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 22 Ka 4500
controlfield tag 007 cr-bnu---uuuuu
008 s2010 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0003401
Predicting fear of recurrence and protective health behaviors using protection motivation theory
h [electronic resource] /
by Heather McGinty.
[Tampa, Fla] :
b University of South Florida,
Title from PDF of title page.
Document formatted into pages; contains X pages.
Thesis (M.A.)--University of South Florida, 2010.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
ABSTRACT: Prior research suggests that fear of cancer recurrence is very common among cancer survivors. This study examined the extent to which Protection Motivation Theory variables of threat appraisal and coping appraisal accounted for differences in fear of recurrence and performance of health behaviors in cancer patients who recently completed treatment. It was hypothesized that greater fear of recurrence would be related to a combination of high threat appraisal and low coping appraisal. Also, it was hypothesized that higher rates of health behaviors would be related to higher threat appraisals for cancer recurrence and higher coping appraisals for reducing risk of recurrence by improving diet or exercising. A sample of 155 early-stage breast cancer patients (mean age = 59 years) who completed surgery, chemotherapy, and/or radiotherapy between 6-24 months previously (mean = 12 months) completed measures of fear of recurrence, threat appraisal (perceived risk and severity of a potential cancer recurrence), fruit and vegetable intake in the past month, exercise for the past week, and coping appraisal (perceived response efficacy and self-efficacy to perform diet and exercise recommendations to reduce recurrence risk). Basic demographic and clinical information was also collected. The study findings supported the hypothesis that the combination of threat and coping appraisal beliefs explain which breast cancer survivors report higher fear of recurrence. However, the observed results did not support the hypothesized interaction between threat and coping appraisal for predicting either diet or exercise habits. Instead, coping appraisal alone predicted both fruit and vegetable consumption and exercise habits. Future research should focus on examining these relationships longitudinally and further assess coping appraisal and how it impacts fear of recurrence.
Advisor: Paul B. Jacobsen, Ph.D.
Behavior change theories
t USF Electronic Theses and Dissertations.
Predicting Fear of Recurrence and Protective Health Behaviors Using Protection Motivation Theory b y Heather L. McGinty A thesis submitted in partial fulfillment of the requirements for the degree of Master of Arts Department of Psychology Colleg e of Arts and Sciences University of South Florida Major Professor: Paul B. Jacobsen, Ph.D. Thomas H. Brandon, Ph.D. Jamie L. Goldenberg, Ph.D. Date of Approval: February 23, 2010 Keywords: breast cancer, diet, physical activity, survivorship behavior change theorie s Copyright 2010 Heather L. McGinty
i Table of Contents List of Tables iii List of Figures iv Abstract v Introduction 1 Fear of Recurrence in Cancer Survivors: Conceptualization and Measurement 1 Characteristics of Fear of Recurrence 4 Predictors of Fear of Recurrence 5 Health Behaviors in Cancer Sur vivors: American Cancer Society Recommendations 6 Me asuring Health Behavior Ch ange 8 Prevalen ce of Health Behavior Changes 9 Predicto rs of Health Behavior Changes 1 2 Overview of the Protection Motivati on Theory Model 14 Overview of the Current Study 16 Relationship of Coping and Threat Appraisal to Fear of Recurrence 17 Hypothesis 1 17 Relationship of Coping and Threat Appraisal to Health Behaviors 1 9 Hypothesis 2 1 9 Exploratory Analyses 2 1 Method 2 3 P articipants 2 3 Measures 2 3 Demographic Characteristics 2 3 Clinical Characteristics 2 3 Fear of Recurrence 2 3 Vulnerability 2 4 Severity 2 4 Self Efficacy 2 5 Response Efficacy 2 6 Dietary Behavior 2 6 Exercise Behavior 2 6 Depression 2 7 Procedure 2 7 Statistical A nalyses 2 8 Results 3 2 Participants 3 2 Descriptive S tatistics 3 5 Preliminary Analyses 3 7
ii Evaluation of Hypothesis 1 3 7 Evaluation of Hypothesis 2 4 4 Exploratory Analyses 4 7 Discussion 4 8 Relationship of Threat and Coping Ap praisal to Fear of Recurrence 4 8 R elationship of Threat and Coping Appraisal to Health Behaviors 50 Evide nce for Meditational Pat hways 5 2 Limitations 5 4 Clinical Imp lications & Future Directions 5 5 List of References 5 7
iii List of Tables Ta ble 1. Sample Characteristics 3 4 Table 2. Rel ationships of Demographic and C linical Variables to Outcomes 3 6 Table 3. Correlations of Protection Motivation Theory Variables with Fear of R ecurrence, Diet, and Exercise 3 8 Table 4. Hierarchical Multiple Regression Analyses Predicting Fear of Recurrenc e From Threat, Coping, an d Threat X Coping Interaction 3 9 Table 5. Hierarchical Multiple Regression Analyses Predicting Fear of Recurrence From Vulnerability, Coping, and Vulne rability X Coping Interaction 41 Table 6. Hierarchical Multiple Regression Ana lyses Predicting Fear of Recurrence From Vulnerability, Diet Coping, and Vulnerabil ity X Diet Coping Interaction 4 2 Table 7. Hierarchical Multiple Regression Analyses Predicting Fruit and Vegetable Intake From Threat, Diet Coping, and Threat X Diet Copin g Interaction 4 5 Table 8. Hierarchical Multiple Regression Analyses Predicting Weekly Exercise From Threat, Exercise Coping, and Threat X Exercise Coping Interaction 4 6
iv List of Figures Figure 1. Protec tion Motivation Theory Model 1 5 Figure 2. Rel ationship between PMT vari ables and fear of recurrence. 1 7 Figure 3 Hypothesized relationship between PMT variables and fear of recurrence. 18 Figure 4. Relationship between PMT variables and recommended health behaviors. 19 Figure 5. Hypothesized rela tionship between PMT va riables and health behaviors. 2 0 Figure 6 A. Behavior mediates relationship betwee n PMT and fear of recurrence. 2 2 Figure 6 B. Fear of recurrence mediates relations hip between PMT and behavior. 2 2 Figure 7. Response rate throughou t recruitment and surveying process. 33 Figure 8. Threat Appraisal X Coping Appraisal Interaction. 4 0 Figure 9. Vulnerability X Coping Appraisal Interaction. 4 1 Figure 10. Vulnerability X Diet Coping Appraisal Interaction. 4 3
v Predicting Fear of Re currence and Protective Health Behaviors Using Protection Motivation Theory Heather L. McGinty ABSTRACT Prior research suggests that fear of cancer recurrence is very common among cancer survivors. This study examined the extent to which Protection Moti vation Theory variables of threat appraisal and coping appraisal accounted for differences in fear of recurrence and performance of health behaviors in cancer patients who recently completed treatment. It was hypothesized that greater fear of recurrence w ould be related to a combination of high threat appraisal and low coping appraisal. Also, it was hypothesized that higher rates of health behaviors would be related to higher threat appraisals for cancer recurrence and higher coping appraisals for reducin g risk of recurrence by improving diet or exercising. A sample of 155 early stage breast cancer patients (mean age = 59 years) who completed surgery, chemotherapy, and/or radiotherapy between 6 24 months previously (mean = 12 months) completed measures of fear of recurrence, threat appraisal (perceived risk and severity of a potential cancer recurrence), fruit and vegetable intake in the past month, exercise for the past week, and coping appraisal (perceived response efficacy and self efficacy to perform d iet and exercise recommendations to reduce recurrence risk). Basic demographic and clinical information was also collected. The study findings supported the hypothesis that the
vi combination of threat and coping appraisal beliefs explain which breast cance r survivors report higher fear of recurrence. However, the observed results did not support the hypothesized interaction between threat and coping appraisal for predicting either diet or exercise habits. Instead, coping appraisal alone predicted both fru it and vegetable consumption and exercise habits. Future research should focus on examining these relationships longitudinally and further assess coping appraisal and how it impacts fear of recurrence.
1 Introduction As the medical treatment of cancer advances and survival rates improve, the number of cancer survivors continues to grow. This trend should motivate clinicians and researchers to take a closer look at the long term consequences of diagnosis and treatment in order to be able to handle the i ncreasing demands for post treatment services. A commonly reported concern of cancer survivors is fear of recurrence, the lingering thoughts and concerns that despite successful treatment, the cancer may return as unexpectedly as it had at the original di agnosis. Elucidating the underlying factors that come together to shape these fears would help guide interventions designed to reduce these fears. Typically, risk reducing health behaviors are expected to reduce the risk of recurrence, which in turn may and coping appraisal variables to both fear of recurrence and engagement in health behaviors. The goal is to establish how well these variables explain which survivors are more likely to report both fear of recurrence and adhere to recommended health behaviors. Fear of Recurrence in Cancer Survivors: Conceptualization and Measurement While cancer survivors ca n readily point out fear of recurrence as a prominent concern following completion of active treatment, researchers have had a harder time pin pointing what fear of recurrence consists of as a measurable construct. The concept
2 can be as broad as over arch ing cancer related concerns or fears, involving both current difficulties and worries about the future and changes in prognosis, or as narrow as the specific fear of being diagnosed with the same type of cancer at the same site at some point in the future. More commonly, though, fear of recurrence concerns fears related to being diagnosed with the same type of cancer again at any point after initial diagnosis and treatment. Although the number of studies focusing on fear of recurrence has increased over t he years, there is no preferred method for measuring the construct. Some studies have relied on study specific single item measures that ask cancer survivors how concerned they are that the cancer may recur. An example of this type of single item measure would be a question that asks participants to rate, on a 5 point scale from poor to excellent Lash, Clough Gorr & Silliman, 2005). Some studies have taken fear of recurrence items from larger measures such as the Cancer Problems In Living Scale (CPILS) and analyzed them separately (Baker, Denniston, Smith, & West, 2005). Two CPILS items tha t relate to fear problem, or not a problem (Baker et al., 2005). Similarly, the Question naire on Stress in Cancer Patients revised version (QSC R23) has a single item that may capture fear of evalence, and the extent to which this fear causes distress (Herschbach et al., 2004). These types of single item measures avoid combining fears of recurrence with other more general worries
3 about health. They are limited, however, in that they do not co llect enough information to help determine when fears reach clinically significant levels. Another measurement strategy is to use multi item scales such as the commonly used Fear of Recurrence Scale (Northouse, 1981). This scale is said to be unidimensio nal, but this claim is suspect due to inclusion of items ranging from concerns about current health, to triggers of worry, to recurrence fears, uncertainty, and worries about the health of others (Lee Jones, Humphris, Dixon, & Hatcher, 1997). In light of these difficulties, several scales elect instead to purposely divide cancer related worries into more than one dimension. One such measure is the six item Assessment of Survivor Concerns (Gotay & Pagano, 2007). It contains two factors: one measuring fear of recurrence and another measuring general future health fears that may or may not be linked to cancer, like worrying about death or personal health (Gotay & Pagano, 2007). Another multidimensional scale is the Concerns About Recurrence Scale (CARS; Vic kburg, 2003), designed for women with breast cancer. This scale is made up of five factors that cover a wide range of concerns that could be tied to fear of recurrence: overall fear, health worries, womanhood worries, role worries, and death worries (Vick berg, 2003). Multidimensional scales provide far more information about fear of recurrence, but also risk including items that relate to concerns that are not specific to functioning may change should one become sick. Finally, Rabin, Leventhal and Goodin (2004) devised a measure that modified item Cancer Worry Scale (1991) to capture worry about getting diagnosed with cancer again Rather than focusing on th e consequences and meaningfulness of the
4 fear of recurrence like other multi item scales, this scale focuses on how often these concerns occur and affect the mood and daily functioning of cancer survivors (Rabin et al., 2004). This type of scale has the b enefit of providing more clinically meaningful information about the prevalence and extent of fears while still focusing exclusively on fears related to recurrence and not general health or other consequences of a cancer diagnosis. Unfortunately, this sca le has rarely been used to date in research on fear of recurrence (McGinty, Andrykowski, & Jacobsen, 2008; Rabin et al., 2004). Characteristics of Fear of Recurrence Fear of recurrence is a common concern among all types of cancer patients. Data from the Survivors of Cancer Study I (SCS I) reported that the overwhelming majority of patients were experiencing concerns related to their cancer (Baker et al., 2005). Survivors surveyed one year after diagnosis expressed concerns about their illness returning ( 68.1% of those reporting), having a recurrence (59.8%), and fears for the future (57.7%) (Baker et al., 2005). In a study of women with breast cancer who had completed treatment and were transitioning into the so fear of recurrence as a dominant concern and nearly half felt that they had moderate to high unmet needs about addressing these fears (Stanton et al., 2005). Even long term survivors continue to have fears about their health. In one study, roughly one third of breast cancer survivors averaging ten years since diagnosis reported worries about a future recurrence, concerns that their current physical symptoms may signal a recurrence, concerns about developing another type of cancer, or worry about future diagnosti c tests (Deimling, Bowman, Sterns, Wagner & Kahana, 2006). A study of breast cancer patients using the modified Cancer Worry Scale only reported the average item score for breast
