UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The relationship between pain appraisals and coping strategy use and adaptation to chronic low back pain:… Grant, Lynda D. 1997

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1997-25061x.pdf [ 7.9MB ]
Metadata
JSON: 831-1.0053995.json
JSON-LD: 831-1.0053995-ld.json
RDF/XML (Pretty): 831-1.0053995-rdf.xml
RDF/JSON: 831-1.0053995-rdf.json
Turtle: 831-1.0053995-turtle.txt
N-Triples: 831-1.0053995-rdf-ntriples.txt
Original Record: 831-1.0053995-source.json
Full Text
831-1.0053995-fulltext.txt
Citation
831-1.0053995.ris

Full Text

T H E RELATIONSHIP B E T W E E N PAIN APPRAISALS A N D COPING S T R A T E G Y U S E A N D A D A P T A T I O N T O C H R O N I C L O W B A C K PAIN A D A I L Y D I A R Y S T U D Y by L Y N D A D. G R A N T B.A. , Northeastern College, NJ. 1974 M . A . , Simon Fraser University, 1987 A THESIS S U B M I T T E D IN P A R T I A L F U L F I L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F D O C T O R O F PHILOSOPHY in T H E F A C U L T Y O F E D U C A T I O N (Department of Counselling Psychology) T H E U N I V E R S I T Y O F BRITISH C O L U M B I A October 7, 1997 © Lynda D. Grant, 1997 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that . copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department The University of British Columbia Vancouver, Canada Date OcV l^/q7 DE-6 (2/88) 11 A B S T R A C T Data from daily diaries were used to examine the relationships between daily pain appraisals (Catastrophizing, Self-Efficacy, and perceived control over pain) and coping strategy use (Distraction, Ignoring Pain, Praying and Hoping, and Reinterpreting Pain Sensation) and nighttime negative mood and pain intensity for 88 women (mean age 46.83 years, SD 11.90) with chronic low back pain who were not attending a specialized pain treatment program. These relationships were examined at two levels using the Hierarchical Linear Modeling program (Bryk & Raudenbush, 1992). The first level of analyses examined whether pain appraisals and coping strategy use during the day predicted levels of nighttime depressed and anxious mood, and pain. This analysis was based on 30 days of monitoring for each participant. The second level of analyses examined whether these daily processes could be predicted by psychosocial and functional variables important to the experience of chronic pain. This analysis was based on the Mutidimensional Pain Inventory (Kerns, Turk, & Rudy, 1985) completed prior to participants beginning the daily monitoring. There were four major findings in this study. First, pain appraisals were more predictive of negative mood and pain intensity than coping strategy use, with Catastrophizing the strongest predictor of depressed and anxious mood, and control the strongest predictor of pain intensity. Second, general affective distress predicted higher levels of negative mood on a daily basis. Third, women who perceived their pain to be interfering a great deal in their lives were more anxious on a daily basis. Fourth, punishing spousal responses predicted nightly negative mood and pain more than solicitous or distracting spousal responses. These results are similar to findings based on patients attending pain treatment programs. This suggests that some of the same processes identified in clinical pain patients may apply to low back pain sufferers in the community who are comparable to study participants. The implications of these findings for pain research and treatment are discussed. IV T A B L E O F C O N T E N T S A B S T R A C T ii T A B L E O F C O N T E N T S iv LIST O F T A B L E S vii LIST O F FIGURES ix A C K N O W L E D G M E N T S x I N T R O D U C T I O N 1 R E V I E W O F T H E L I T E R A T U R E 9 Adaptation 9 Stress 12 Appraisal 15 Critique 21 Coping 30 Critique 33 Coping Self-Statements Subscale 35 Increased Behavioral Activities Subscale 35 Reinterpreting Pain Sensation and Praying and Hoping Subscales 36 Ignoring Pain and Diverting Attention Subscales 36 Pain Duration 42 Summary 43 H Y P O T H E S E S 48 The Relationship Between Daily Nighttime Depressed Mood, Nighttime Anxious Mood, and Nighttime Pain Intensity, Appraisals, and Coping Strategy Use 48 Hypothesis 1 48 The Relationship Between the MPI Subscales and Daily Responses to Pain 49 Hypothesis 2 49 Exploratory Analyses 50 Exploratory Analyses 1 50 Exploratory Analyses 2 50 Exploratory Analyses 3 50 V M E T H O D 51 Design 51 Participants 51 Measures 53 Coping Strategy Questionnaire 54 Catastrophizing 55 Perceived Control Over Pain 56 Self-Efficacy 57 State-Trait Personality Inventory - Anxiety and Depression Subscales 58 Pain intensity.... 60 Multidimensional Pain Inventory 60 McGi l l Pain Questionnaire 65 Procedure 67 Data Analyses 68 Missing data 68 Preliminary Analyses 69 Analyses of Diary Data: Hierarchical Linear Models 70 Steps in H L M Model Building 75 Null Model 75 Level-1 Models 76 Level-2 Models '.....77 Interaction Effects 78 R E S U L T S 79 Completers Versus Non-Completers Comparability 79 Descriptive Statistics 80 Means, standard deviations, and Cronbach's alphas for the daily variables 80 Correlations for the Daily Variables and the Time 1 MPI 82 Comparisons With Selected Pain Populations on the MPI and M P Q 86 Main Analyses 87 Overview 87 Null Models... 88 Nighttime depressed mood 90 Level-1 analyses 90 Level-2 analyses 95 Nighttime anxious mood 101 Level-1 analyses 101 Level-2 analyses 105 vi Nighttime pain intensity I l l Level-1 analyses I l l Level-2 analyses 116 Exploratory Analyses of the Interaction Effects ...121 DISCUSSION 125 Predicting Daily Negative Mood and Pain 126 Appraisals 126 Coping Strategy Use 128 The Role of Psychosocial and Functional Variables on Daily Responses to Pain 133 Conclusions 138 Limitations of the Study 139 Research Implications 143 Clinical Implications 145 REFERENCES 147 APPENDICES Appendix A: Summary Tables of the Literature Review 158 Appendix B: Demographic Questionnaire and Data Summary 176 Appendix C: Consent Form 185 Appendix D: Equations for the H L M Models 187 Appendix E: Completers Versus Non-Completers Data 192 Appendix F: Correlations for Days 15 and 30, and Comparisons With Clinical Pain Studies 196 Appendix G: Results of the H L M Level-2 Analyses for the MPI Functional and Interpersonal Subscales 205 Appendix H: Tables and Figures for the Significant Interactions 208 Appendix I: Examples of Questions For Each of The Multi-Item Measures 215 LIST O F T A B L E S Table 1. Means and Standard Deviations for the MPI Subscales for Time 1 and Time 2 and Cronbach's alphas for Time 1 64 Table 2. Means, Standard Deviations, and Cronbach's Alphas for the Daily variables for days 1, 15, and 30 81 Table 3. Pearson-Product Moment Correlation matrix of Demographic and Daily Variables for Day 1 83 Table 4. Pearson-Product Moment Correlation Matrix for Time 1 MPI, Demographic, and Daily Variables for Day 1 84 Table 5. H L M Results for the Null Models for Nighttime Depressed Mood, Nighttime Anxious Mood, and Nighttime Pain Intensity .. 89 Table 6. Level-1 H L M Analyses of the Effects of Morning Depressed Mood, the Three Pain Appraisals, and Distraction, Ignoring Pain, and Praying and Hoping on Nighttime Depressed Mood 91 Table 7. Level-2 H L M Results for Nighttime Depressed Mood. The Relationships Between the MPI Functional Subscales and the Level-1 Intercepts and Morning Depressed Mood Slope 97 Table 8. Level-2 H L M Results for Nighttime Depressed Mood. The Relationships Between the MPI Interpersonal Subscales and the Level-1 Intercepts and Morning Depressed Mood Slope 99 Table 9. Level-1 H L M Analysis of the Effects of Morning Anxious Mood, the Three Pain Appraisals, and Distraction, Ignoring Pain, and Praying and Hoping on Nighttime Anxious Mood 102 Table 10. Level-2 H L M Results for Nighttime Anxious Mood. The Relationships Between the MPI Functional Subscales and the Level-1 Intercepts and Morning Anxious Mood Slope 107 Table 11. Level-2 H L M Results for Nighttime Anxious Mood. The Relationships Between the MPI Interpersonal Subscales and the Level-1 Intercepts and Morning Anxious Mood 110 Table 12. Level-1 H L M Analysis of the Effects of Morning Pain Intensity, the Three Pain Appraisals, and the Four Coping Strategies for Nighttime Pain Intensity 113 Table 13. Level-2 H L M Results for Nighttime Pain Intensity. The Relationships Between the MPI Functional Subscales and the Level-1 Intercepts and Morning Pain Intensity Slope 117 Table 14. Level-2 H L M Results for Nighttime Pain Intensity. The Relationships Between the MPI Interpersonal Subscales and the Level-1 Intercepts and Morning Pain Intensity Slope 120 ix LIST OF FIGURES Figure 1. Graphic Representation of H L M Level-1 Analyses 72 Figure 2. Graphic Representation of H L M Level-2 Analyses 74 A C K O W L E D G E M E N T S I would like to thank Bonita Long, my dissertation supervisor, for her support during the preparation of this thesis. Her dedication to her students, her thoroughness, and her desire to learn and stretch herself has contributed greatly to my personal and scholastic development. I would also like to thank the members of my committee, Ken Craig and Anita DeLongis. Their knowledge and expertise greatly strengthened this research study. Additional thanks are extended to Doug Willms for sharing his statistical expertise in H L M , and to the 88 women whose willingness to give so much time to complete the questionnaires made this project possible. Last, but certainly not least, I would like to thank my husband, Ron Gibson. He has never wavered in his support and encouragement, despite having to take on the additional tasks and burdens that are required of a student's spouse. This Ph.D has been very much of a team effort. 