it is important to note that some patients become confused and weak with elevated temperatures in thi s population. In the non-faller group, patients had a decreased white blood count (WBC) wi th a mean of 0.91 with seven patients becoming febrile of day 11. This may be in dicative of reaching nadir (decreased blood counts) related to chemotherapy. At nadir pat ients are less able to mount a fever, which is why oncology nurses monitor for fevers of 100.5. Future studies should consider
33 temperatures as a study variable in falls research to determine if fevers are a contributing factor to falls. Laboratory Values In this study, platelet count, glucose, blood urea nitrogen (BUN), and phosphorous showed significant differences between fallers and non-fallers. The mean platelet count for fallers was found to be higher i n fallers as compared to non-fallers with M = 90.11 and M = 31.83, respectively. The reasons for these diffe rences are thought to be that fallers were most likely further out from t reatment (i.e. chemotherapy or bone marrow transplant) as compared to the non-fallers. The non-fallers were more likely to be at their nadir (lowest point in their blood counts) and may have been monitored and assessed more frequently. The lower platelet count in the non-faller group may illustrate that these patients were more likely to be receivin g active treatment, which can cause a decrease in platelet count known as thrombocytopeni a. This may be confirmed by the difference in white blood count (WBC) in both group s. The fallers had recovered their WBC with a mean of 9.16 as compared to the non-fall er group whose mean was 0.91. A lower white count can be indicative of current trea tment and more frequent nursing assessments. Glucose levels were found to be higher in fallers ( M =135.7) as opposed to nonfallers ( M =110.7). The reasons for this are unclear; however, higher glucose levels are typically found in sicker patients (i.e. infection) or patients who have steroids prescribed for them. Blood urea nitrogen (BUN) was also found to be mor e elevated in fallers ( M =15.48) as compared to non-fallers ( M =10.37). These levels are within the normal
34 ranges of BUN, 7 to 25. Patients with elevated BUN levels often are dehydrated or have compromised renal function. It is unclear what role this might have played in falls; further research is indicated. Phosphorous levels were found to be decreased in f allers ( M =2.54) when compared to non-fallers ( M =3.85). Potential reasons for a lower phosphorous l evel include nausea, vomiting, and diarrhea. These sympt oms often result in the need to ambulate frequently to a bedside commode or bathroo m. Consistent with findings from OHagan and colleagu es (2005) and Morse and colleagues (1989), this study revealed that there i s no statistical difference in hemoglobin levels of fallers and non-fallers in the bone marro w transplant population. Hemoglobin levels were found to be relatively similar in falle rs and non-fallers, M =9.86 and M =9.38, respectively. These values both represent anemia; h owever, this is a common finding in the oncology population. Future studies should cons ider having non-oncology patients as a control group to determine hemoglobin significanc e. Future findings may shed light on future nursing implications of monitoring laborator y values more closely. Vital Signs and Medications In this retrospective study, it was difficult to g ather vital signs and medications taken within 24 hours. During the time frame of Jan uary 2007-January 2009 this hospital changed electronic charting systems, thus making it more difficult to obtain certain information needed for the chart review. Pain medications and sedating anti-nausea medicati on, such as Ativan, are commonly prescribed on this unit. In addition, ster oids are also used in conjunction with
35 chemotherapy regimens and have the tendency to rais e blood glucose levels in these patients; which often requires additional therapy t o maintain normal glucose levels. Inter-Rater Reliability Inter-rater reliability was not ideal and might ha ve been improved by educating the lead nursing research scientist at this hospita l on the use of this chart audit or have a sample chart audit pre-completed as a reference. Di fficulties arose with this nursing research scientist not being familiar with the abbr eviated version of the drawing of part of the CMP. This abbreviated version allows health car e providers to save time. However, it apparently is not universally well known. This part was missed on both the fallers and non-fallers forms. In addition, the age of the pati ent was missed a lot. The principal investigator used the age of the patient at the tim e of the initial fall or during the time of day 11. However, this research scientist used the c urrent age of the patient as she was completing the chart audit. The low scores for the non-fallers resulted from t he research scientist looking at Morse Fall score information and laboratory values from the date of admission. The 11 days were not added to the date of admission to ach ieve accurate results. Subsequently, the directions on the chart audit were amended from stating 24 hours within day 11 for non-fallers to add two words for clarification. The chart audit now states 24 hours within day 11 of admission for non-fallers. The investigat or welcomes future studies to utilize her chart audit or modify it for the purposes of pa tient populations in their setting. It is recommended that all raters discuss the form prior to utilizing it in a study instead of assuming that it is self-explanatory.