5 cancer patients three weeks before, one month after, and three months after cessation of treatment (Rabin et al., 2004). The average item score for the sample fell between not at all or rarely and sometimes for the previous month across the three assessments (Rabin et al., 2004). In another study using the modified Cancer Worry Scale, 18% of breast cancer survivors surveyed three years following treatment reported worrying about recurrence often or a lot in the past month (McGinty et al., 2008). Predictors of Fear of Recurrence Several factors have been found to be related to d ifferences in fear of recurrence among people diagnosed with cancer. With regard to demographic variables, older age and being African American have been found to be related to less reported fear of recurrence (Deimling et al., 2006; Stanton et al., 2006; van den Beuken van Everdingen et al., 2008). With regard to clinical variables, less time since diagnosis, past mastectomy, past chemotherapy, more symptoms since diagnosis, more current symptoms, and pain were found to be related to more fear of recurre nce using various measures (Rabin et al., 2004; Deimling et al., 2006; Stanton et al., 2006; van den Beuken van Everdingen, et al., 2008). Various psychosocial variables also appear to play a role in recurrence fears. Positive relationships have been rep orted between fear of recurrence and general levels of anxiety and depression (Deimling et al., 2006). Patients who used avoidance oriented coping styles or who were less optimistic have also been found to report more fear of recurrence (Deimling et al., 2006; Stanton, Danoff burg, & Huggins, 2002; Stanton et al., 2006). In addition, conceptualizing cancer as being an acute illness, rather than either a chronic or cyclic condition is related to less fear of recurrence (Rabin et al., 2004). The influence of family caregivers may also shape fears of recurrence, as
6 the fears of the survivor are related to those of the caregiver and vice versa (Mellon, Kershaw, Northouse, & Freeman Gibb, 2007). More family stressors, finding less positive meaning in the canc er experience, and younger caregivers paired with older survivors are additional factors related to greater fear of recurrence in survivors and caregivers alike (Mellon et al., 2007). As noted above, the existing literature has identified several demograph ic, clinical, and psychosocial variables that are related to fear of recurrence and may assist in identifying which patients are at higher risk for developing these fears. However, there are also several limitations with this body of research. Due to the various types of measures employed to study recurrence fears, it is difficult to generalize findings from one study to the next. Each measure seems to identify different aspects of recurrence fears, from direct measures of how much one is distressed by t he fear to the myriad fears that accompany any diagnosis of a severe medical problem. In addition, the study samples differ widely in terms of their disease characteristics, objective risk of recurrence, and time since treatment completion. Also, it is r are for a study examining correlates of fear of recurrence to be informed by a conceptual model. Before describing a conceptual model of considerable relevance to the study of fear of recurrence (i.e., Protection Motivation Theory), this proposal will fir st consider the issue of how fear of recurrence may have an influence on the health behaviors of cancer survivors. Health Behaviors in Cancer Survivors: American Cancer Society Recommendations A major aim of the current study is to identify variables relat ed to engagement in health behaviors that could potentially reduce the risk of recurrence among cancer survivors. To address this issue, it is important to identify health behaviors that have
7 been recommended for cancer survivors and their potential to re duce the risk of new malignancies. The American Cancer Society (ACS) provides recommendations for cancer survivors based primarily on their recommendations for cancer prevention in healthy populations (Doyle et al., 2006). Reflecting evidence that many p rimary and secondary cancers are linked to being overweight, especially in the case of breast cancer, the guidelines focus on factors that would help maintain a healthy weight, such as diet and exercise recommendations (Doyle et al., 2006). Specifically, survivors are moderate to vigorous activity (not including usual activities) five or more days each week, with 45 60 minute sessions preferable (Doyle et al., 2006). Die t is also at the heart of the ACS recommendations, with the focus on a plant based diet. Survivors should strive to eat five or more servings of fruits and vegetables per day, substitute whole grains for refined or processed grains, and reduce processed a nd red meats in the diet (Doyle et al., 2006). Few studies to date have demonstrated a connection between physical activity (PA) and cancer survival, but associations have been shown between physical activity and quality of life (Blanchard, Courneya, & St ein, 2008; Courneya, 2003; Knols, Aaronson, Uebelhart, Fransen, & Aufdemkampe, 2005;), reduced treatment related symptoms such as fatigue (Rabin, Pinto, Dunsiger, Nash, & Trask, 2008; Schmitz et al., 2005), and reduced risk of other life threatening comorb idities (i.e., diabetes and cardiovascular disease) (Doyle et al., 2006). At least one study has found that survival improves with moderate weekly PA levels. Breast cancer patients who reported the equivalent of 3 5 hours of walking each week had better survival rates than those who
8 were more sedentary; interestingly, more exercise each week over and above moderate levels did not provide increasing benefits for survival (Holmes, Chen, Feskanich, Kroenke, & Colditz, 2005). The connection between obesity a nd survival (Bianchini, Kaaks, & Vainio, 2002; Kroenke, Chen, Rosner, & Holmes, 2005; Norman et al., 2007) and evidence that breast cancer survivors may be at risk for weight gain following diagnosis and treatment (Rock et al., 1999) makes PA recommendatio ns logical contributors to overall health recommendations for cancer survivors. Similarly, the diet recommendations are based on broad health improvement rather than data that confirm direct survival benefits (Doyle et al., 2006). At one point, reducing f at intake was assumed to be the most important factor in increasing survival after a cancer diagnosis, but there has not been sufficient evidence to support this claim (Greenwald, Sherwood, & McDonald, 1997). Recent research is turning away from dietary f at and towards evidence for the benefits of fruit and vegetables (Willett, 2005). Some reviews cite evidence across several studies that fruit and vegetable consumption is associated with better survival rates in breast cancer survivors (Ewertz, Gillander s, Meyer, & Zedeler, 1991; Holmes et al., 1999; Ingram, 1994; Jain, Miller, & To, 1994; Rohan, Hiller, & McMichael, 1993). Measuring Health Behavior Change Research on health behaviors in cancer patients has tended to use one of three designs: comparing su rvivors to healthy control subjects, comparing survivors to themselves over time, and a combination of the two (i.e., following both survivors and controls over time). Behavioral data are analyzed as either rates of activity over a particular time period or are used to classify participants as complying with certain
9 guidelines or study specific criteria for eating healthy or being physically active. Self reported changes often serve as a proxy for collecting behavioral data prior to diagnosis, with cancer survivors being asked to indicate any changes they have made since being diagnosed and the direction of change. Studies of PA and diet behaviors in cancer survivors have typically relied on self report measures. Reports of frequency and duration of bout s of different types of exercise provide information about regular exercise what does and does not count as exercise. Diet is often measured through the use of food f requency questionnaires (FFQ) where participants are asked to recall specific amounts of particular foods over a specified time period, typically the past week or past month. This allows researchers to focus on the patterns of consumption for particular f oods and helps people do or do not meet certain quantities of particular nutrients. Prevalence of Health Behavior Changes Changes in health behaviors appear to be c ommon among cancer survivors, thirds of survivors interviewed saw themselves as having some personal control over their chances of a cancer recurrence. However, a study of global health changes since diagnosis in cancer survivors across the major types of cancer found that fewer patients than might be expected made specific changes based on advice from their doctors (Blanchard, Denniston, et al., 2003). For example, only 46% of smokers h ad quit smoking following diagnosis, while 47% of the total sample improved their diet (including 50.6% of respondents reducing fat, 43.5% increasing fiber, and 42.9% reducing red meat) with
10 influence from their doctor (Blanchard, Denniston, et al., 2003). Another study found that 48% of cancer patients reported making positive changes in their diet following diagnosis consistent with health guidelines (Maskarinec, Murphy, Shumay, & Kakai, 2001). Forty one percent of breast cancer patients in another stud y reported making some type of diet change in the year following diagnosis; decreasing red meat (77% of those who reported diet change) and increasing fruits and vegetables (72%) were the most common changes (Maunsell, Drolet, Brisson, Robert, & Deschenes, 2002). Participants in the changes in diet following breast cancer diagnosis, including eating more fruits (57.9%), vegetables (60.4%), and fiber rich foods (38.8%) (Thomson et al., 2002). Turning to physical activity levels, one study found that relatively few cancer patients had started exercising more since their diagnosis (15.7%), while nearly twice as many (30.6%) were exercising less than they had before getting cancer (Blanchard, Denn iston, et al., 2003). However, in another study 58% of a sample of patients with early stage breast or prostate cancer reported participating in routine exercise on a regular basis for an average of 40 minutes, 4 times a week (Denmark Wahnefried, Peterson McBride, Lipkus, & Clipp, 2000 ) Several studies have compared the physical activity patterns of cancer patients with healthy controls. Baseline data from the Life after Cancer Epidemiology (LACE) study found PA patterns to be similar between breast ca ncer patients two years after diagnosis and healthy comparison samples from other studies (Caan et al., 2005). However, another study using a nationally representative sample found that physical inactivity was more common in cancer patients than healthy c ontrols, with nearly three
11 fourths of cancer patients considered physically inactive; there were no measured dietary differences between groups (Coups & Ostroff, 2005). Similarly, data from the Health Information National Trends Study (HINTS) revealed onl y 45.3% of cancer patients reported being active at least weekly compared to 53% of non patients (Mayer et al., 2007). Perhaps most concerning, there was a low reporting of beliefs that health behaviors could reduce the risk of cancer (Mayer et al., 2007) A potential reason for the conflicting evidence on PA following diagnosis is the finding that PA levels decrease in the year following diagnosis compared with levels one year prior, with an average decrease being two hours less activity per week (Irwin et al., 2003). Pinto, Trunzo, Reiss, and Shiu (2002) also followed women in their first year after diagnosis and found exercise levels did not increase on average. Other research has examined whether cancer survivors are meeting recommendations for healt hy behaviors set by various organizations. For physical activity, Bellizzi, Rowland, Jeffrey and McNeel (2005) found that survivors were 9% more likely to be meeting PA recommendations compared to healthy controls in a nationally representative sample, wi th survivors 2 9 years post diagnosis the most likely to meet recommendations. Unfortunately, fully three fourths of survivors were not meeting PA recommendations (Bellizzi et al., 2005). Another study comparing healthy women to breast cancer survivors f ollowing treatment found survivors were more likely to meet recommended PA levels after controlling for demographic variables than controls, with both frequency and duration of exercise being higher in survivors for activities like stretching (Blanchard, C okkinides, et al., 2003). Data from the Cancer Survivors Study II (SCS II) confirmed that only 29.6 47.3% of all cancer survivors
12 (representing a variety of cancer types) were meeting PA recommendations (Blanchard et al., 2008). Breast cancer patients re ported adhering to PA recommendations 37.1% of the time (Blanchard et al., 2008). In terms of adherence to dietary recommendations, Caan et al. (2005) found early stage breast cancer survivors two years after diagnosis reported a mean of fewer than four s ervings of fruits and vegetables a day, just missing the five a day recommendation. In another study, 55% of survivors of breast or prostate cancer were meeting the five a day guideline and 69% reported adhering to a low fat diet (Denmark Wahnefried et al ., 2000). Reports from the HINTS and SCS II data were less positive, with 18% and 14.8 19.1% of survivors meeting recommendations respectively (Mayer et al., 2007; Blanchard et al., 2008). Comparisons by cancer type revealed melanoma and prostate cancer patients were the most likely to meet fruit and vegetable recommendations out of the most common cancer types (Coups & Ostroff, 2005). Predictors of Health Behavior Changes Receiving a cancer diagnosis alone may not predict behavior change, so it is impor tant to determine what other factors predict health behavior change in cancer patients. In a qualitative research study, Maskarinec et al. (2001) found themes for diet change following diagnosis of cancer that ranged from desire to improve overall physica l well being through better nutrition, maintaining health, preventing recurrence, and beliefs that foods that determine cancer risk should be changed to reduce risk. A study of breast cancer patients assessed during the first year after treatment completi on found that the following were related to greater changes in dietary behavior: younger age, hormone positive status, history of adjuvant treatment, and higher levels of distress (Maunsell et