1 CHAPTER ONE Introduction Adaptation to pain is considered to be a multidimensional phenomena that involves a broad range of psychosocial and functional variables (Turk & Rudy, 1988). These include perceptions of pain severity, interference of pain on life, psychological well-being, social support, ability to control pain, and functional abilities. Although most people with chronic pain adapt well to their condition (Jensen, Turner, Romano, & Karoly, 1991), a certain proportion of sufferers do not and find their pain both emotionally and physically disabling. These people are responsible for the largest percentage of annual pain related work days lost, health care costs, and compensation payments (Nachemson, 1984). Consequently, the identification of factors that help or hinder peoples' abilities to adapt to chronic pain may help us understand the wide variation in response to pain, and increase our clinical abilities to assist those who are not adapting well. Thus, the purpose of this study was to examine the relationships between daily pain appraisals and coping strategy use and nighttime negative mood and pain, and whether these daily processes were predicted by psychosocial and functional variables important to the experience of chronic pain. Over the past 10 years there has been a noticeable increase in research focused on chronic pain and adaptation. Recently, models of stress and coping 2 have been called upon to explain adaptational differences among chronic pain sufferers (Jensen & Karoly, 1991; Keefe et al., 1987). One of the models used most frequently is Lazarus's transactional model of stress and coping (Lazarus & Folkman, 1984). Within this model, pain is the stressor, and pain appraisals (personal judgments about pain) and coping strategies (purposeful efforts to manage the pain) are believed to play a central role in how well people adapt to pain. There is a substantial literature supporting the relationship between pain specific appraisals and adaptation to chronic pain. Catastrophizing, perceived control over pain, and self-efficacy judgments are three appraisals that have been shown to be most central to adaptation (Jensen & Karoly, 1991). Catastrophizing appraisals are defined as frequent negative and worrying thoughts about pain and the prognosis for the future (Turner, 1991). They have been shown to have significant positive associations with pain intensity (Flor & Turk, 1988; Hill , 1993; Keefe, Browne, Wallston, & Caldwell, 1989; Romano, Turner, Syrjala, & Levy, 1987), anxiety and depression (Hill, 1993; Keefe et al., 1989; McCracken & Gross, 1933; Romano et al., 1987), and functional limitations (Flor & Turk, 1988, Reesor & Craig, 1988). Appraisals of perceived control over pain have been found to be positively associated with physical activity (Spinhoven, Ter Kuile, Linssen, & Gazendam, 3 1989) and better psychological functioning (Jensen & Karoly, 1991), and negatively associated with pain intensity (Spinhoven et al., 1989; Spinhoven & Linssen, 1991), and depression, anxiety, and poorer psychological functioning (Affleck, Tennen, Pfeiffer, & Fifield, 1987; Keefe, et al., 1989; Keefe & Williams, 1990) . Self-efficacy appraisals have shown significant positive relationships with the amount of exercise and physical activity people engage in (Council, Ahern, Follick, & Kline, 1988; Dolce, Crocker, & Doleys, 1986; Dolce, Crocker, Moletteire, & Doleys, 1986), and the frequency of coping strategy use (Jensen, Turner, & Karoly, 1991). Self-efficacy appraisals for pain, functioning, and the ability to manage non-arthritis symptoms have all been shown to be significantly and negatively correlated with pain intensity, depression, and disability for arthritis patients (Lorig, Chastain, Ung, Shoor, & Holman, 1989; O'Leary, Shoor, Lorig, & Holman, 1988; Regan, Lorig, & Thorensen, 1988). A substantial number of studies have examined the relationship between coping strategy use and adaptation. One of the most frequently used measures of coping with pain is the Coping Strategy Questionnaire (CSQ; Rosentiel & Keefe, 1983). However, factor analytic studies of the CSQ have failed to find a reliable factor structure, or consistent relationships between individual factors and adaptational outcomes (Hill, 1993; Jensen, Turner, & Romano, 1994; Spinhoven & 4 Lissen, 1991). Studies using the individual C S Q subscales have shown more promising results, although the results have been somewhat weak and are not always consistent. Taken as a whole, these studies suggest that coping strategy use is related to psychological and physical functioning, pain intensity, and interference of pain on life (Hill, 1993; Jensen & Karoly, 1991; Keefe & Williams, 1990; Romano et al., 1987). In summary, there is evidence that the appraisals of Catastrophizing, perceived control over pain, and Self-Efficacy, and coping strategy use are associated with adaptation to chronic pain. However, we are still left with at least three unanswered questions. First, what is the daily relationship between pain appraisals and coping strategy use and responses to pain such as depression, anxiety, and pain intensity? A sub-component of this question is how independent are the appraisals and the coping strategies in their ability to predict adaptation to pain when they are considered together in the analysis? Second, what is the relationship between these daily processes and psychosocial and functional variables important to the experience of chronic pain? Third, are the relationships identified in the literature similar for non-clinical pain populations, which are the vast majority of chronic pain sufferers (Sternbach, 1986)? Part of the reason for these gaps in our knowledge is due to methodological issues such as retrospective recall and cross-sectional data collection methods, the tendency to use 5 unidimensional measures of the pain experience, and the almost exclusive reliance on clinical pain populations. The use of retrospective recall data collection methods, where participants are asked to remember what they did or thought over a number of days or weeks, and then summarize this temporal experience into one response creates concerns about measurement error. A number of authors have expressed concern about recall error and bias, as well as responses that are confounded by person or situational variables that are unique to the time of measurement (Affleck, Urrows, Tennen, & Higgins, 1992; DeLongis, Hemphill, & Lehman, 1992; Epstein, 1986; Stone & Kennedy-Moore, 1992; Wethington & Kessler, 1993). The reliance on cross-sectional research designs creates a number of difficulties. First, Lazarus and Folkman (1894) emphasize that the coping process is best understood as an ongoing interaction with the environment rather than a static event. One or few measurement points do not allow for the examination of the variation in the relationship between appraisals and coping strategy use and same day responses to pain, or how these daily processes relate to overall adaptation. Second, even though the coping process is variable, the importance of some stability in this process to psychological well-being and social functioning is also recognized (Folkman, 1992; Lazarus, 1990). Cross-sectional designs can not 6 identify these more stable processes that occur in responses over time (Epstein, 1986). The most common methods of assessing adaptation to pain in the chronic pain literature have been based on the concepts of psychological well-being and psychopathology, especially anxiety and depression (Jensen, Turner, Romano, & Karoly, 1991; Turk & Rudy, 1992). However, some authors have argued that this is a unidimensional conceptualization of the pain experience, and not sufficient to represent the midtidimensionality of this complex phenomena (Turk & Rudy, 1988). These researchers suggest that pain studies need to examine a wide range of psychosocial and functional variables important to the experience of chronic pain (Kerns, Turk, & Rudy, 1985; Turk & Rudy, 1988). The almost exclusive use of clinical pain populations (i.e., those attending specialized pain treatment programs) creates a number of difficulties. First, there is evidence to suggest that these patients are more depressed, have higher levels of pain intensity, and use more social withdrawal as a response to pain than chronic pain sufferers in the general community (Crooks, Rideout, & Browne, 1984). These findings suggest that we should be cautious about generalizing research results obtained from participants actively involved in pain treatment programs to non-clinical pain populations. 