36 Implications for Nursing Nursing curriculum should emphasize the importance of identifying and preventing patient falls. Nursing students should b e introduced to the Morse Fall Scale or other falls assessment scales currently in use in t heir local communities. Introduction to such scales brings familiarity, a level of comfort, and the ability to accurately assess patients at high risk for falling. Further, unit-sp ecific education on patient safety and interventions to decrease falls should be the focus of all nursing units. In addition, hospital orientation for new nurses should include case studies on how to correctly score patients. Educating staff closes the gap on leaving fall scales open for interpretation by the user. It is possible that nurses check on patients less frequently during the night shift because they are expected to be asleep. However, 77 % of falls on the night shift occurred between 1:30am and 6:20am, it is suggested that nur ses and nurses aides arrange beds in the lowest position, keep call lights within reach, offer frequent toileting and assistance, and make frequent comfort rounds during these times to help decrease and/or prevent patient falls. Laboratory findings from this study suggest that a ll patients regardless of treatment modality must be monitored more closely. These values can indicate the extent of treatment a patient is receiving and some of the side effects a patient may experience. For instance, patients with lower phosphorous and e lectrolyte levels may present this way from side effects of active treatment or simply fro m nausea, vomiting, or diarrhea while hospitalized. Laboratory values share volumes of in formation with nurses. Nurses need to
37 become familiar with laboratory values and the sign s and symptoms of laboratory values outside of these ranges to truly assess and keep th eir patients safe and free from injury. In this sample, the mean number of days between ad mission and falling in the bone marrow transplant unit was M =14.5, the time when blood counts are lowest. It should be a priority of nurses and nurses aides to encourage ambulation and/or assist with exercise at least twice daily to maintain lowe r extremity strength. Observing ambulatory patients can indicate whether patients m ay have difficulty with their gait and if they could benefit from physical therapy. Being proactive is essential in patient safety. Conclusions This is the first study to document falls in the b one marrow transplant population. This study should serve as a stimulus for future st udies conducted in the oncology and/or bone marrow transplant population. Further studies should explore these patients prospectively utilizing known group comparisons of non-oncology patients with a larger sample. Studying falls in these patients is essent ial to understanding the physiological risk factors that may contribute to patient falls. Findings lay the foundation for studying falls in the bone marrow transplant population. It is crucial to study falls in this population in order to make appropriate assessments and interventions to keep this population free from injury. Recommendations for Future Research Nurses wishing to study falls in the bone marrow t ransplant population should consider doing a prospective study of fallers. A pr ospective study will enable researchers to examine patient characteristics such as co-morbi dities, medications taken within a 24
38 hour period, vital signs, temperature, laboratory v alues, fall assessment scores, and other variables of interest such as nadir.
39 References Burns, N. & Grove, S. (2005). The practice of nursing research: Conduct, critique and utilization (5th ed.). St. Louis, MO: Elsevier. Centers for Medicare and Medicaid Services (2008). Medicare takes new steps to help make your hospital stay safer. Retrieved September 27, 2008, from http://www.cms.hhs.gov/apps/media/press/factsheet.a sp?Counter=3227&intNum PerPage=10&checkDate=&checkKey=&srchType=1&numDays= 3500&srchOpt= 0&srchData=&keywordType=All&chkNewsType=6&intPage=& showAll=&pYe ar=&year=&desc=&cboOrder=date Dharmarajan, T., Avula, S., & Norkus, E. (2007). An emia increases risk for falls in hospitalized older adults: An evaluation of falls i n 362 hospitalized, ambulatory, long-term care, and community patients. Journal of the American Medical Directors Association 8, E.9-E.14. Hitcho, E., Krauss, M., Birge, S., Dunagan, W., Fis cher, I., Johnson, S., et al. (2004). Characteristics and circumstances of falls in a hos pital setting: A prospective study. Journal of General Internal Medicine 19, 732-739. Holley, S. (2002). A look at the problem of falls a mong people with cancer. Clinical Journal of Oncology Nursing (6), 4, 193-197. Hsu, S., Lee, C., Wang, S., Shyu, S., Tseng, H., Le i, Y., et al. (2004). Falls risk factors assessment tool: Enhancing effectiveness in falls s creening. Journal of Nursing Research (22), 3, 169-178. Joint Commission. (2008). 2008 National patient safety goals hospital program Retrieved February 3, 2008, from http://www.jointcommission.org/PatientSafety/Nation alPatientSafetyGoals/08_ha p_npsgs.htm Kim, E., Mordiffi, S., Bee, W., Devi, K., & Evans, D. (2007). Evaluation of three fall-risk assessment tools in an acute care setting. Journal of Advanced Nursing (60) 4, 427-435. Krauss, M., Evanoff, B, Hitcho, E., Ngugi, K, Dunag an, W., Fischer, I., et al. (2005). A case-control study of patient, medication, and care -related risk factors for inpatient falls. Journal of General Internal Medicine (20), 2, 116-122.