13 al., 2002). On average, less than 9% of those who made changes made negative dietary changes (Maunsell et al., 2002). The relationship of distress to diet change was particularly strong. Women who changed their diets had the greatest reductions in distress after one year compared to women who did not change their d iets (in either a positive or negative direction) (Maunsell et al., 2002). A study of women during the first three years after breast cancer diagnosis found that those treated with radiation plus chemotherapy experienced greater declines in PA compared th ose women treated with surgery only or radiation only (Irwin et al., 2003); additional findings showed that obese patients decreased PA immediately after diagnosis to a greater extent than normal weight patients (Irwin et al., 2003). Another study of wome n with breast cancer found that younger age, having a significant other, a longer time since diagnosis, higher social support, and higher initial depression predicted greater increases in PA over a 12 month period following treatment completion (Pinto et a l., 2002). Few studies to date have examined the relationship between fear of recurrence and protective health behaviors. Previous work in this area has focused on the conceptualizations of cancer that serve as the source of motivating fears, often relyi ng on 2005; Lee Jones et al., 1997; Rabin & Pinto, 2006). This model elaborately details the formation of illness representations based on several features of the target i llness that include illness symptoms, timeline, consequences, causes, and controllability (Leventhal, Diefenbach, & Leventhal, 1992). Using this model, Rabin and Pinto (2006) found that women with breast cancer who attributed cancer to diet were more like ly to report changing their diet and alcohol consumption than those who did not think diet caused
14 their cancer (Rabin & Pinto, 2006). Also, beliefs that healthy behaviors such as diet, exercise, and reduced alcohol consumption could control cancer risk we re found to predict subsequent changes in diet and alcohol consumption (Rabin & Pinto, 2006). Overview of the Protection Motivation Theory Model As noted previously, a major limitation of previous research examining predictors of fear of recurrence and he alth behaviors in cancer survivors is the general lack of use of conceptual models to select variables and formulate hypotheses. In this section, we seek to show the relevance of Protection Motivation Theory (PMT) to the study of these issues. Protection Motivation Theory (see Figure 1), first proposed by Rogers (1975), states that there are two processes that determine intentions to adopt protective health behaviors (Maddux & Rogers, 1983; Ripptoe & Rogers, 1987; Rogers, 1975). The first component is re to as threat appraisal (Rogers, 1975). Threat appraisal is comprised of vulnerability the perceived personal risk that the health threat will occur, and severity the inherent d angerousness of the health threat if it were to happen (Ripptoe & Rogers, 1987). Together, vulnerability and severity increase both the level of fear arousal and the likelihood of performing protective health behaviors (Ripptoe & Rogers, 1987). The secon d component, called coping appraisal their ability to reduce or even eliminate the threat (Maddux & Rogers, 1983; Rogers, 1975). Coping appraisal is made up of response efficacy the expectation that a given b ehavior will successfully reduce the health threat and self efficacy the belief that one can successfully perform a given behavior (Ripptoe & Rogers, 1987). Response efficacy and self efficacy combine to improve the motivation to adopt certain behaviors (Ripptoe
15 & Rogers, 1987). Threat and coping appraisal combine to form protection motivation the intention to adopt a particular set of protective health behaviors, which serves as the theoretical mediator between cognitions and health protective behavior s (Maddux & Rogers, 1983; Rogers, 1975). Figure 1 Protection Motivation Theory Model. PMT has an advantage over other health theories in that it identifies predictors of both fears and behaviors; other health belief theories focus more stri ctly on prediction of health behaviors (Helmes, 2002; Ripptoe & Rogers, 1987). While models such as the Cognitive Social Health Information Processing (C SHIP) model also factor in fear and anxiety in the prediction of health behaviors from perceived thre ats, PMT is more parsimonious, relying on a handful of predictive variables in contrast to the myriad of predictors in C SHIP models (Miller, Shoda, & Hurley, 1996). Protection Motivation also explores the influence of self efficacy, which has repeatedly been shown to be a strong predictor of adoption of new behaviors (Floyd, Prentice Dunn, & Rogers, 2000; Maddux & Rogers, 1983; Milne, Sheeran, & Orbell, 2000; Schwarzer, 2008; Seydel, Taal, & Wiegman, 1990; Sheeshka, Woolcott, & MacKinnon, 1993; Stanley & Maddux, Protection Motivation Behavior Fear Arousal Coping Appraisal: Perceived Response Efficacy Perceived Self efficacy Threat Appraisal: Perceived Vulnerability Perceived Severity
16 1986). However, there have been criticisms of the PMT framework (McCaul & Mullens, 2003; Salovey, Rothman, & Rodin, 1998; Witte, 1992; Witte, 1998). The effects of threat and coping appraisal on behaviors is predicted to be summative in the origi nal model, however, interactions between at least one threat appraisal variable and at least one coping appraisal variable predicting intentions to perform behaviors are commonly found in research using PMT (Maddux & Rogers, 1983; Wurtele & Maddux, 1987; W itte, 1998). Also, the original model suggests that fear is only predicted by threat appraisal and that coping appraisal does not influence fear at all; recent findings contradict this view (McMahan, Witte, & Meyer, 1998; Witte, 1992; Witte, 1998). McMah an et al. (1998) found that coping appraisal variables such as self efficacy predicted so called fear control responses in participants presented with messages about protecting themselves from a possible environmental health threat. These findings demonst rate that fear responses such as defensive avoidance or minimization of the threat are influenced by perceived ability to control and reduce the threat; threat appraisal alone does not predict fear control responses (McMahan et al., 1998). Reflecting thes e findings, the Extended Parallel Process Model uses the same variables as PMT, but hypothesizes that interactions between threat and coping appraisal determine the level of fear and the likelihood of performing the target health behavior (Witte, 1992; Wit te, 1998). Overview of the Current Study As noted previously the purpose of the present study is to examine the engagement in health behaviors. Toward this end, thi s study will administer
17 questionnaires to women who have completed treatment for breast cancer between six and twenty four months previously and currently have no clinical evidence of disease. In the following section, hypotheses are presented based on PMT regarding fear of recurrence and engagement in health behaviors. Relationship of Coping and Threat Appraisal to Fear of Recurrence Figure 2 Relationship between PMT variables and fear of recurrence. Hypothesis 1: It is hypothesized that the intera ction between threat appraisal and coping appraisal will be related to fear of recurrence (see Figure 2). Specifically, it is hypothesized that survivors who report high levels of threat appraisal in combination with low levels of coping appraisal will r eport greater fear of recurrence than survivors who report low levels of both threat and coping appraisal, low levels of threat appraisal but high levels of coping appraisal, or high levels of threat appraisal and high levels of coping appraisal (see Figur e 3). Reporting high threat appraisal, but low coping appraisal means that one anticipates a health threat, but perceives there to be few protective efforts available to reduce the risk. These predictions are consistent with the Expanded Parallel Process Model which builds upon the PMT framework and which proposes that high threat appraisal in combination with low coping appraisal lead to fear responses (McCaul & Mullens, 2003; Witte, 1992; Witte, 1998). Fear of Recurrence Threat Appraisal: Perceived Vulnerability Perceived Severity Coping Apprais al: Perceived Response Efficacy Perceived Self efficacy +
18 Figure 3 Hypothesized relatio nship between PMT variables and fear of recurrence. Existing research provides some support for this hypothesis. A study by Helmes (2002), examining the relationship of PMT variables to prediction of genetic testing in high risk women, found that vulnerab ility beliefs were related to fears and worries about getting cancer and intrusive thoughts about cancer. Women who were high in vulnerability reported more cancer worry and more intrusive thoughts about cancer (Helmes, 2002). The impact of severity perc eptions on fear has also been demonstrated; with high severity perceptions being related to more fear than low severity (Maddux & Rogers, 1983). In a study directly relevant to the current one, Ripptoe and Rogers (1987) found that healthy women who were t old that breast self examination was both an easy (high self efficacy) and effective (high response efficacy) method of reducing the threat of breast cancer reported the lowest levels of fear (Ripptoe & Rogers, 1987). High Threat Appraisal Low Threat Appraisal Coping Appraisal Fear of Recurrence low low high high
19 Relationship of Coping and Threat Appraisal to Health Behaviors Figure 4 Relationship between PMT variables and recommended health behaviors. Hypothesis 2: It is hypothesized that the interaction between threat appraisal and coping appraisal will be related to engagement in health related behaviors (see Figure 4). Specifically, it is hypothesized that survivors who report high levels of both threat appraisal and coping appraisal will report greater engagement in behaviors than survivors who report low levels of both threat and cop ing appraisal, low levels of threat appraisal but high levels of coping appraisal, or high levels of threat appraisal but low levels of coping appraisal (see Figure 5). Those who anticipate a threat and feel confident in their ability to perform behaviors they believe will reduce the possibility of the threat will be motivated the most to act accordingly. Those who are low in threat appraisal will be less motivated to alter their current behaviors since they do not expect that there will be any related he alth consequences for failing to act. Also, those who are high in threat appraisal, but low in coping appraisal will be less motivated to change their behaviors to protect against a perceived health threat since they will not feel capable of performing an y actions that might reduce the threat. Threat Appraisal: Perceived Vulnerability Perceived Severity Copi ng Appraisal: Perceived Response Efficacy Perceived Self efficacy Behavior + +
20 Figure 5 Hypothesized relationship between PMT variables and health behaviors. The influence of PMT variables on health behavior intentions and ultimately, performance of protective health measu res has been widely documented. Reviews of the literature have found that all four PMT variables positively relate to both current behaviors and intentions to perform protective behaviors, with severity having the smallest influence on intentions (Milne e t al., 2000). A review by Floyd et al. (2000) found that vulnerability and self efficacy produced the strongest effects in cancer prevention behaviors such as sunscreen use and breast self exams. In general, threat appraisal variables had stronger relati ons to subsequent behaviors than coping appraisal variables, but both were still important predictors of behaviors (Floyd et al., 2000). Studies with healthy participants have also found both vulnerability and self efficacy beliefs to be related to intent ions to exercise (Wurtele & Maddux, 1987). In addition, self efficacy and fear of coronary heart disease were related to intentions to increase exercise in recovering cardiac patients (Plotnikoff & Higginbotham, 1998). There are also studies demonstratin g the effectiveness of the PMT based interventions to promote health behaviors. By increasing both threat and coping appraisal McClendon and High Threat Appraisal Low Threat Appraisal Coping Ap praisal Performance of Health Behaviors low low high high
21 Prentice skin cancer risk in a samp le of college students who had a history of purposely tanning. Beyond simply reducing risky behaviors by influencing threat and coping appraisals, other studies were able to increase intentions to perform novel health behaviors. Courneya and Hellsten (20 01) manipulated PMT variables and measured college Again, both threat and coping appraisal were positively related to intentions to perform the recommended exercise regimen (Courneya & Hellsten, 2001). Exploratory Analyses Finally, exploratory analyses will be performed to determine the interrelationships among threat and coping appraisal, health behaviors, and fear of recurrence (see Figure 6). In particular, three possible ways that these variables could be related will be examined. The first two ways assume that a relationship will be observed between fears of recurrence and health behaviors. Prior research suggests that an inverse relationship between health be haviors and fear of recurrence in cancer survivors will be observed. Along these lines, Maunsell et al. (2002) found that women with breast cancer initially high in distress were more likely to change their diet for the better during the first 12 months a fter surgery than women low in distress. Assuming that a relationship between health behaviors and fear of recurrence is observed and that coping and appraisal variables are related to both fear of recurrence and health behaviors, the proposed analyses wi ll examine whether health behaviors mediate the relationship between coping and appraisal variables and fear of recurrence (Figure 6 A) or whether fear of recurrence