7 This study sought to address these methodological issues in several ways. First, data were collected via daily diaries for 30 days. Daily diaries are self-report instruments that are completed one or more times a day over a period of days or weeks. They are designed to pick up day-to-day variations in participants' responses to pain, as well as more stable patterns of responding. This type of data collection typically yields stronger predictive validity as it greatly reduces the recall period, thereby reducing recall error and increasing reports of prevalence of coping responses (Bolger, DeLongis, Kessler, & Schilling, 1989; DeLongis, Folkman, & Lazarus, 1988; Suls, Wan, & Blanchard, 1994; Verbrugge, 1980). Second, psychosocial and functional variables, measured by the Multidimensional Pain Inventory (MPI; Kerns et al., 1985), were also examined to see whether they predicted the daily pain processes identified in the daily diary analyses. Third, participants were recruited for the study via media advertisements from the community rather than from specialized pain clinics. I expected to find that (a) higher levels of negative mood at night would be associated with higher levels of catastrophic thinking and the use of Praying and Hoping coping strategies during the day, whereas lower levels of negative mood would be associated with higher levels of Self-Efficacy and control appraisals, and the more frequent use of Distraction and Ignoring Pain coping strategies; (b) higher levels of pain intensity at night would be associated with higher levels of 8 catastrophic thinking and the use of Praying and Hoping, Distraction, and Ignoring Pain coping strategies during the day, whereas lower levels of pain would be associated with higher levels of Self-Efficacy and control appraisals; and (c) people who generally perceived themselves to be adapting well to their pain from a psychosocial and functional perspective would report lower levels of negative mood and pain intensity on a daily basis. There is also some suggestion in the literature that there may be interaction effects between pain appraisals and pain intensity, and coping strategy use and pain intensity. Estlander and Harkapaa (1989) found that chronic pain patients engaged in significantly higher levels of catastrophizing when experiencing increased levels of pain, and Jensen and Karoly (1991) found that control appraisals had a significant positive effect on activity levels only when their interaction with lower levels of pain intensity was taken into account. Both of these studies also found that the use of more active coping strategies appeared to increase during lower levels of pain, and the use of more passive strategies increased during higher levels of pain. Because so few studies have examined these types of relationship, I included them in this study for exploratory purposes only. 9 C H A P T E R T W O Review of the Literature In order to understand more clearly the relationships between pain appraisals, coping strategy use, overall adaptation, and daily responses to pain, I review the relevant conceptual and empirical literature on adaptation to chronic pain. First, I provide an overview of the theoretical constructs that form the basis for this research project. Second, I review research on how these constructs pertain to adaptation to chronic pain. Finally, I discuss the limitations of our current procedures for examining adaptation to chronic pain. Richard Lazarus is recognized as a leader in the field of stress and coping. His emphasis on complex mental processes has advanced our understanding of how people cope with stressful events. His cognitive-affective, transactional, process-oriented conceptualization of stress and coping has been widely adopted within the chronic pain and coping literature (Jensen, Turner, Romano, & Karoly, 1991). Four constructs from Lazarus's stress and coping model that relate to chronic pain are reviewed: adaptation, stress, appraisal, and coping. Adaptation As Lazarus states, no matter how appraisals and coping processes are defined or conceptualized, their primary task is to affect adaptational outcomes (Lazarus & Folkman, 1984). In the chronic pain literature, the most common 10 methods of assessing adaptation have been based on the concepts of psychological well-being and psychopathology, especially anxiety and depression (Jensen, Turner, Romano, & Karoly, 1991; T u r k & Rudy, 1992). Recently, some researchers have been calling for an abandonment of this type of unidimensional conceptualization. They argue that it is not sufficient to represent the multidimensionality of the complex phenomena of pain (Turk & Rudy, 1988). Kerns et al. (1985) developed the Multidimensional Pain Inventory (MPI) in order to address the above issue. The MPI is a multifactor instrument designed to measure a broad domain of psychosocial and functional variables important to the experience of chronic pain. The MPI functional subscales are: Life Control, Pain Severity, Interference, Affective Distress, Support, and General Activity Level. The MPI interpersonal subscales are: Punishing Responses, Solicitous Responses, and Distracting Responses. With regards to the MPI functional subscales, Life Control has been found to be negatively related to measures of catastrophizing and helplessness, rs = -.62 and -.65, respectively, p_ < .01 (Flor & Turk, 1988), and depression, r = -.44, p < .01 (Turk, Okifuji, & Schaff, 1995). Pain Severity has shown positive correlations with catastrophizing, r = .37, p<.01, and negative correlations with control over pain, r = -.39, p < .01 (Geisser, Robinson, & Henson, 1994). For arthritis patients, General Activity Level has been found to be positively associated with self-11 efficacy for symptoms other than arthritis, r = .26, g < .01, and negatively associated with the use of more dependent types of coping strategies, r = -.21, p < .01 (Regan et al., 1988). Interference of pain on life has demonstrated positive relationships with catastrophizing, r = .52, p < .01 (Geisser, Robinson, & Henson, 1994) and depression, r = .27, p < .01 (Turk et al., 1995), and negative relationships with the perception of ability to decrease pain, r = -. 19, p_ < .05 (Geisser, Robinson, & Henson, 1994) With regards to the MPI interpersonal subscales, Schwartz, Slater, and Birchler (1996) found that Punishing Responses accounted for 12% of the variance in the prediction of increased average pain, 15% of the variance for increased functional impairment, and 15% of the variance for increased psychosocial impairment. Kerns et al. (1991) found that this subscale accounted for 12% of the variance for increased affective distress. Kerns et al. (1991) found that the Solicitous Responses subscale accounted for 3% of the variance in increased help-seeking behaviour and 5% of the variance in increased pain behaviour. Lousberg, Schmidt, and Groenman (1992) found that patients who perceived their spouses to be solicitous reported greater increases in pain when their spouses were observing them than patients with non-solicitous spouses, F(l,36) = 4.69, p_ < .01. Flor, Turk, and Rudy (1989) found this subscale correlated moderately with Pain Impact, which was a composite score of the MPI 12 Pain Severity and Pain Interference subscales, r = .45, rj < .01, and accounted for 35% of the variance in increased marital satisfaction. In contrast, Schwartz et al. (1996) and Burns et al. (1996) found no relationship between Solicitous Responses and poorer adaptation to pain. Finally, the Distracting Responses subscale has accounted for 6% of the variance in the prediction of increased pain behaviour, and 5% of the variance in increased help-seeking behaviour (Kerns et al., 1991). Distracting Responses have also been found to be significantly and positively related to pain severity (Kerns et al., 1990). In summary, adaptation to chronic pain is a complex phenomena that includes psychological, social, and functional components. The MPI has shown utility in assessing these components for clinical pain populations. The results suggest that people who perceive themselves to be functioning well, both personally and interpersonally, report lower levels of affective distress and pain intensity. Stress Lazarus defines stress as "a relationship between the person and the environment that is appraised by the person as taxing or exceeding his or her resources and as endangering well-being" (Lazarus & Folkman, 1984, p. 19). Consequently, stress is not the result of the property of a person or the 13 environment, but the product of the relationship between the two. This definition recognizes the multifactorial nature of the stress process. The relationship between pain and stress has been recognized for a number of years, either from the perspective of pain as a stressor or as a symptom of increased stress. A number of authors have postulated that the strain of living with daily pain and the physical limitations this may impose can lead to physical and psychological distress in and of itself, and create an increased vulnerability to further increases in stress and decreases in personal resources (Turner, 1991). Turner, Clancy, and Vitaliano (1987) studied community residents who were recruited by offering them a behavioural treatment program for chronic low back pain. Even though these participants were only mildly distressed and disabled by their pain, 35% of them identified pain as the primary stressor in their lives, and believed that pain required them to hold back from doing what they wanted. A number of studies have identified a relationship between stressful life events or daily hassles and a variety of pain problems (De Benedittis, Lorenzetti, & Pieri, 1990; Feuerstein, Suit, & Houle, 1985; Holm, Holroyd, Hursey, & Penzien, 1986). In one of the largest studies, questions concerning stress and pain were included in a nation wide survey of over 1,000 people in the United States (Sternbach, 1986). Strong associations were found between common pain problems, such as headaches and back pain, and stress (assessed by asking "How 14 often do you feel under great stress?") and daily hassles (i.e., a quarrel, argument, fuss; a difficulty, problem, trouble). Stress was voluntarily mentioned by respondents as a major contributor to the pain of headaches, backaches, stomach pain, and menstrual pains, but not for joint or dental pains. Sternbach concluded that the greater the daily stress and hassles, the greater the incidence and severity of the pain problems reported. Flor, Turk, and Birnaumer (1985) investigated the pschophysiological responses of chronic low back pain patients in response to stress. They found that patients responded with abnormal back muscle reactivity to personal stressors, and exhibited prolonged delay returning to baseline levels. This muscular reactivity is important as increased muscle tension is known to be associated with increased pain (Turk & Holzman, 1986). Collectively, these studies suggest that greater pain may be a consequence of increased stress. Consequently, pain intensity is viewed as a measure of stress in this study. However, individual responses to particular events vary widely. Whether a specific person-environment relationship is judged as stressful, and how this relationship is responded to, depends on appraisal and coping processes that mediate this relationship (Lazarus & Folkman, 1984). 