40 Morse, J., Black, C., Oberle, K., & Donahue, P. (19 89). A prospective study to identify the fall-prone patient. Social Sciences and Medicine (28), 1, 81-86. OConnell, B., Baker, L., Gaskin, C., & Hawkins, M. (2007). Risk items associated with patient falls in oncology and medical settings. Journal of Nursing Care Quality (22) 2, 130-137. OConnell, B., Cockayne, M., Wellman, D., & Baker, L. (2005). Fall risk factors and the nature of falls in inpatient oncology and palliati ve care settings. Contemporary Nurse (18) 3, 247-257. OHagan, C. & OConnell, B. (2005). The relationshi p between patient blood pathology values and patient falls in an acute-care setting: A retrospective analysis. International Journal of Nursing Practice 11, 161-168. Overcash, J. (2007). Prediction of falls in older a dults with cancer: A preliminary study. Oncology Nursing Forum (34), 2, 341-346. Pearse, H., Nicholson, L., & Bennett, M. (2004). Fa lls in hospices: A cancer network observational study of fall rates and risk factors. Palliative Medicine (18) 478481. Shever, L., Titler, M., Kerr, P., Qin, R., Kim, T., & Picone, D. (2008). The effect of high nursing surveillance on hospital cost. Journal of Nursing Scholarship (40) 2, 161169. Stevenson, B., Mills, E., Welin, L., & Beal, K. (19 98). Falls risk factors in an acute-care setting: A retrospective study. Canadian Journal of Nursing Research (30) 1, 97111. Tommasini, C., Talamini, R., Bidoli, E., Sicolo, N. & Palese, A. (2008). Risk factors of falls in elderly population in acute care hospitals and nursing homes in North Italy: A retrospective study. Journal of Nursing Care Quality (23) 1, 43-49.
42 Appendix A Chart Audit for Fallers & Non-Fallers (Information from 24 hours prior to fall or 24 hour s within day 11 of admission for nonfallers) Hospital: BMTU Day __________ Date __________ Patient Unique ID ________________ Circle: F aller Non-Faller Days from Admission: ___________ Date & Time of Fall: ___________ 1. Age: 2. Gender : Male Female 3. Diagnosis: 4. Blood Pressure prior to Fall (or ranges on day 11 f or non-fallers): 5. Temperature prior to Fall (or Tmax on day 11 for no n-fallers): The following information should be found on the Mo rse Fall Assessment Scale. 6. History of Falls: Yes No 7. Presence of Secondary Dx: Yes No 8. Use of Ambulatory Aids: Yes No 9. Current IV Administration: Yes No 10. Confusion: Yes No 11. Muscle Weakness: Yes No 12. Morse Fall Score:
43 The following laboratory values and medication shou ld be collected within 24 hours of the patient fall (or day 11 for non-fallers). ALT ______ AST ______ ALP ______ Bilirubin _____ Albumin _______ Calcium ______ Magnesium ______ Phosphorous ______ 13. Chemotherapy Medication: Ordered Information Administered Inform ation Medication Dose Route Schedule Dose Route Times Giv en 14. Adjunct Medication (e.g. anti-depressants, sedatives, anti-emetics, di uretics, antiarrhythmics, antihistamines, opioids) Ordered Information Administered Inform ation Medication Dose Route Schedule Dose Route Times Giv en
44 15. Diabetic Medication Ordered Information Administered Informa tion Medication Dose Route Schedule Dose Route Times Giv en
45 Appendix B Morse Fall Scale (MFS) Florida Hospital Orlando (Adapted from VA National Center for Patient Safety Fall Prevention & Management) Variables Numeric Values Score 1. History of falling (immediate or within 3months) Yes 25 No 0 _______ 2. Secondary diagnosis (more than 1 diagnosis is listed on the patients chart) Yes 15 No 0 _______ 3. Ambulatory aid None/bed rest/wheelchair/nurse assist (0) Crutches/cane/walker (15) Furniture (30) (clutching as support) 0 15 30 _______ 4. IV or IV Access Yes 20 No 0 _______ 5. Gait/Transferring Normal/bed rest/immobile (0) Weak (10) Impaired (20) 0 10 20 _______ 6. Mental status (Patient self-assessment of ambulation) Oriented to own ability (0) (realistic to mobility level) Forgets limitations (15) (unrealistic self-evaluation) 0 15 _______ Morse Fall Scale Score = Total ______
46 Morse Fall Scale Scoring: 0-24 = Low Risk Fall Prevention Interventions 25-49 = Moderate Risk Fall Prevention Interventions >=50 High Risk Fall Prevention Interventions
47 Appendix C Letter of Approval from Florida Hospital
48 Appendix D IRB Approval from Florida Hospital
49 Appendix E IRB Approval from the University of South Florida