22 mediates the relationship between coping and appraisal variables and health behaviors (Fig ure 6 B) or whether there is no mediation. Figure 6 A Behavior mediates relationship between PMT and fear of recurrence. Figure 6 B Fear of recurrence mediates relationship between PMT and behavior. Threat Appraisal Behavior Fear of Recurrence Coping Appraisal Threat Appraisal Behavior Fear of Recurrence Coping Appraisal
23 Method Participant s The sample was comprised of women who had been treated for early stage breast cancer (stage I II) six to twenty four months prior to enrollment in the study. Additional eligibility criteria were: a) ability to give informed consent, b) ability to speak and read English, c) no history of other cancers except non melanoma skin cancer, d) no recurrence of breast cancer since original diagnosis and, e) age between 18 and 90. Measures Demographic characteristics. A standardized self report form was used to collect the following demographic information: age, height, weight, race, ethnicity, marital status, and highest degree attained. Clinical characteristics. Clinical characteristics collected via chart review included stage at diagnosis, treatment(s) r eceived (including any surgeries or adjuvant treatments), and time since diagnosis and treatment completion. Fear of recurrence. The four item Cancer Worry Scale (Lerman et al., 1991) ence by adding the (Rabin et al., 2004). This measure assesses the frequency of recurrence worry over the course of the past month. Response options for each item were 1 ( N ot at all or rarely ), 2 ( Sometimes ), 3 ( Often ), or 4 ( A lot ). The revised version has demonstrated acceptable
24 internal consistency in a sample of breast cancer survivors ( = .78) and is positively related to measures of psychological distress such as the CES D ( r = 0.49) and negatively related to mental health ( r = 0.43) (McGinty, et al., 2008). This measure had good internal consistency in the current sample ( = .87). Vulnerability. Perceived vulnerability to a cancer recurrence was measured by prior research (Valdimarsdottir et al., 1995). To assess absolute risk participants were again during your chances are of having breast cancer again in your lifetime compared to other women your age with breast cancer who received the same treatment for the sam e type of breast extremely unlikely to extremely likely on a 6 point scale for the absolute risk item and from much higher to much lower on a 5 point scale for the comparative risk item. Absolute and comparative ri sk estimates were converted to the same metric and then combined to create a total vulnerability score ( = .75). Severity. The Consequences Subscale of the Revised Illness Perception Questionnaire (Moss Morris et al., 2002) was used to determine perceived severity of the consequences of a potential breast cancer recurrence. All items were adapt ed to refer to a potential medical, social, financial, and psychological consequences of a breast cancer recurrence. Participants were asked to respond how much they agreed with each consequence on a Likert type scale ranging from 1 ( strongly disagree ) to 5 ( strongly
25 agree ). This measure showed acceptable validity and good internal consistency for the selected subscale ( = .84) in previous research (Moss Morris et al., 2002). Internal consistency was also good for the current study sample ( = .81). Self efficacy. Participants indicated the extent to which they felt capable of successfully adherin g to both diet and exercise recommendations from the American Cancer Society (ACS) ( Kushi et al., 2006 ) to provide information about self efficacy beliefs regarding these health behaviors. With regard to diet, the ACS recommends eating five or more servin gs of fruits and vegetables a day, whole grains in place of refined grains, and limiting consumption of red meats. With regard to exercise, the ACS recommends at least 30 minutes of moderate to vigorous physical activity over and above usual activities at least 5 days a week. After reading these recommendations, participants were asked to rate their degree of confidence that they could adhere to the recommendations given 18 different hypothetical circumstances that may limit regular activity using a scale from 0 ( cannot do at all ) to 10 ( highly certain can do ) (Bandura, 2006). Similarly, participants also responded to 30 items detailing circumstances that may limit adherence to ACS diet recommendations (Bandura, 2006). The approach described here, which is modified after that recommended by Bandura (2006), was found to yield acceptable psychometric properties for the exercise measure in a Korean sample of patients with chronic illness (Shin, Jang, & Pender, 2001). In that study, the scale overall had exc = .94. Test retest reliability was evaluated after a two week period, resulting in a strong correlation between testing intervals ( r = 0.77) (Shin et al., 2001). For the current study sample, internal consistency
26 was excellent for both the exercise sel f efficacy measure ( = .97) and the diet self efficacy measure ( = .97). Response efficacy. Participants were also asked about their perceptions of the response efficacy of the ACS diet and exercise recommendations using a format a dapted from previous research (Plotnikoff & Higgenbotham, 1995; 1998 cited in Rhodes & Plotnikoff, 2005). Similar items were used to assess efficacy of diet and exercise To what extent do you agree or disagree with physically active lifestyle would reduce your chances of having a breast cancer recurrence To what extent do you agree or disagree with the healthy diet would reduce my chances of having a breast cancer recurrence recommendations. Response options ranged from 1 ( strongly disagree ) to 5 ( strongly agree ). Diet ary behavior. To assess adherence to a diet rich in fruits and vegetables, participants first completed the All Day Screener, which details number of servings and serving sizes for a variety of fruits and vegetables consumed on a regular basis for the pas them as adherent or nonadherent to ACS recommendations that survivors consume five or more servings of fruits and vegetables every day Exercise behavior. The Leisure Score Index (LSI) was used to record the amount of exercise participants reported for the previous week above usual activities (Godin & Shephard, 1985). Exercise was divided into three categories based on intensity
27 of exercise: strenuous, moderate, and mi ld exercise. Both the frequency and duration of exercise for each class of physical activity was recorded. Weekly metabolic equivalents (METS) were calculated by the following formula: total METS = (total minutes of strenuous exercise X 9) + (total minut es of moderate exercise X 5) + (total minutes of to categorize them as adherent or nonadherent to ACS recommendation that survivors engage in at least 30 minutes of mode rate to vigorous activity at least five days a week (Kushi et al., 2006). Depression. The Center for Epidemiologic Studies Depression Scale (CES D) was used to assess depressive symptoms in this sample (Radloff, 1977). This 20 item measure assessed com mon depressive symptoms that are not attributable to common health related problems typically found in medical patient populations, such as sleep trouble keeping my mind o (reverse scored). Response options ranged from 0 ( rarely or none of the time ) to 3 ( most or all of the time ). Depressive symptomatology was examined as a potential confounding variable in the sta tistical analyses using this measure. The internal consistency for this sample for the CES D was excellent ( = .91). Procedure Following chart review for initial eligibility, potential participants were mailed a letter describing the study and a postcard to return if they wish to decline participation. Women were also mailed an informed consent form additional eligibility screening measures (including questions regarding ability to read and write in English and
28 confirmation that they had not have a breast cancer recurrence or other cancer diagnosis), questionnaire packets, and postage paid return en velopes. Women who did not meet full eligibilty criteria were instructed to only complete the screening measure and consent form and return the rest of the packet blank. Participants who did not return completed forms via mail within the one month deadli ne were contacted via phone three times with reminders to complete and return the survey within the next month at which point non responders were considered lost to follow up. Participants who met full screening criteria first provided basic demographic in formation and then responded to the modified Cancer Worry Scale to provide their baseline level of personal fear of recurrence. Next, threat appraisal was assessed by vulnerability and severity perceptions. Participants were then asked to give informatio n on their current exercise and diet habits followed by a description of ACS recommended diet and exercise guidelines to assess adherence. Current health behaviors were assessed prior to outlining ACS recommendations to reduce demand characteristics for c urrent habits. Women were then asked about their perceived response efficacy and perceived self efficacy for ACS diet and exercise recommendations regardless of whether they have ever attempted these or not. Participants first responded to items about th eir current exercise and perceived response efficacy and self efficacy of adhering to ACS exercise recommendations, and then completed the same measures for diet. Statistical Analyses Hypothesis 1 states that there will be an interaction between threat and coping appraisal such that those with high threat appraisal and low coping appraisal will report greater fear of recurrence than those with overall low threat appraisal or high coping
29 appraisal. To test this hypothesis, the data were analyzed using multi ple regression analysis. The data for the independent variables were centered following procedures recommended by Aiken and West (1991). Composite threat appraisal and coping appraisal variables were first computed. To form the threat appraisal variable scores for perceived vulnerability and severity were transformed to z scores and then summed. For coping appraisal, three different composite variables were created: diet coping appraisal, exercise coping appraisal, and overall coping appraisal. Diet c oping appraisal represents the sum of the z scores for diet self efficacy and diet response efficacy. Exercise coping appraisal represents the sum of the z scores for exercise self efficacy and exercise response efficacy. To compute the overall coping ap praisal variable, z scores for diet self efficacy, exercise self efficacy, diet response efficacy, and exercise response efficacy were combined. After entering the threat appraisal variable and the appraisal variable on the first step, the possible presen ce of an interaction was examined by entering the multiplicative product of these two variables on the second step. A significant change in variance accounted for ( R 2 ) on the second step would support the study hypothesis. If a significant interaction wa s obtained, procedures outlined by Aiken and West (1991) were then conducted to determine whether the nature of the interactive effect conformed to expectations. Hypothesis 2 states that there will be an interaction between coping and threat appraisal such that those who are both high in threat and high in coping will report more health promoting behaviors than those who are either low in both or low in either threat or coping appraisal. Procedures similar to those described above for hypothesis 1 were use d to test hypothesis 2. The threat appraisal variable was the same as that used to test
30 hypothesis 1. For coping appraisal, only the diet coping appraisal term was used to predict daily fruit and vegetable intake, and only the exercise coping appraisal t erm was used to predict weekly exercise. Diet and exercise outcomes were predicted as continuous variables (total average number of daily fruit and vegetable servings, total weekly METS) in hierarchical regressions. Separate analyses were conducted to ev aluate dietary and exercise behavior. A similar set of analyses was conducted using multiple logistic regressions to examine interactive effects on adherence to ACS diet and exercise recommendations (a dichotomous variable). Finally, a set of exploratory analyses were planned to examine relationships among threat and coping appraisal, health behaviors, and fear of recurrence. First, univariate analyses were conducted to determine whether fear of recurrence and performance of the two different health beha viors were related to each other. If there were a significant relationship between reported fear of recurrence and current performance of healthy behaviors, the mediational analyses would then be conducted. The first analysis would examine whether the co mbination of high coping appraisal and high threat appraisal is associated with greater engagement in health behaviors that, in turn, should be associated with lo wer fear of recurrence (Figure 6 A). To evaluate this possibility, a series of regression ana lyses and a Sobel test would be conducted to determine whether health behaviors mediate the expected relationships of threat and coping appraisal with fear of recurrence. The alternative possibility, that fear of recurrence mediates the relationship betwe en PMT variables and performance of recomm ended health behaviors (Figure 6 B), would also be tested in a second analysis.
31 Power analyses were computed for both univariate correlational analyses and multiple regression analyses. A sample size of 155 part icipants with complete data yields power of .80 to detect an expected medium effect size of 0.22 at = .05 (two tailed). A second power analysis for the proposed regression analyses was based on two steps with one variable each where each step explains 10% of the variance, followed by a two variable interaction term that explained an additional 5% of t he variance. A sample size of 155 participants with complete data yields power of .89 to detect the 5% increase on the interaction step at = .05 (two tailed).