15 Appraisal Cognitive-affective appraisal is an evaluative process through which people assess what is at stake in a particular encounter (primary appraisal) and what coping resources and options they have available in order to remedy the situation (secondary appraisal). These appraisals are then modified (i.e., reappraisals) according to any new information from the environment, the person, or both (Lazarus & Folkman, 1984). Appraisal is a phenomenological experience through which people come to understand the meaning of what is happening to them and how their sense of personal well-being may be affected. Consequently, appraisals determine how a person will adapt to a particular event because these self-assessments influence the amount of stress experienced as well as cognitive and behavioural coping responses. Stress and responses to the stressor depend more on inferential meanings about what is happening than on what actually happens, therefore, individual variations in response to stressors are to be expected. Appraisals are hypothesized to affect adaptation to chronic pain in three ways. First, maladaptive appraisals about the experience of pain, personal efficacy, or both may be associated with depression, inactivity, and increased sensitivity to physiological processes (Turk & Holzman, 1986). Second, appraisals are thought to impact adaptation by influencing the type of coping 16 strategies people engage in (Jensen, Turner, & Romano, 1991). Finally, appraisals may influence adherence and response to pain management treatment (DeGood & Shutty, 1992). Lazarus has been criticized for ignoring the impact of dispositional tendencies in his transactional theory (Ben-Porath & Tellegen, 1990; Costa & McCrae, 1990). However, he argues that he has always contended that appraisals are influenced by these more global beliefs as they establish the basis of the relationship between the person and the environment. He states that what he classifies as person variables (i.e., commitments and beliefs) are actually dispositional factors (Lazarus, 1990; Lazarus & Folkman, 1984). However, he also believes that in order to understand the coping process, situation specific appraisals should be a primary focus of study. DeGood and Shutty (1992) support this view, and state that the investigation of highly specific beliefs about pain has the greatest potential for clinical application. Catastrophizing, perceived control over pain, and self-efficacy judgments are three appraisals that have been shown to be most central to adaptation (Jensen & Karoly, 1991). Presented in Table A of Appendix A is a tabulated summary of the general findings of studies that have examined the relationship between these appraisals and adaptation to chronic pain. Prior to discussing the results of these studies, the rationales for treating catastrophizing as a separate construct from 17 depression, as well as classifying it as an appraisal rather than a coping strategy are given. Then, the samples and criterion measures are described in order to put the results into context. There has been some debate in the literature as to whether catastrophizing, which is defined as having frequent negative and worrying thoughts about pain and the prognosis for the future (Turner, 1991), is a just a symptom of depression, or whether it is representative of a separate construct (Jensen, Turner, Romano, & Karoly, 1991). Sullivan and D'Eon (1990) studied the relationship between the Coping Strategy Questionnaire (CSQ; Rosentiel & Keefe, 1983) subscales, which include a measure of Catastrophizing, and depression, as measured by the Beck Depression Inventory (BDI; Beck, Ward, Mendelson, Mock, & Erbaugh, 1961). Using hierarchical regression analysis, the authors found that the C S Q subscales accounted for 16% of the variance in the depression scores. However, when the Catastrophizing subscale was removed, the remaining subscales accounted for only 3% of the variance in depression scores. The authors concluded that since only the Catastrophizing scale was significantly and positively associated with depression, catastrophizing and depression were most likely not separate constructs. In response to the Sullivan and D'Eon (1990) article, Haaga (1992) points out that their results could just as easily be interpreted as indicating that although a 18 positive correlation between catastrophizing and depression has a very high probability, this does not make them indistinguishable constructs. He argues that Sullivan and D'Eon (1990) use an "all-or-none" decision making approach. That is, catastrophizing about pain is either a unique construct or symptomatic of depression. Haaga (1992) states that the C S Q Catastrophizing subscale and the BDI may be measures of constructs that are positively, but not perfectly, correlated. He advocates taking a more continuous variable approach to the question of construct distinction, rather than Sullivan and D'Eon's (1990) dichotomous approach. Using path analysis, Geisser, Robinson, Keefe, and Weiner (1994) examined whether catastrophizing and depression predicted different dimensions of the pain experience. Catastrophizing was measured by the CSQ, depression by the BDI (Beck et al., 1961), and pain experience by the McGi l l Pain Questionnaire (MPQ; Melzack, 1975). The M P Q was designed to measure three major psychological dimensions of pain: sensory-discriminative, motivational-affective, and cognitive-evaluative. Patients' scores on the C S Q Catastrophizing subscale and the BDI were included in three path analyses predicting scores on either the Sensory, Affective, or Evaluative subscales of the M P Q . The path coefficient between BDI scores and the M P Q Sensory subscale was statistically significant, t(82)=2.0, p_< 05, but the path coefficients between the BDI and the M P Q 19 Evaluative and Affective subscales were not. On the other hand, although the path coefficient between Catastrophizing and the M P Q Sensory subscale was not significant, the paths between Catastrophizing and the M P Q Affective and Evaluative subscales were, t(82)=2.8, p<01, and t(82)=2.4, p<.05, respectively. Geisser et al. (1994) concluded that these results lend support to the hypothesis that catastrophizing and depression are separate constructs. Further evidence for this hypothesis is provided by studies that found that catastrophizing was related to pain intensity and functional impairment, even after controlling for depression (Keefe et al., 1989; Romano et al., 1987). Based on the arguments put forward by Haaga (1992) and the research findings described above, it is reasonable to assume that catastrophic appraisals about pain may be important in the understanding of adaptation to chronic pain. Consequently, it was treated as a separate construct in this research project. Some authors have argued that even though the C S Q Catastrophizing subscale is included in a measure of coping, it should be considered an appraisal rather than a coping strategy because of the way in which the questions are phrased (Jensen & Karoly, 1991; Jensen, Turner, Romano, & Karoly, 1991; Turner, 1991). They contend that this type of phraseology assesses the degree to which people worry and display negative thinking in response to pain (e.g., "It is awful and I feel that it overwhelms me") rather than the degree to which people 20 use these thoughts as a strategy for coping with pain. Because this line of reasoning appears logical, and the face validity of this scale supports it being used as an appraisal measure, it was used in this manner in this study. Sample sizes for the studies reviewed in Table A in Appendix A ranged from 14 to 223, with a median of 88 participants. The percentage of female, compared with male participants, ranged from 47% to 100%, with a median of 57%. For those studies that reported mean pain duration the range was 1.8 to 13.2 years, with a median of 8 years. One study examined phantom limb pain, 39% examined primarily back pain patients, 26% examined patients with some type of arthritis, and 30% examined mixed pain type patients. All of the participants were receiving some kind of pain treatment either at specialized pain clinics, physiotherapy clinics, or rheumatology practices. Adaptational