32 Results Participants A total of 498 breast cancer patients were screened for eligibilit y. Of these patients, 163 were ineligible (see Figure 7). Questionnaires and consent forms were mailed to the remaining 326 women; of these women 66 refused to participate, 60 could not be reached, 60 indicated interest in participating but did not retur n study materials after 3 reminder phone calls, and an additional 10 were ineligible before consent (e.g., indicated that they did not speak English, could not provide consent, etc.). Consent forms were signed by 160 women. Three were ineligible after co nsent due to other cancer diagnoses and two did not provide complete data, leaving 155 participants in the study sample. The overall response rate of the patients who were mailed study materials was 47.5%. Participants ranged in age from 30 to 87 years o ld ( M = 58.83, SD = 11.83). The majority of the participants had completed at least some college or specialized training (80%), were married (69%), had a gross annual income greater than $40,000 (71%), and were Caucasian (90%). The sample included both S tage I (61%) and Stage II (39%) patients who were an average of 1.45 years ( SD = 0.34) since diagnosis and 1.04 years ( SD = 0.36) since treatment completion. See Table 1 for complete demographic and clinical information. Participating patients were compa red to non responders on clinical characteristics. There was a trend for responders to have more time passed since diagnosis ( p = .06). Also, responders had significantly ( p
33 Figure 7 Response rate throughout recruitment and surveying process. 160 consented 3 ineligible after consent 155 surveys com pleted 2 missing data 3 ove r age 90 5 male 2 live outside US 3 no current mailing address 65 treatment dates unclear 42 > 6 months since treatment 9 disease stage 8 primary cancer recurrence 7 other cancers 18 non English speakers 489 screened for eligibility 326 eligible to be mailed 296 reached 163 ineligible 30 unreachable 10 ineligible before consent 1 cannot provide consent 1 disease stage 4 other cancers 2 deceased 5 non English speakers 126 refused 60 soft refusals: multiple reminder phone calls 66 refused after screening
34 Table 1 Sample Characteristics (N = 155) Age ( M SD ) 58.8 11.83 Body Mass Index ( M SD ) 27.3 5.23 n % Race Caucasian 140 90.3 African American 7 4.5 Asian 2 1.3 Pacific Islander 2 1.3 American Indian 1 0.7 Other 3 1.9 Ethnicity Hispanic 10 6.5 Non Hispanic 145 93.5 Marital Status Single 11 7.1 Married 107 69.1 Separated 5 3.2 Divorced 18 11.6 Widowed 14 9.0 Education Less than high school 6 3.8 High school 25 16.1 Some college 59 38. 1 College graduate 37 23.9 Graduate degree 28 18.1 Annual Income < $40,000 43 28.9 > $40,000 106 71.1 Treatment Radiation 47 30.3 Chemotherapy 31 20.0 Chemo & Radiation 34 21.9 Surgery Only 43 2 7.8 Adjuvant Hormone Therapy 126 81.3 Surgery Type Excisional Biopsy 1 0.7 Lumpectomy 69 44.5 Mastectomy 85 54.8 Stage at diagnosis Stage I 95 61.3 Stage II 60 38.7
35 treatment completion, were more like ly to have used hormone therapy for breast cancer, and were less likely to have had a lumpectomy as surgical treatment than non responders. Descriptive Statistics Overall, the sample reported a mean fear of recurrence of 1.74 ( SD = 0.71), which corresponds to reporting experiencing thoughts or difficulties due to concerns about cancer coming back between not at all or rarely (1) and sometimes (2) For perceived risk, responses for absolute and relative risk estimates were combined on a scale that ranged fr om 11 60, with higher scores indicating more perceived risk. Participants overall indicated moderate perceived risk of recurrence, with the majority of responses (51%) falling between somewhat unlikely (3) to somewhat likely (4) for absolute risk. The ma jority of responses (65%) were about the same (3) for relative risk when comparing risk of recurrence to other breast cancer patients. Perceived severity was also moderate ( M = 22.15, SD = 4.54) in this sample as scores fell near the midrange on this meas ure (possible range = 9 to 30). Mean self efficacy scores fell near the midrange for diet, ( M = 178.90, SD = 57.47; possible range = 0 to 300) and for exercise ( M = 96.96, SD = 42.63; possible range = 0 to 180). The majority of participants reported agre eing that the American Cancer Society guidelines for diet ( 68% agree or strongly agree ) and exercise (57% agree or strongly agree) could reduce their risk of a cancer recurrence. Finally, 41% of participants reported that they met ACS guidelines for daily fruit and vegetable intake and 37% reported that they met ACS guidelines for weekly physical activity.
36 Table 2 Relationships of Demographic and Clinical Variables to Outcomes Fear of Recurrence Daily Fruit & Vegetable Intake Weekly Exercise r r r Age .25** .17* .13 Body Mass Index .05 .02 .20* M SD M SD M SD Race Caucasian 6.89 2.72 5.55 a 4.84 26.29 21.44 Other 7.73 3.77 9.14 10.04 19.13 17.68 Ethnicity Hispanic 6.60 2.27 6.15 4.57 25.70 20.93 Non Hispanic 6.99 2.87 5.88 5.40 25.59 21.25 Marital Status Married 7.14 2.86 5.56 4.71 24.95 20.58 Not married 6.58 2.77 6.63 6.52 27.04 22.56 Education Less than college 7.23 2.87 6.03 4.50 20.97 23.29 College graduate 6.90 2.83 5.86 5.54 26.76 20.53 Annual Income < $40,000 6.86 2.96 6.73 6.64 27.34 25.51 > $40,000 6.92 2.75 5.66 4.84 24.98 19.64 Treatment Radiation 6.70 2.62 5.87 6.47 24.55 22.72 Chemotherapy 7.35 3.16 6.1 1 4.15 24.58 21.80 Chemo & Radiation 7.15 2.91 5.70 4.50 24.88 19.56 Surgery alone 6.84 2.81 5.92 5.49 28.05 20.69 Surgery Type Bilateral Mastectomy 7.43 3.10 6.08 5.66 24.06 20.26 Other surgery 6.75 2.69 5.81 5.20 26.31 21.6 2 Stage at diagnosis Stage I 6.59 b 2.52 5.66 5.19 26.41 21.43 Stage II 7.57 3.20 6.26 5.58 24.31 20.84 Note a F (1, 153) = 6.37, p = .02. b F (1, 153) = 4.47, p = .04. p < .05, ** p < .01.
37 Preliminary Analyses Relationships between th e outcome variables and demographic and clinical variables are presented in Table 2. Higher fear of recurrence was related to younger age and more advanced disease stage. In addition, more depressive symptoms were related to higher fear of recurrence, r ( 153) = .61, p < .001. Consuming more daily fruit and vegetable servings was associated with younger age and being Caucasian, but the race differences were likely due to outliers in the non Caucasian group on this measure. Engaging in more weekly exercise was related to lower body mass index. No other demographic or clinical variables were related to either fear of recurrence or the diet or PA measures. Interrelationships between Protection Motivation Theory variables and outcome variables are presented in Table 3. Fear of recurrence was related to greater threat appraisal overall, and higher perceived vulnerability and higher perceived severity of a breast cancer recurrence. Higher fruit and vegetable intake was related to greater diet coping appraisal overall and higher diet self efficacy. Greater exercise was related to greater exercise coping appraisal overall, higher exercise self efficacy, and higher exercise response efficacy. Evaluation of Hypothesis 1 Hypothesis 1 stated that the combination of high threat appraisal and low coping appraisal would predict high fear of recurrence. To test hypothesis 1, composite threat appraisal and coping appraisal variables were first computed as described in the statistical analysis section. Hierarchical re gression analysis was then conducted to test whether the interaction of threat appraisal and coping appraisal predicted fear of recurrence (see Table 4). Threat appraisal and coping appraisal were entered in the first step, followed by the
38 Table 3 Correl ations of Protection Motivation Theory Variables with Fear of Recurrence, Diet, and Exercise Fear of Recurrence Daily Fruit & Vegetable Intake Weekly Exercise Coping .11 Exercise Coping .09 .26** Diet Coping .11 .23** Response Efficacy .04 Exercise Response Efficacy .03 .21** Diet Response Efficacy .04 .12 Self Efficac y .12 Exercise Self Efficacy .10 .18* Diet Self Efficacy .11 .22** Threat .60*** .03 .03 Vulnerability .53*** .12 .08 Severity .48*** .07 .02 Weekly Exercise .03 .15 Daily Fruit & Vegetable Intake .14 .15 Fear of Recurrence .03 .14 Note p < .05, ** p < .01, *** p < .001.
39 T able 4 Hierarchical Multiple Regression Analyses Predicting Fear of Recurrence From Threat, Coping, and Threat X Coping Interaction Predictor R 2 Cumulative R 2 p Step 1 .37 .37 < .001 Threat .63*** C oping .01 Step 2 .02 .39 .03 Threat X Coping .14* Note. F (3, 151) = 31.65, p < .001. p < .05, ** p < .01, *** p < .001. threat X coping interaction term in the second step. In the first step, threat appraisal and overall c oping appraisal accounted for 37% of the variance in fear of recurrence ( p < .001). Threat alone accounted for a significant portion of variance ( p < .001) while coping alone did not. In the next step, the threat X coping interaction term accounted for 2% of the remaining variance ( p = .03). The slopes of the lines were significantly different from zero and significantly different from each other ( p direction of the interaction supported hypothesis 1 (see Figure 8). That is, p articipants who reported both high threat appraisal and low coping appraisal had the highest reported levels of fear of recurrence. When controlling for age and disease stage by entering these variables in the first step, the threat X coping interaction r emained significant ( p = .04).
40 It should be noted that when depression was added as a control variable in the first step, the threat X coping interaction was no longer significant ( p > .05). Figure 8 Threat Appraisal X Coping Appraisal Interaction. Be cause the threat X coping interaction term contained composite scores, additional analyses were conducted that involved deconstructing the threat appraisal and coping appraisal terms to determine which components were driving the significant interaction. First, exploratory multiple regression analyses were conducted to determine if the threat appraisal portion of the threat X coping interaction term was attributable to perceived vulnerability or perceived severity. Findings were significant only for the p erceived vulnerability X coping interaction term. Perceived vulnerability and coping appraisal accounted for 29% of the variance in fear of recurrence ( p < .001). The addition of the vulnerability X coping interaction term accounted for another 2% of the remaining variance ( p = .02) (see Table 5). Participants who reported high perceived
41 Table 5 Hierarchical Multiple Regression Analyses Predicting Fear of Recurrence From Vulnerability, Coping, and Vulnerability X Coping Interaction Predictor R 2 Cumulative R 2 p Step 1 .29 .29 < .001 Vulnerability .59*** Coping .02 Step 2 .02 .31 .02 Vulnerability X Coping .16* Note. F (3, 151) = 22.80, p < .001. p < .05, ** p < .01, *** p < .001. Figure 9 Vulnerability X Coping Appraisal Interaction.
42 vulnerability and low coping appraisal had the highest levels of fear of recurrence (see Figure 9). Again, when controlling for age and disease stage in the first step, the vulnera bility X coping interaction remained significant ( p = .02). Exploratory hierarchical multiple regression analyses were then conducted to determine if coping portion of the vulnerability x coping interaction reflected a vulnerability X diet coping interacti on or a vulnerability X exercise coping interaction. Only the vulnerability X diet coping interaction was significant. In step 1, vulnerability and diet coping accounted for 29% of the variance in fear of recurrence ( p < .001). In step 2, the addition o f the vulnerability X diet coping interaction term accounted for an additional 3% of the variance ( p = .01) (see Table 6). Participants who reported high Table 6 Hierarchical Multiple Regression Analyses Predicting Fear of Recurrence From Vulnerability, Diet Coping, and Vulnerability X Diet Coping Interaction Predictor R 2 Cumulative R 2 p Step 1 .29 .29 < .001 Vulnerability .59*** Diet Coping .02 Step 2 .03 .32 .01 Vulnerability X Diet Coping .18* Note. F (3, 151) = 23.22, p < .001. p < .05, ** p < .01, *** p < .001.
43 Figure 10 Vulnerability X Diet Coping Appraisal Interaction. perceived vulnerability and low diet coping appraisal had the highest levels of fea r of recurrence (see Figure 10). The interaction remained significant after controlling for age and disease stage ( p = .01). Finally, the last set of hierarchical regression analyses determined whether the vulnerability X diet coping interaction was due to the interaction of vulnerability beliefs with diet self efficacy or with exercise self efficacy. For diet self efficacy, the first step was significant with vulnerability and diet self efficacy accounting for 29% of the variance ( p < .001). However, the vulnerability X diet self efficacy interaction was not significant with ( p = .11) or without age and stage as control variables ( p = .07), and so did not contribute additional variance to the total variance in fear of recurrence. For diet response eff icacy, again, the first step was significant ( p < .001); vulnerability and diet response efficacy accounted for 29% of the variance in fear of recurrence. However, the
44 vulnerability X diet response efficacy interaction did not contribute significant varia nce to the model with ( p = .06) or without age and stage as control variables in the first step ( p = .07). In summary, these findings indicate that the diet coping portion of the vulnerability X diet coping interaction cannot be further dismantled into si gnificant subcomponents. Evaluation of Hypothesis 2 Hypothesis 2 stated that the combination of high threat appraisal and high coping appraisal would predict high performance of both healthy diet and PA recommendations from the ACS. To test hypothesis 2, composite threat appraisal and coping appraisal variables were first computed as described in the statistical analyses section. For diet outcomes, hierarchical regression analysis was conducted to test whether the interaction of threat appraisal and diet coping appraisal predicted daily fruit and vegetable intake. Threat appraisal and diet coping appraisal were entered in the first step, followed by a threat X diet coping interaction term in the second step (see Table 7). In the first step, threat apprai sal and diet coping appraisal accounted for 11% of the variance in daily fruit and vegetable intake ( p < .001). Diet coping alone contributed a significant portion of the variance ( p < .001) while threat alone did not. The threat X diet coping interactio n term was not significant ( p = .98). There was no difference in the significance level when age was controlled for statistically by adding it in the first step. Exclusion of outliers two or more standard deviations from the mean number of fruit and veg etable servings provided no differences in the significance of the findings.