measures used in these studies can generally be classified into four main categories. Psychological adaptation was assessed via measures of depression, anxiety, general mood, or personality/psychopathology inventories. Behavioural adaptation was assessed through measures of medication use, exercise and activity levels, functional limitations, professional service utilization, non-organic signs, and observed pain behaviours. The impact of pain on the person's life was assessed via stress and satisfaction with life scales, and by measures of 21 perceived interference of pain on work, home, and interpersonal activities. Finally, pain intensity was measured via various types of analogue scales. Critique. The results indicate that the CSQ Catastrophizing subscale demonstrates a consistent relationship with adaptation, and that the more a person catastrophized about pain the more physically and psychologically distressed he or she is. Catastrophizing has been found to be positively associated with pain intensity (Flor & Turk, 1988; Hill, 1993; Keefe et al., 1989; Romano et al., 1987), anxiety (McCracken & Gross, 1993), depression (Hill, 1993; Keefe et al, 1989; Romano et al., 1987), and interference with life activities and functional limitations (Flor & Turk, 1988; Hadjistavropoulos & Craig, 1994; Reesor & Craig, 1988). Catastrophizing has also been found to be predictive of depression, pain intensity, and functional impairment for up to 6 months (Keefe et al., 1989). Consistent relationships have also been found between the appraisals of perceived control over pain and adaptational outcomes. Research findings indicate that the more control people believe they have over pain, the less psychologically and physically distressed they are. Control appraisals have demonstrated negative relationships with depression, anxiety, and poorer psychological functioning (Affleck et al., 1987; Jensen & Karoly, 1991; Keefe et al., 1989; Keefe & Williams 1990), pain intensity (Spinhoven et al., 1989; Spinhoven & Linnsen, 1991), functional limitations, and low activity levels (Spinhoven et al., 1989). 22 There is also some indication that appraisals of control influence the type of coping strategies people engage in. Higher levels of perceived control over pain have been found to be positively related to the use of relaxation strategies (Affleck et al., 1992; Jensen, Karoly, & Huger, 1987), and negatively related to the use of emotional expression as a method for managing pain (Affleck et al., 1992). The concept of self-efficacy refers to appraisals concerning one's capabilities to successfully execute the behaviour required to produce a specific outcome (Bandura, 1986, 1997). It does not refer to personality characteristics or global traits that operate across situations, independent of contextual factors (Strecher, DeVillis, Backer, & Rosenstock, 1986). Self-efficacy is also different from self-esteem, which is an evaluation of self-worth rather than an appraisal about specific capabilities in specific situations (Strecher et al., 1986). As can be seen from the results presented in Table A in Appendix A , there is evidence that higher levels of perceived self-efficacy are associated with better adaptation to chronic pain. Self-efficacy for exercise appraisals have shown a positive relationship with actual engagement in physical movement and exercise (Council et al., 1988; Dolce, Crocker, & Doleys, 1986; Dolce, Crocker, Moletteire, & Doleys, 1986). Self-efficacy for coping is positively related to the frequency of coping strategy use (Jensen, Turner, Romano, & Karoly, 1991). Self-efficacy for pain, physical functioning, and managing symptoms is negatively 23 associated with pain intensity, depression, and disability for arthritis patients (Lorig & Holman, 1993; Lorig et a l , 1989; O'Leary et al., 1988; Regan et al., 1988). In summary, there is evidence that pain specific catastrophizing, perceived control over pain, and self-efficacy appraisals are associated with adaptational outcomes. Specifically, the results suggest that these pain appraisals are related to (a) short- and long-term adaptation, (b) psychological well-being, (c) perceived pain intensity, (d) the types of pain coping strategies people engage in, and (e) functional impairment and exercise levels. Despite the consistency of these findings, we are still left with at least two unanswered questions: (a) What are the relationships between Catastrophizing, perceived control over pain, and self-efficacy appraisals, daily responses to pain such as anxiety, depression, pain intensity, and overall psychosocial and functional adaptational outcomes? and (b) Are the relationships identified in the literature similar for non-clinical pain populations? With regards to the first question, a problem with current research findings is that the vast majority of the studies reviewed used a cross-sectional research design. Respondents were asked to report on their appraisals at a specific point in time, or to remember what their appraisals were over days, weeks, or an unspecified period of time, and then summarize this temporal experience into one 24 response. This type of data collection is problematic for three reasons. First, although examining the relationship between current appraisals and current experience allows for the exploration of their correlational relationship, it is a point-in-time "snapshot" of the process, and does not allow for the examination of the variation in the daily relationships between appraisals and same day adaptational outcomes, or how these daily processes relate to overall adaptation to pain. Lazarus emphasizes that cognitions and behaviours are best understood as part of ongoing interactions with the environment rather than static events. That is, appraisals initiate coping efforts, which impact the person-stressor relationship, which then leads to reappraisal and further coping efforts and reappraisals (Lazarus & Folkman, 1984). The effects of this ongoing interaction are thought to occur over time, and cross-sectional designs that rely on one or few assessment points may miss important aspects of this interaction and lead to conclusions that are erroneous or incomplete (DeLongis et al., 1992). Second, cross-sectional designs do not allow for the identification of what Epstein (1986) calls "threads of stability and generality" (p. 1201) that occur in behaviour across time. Even though Lazarus advocates assessing appraisals and coping responses from a situation specific contextual model (Lazarus & Folkman, 1984), the importance to psychological well-being and social functioning of having some stability in coping processes over occasions and time is also 25 recognized (Folkman, 1992; Lazarus, 1990). By repeatedly assessing responses over situations and time these stable aspects of the coping process can be identified (Folkman, 1992). Third, cross-sectional designs generally ask respondents to retrospectively recall days or weeks of pain appraisals, which runs the risk of measurement error. Generally, the literature suggests that people are not particularly good at remembering past events, moods, or cognitions (Cohen, 1987; DeLongis et al., 1992; Wethington & Kessler, 1993). Epstein (1986) also argues that the results of cross-sectional designs can be confounded by person and/or situational variables that are unique to the point in time that the measures are taken. Participants may respond for a variety of reasons in a way that is unrepresentative of their more usual way of responding, or there may be current aspects of the situation that are influencing their responses. These variables could contribute to measurement error and decrease the reliability and generalizability of the findings. There is some indication in the pain literature that specific conditions associated with the time of recall can influence recall accuracy. As far as I am aware, researchers have not examined accuracy of recall for the pain specific appraisals of catastrophizing, perceived control, or self-efficacy. However, research in other areas indicates that memory for the experience of pain can be flawed. For example, there is some evidence that memory for pain intensity is 26 subject to distortion. Linton and Melin (1982) studied 12 chronic pain patients referred to a pain treatment program. Participants rated their level of pain on an 101-point Likert scale prior to starting treatment, and then re-rated their pre-admission pain 3 to 11 weeks later at the end of treatment. Eleven of the 12 patients significantly overestimated their pre-treatment pain levels, and the amount of overestimation was unrelated to overall improvement in pain levels during the treatment program. Memory for pain may also be pain state dependent. Eich, Reeves, Jaeger, and Graff-Radford (1985) asked patients to recall pain levels they had reported 1 week prior to this assessment point. They found that when present pain intensity was high, patients with chronic headaches rated their maximum, usual, and minimum levels of prior pain as being more severe than their hourly pain diaries indicated. When their present pain intensity was low, these same patients remembered all three levels of prior pain as being less severe than they actually were. These are crucial issues when examining pain correlates. By collecting data via multiple measurement points across participants, situations, and time, these unique person or situational components can be minimized and the more broad, crpss-situational pain-appraisal-adaptation processes are more likely to be identified. 27 With regards to the second question concerning whether research findings would be similar for non-clinical pain populations, one of the reasons for this gap in our knowledge is that all of the studies reviewed have focused on clinical pain populations, i.e., participants who were attending some kind of pain treatment program. This creates two difficulties. First, clinical pain patients may not be representative of the larger population of chronic pain sufferers. Studies that compared chronic pain patients involved in active pain treatment programs with chronic pain sufferers in the general community have found differences. Crook et al. (1984) conducted telephone surveys of 90 community residents suffering from persistent pain but not attending any pain clinic, and 48 people who had attended active treatment at a local university pain clinic during the previous 6 months and who were still experiencing persistent pain. Persistent pain was defined as "often troubled by pain" and "noteworthy pain in the last 2 weeks." The authors found no difference between the groups on pain duration or type of pain problem. However, the pain clinic group reported significantly higher levels of pain intensity, utilization of health services, somatic symptoms, depression, and psychosocial and functional impairment, and were more frequently unemployed as a consequence of pain. The authors concluded that chronic pain patients who attend pain clinics are not representative of individuals in the general population who also have chronic pain but are not referred to such clinics. 28 In contrast, Toomey, Mann, Abahsian, Carnike, and Hernandez (1993) found little difference between 28 outpatient pain patients being treated by a variety of medical specialists and 48 patients attending an outpatient pain clinic. The only significant finding was that the pain clinic patients demonstrated lower "powerful others" scores as measured by the Pain Locus of Control Scale, which is a measure the authors developed by revising the Multidimensional Health Locus of Control Scale (Wallston & Wallston, 1978). The authors interpreted this result as indicating that the pain clime patients felt less predictable control over pain than medical clinic patients. The contradictory findings in these two studies may be due to the differences in the populations studied. One of the major differences between the samples was that in the Crook et al. (1984) study only 30% of the community pain sufferers had consulted a physician in the previous 2 weeks, compared to 68% of the pain clinic patients. In the Toomey et al. (1993) study, although the medical clinic patients were not attending a pain clinic, they were receiving on-going active pain treatment from medical specialists. It may be that patients who attend specialized pain programs and patients who are actively being treated by a medical specialist experience higher levels of pain and affective distress, and cope less well than pain sufferers not receiving such treatment. Although it is difficult to draw definitive conclusions from such a small number of studies, there seems to be 29 enough evidence to make one cautious about generalizing research results obtained from clinical pain populations to non-clinical pain populations. One additional question that remains unanswered is whether there is an interaction effect between pain intensity and pain appraisals. Research indicates that there might be. Estlander and Harkapaa (1989) divided participants into four groups based on pain intensity and disability levels, and asked them to rate how much they engaged in different strategies listed on the C S Q when in mild or high levels of pain (the authors included Catastrophizing as a coping strategy). No intergroup differences were found, but there were within group differences. Participants engaged in higher levels of catastrophizing when experiencing increased levels of pain. Jensen and Karoly (1991) examined the patterns of relationships between appraisals of control, coping efforts, and adaptation to pain. A statistically significant interaction effect was found, where perceived control and activity level were positively related only for those respondents reporting lower levels of pain. Finally, the results of a study by Affleck and his colleagues indicates pain intensity may influence responses to arthritis (Affleck et al., 1987). They reported two interaction effects: (a) perceived control over aithritis symptoms (including pain) correlated positively with mood for patients with moderate and severe symptoms, and (b) perceived control over disease course correlated negatively 30 with mood and psychosocial adjustment for those with severe symptoms. The results of these studies provide enough evidence to warrant further investigation of the potential interaction effects of pain intensity and pain appraisals. Even though few studies have investigated potential interaction effects between pain intensity and appraisals and coping strategy use, the results suggest they are worth examining. However, because of the lack of empirical research and a theoretical rationale for predicting specific interactions, these effects were examined in an exploratory manner in this study. In summary, pain specific catastrophizing, control, and self-efficacy appraisals appear to be related to adaptational outcomes. However, further understanding of these relationships is limited by the reliance on one-point-in-time retrospective recall data collection methods, and the almost exclusive use of clinical pain populations. Coping Coping is defined as "the constantly changing cognitive and behavioral efforts to manage specific external and/or internal demands that are appraised as taxing or exceeding the resources of the person" (Lazarus & Folkman, 1984, p. 141). Consequently, the process of coping is situation specific. That is, a person responds to a particular person-environment relationship according to the unique context in which it occurs. 31 This definition is broad enough to allow for the inclusion of internal (emotional reactions) and external (the situation) coping targets, as well as more specific dimensions of the event (e.g., uncertainty, important consequences) (Stone, Kennedy-Moore, Newman, & Greenberg, 1992). For example, a person may attempt to lessen or manage the emotional distress of pain by reinterpreting the pain sensation, such as "I don't think of it as pain but rather as a dull or warm feeling" (Rosentiel & Keefe, 1983). Separately, or in conjunction with this cognitive strategy, a behavioural coping strategy may also be implemented, such as going to a movie. Lazarus emphasizes that coping is defined by the a person's cognitive and behavioural responses, and not by the success of these responses (Lazarus, 1990; Lazarus & Folkman, 1984). In other words, reinterpreting pain sensations or going to a movie are forms of coping, regardless of their effectiveness. One of the most frequently used measures of coping in pain research is the Coping Strategy Questionnaire (CSQ; Rosentiel & Keefe, 1983). This self-report instrument is designed to assess four cognitive and two behavioural categories of coping strategies: (a) Diverting Attention, (b) Reinterpreting Pain Sensations, (c) Ignoring Pain Sensations, (d) Praying or Hoping, (e) Coping Self-Statements, and (f) Increasing Behavioral Activities. The Catastrophizing subscale is not included in this section of the literature review as it is considered to be a measure of 32 appraisal, as are two CSQ Numerical Analogue Scales (NAS) that assess respondents appraisals of ability to control and decrease pain. Table B in Appendix A is a tabulated summary of the general findings of studies that have examined the relationship between the CSQ and adaptation to chronic pain. This table also identifies which studies used factor derived scales and which used the original CSQ subscales. A description of the samples and the criterion measures is presented prior to summarizing research findings. Sample sizes ranged from 34 to 125, with a median of 76 participants. The percentage of female compared to male participants ranged from 0% to 75%, with a median of 57%. For those studies that reported average pain duration, the range was 1 to 13.2 years, with a median of 8.1 years. One study examined phantom limb pain and another headache patients, 53% examined primarily back pain patients, 21% examined patients with some form of arthritis, and 16% examined mixed pain type patients. All participants were involved in some form of active pain treatment either at specialized pain centres or rheumatology practices. Adaptational measures fell into four broad categories. Psychological adaptation was assessed via self-report measures of depression, anxiety, helplessness, and personality/ psychopathology inventories. Behavioural adaptation was assessed via self-report measures of disability and functional impairment, uptime/downtime, medication use, professional service utilization, 33 and observed pain behaviours. Interference of pain on life was assessed via measures of perceived impact of pain on activities and sleep, and satisfaction with life scales. Finally, pain frequency and intensity was assessed by analogue scales, and pain quality by the McGi l l Pain Questionnaire (Melzack, 1975). Critique. The results summarized in Table B in Appendix A indicate that the relationships found between the C S Q