45 Table 7 Hierarchical Multiple Regression Analyses Predicting Fruit and Vegetable Intake From Threat, Diet Coping, and Threat X Diet Coping Interaction Predictor R 2 C umulative R 2 p Step 1 .11 .11 < .001 Threat .03 Diet Coping .33*** Step 2 .00 .11 .98 Threat X Diet Coping .00 Note. F (2, 152) = 9.19, p < .001. p < .05, ** p < .01, *** p < .00 1. Similarly, for the logistic regression analysis for adherence to ACS diet guidelines, threat appraisal and diet coping were entered in the first step, followed by a threat X diet coping interaction term in the second step. Threat appraisal and diet co ping appraisal were not significant predictors on the first step 2 (2, N = 155) = 3.73, p = .16. The threat X diet coping interaction term in the second step was also not significant 2 (1, N = 155) = 2.01, p = .16. Hence, the threat X diet coping interac tion did not predict ACS diet guideline adherence. For PA outcomes, hierarchical regression analysis was used to test whether the interaction of threat appraisal and exercise coping appraisal predicted weekly exercise. Threat appraisal and exercise coping appraisal were entered in the first step, followed by a threat X exercise coping interaction term in the second step (see Table 8). Threat
46 appraisal and exercise coping appraisal accounted for 19% of the variance in weekly exercise ( p < .001). Exercise coping alone added significant variance to the model ( p < .001) while threat alone did not. The threat X exercise coping interaction term was not significant ( p = .58). Again, excluding outliers (values two or more standard deviations from the mean) and adding BMI as a potential control variable in the first step did not change the significance of the findings. Table 8 Hierarchical Multiple Regression Analyses Predicting Weekly Exercise From Threat, Exercise Coping, and Threat X Exercise Coping Interacti on Predictor R 2 Cumulative R 2 p Step 1 .19 .19 < .001 Threat .06 Exercise Coping .45*** Step 2 .00 .19 .58 Threat X Exercise Coping .04 Note. F (2, 152) = 17.85, p < .001. p < .05, ** p < .01, *** p < .001. The logistic regression analysis for adherence to ACS PA guidelines entered threat appraisal and exercise coping appraisal in the first step followed by a threat X exercise coping interaction term in the second step. Threa t appraisal and exercise coping appraisal were significant predictors in the first step 2 (2, N = 155) = 29.42, p < .001. Exercise coping accounted for significant predictive value ( p < .001) while threat did not
47 ( p = .10). The threat X exercise coping i nteraction term in the second step was not significant 2 (1, N = 155) = 0.38, p = .54. Hence, the threat X exercise coping interaction did not predict ACS PA guideline adherence. Exploratory Analyses Exploratory analyses were conducted to test for interre lationships among threat and coping appraisal, health behaviors, and fear of recurrence to determine whether further analyses should be conducted to evaluate mediation pathways between these variables. Initial univariate analyses showed that there was no correlation between fear of recurrence and performance of health behaviors for diet r (153) = .14, p = .07, or PA, r (153) = .03, p = .72. Hence, conditions did not exist to test whether health behaviors mediated the relationship between threat and coping a ppraisal and fear of recurrence, or whether fear of recurrence mediated the relationship between threat and coping appraisal and performance of health behaviors.
48 Discussion Relationship of Threat and Coping Appraisal to Fear of Recurrence Results f rom this study supported the hypothesis that the combination of high threat appraisal and low coping appraisal would predict greatest fear of recurrence in early stage breast cancer survivors. To better understand which variables were driving this signifi cant interaction, the threat and coping variables were broken down into their component variables. Based on the analyses breaking down threat appraisal, perceived vulnerability to a cancer recurrence appeared to contribute more to the interaction than per ceived severity of a recurrence. Also, after dismantling coping appraisal, diet coping was more influential than exercise coping and significantly interacted with perceived vulnerability to predict cancer recurrence fears. When the two subcomponents of d iet coping were further dismantled, neither diet self efficacy nor diet response efficacy had significant interactions with perceived vulnerability. These findings provide new insight into the prediction of fear of recurrence. Most fear of recurrence re search to date has been guided by demographic, clinical, or psychosocial variables as predictors, often neglecting how these various separate factors interact to moderate the levels of fear experienced by cancer survivors. The observed relationship betwee n threat appraisal and fear of recurrence is consistent with Protection Motivation Theory, which stipulates that perceived threats to health should predict fear associated with that health threat (Rippetoe & Rogers, 1987). It is also consistent with
49 previ ous research using PMT that found direct relationships between both vulnerability and severity predicting health fears (Helmes, 2002; Maddux & Rogers, 1983). However, the current study also found evidence of an interaction between threat appraisal and cop ing appraisal predicting fear of recurrence. This finding is consistent with the Extended Parallel Process Model which predicts that coping appraisals moderate the effects of threat appraisal on fear with the highest levels of fear associated with high th reat appraisal in combination with low coping appraisal (McCaul & Mullens, 2003; Witte, 1992; Witte, 1998). As can best be determined, this is the first research to show that the combination of high threat and low coping appraisal is related to greater fe ar of cancer recurrence. One issue complicating the interpretation of the observed interaction is the relationship of depression to fear of recurrence. Analyses indicated that higher levels of depressive symptomatology were strongly related to greater fear of recurrence. Accordingly, when depression was added as a predictor of fear of recurrence, the interaction between threat and coping appraisal was no longer significant. More research is needed to understand the connection between recurrence fears and depression in cancer survivors. It may be the case that depression influences the occurrence of cancer related fears. A recently published longitudinal study of cancer related fear of the future (a concept closely related to, but more broad in scope than fear of cancer recurrence) found that distress levels predicted later fear, but that fears did not affect later distress (Lebel, Rosberger, Edgar, & Devins, 2009). The current pattern of results might also be interpreted to suggest that fear of recur rence is not distinct from depression or that current
50 measures of fear of recurrence do not do a good job of discriminating between fear of recurrence and depression. Overall, the reports of fear of recurrence in the present sample of breast cancer surviv ors were similar to those found in previous studies, with most survivors reporting fears or concerns at least sometimes (Baker et al., 2005; Bluman, Borstelmann, Rimer, Iglehart & Winer, 2001; Deimling et al., 2006; Herschbach et al., 2004; Schroevers et a l., 2006; Stanton et al., 2006; van den Beuken van Everdingen et al., 2008). Therefore, the data seem to reflect typical levels of fear of recurrence in early stage cancer survivors and suggest that the findings are generalizable to other breast cancer su rvivors. Relationship of Threat and Coping Appraisal to Health Behaviors Results did not support the second study hypothesis that the combination of high threat appraisal and high coping appraisal would predict more consumption of fruits and vegetables, an d greater exercise, as well as better adherence to ACS recommendations for cancer survivors regarding diet and exercise. There was evidence, however, of a main effect for coping appraisal but not threat appraisal on these outcomes. Specifically, greater coping appraisal was related to greater daily fruit and vegetable intake and greater weekly exercise, as well as to higher adherence rates for both behaviors. Results of univariate analyses suggested that these relationships reflect positive relationships of response efficacy and self efficacy with exercise and a positive relationship of self efficacy but not response efficacy with fruit and vegetable intake. There are several possible explanations for the lack of support for the hypothesis that the comb ination of high threat appraisal and high coping appraisal would predict healthier behavior. In the case of the current study, where the focus was on general
51 health behaviors that are not specific to cancer, participants may have had several reasons for their patterns of health habits other than reducing their risk of cancer. That is, people may follow healthy diets or exercise regimes for several reasons unrelated to concerns about cancer. Possibilities include weight and appearance concerns, convenien ce, general health knowledge, comorbid illnesses that might make certain activities more difficult, or simply established health habits (Baranowski, Cullen, & Baranowski, 1999; Sherwood & Jeffery, 2000; Trost, Owen, Bauman, Sallis, & Brown, 2002). Because we did not assess these habits before cancer diagnosis, we cannot ascertain how much these behaviors were influenced by the experience of cancer itself. Our finding that neither perceived vulnerability nor perceived severity of a cancer recurrence predic ted either health activity suggests a limited impact of the threat of cancer in predicting cancer agreeing that following ACS guidelines can help reduce their risk of future cancer, it was not clear how much they expected their risk to be reduced if they followed those recommendations. If they only anticipated a marginal reduction in risk, patients may not be persuaded to engage the effort necessary to change establis hed health habits. Another possible explanation for the negative findings may be differences in adherence rates between the current sample and prior study samples. Adherence to recommended fruit and vegetable consumption appeared to be more common in this sample than what was found in earlier studies of breast cancer survivors; roughly 18% of survivors were classified as adherent in previous studies versus 41% adherent in the present study (Blanchard et al., 2008; Meyer et al., 2007). Higher rates of adhe rence in this study sample might indicate measurement error or a sample that may not represent
52 typical breast cancer survivors. However, adherence to recommended levels of exercise in the current sample were similar to other studies of cancer survivors, w ith the rate in the current study (37%) falling somewhere between the 30% to 45% adherence rates reported in previous research (Blanchard et al., 2008; Coups & Ostroff, 2005; Meyer et al., 2007). Therefore, systematic differences in adherence rates in the study sample do not appear to explain the lack of support for an interaction between threat and coping appraisal predicting exercise. The relationships observed in the current study between Protection Motivation Theory variables and health behaviors co nfirm some findings from previous studies. Consistent with the present study, several studies have found that coping appraisal beliefs, such as self efficacy and response efficacy beliefs, predicted health behaviors (Floyd et al., 2000; Maddux & Rogers, 1 983; Milne et al., 2000; Lewis et al., 2002; Rippetoe & Rogers, 1987; Seydel et al., 1990; Sheeshka et al., 1993; Stanley & Maddux, 1986). However, other findings are inconsistent with prior research. Although several studies have found associations betw een threat appraisal and health behaviors such as smoking cessation and increased exercise in healthy subjects (Courneya & Hellsten, 2001; Greenwald, 1997), threat appraisal was not related to diet or exercise behaviors in this study of cancer survivors. As noted previously, the lack of a relationship between threat appraisal and health behaviors in the present study may be due to health behaviors having already changed after cancer diagnosis. Evidence for Meditational Pathways Finally, this investigatio n examined possible interrelationships between PMT variables, fear, and health behaviors. The proposed meditational models were based on
53 the expectation that there would be significant positive relationships between fear of recurrence and health behaviors Results provided no evidence of these relationships, and, therefore, full testing of the models was not undertaken. There are several possible reasons why relationships between fear of recurrence and hea lth behaviors were not found. One possibility i s related to the fact that most survivors in the study endorsed only occasionally thinking about or being concerned that their cancer would return. Because the levels of fear of recurrence were fairly low overall, they may have been too low to motivate pa rticipants to engage in greater exercise or fruit and vegetable intake. The possibility also exists that participants higher in fear of recurrence may have been more likely to engage in other types of health behaviors that were not studied. These include activities like increased contact with medical professionals and better adherence to cancer screening regimens. Prior literature has also often failed to establish links between the fear or arousal associated with health threats and performance of behavio rs aimed at reducing the likelihood of the threat (Bowen et al., 2004; McCaul, Branstetter, positive relationship between behavior change and distress, especially for diet. Also, Pinto et al. (2002) found that initial depression predicted greater increases in physical activity over the first year after diagnosis, indicating that psychological distress can be related to later health behaviors in cancer survivors. We did not assess health behavior change, however, so further study is required to replicate these findings that link distress to health behaviors.
54 Limitations There are several limitations to note in this study. The sample included only breast cancer survivors wi th early stage disease; consequently, findings may not be generalizable to survivors of other types of cancer and survivors with more advanced disease. Also, because the data were cross sectional, the direction of the observed relationships remains unclea r. Although the data suggest that threat and coping appraisal influenced fear of recurrence, it is also possible that fear of recurrence influenced threat and coping appraisal. Similarly, for health activities, it is not clear if coping appraisal affecte d performance of healthy behaviors or if performance of healthy behaviors affected coping appraisal. In addition, this study had some limitations with regard to measurement. First, the use of single item measures to assess absolute risk, relative risk, and response efficacy raises concerns about the reliability of the information obtained. Second, the strong observed correlation between depression and fear of recurrence suggests that these constructs may be difficult to distinguish with the measures use d. Third, there are several problems associated with the use of retrospective self reports of diet and exercise, including recall bias, confusion understanding serving portion sizes in reporting diet, and positive impression management to report better he alth habits. Fourth, while fruit and vegetable intake were evaluated to assess diet recommendation adherence, participants answered items about response efficacy and self efficacy to follow a diet that included limited amounts of red and processed meats, and high levels of whole wheat consumption. Finally, change in health behaviors since diagnosis was not evaluated, so it
55 is unclear whether the health habits reported reflected changes influenced by cancer related concerns or were just part of life long h abits. Clinical Implications & Future Directions There are several clinical implications of this investigation. First, based on the high correlation between fear of recurrence and depression, patients who report moderate levels of fear of recurrence s hould also be assessed for possible depression. Second, the interaction of threat and coping appraisal predicting fear of recurrence highlights the importance of both perceptions of threat and perceptions of potentially adaptive coping strategies in deter mining which survivors will report the highest levels of fear. Future interventions should incorporate strategies to both reduce perceived threat to more efficacy to perform various healthy behaviors. Third, when creating interventions to promote healthy behaviors in cancer survivors, clinicians should focus on increasing self efficacy and response efficacy for the target behaviors. Because there was no link betwe en fear of recurrence and health behaviors, the use of scare tactics to encourage behavior change may not be effective. Instead, enhancing competence to perform the behaviors and providing information to help survivors understand what they could be doing to reduce their cancer risk are potentially better strategies. Future research should also determine whether the observed relationships hold in other samples of cancer survivors including different cancer types and survivors of more advanced disease. Rese arch should also examine relationships between PMT variables and health behavior change over time in order to determine if the same relationships hold when predicting rates of behavior change at various points in cancer survivorship.
56 Additional health beh aviors should be included as outcomes variables, including both those with empirical support that they reduce cancer mortality risk (e.g., cancer screening) and those without (e.g., popular herbal remedies and other alternative medicines). It is important to elucidate the causal nature of these relationships to further determine which modifiable factors lead to both increased fears and increased performance of health behaviors in cancer survivors. Enhanced knowledge of the variables that contribute to the se outcomes could assist in the development of interventions to improve quality of life during survivorship.
57 List of References Aiken L. S. & West, S. G. (1991). Multiple regression: Testing and interpreting interactions Newbury Park, CA: Sage. Baker, F., Denniston, M., Smith, T., & West, M. M. (2005). Adult cancer survivors: How are they faring? Cancer, 104 (11 Suppl.), 2565 2576. doi: 10.1002/cncr.21488 Bandura, A. (2006). Guide for constructing self efficacy scales. In F. Pa jares & T. Urdan (Eds.), Self efficacy beliefs of adolescents (pp. 307 337). Greenwich, CT: Information Age Publishing. Baranowski, T., Cullen, K. W., & Baranowski, J. (1999). Psychosocial correlates of dietary intake: Advancing dietary in tervention. Annu al Review of Nutrition, 19 17 40. doi: 10.1146/annurev.nutr.19.1.17 Bellizzi, K. M., Rowland, J. H., Jeffery, D. D., & McNeel, T. (2005). Health behaviors of cancer survivors: Examining opportunities for cancer control intervention. Journal of Clinical On cology, 23 (34), 8884 8893. doi:10.1200/JCO.2005.02.2343 Bianchini, F., Kaaks, R., & Vainio, H. (2002). Weight control and physical activity in cancer prevention. Obesity Reviews, 3 (1), 5 8. doi: 10.1046/j.1467 789X.2002.00046.x Blanchard, C. M., Cokkinide s, V., Courneya, K. S., Nehl, E. J., Stein, K., & Baker, F. (2003). A comparison of physical activity of posttreatment breast cancer survivors and noncancer controls. Behavioral Medicine, 28 (4), 140 149. Blanchard, C. M., Courneya, K. S., & Stein, K. (200 8). Cancer survivors' adherence to lifestyle behavior recommendations and associations with health related quality of life: Results from the American Cancer Society's SCS II. Journal of Clinical Oncology, 26 (13), 2198 2204. doi:10.1200/JCO.2007.14.6217 Bl anchard, C. M., Denniston, M. M., Baker, F., Ainsworth, S. R., Courneya, K. S., Hann, cancer diagnosis? American Journal of Health Behavior, 27 (3), 246 256. Bluman, L. G., B orstelmann, N. A., Rimer, B. K., Iglehart, J. D., & Winer, E. P. (2001). Knowledge, satisfaction, and perceived cancer risk among women diagnosed with ductal carcinoma in situ. Journal of Women's Health & Gender based Medicine, 10(6), 589 598. doi:10.1089/ 15246090152543175
58 Caan, B., Sternfeld, B., Gunderson, E., Coates, A., Quesenberry, C., & Slattery, M. L. (2005). Life After Cancer Epidemiology (LACE) Study: A cohort of early stage breast cancer survivors (United States). Cancer Causes & Control, 16 (5), 5 45 556. doi:10.1007/s10552 004 8340 3 Costanzo, E. S., Lutgendorf, S. K., Bradley, S. L., Rose, S. L., & Anderson, B. (2005). Cancer attributions, distress, and health practices among gynecologic cancer survivors. Psychosomatic Medicine, 67 (6), 972 980. d oi:10.1097/01.psy.0000188402.95398.c0 Coups, E. J., & Ostroff, J. S. (2005). A population based estimate of the prevalence of behavioral risk factors among adult cancer survivors and noncancer controls. Preventive Medicine, 40 (6), 702 711. doi: 10.1016/j.y pmed.2004.09.011 Courneya, K. S. (2003). Exercise in cancer survivors: An overview of research. Medicine & Science in Sports & Exercise, 35 (11), 1846 1852. doi:10.1249/01.MSS.0000093622.41587.B6 Courneya, K. S., & Hellsten, L. A. M. (2001). Cancer preven tion as a source of exercise motivation: An experimental test using protection motivation theory. Psychology, Health & Medicine, 6 (1), 59 64. doi:10.1080/13548500125267 Deimling, G. T., Bowman, K. F., Sterns, S., Wagner, L. J., & Kahana, B. (2006). Cancer related health worries and psychological distress among older adult, long term cancer survivors. Psycho Oncology, 15 (4), 306 320. doi:10.1002/pon.955 Denmark Wahnefried, W., Peterson, B., McBride, C., Lipkus, I., & Clipp, E. (2000). Current health behavi ors and readiness to pursue life style changes among men and women diagnosed with early stage prostate and breast carcinomas. Cancer, 88 (3), 674 684. doi:10.1002/(SICI)1097 0142(20000201)88:3<674::AID CNCR26>3.0.CO;2 R Doyle, C., Kushi, L. H., Byers, T., Courneya, K. S., Demark Wahnefried, W., Grant, cancer treatment: An American Cancer Society guide for informed choices. CA: A Cancer Journal for Clinicians, 56 (6), 323 353. doi: 10 .3322/canjclin.56.6.323 Ewertz, M., Gillanders, S., Meyer, L., & Zedeler, K. (1991). Survival of breast cancer patients in relation to factors which affect the risk of developing breast cancer. International Journal of Cancer, 49 (4), 526 530. doi:10.1002/ ijc.2910490409 Floyd, D. L., Prentice Dunn, S., & Rogers, R. W. (2000). A meta analysis of research on protection motivation theory. Journal of Applied Social Psychology, 30 (2), 407 429. doi: 10.1111/j.1559 1816.2000.tb02323.x
59 Godin, G., & Shephard, R. J. (1985). A simple method to assess exercise behavior in the community. Canadian Journal of Applied Sport Sciences, 10 (3), 141 146. Gotay, C. C., & Pagano, I. S. (2007). Assessment of Survivor Concerns (ASC): A newly proposed brief questionnaire. Health and Quality of Life Outcomes, 5 15. doi:10.1186/1477 7525 5 15 Greenwald, P., Sherwood, K., & McDonald, S. S. (1997). Fat, caloric intake, and obesity: Lifestyle risk factors for breast cancer. Journal of the American Dietetic Association, 97 (7 Suppl), S24 30. doi: 10.1016/S0002 8223(97)00726 8 Helmes, A. W. (2002). Application of the protection motivation theory to genetic testing for breast cancer risk. Preventive Medicine, 35 (5), 453 462. doi: 10.1006/pmed.2002.1110 Herschbach, P., Keller, M., Knight, L., Brandl, T., Huber, T., Henrich, B., & Marten Mittag, B. (2004). Psychological problems of cancer patients: A cancer distress screening with a cancer specific questionnaire. British Journal of Cancer, 91 (3), 504 511. doi:10.1038/sj.bjc.6601986 Holmes, M. D., Chen, W. Y., Feskanich, D., Kroenke, C. H., & Colditz, G. A. (2005). Physical activity and survival after breast cancer diagnosis. Journal of the American Medical Association, 293 (20), 2479 2486. doi:10.1001/jama.293.20.2479 Holmes, M. D., Stampfer, M J., Colditz, G. A., Rosner, B., Hunter, D. J., & Willett, W. C. (1999). Dietary factors and the survival of women with breast carcinoma. Cancer, 86 (5), 826 835. doi:10.1002/(SICI)1097 0142(19990901)86:5<826::AID CNCR19>3.0.CO;2 0 Ingram, D. (1994). Die t and subsequent survival in women with breast cancer. British Journal of Cancer, 69 (3), 592 595. Irwin, M. L., Crumley, D., McTiernan, A., Bernstein, L., Baumgartner, R., Gilliland, F. Barbash, R. (2003). Physical activity levels before and after a diagnosis of breast carcinoma: The Health, Eating, Activity, and Lifestyle (HEAL) study. Cancer, 97 (7), 1746 1757. doi: 10.1002/cncr.11227 Jain, M., Miller, A. B., & To, T. (1994). Premorbid diet and the prognosis of women with breast cancer. Jour nal of the National Cancer Institute, 86 (18), 1390 1397. doi:10.1093/jnci/86.18.1390