and adaptational outcomes have not always been consistent. Factor-analytic studies of the C S Q have failed to find a reliable factor structure (Hill, 1993; Rosentiel & Keefe, 1983; Spinhoven et al., 1991), and either no significant relationships or weak relationships with adaptational outcomes (Jensen et al., 1994; Parker et al., 1989; Spinhoven & Linssen, 1991; Spinhoven, Jochems, Linssen, & Bogaards, 1991; Spinhoven et al., 1989). Recently, researchers have begun to examine the relationship between the individual subscales of the C S Q and adaptation to chronic pain. As Jensen, Turner, Romano, and Karoly (1991) point out, although there are advantages to using composite scores derived through factor analysis, their use does not allow for an examination of the relationship between specific pain coping strategies and adaptation. Jensen, Turner, and Romano (1991) hypothesize that it may be only a limited number of coping strategies that actually enhance adaptation, and the 34 exclusive use of composite measures could make the identification of these strategies more difficult. Following this line of reasoning, Geisser, Robinson, and Hensen (1994) attempted to determine whether individual C S Q subscale scores were better predictors of adaptation than factor composite scores. They examined the factor structure of the CSQ subscales that reflected cognitive coping, and compared these factors and the individual subscales to measures of adaptation to chronic pain. The Catastrophizing and Increasing Behavioral Activities subscales and the two N A S pain control ratings were excluded from the factor analysis as they were seen as conceptually different from the cognitive strategies. Two basic factors were identified, Conscious Cognitive Coping (containing the Coping Self-Statements, Ignoring Pain Sensations, and Reinterpreting Pain Sensations subscales) and Pain Avoidance (containing the Praying and Hoping and Diverting Attention subscales). The authors found that neither of these factors were associated with any of the measures of adaptation examined in the study. However, an analysis of the individual subscales showed that Praying and Hoping was positively correlated with pain severity and interference of pain on daily life (rs = .29 and .30, p_<001), and affective distress (r_=-25, p_<.05). Ignoring Pain Sensations was weakly and negatively correlated with pain intensity (r = -.18, E<05). The other cognitive coping scales were unrelated to the 35 dependent variables. These authors concluded that the individual CSQ subscales may have greater utility in terms of examining coping and adjustment to chronic pain. Based on the arguments proffered by Jensen, Turner, Romano, and Karoly (1991) and the findings of Geisser, Robinson, and Henson (1994), the individual CSQ subscales were used in this study rather than factor composite scores. Coping Self-Statements Subscale. The Coping Self-Statements subscale has been found to be positively associated with psychological well-being, although the strength of these relationships is not strong. Jensen and Karoly (1991) found that, after controlling for pain severity, this subscale accounted for 10% of the variance (p_< 01) in better psychological functioning. Hill (1993) found that Coping Self-Statements correlated negatively with psychological distress (r = -.24, p< 05) and pain intensity (r = -.21, p<05), and Keefe and Williams (1990) found a negative correlation between this subscale and depression (r = -.29, p_<01). Rosentiel and Keefe (1983) also found that Coping Self-Statements correlated positively with functional capacities, although once again the relationship was fairly weak (r = .18,p<05). Increased Behavioral Activities Subscale. The results for the CSQ Increased Behavioral Activities subscale are inconsistent. Keefe and Williams (1990) found this subscale to be negatively associated with depression (r = -.25, P<.01), and Jensen and Karoly (1991) found it accounted for 12% of the variance 36 (p_< 001) in better psychological functioning. However, Hill (1993) found a positive association between increased activities and psychological distress (r = .21, p_<.05). Other studies found no significant correlations between this subscale and measures of psychological and physical functioning (Geisser, Robinson, & Henson, 1994; Rosentiel & Keefe, 1983; Turner & Clancy, 1986). Finally, both Hill (1993) and Keefe and Williams (1990) found the use of this strategy to be positively associated with increased pain (rs = .21 and .23, ps<05, respectively). Reinterpreting Pain Sensations and Praying and Hoping Subscales. Most of the studies examining the relationship between these two subscales and measures of psychological and functional adaptation to pain have generally found no significant association (Jensen & Karoly, 1991; Jensen, Turner, & Romano, 1994; Keefe & Williams, 1990). However, Geisser, Robinson, and Henson (1994) found that Praying and Hoping was significantly related to increased pain severity and greater interference of pain (rs = .29 and .30, p_s<001), and higher levels of affective distress (r =.25, p_<05). Hil l (1993) found that Praying and Hoping was positively correlated with pain intensity, whereas Turner and Clancy (1986) found a negative association with pain levels (rs = .27 and -.21, ps <05 respectively). Ignoring Pain and Diverting Attention Subscales. These last two subscales have demonstrated weaker relations to adaptation than the other scales. Most studies have found no significant relationship between these subscales and 37 adaptational outcomes (Geisser, Robinson, & Henson, 1994; Jensen & Karoly, 1991; Turner & Clancy, 1986). In contrast, Romano et al. (1987) found that Ignoring Pain was moderately correlated with increased disability (r=-.41, p_<01), and Jensen and Karoly (1991) found it accounted for 7% of the variance (p_<.05) in better psychological functioning. In summary, the results from studies using a factor breakdown of the C S Q have generally found no significant associations between composite scales and measures of adaptation to chronic pain. The results from studies using the individual C S Q subscales have been somewhat more promising and have shown that coping strategies are related to adaptational outcomes, although the relationships are somewhat inconsistent. Specifically, Coping Self-Statements has demonstrated positive relationships with psychological and physical functioning, and negative relationships with pain intensity. Praying and Hoping has demonstrated positive relationships with pain intensity, interference of pain on life, and psychological distress. Inconsistent results have been found for the Increasing Behavioral Activities subscale, with some studies finding it had negative relationships with better psychological functioning, and others finding it had positive relationships with poorer psychological functioning and pain intensity. Most studies have generally found no significant relationships between the subscales of Reinterpreting Pain Sensation, Ignoring Pain, and Diverting 38 Attention and adaptational measures. However, two studies found positive relationships between the latter two subscales and physical and psychological distress. The reasons for these inconsistencies are unclear, but may be partly related to the same issues that were raised in the previous section; that is, the use of cross-sectional designs, retrospective data collection methods, and the reliance on clinical pain populations. Another reason may be that few studies have examined the question of whether there is an interaction effect between pain intensity and coping strategy use. For example, some studies have found that coping strategies, as measured by the CSQ, change in accordance with different pain levels. Estlander and Harkapaa (1989) found that during mild pain episodes, participants tended to use more Diverting Attention, t(102) = 4.71, p< 001, and Ignoring Pain, t(102) = 8.26, p< 001. During more severe pain, these strategies decreased and the use of Praying and Hoping increased. Jensen and Karoly (1991) found that people engaged in fewer coping strategies (measured by the CSQ) as pain levels increased. They also found that the negative relationships between the strategies of Ignoring Pain Sensation, Diverting Attention, and Coping Self-Statements, and activity levels were only significant when their interaction with pain intensity was taken into account. That is, these strategies seemed to be important to functioning only for those individuals with relatively low levels of pain intensity. 