60 Knols, R., Aaronson, N. K., Uebelhart, D., Fransen, J., & Aufdemkampe, G. (2005). Physical exercise in cancer patients during and after medical treatment: A systematic review of randomized and controlled clinical trials. Journal of Clinical Oncology, 23 (16), 3830 3842. doi:10.1200/JCO.2005.02.148 Kroenke, C. H., Chen, W. Y., Rosner, B., & Holmes, M. D. (2005). Weight, weight gain, and survival after breast cancer diagno sis. Journal of Clinical Oncology, 23 (7), 1370 1378. doi:10.1200/JCO.2005.01.079 American Cancer Society 2006 Nutrition and Physical Activity Guidelines Advisory Commit tee. (2006). American Cancer Society Guidelines on Nutrition and Physical Activity for cancer prevention: Reducing the risk of cancer with healthy food choices and physical activity. CA: A Cancer Journal for Clinicians, 56 (5), 254 281; quiz 313 254. doi: 10 .3322/canjclin.56.5.254 Lash, T. L., Clough Gorr, K., & Silliman, R. A. (2005). Reduced rates of cancer related worries and mortality associated with guideline surveillance after breast cancer therapy. Breast Cancer Research and Treatment, 89 (1), 61 67. d oi:10.1007/s10549 004 1472 z Lee Jones, C., Humphris, G., Dixon, R., & Hatcher, M. B. (1997). Fear of recurrence: A literature review and proposed cognitive formulation to explain exacerbation of recurrence fears. Psycho Oncology, 6 95 105. doi:10.1002/( SICI)1099 1611(199706)6:2<95::AID PON250>3.0.CO;2 B Lebel, S., Rosberger, Z., Edgar, L., & Devins, G. M. (2009). Emotional distress impacts fear of the future among breast cancer survivors not the reverse. Journal of Cancer Survivorship, 3 117 127. doi: 1 0.1007/s11764 009 0082 5 Lerman, C., Trock, B., Rimer, B. K., Jepson, C., Brody, D., & Boyce, A. (1991). Psychological side effects of breast cancer screening. Health Psychology, 10 (4), 259 267. Leventhal, H., Diefenbach, M., & Leventhal, E. A. (1992). I llness cognition: Using common sense to understand treatment adherence and affect cognition interactions. Cognitive Therapy and Research, 16 (2), 143 163. doi:10.1007/BF01173486 Maddux, J. E., & Rogers, R. W. (1983). Protection motivation and self efficacy : A revised theory of fear appeals and attitude change. Journal of Experimental Social Psychology, 19 469 479. doi:10.1016/0022 1031(83)90023 9
61 Maskarinec, G., Murphy, S., Shumay, D. M., & Kakai, H. (2001). Dietary changes among cancer survivors. Europea n Jounral in Cancer Care (English), 10 (1), 12 20. doi: 10.1111/j.1365 2354.2001.00245.x Maunsell, E., Drolet, M., Brisson, J., Robert, J., & Deschenes, L. (2002). Dietary change after breast cancer: Extent, predictors, and relation with psychological distr ess. Journal of Clinical Oncology, 20 (4), 1017 1025. Mayer, D. K., Terrin, N. C., Menon, U., Kreps, G. L., McCance, K., Parsons, S. K., & Mooney, K. H. (2007). Health behaviors in cancer survivors. Oncology Nursing Forum, 34 (3), 643 651. doi:10.1188/07.ON F.643 651 McCaul, K. D., Branstetter, A. D., O'Donnell, S. M., Jacobson, K., & Quinlan, K. B. (1998). A descriptive study of breast cancer worry. Journal of Behavioral Medicine, 21(6), 565 579. doi:10.1023/A:1018748712987 McCaul, K. D. & Mullens, A. B. ( 2003). Affect, thought and self protective health behavior: The case of worry and cancer screening. Social psychological foundations of health and illness. In J. Suls & K. A. Wallston (Eds.), Social Psychological Foundations of Health and Illness (pp. 137 168). Malden, MA: Blackwell Publishing. McClendon, B. T., & Prentice Dunn, S. (2001). Reducing skin cancer risk: An intervention based on protection motivation theory. Journal of Health Psychology, 6 (3), 321 328. doi:10.1177/135910530100600305 McGinty, H. L., Jacobsen, P. B., & Andrykowski, M. (2008, June). Mental health and worries of cancer recurrence in breast cancer survivors Poster session presented at the 4th Biennial Cancer Survivorship Research Conference, Atlanta, GA. McMahan, S., Witte, K., & Meyer, J. (1998). The perception of risk messages regarding electromagnetic fields: Extending the Parallel Process Model to an unknown risk. Health Communication, 10(3), 247 259. doi:10.1207/s15327027hc1003_4 Mellon, S., Kershaw, T. S., Northouse, L. L. & Freeman Gibb, L. (2007). A family based model to predict fear of recurrence for cancer survivors and their caregivers. Psycho Oncology, 16 (3), 214 223. doi: 10.1002/pon.1074 Miller, S. M., Shoda, Y., & Hurley, K. (1996). Applying cognitive social theor y to health protective behavior: Breast self examination in cancer screening. Psychological Bulletin, 119 (1), 70 94. doi:10.1037/0033 2909.119.1.70
62 Milne, S., Sheeran, P., & Orbell, S. (2000). Prediction and intervention in health related behavior: A me ta analytic review of protection motivation theory. Journal of Applied Social Psychology, 30 (1), 106 143. doi: 10.1111/j.1559 1816.2000.tb02308.x Moss Morris, R., Weinman, J., Petrie, K. J., Horne, R., Cameron, L. D., & Buick, D. (2002). The Revised Illnes s Perception Questionnaire (IPQ R). Psychology and Health, 17 (1), 1 16. doi:10.1080/08870440290001494 Norman, S. A., Potashnik, S. L., Galantino, M. L., De Michele, A. M., House, L., & Localio, A. R. (2007). Modifiable risk factors for breast cancer recur rence: What can we tell survivors? Journal of Womens Health (Larchmt), 16 (2), 177 190. doi: 10.1089/jwh.2006.0047 Northouse, L. L. (1981). Mastectomy patients and the fear of cancer recurrence. Cancer Nursing, 4 (3), 213 220. Pinto, B. M., Trunzo, J. J., Reiss, P., & Shiu, S. Y. (2002). Exercise participation after diagnosis of breast cancer: Trends and effects on mood and quality of life. Psycho Oncology, 11 (5), 389 400. doi: 10.1002/pon.594 Plotnikoff, R. C., & Higginbotham, N. (1998). Protection motivat ion theory and the prediction of exercise and low fat diet behaviours among Australian cardiac patients. Psychology and Health, 13 (3), 411 429. doi:10.1080/08870449808407300 Rabin, C., Leventhal, H., & Goodin, S. (2004). Conceptualization of disease timel ine predicts posttreatment distress in breast cancer patients. Health Psychology, 23 (4), 407 412. doi:10.1037/0278 618.104.22.1687 Rabin, C., & Pinto, B. (2006). Cancer related beliefs and health behavior change among breast cancer survivors and their first degree relatives. Psycho Oncology, 15 (8), 701 712. doi:10.1002/pon.1000 Rabin, C., Pinto, B., Dunsiger, S., Nash, J., & Trask, P. (2008). Exercise and relaxation intervention for breast cancer survivors: Feasibility, acceptability and effects. Psycho Onc ology doi:10.1002/pon.1341 Radloff, L. S. (1977). The CES D Scale: A self report depression scale for research in the general population. Applied Psychological Measurement, 1 (3), 385 401. doi:10.1177/014662167700100306
63 Rhodes, R. E., & Plotnikoff, R. C. (2005). Can current physical activity act as a reasonable proxy measure of future physical activity? Evaluating cross sectional and passive prospective designs with the use of social cognition models. Preventive Medicine, 40 (5), 547 555. doi:10.1016/j.y pmed.2004.07.016 Ripptoe, P. A., & Rogers, R. W. (1987). Effects of components of protection motivation theory on adaptive and maladaptive coping with a health threat. Journal of Personality and Social Psychology, 52 (3), 596 604. doi: 10.1037/0022 3514.52 .3.596 the Women's Healthy Eating and Living Study Group. (1999). Factors associated with weight gain in women after diagnosis of breast cancer. Women's Healthy Eati ng and Living Study Group. Journal of the American Dietetic Association, 99 (10), 1212 1221. doi: 10.1016/S0002 8223(99)00298 9 Rogers, R. W. (1975). A Protection motivation theory of fear appeals and attitude change. The Journal of Psychology, 91 93 114. Rohan, T. E., Hiller, J. E., & McMichael, A. J. (1993). Dietary factors and survival from breast cancer. Nutrition and Cancer, 20 (2), 167 177. doi:10.1080/01635589309514283 Salovey, P., Rothman, A. J., & Rodin, J. (1998). Health behavior. In D. T. Gilber t, S. T. Fiske, & G. Lindzey (Eds.), The Handbook of Social Psychology (pp. 633 683). New York, NY: McGraw Hill. Schmitz, K. H., Holtzman, J., Courneya, K. S., Masse, L. C., Duval, S., & Kane, R. (2005). Controlled physical activity trials in cancer surv ivors: A systematic review and meta analysis. Cancer Epidemiology Biomarkers & Prevention, 14 (7), 1588 1595. doi: 10.1158/1055 9965.EPI 04 0703 Schwarzer, R. (2008). Modeling health behavior change: How to predict and modify the adoption and maintenance or health behaviors. Applied Psychology: An International Review, 57 (1), 1 29. doi:10.1111/j.1464 0597.2007.00325.x Seydel, E., Taal, E., & Wiegman, O. (1990). Risk appraisal, outcome and self efficacy expectancies: Cognitive factors in preventative behavio ur related to cancer. Psychology and Health, 4 99 109. doi:10.1080/08870449008408144 Sheeshka, J. D., Woolcott, D. M., & MacKinnon, N. J. (1993). Social cognitive theory as a framework to explain intentions to practice healthy eating behaviors. Journal o f Applied Social Psychology, 23 (19), 1547 1573. doi: 10.1111/j.1559 1816.1993.tb01047.x
64 Sherwood, N. E., & Jeffery, R. W. (2000). The behavioral determinants of exercise: Implications for physical activi ty interventions. Annual Review of Nutrition, 20 21 44. doi: 10.1146/annurev.nutr.20.1.21 Shin, Y., Jang, H., & Pender, N. J. (2001). Psychometric evaluation of the exercise self efficacy scale among Korean adults with chronic diseases. Research in Nursing & Health, 24 (1), 68 76. doi:10.1002/1098 240X(20010 2)24:1<68::AID NUR1008>3.0.CO;2 C Stanley, M. A., & Maddux, J. E. (1986). Cognitive processes in health enhancement: Investigation of a combined protection motivation and self efficacy model. Basic and Applied Social Psychology, 7 (2), 101 113. doi:10.1207 /s15324834basp0702_2 Stanton, A. L., Danoff Burg, S., & Huggins, M. E. (2002). The first year after breast cancer diagnosis: Hope and coping strategies as predictors of adjustment. Psycho Oncology, 11 (2), 93 102. doi:10.1002/pon.574 Stanton, A. L., Ganz P. A., Rowland, J. H., Meyerowitz, B. E., Krupnick, J. L., & Sears, S. R. (2005). Promoting adjustment after treatment for cancer. Cancer, 104 (11 Suppl), 2608 2613. doi:10.1002/cncr.21246 Taylor, S. E. (1983). Adjustment to threatening events: A theory of cognitive adaptation. American Psychologist, 38 (11), 1161 1173. doi:10.1037/0003 066X.38.11.1161 Thompson, F. E., Subar, A. F., Smith, A. F., Midthune, D., Radimer, K. L., Kahle, L. L., & Kipnis, V. (2002). Fruit and vegetable assessment: Performance o f 2 new short instruments and a food frequency questionnaire. Journal of the American Dietetic Association, 102 (12), 1764 1772. doi: 10.1016/S0002 8223(02)90379 2 Thomson, C. A., Flatt, S. W., Rock, C. L., Ritenbaugh, C., Newman, V., & Pierce, J. P. (2002). Increased fruit, vegetable and fiber intake and lower fat intake reported among women previously treated for invasive breast cancer. Journal of the American Dietetic Association, 102 (6), 801 808. doi: 10.1016/S0002 8 223(02)90180 X Trost, S. G., Owen, N., Bauman, A. E., Sallis, J. F., & Brown, W. (2002). Correlates of adults' participation in physical activity: Review and update. Medicine & Science in Sports & Exercise, 34 (12), 1996 2001. doi: 10.1249/01.MSS.0000038974 .76900.92 Valdimarsdottir, H. B., Bovbjerg, D. H., Kash, K. M., Holland, J. C., Osborne, M. P., & Miller, D. G. (1995). Psychological distress in women with a familial risk of breast cancer. Psycho Oncology, 4 133 141. doi:10.1002/pon.2960040207
65 van den Beuken van Everdingen, M. H., Peters, M. L., de Rijke, J. M., Schouten, H. C., van Kleef, M., & Patijn, J. (2008). Concerns of former breast cancer patients about disease recurrence: A validation and prevalence study. Psycho Oncology, 17 (11), 1137 1145. d oi:10.1002/pon.1340 Vickberg, S. M. (2003). The Concerns About Recurrence Scale (CARS): A systematic measure of women's fears about the possibility of breast cancer recurrence. Annals of Behavioral Medicine, 25 (1), 16 24. doi:10.1207/S15324796ABM2501_03 Willett, W. C. (2005). Diet and cancer: An evolving picture. Journal of the American Medical Association, 293 (2), 233 234. Witte, K. (1992). Putting the fear back into fear appeals: The extended parallel process model. Communication Monographs, 59, 329 3 49. Witte, K. (1998). Fear as motivator, fear as inhibitor: Using the Extended Parallel Process Model to explain fear appeal successes and failures. In P. A. Andersen & L. K. Guerrero (Eds.), The Handbook of Communication and Emotion: Research, Theory, A pplications, and Contexts (pp. 423 450). San Diego, CA: Academic Press. Wurtele, S. K., & Maddux, J. E. (1987). Relative contributions of protection motivation theory components in predicting exercise intentions and behavior. Health Psychology, 6 (5), 453 466.