39 These results suggest that we can not assume respondents use the same coping strategies in dealing with all aspects of a pain problem. Consequently, the potential interaction effects between pain intensity and coping strategy use need to be examined in the analysis of coping strategy use. However, because of the limited number of studies on interaction effects, and the lack of a theoretical rationale for predicting specific interactions, the examination of these types of processes is still at the exploratory stage. The question of how daily coping strategy use is related to daily responses to pain has also generally been neglected in chronic pain research. No studies that I am aware of have used the C S Q in this manner. One researcher who has examined daily responses to pain in patients with rheumatoid arthritis is Glen Affleck. Affleck et al. (1992) conducted an analysis of how people differed (between-person effects) in their relationships between coping strategies, pain, and mood, via correlations and multiple regression analyses on the aggregated daily diary data (N = 75). They found that increased daily coping strategy use, as measured by the Daily Coping Inventory (Stone & Neale, 1984), was associated with decreased daily pain and increased positive mood. They also found that the frequency of the use of Relaxation was negatively associated with mean daily pain, and positively associated with mean daily mood, and that the strategy of 40 Express Emotion was positively associated with mean daily pain, and negatively associated with mean daily mood. Recently, Keefe et al. (1997) examined the relationships between coping strategy use (measured by the Daily Coping Inventory; Stone & Neale, 1984) and daily negative and positive mood and pain in a group of patients with rheumatoid arthritis (N = 53). The authors not only examined between-person effects on the relationships between the dependent and independent variables (based on aggregated data), they also used pooled regression analysis with (subject) dummy variables and fixed effects to examine variability in these same relationships across days in individual recording (within-person effects). At the between-person level of analysis, they found significant positive correlations between the use of the strategies of Relaxation and Venting Emotion and positive mood, and significant negative correlations between the use of Seek Emotional Support and negative mood. At the within-person level of analysis, they found nighttime pain was predicted by the higher use of the strategies of Pain Reduction Effort, Relaxation, Venting Emotions, and Seek Spiritual Comfort. Negative mood was predicted by higher use of the strategies of Pain Reduction Effort and Venting Emotion. Finally, positive mood was predicted by the higher use of Distraction, Seek Emotional Support, and the lower use of Pain Reduction Effort. 41 The results of these studies are limited by two factors. First, both based the between-person analyses on aggregated data, and the resulting "average effects" may have masked important individual differences in the relationships being examined. Second, Affleck et al. (1992) did not examine within-person effects, which may have provided unique information about the pain-coping-mood relationships not apparent in the between-person analyses. As Affleck states (Tennen & Affleck, 1996), within- and between-person analyses answer different questions, both of which are important: Within-person analyses examines whether there are relations among coping-relevant variables for each individual in a study, whereas between-person analyses examines whether these individual relationships generalize across individuals or relate to differences between individuals. A clearer picture of the relationships between pain, coping, and mood, may have emerged from these studies if the between-person analyses had been based on each individual's pattern of responding rather than on aggregated data, and if both individual patterns of responding and differences between individuals in these patterns had been examined. Because of the paucity of research on the relationship between daily coping strategy use and daily responses to pain, it is difficult to make definitive statements about these relationships. Additional daily diary studies are needed in order to broaden our understanding of how individuals respond to pain on a daily basis, 42 how people may differ in these responses, and how these daily processes are influenced by psychosocial and functional variables important to the experience of chronic pain. In summary, coping strategies, as measured by the CSQ, appear to be related to measures of adaptation to chronic pain. However, our understanding of these relationships has been hampered by retrospective recall and one-point-in-time data collection methods, a reliance on clinical pain populations, the lack of examination of the potential interaction effects between pain intensity and coping strategy use, and for those studies that have used daily diary data, the lack of the consideration of both the within- and between-participant variances. Pain Duration There is strong evidence that pain duration is unrelated to response to pain. Specifically, most studies have found no relationship between pain duration and depression (Sullivan & D'Eon, 1990; Turner & Clancy, 1986), coping strategy use (Flor & Turk, 1988; Keefe, Caldwell, Martinez, Beckham, & Williams, 1991; Rosential & Keefe, 1983; Spinhoven et al., 1989), self-efficacy appraisals (Jensen & Karoly, 1992), pain intensity (Turner & Clancy, 1986), or medically congruent versus medically non-congruent pain (Hadjistavropoulos & Craig, 1994). In contrast, Affleck et al. (1987) found that pain duration was negatively associated with pain severity, and Keefe, Grisson, Urban, and Williams (1990) 43 found that people with pain of less than 2 years were more depressed than those with longer pain duration. However, in both of these studies the relationships found were quite weak (rs = -.22 and .22, p_s<.05, respectively). Hil l (1993) found a moderate relationship between pain duration and psychological distress (r = .44, p<.05). Because this study was based on participants suffering from phantom limb pain, it is possible that this finding reflects more of a characteristic of this specific population than a response of chronic pain sufferers in general. In summary, the vast majority of studies support the hypothesis that pain duration is unrelated to response to pain. Consequently, pain duration was not included as a variable in this study. Summary A review of the chronic pain literature indicates that there is an association between pain specific appraisals and coping strategy use and adaptation to chronic pain. The pain specific appraisals of catastrophizing, perceived control over pain, and self-efficacy have been shown to be related to (a) short- and long-term adaptation, (b) psychological well-being, (c) perceived pain intensity, (d) the types of coping strategies people engage in, and (e) activity and exercise levels. Coping strategies, as measured by the CSQ, have been found to be related to (a) psychological well-being, (b) physical functioning, and (c) perceived pain intensity. There is also some evidence to suggest that both the frequency and 44 strength of appraisals, and the frequency of coping strategy use change according to different pain levels. Research to date has aided our understanding of the process of adapting to chronic pain. However, our knowledge of this process have been hampered by a number of methodological short-comings: (a) retrospective recall and cross-sectional data collection methods have limited our ability to understand daily responses to pain; (b) many studies have limited themselves to unidimensional measures of the pain experience, and not have not considered the relationships between these measures and a broader range of psychosocial and functional variables important to the experience of chronic pain; (c) the reliance on clinical pain populations has not allowed us to examine whether these findings are similar for the greatest proportion of chronic low back pain sufferers, that is, those not attending a specialized pain clinic; and (d) the failure to consider the potential interaction effects between pain intensity and appraisals, and pain intensity and coping strategy use. This research project attempted to address these methodological shortcomings. First, participants recorded data every day for 30 days (i.e., daily diaries). Second, they completed the MPI (Kerns et al., 1985) prior to their daily monitoring as a measure of psychosocial and functional variables important to the experience of chronic pain. Third, instead of recruiting participants for this study 45 from specialized pain clinics, they were recruited via media advertisements from the local community. Fourth, interaction effects between pain intensity and pain appraisals, and pain intensity and coping strategy use were examined in exploratory analyses. There are two other aspects to this study that make it different from previous research. First, the appraisals and coping strategies were examined in the same analysis. This makes the design of the study closer to Lazarus and Folkman's (1984) model of stress and coping, and accounts for the overlapping variance between the appraisals and coping strategies. Second, the morning score for the dependent variable was controlled for in each of the analyses. Thus, the results indicate the relationships between daily appraisals and coping strategy use and the change between morning and nighttime negative mood and pain intensity. This study only involved female participants because analysis of gender differences in responses to pain is complex and involves not just an analysis of male and female mean differences on study variables, but also the consideration of factors such as the interaction effects of gender and situational contextual factors that may influence these responses (i.e., socioeconomic status, power imbalances, and different familial demands and responsibilities). Because this study is already considering a large number of variables, it was felt that involving both sexes and 46 trying to control for gender differences would make the design and analysis too complex. For the daily process analysis, the dependent variables were nighttime depressed mood, nighttime anxious mood, and nighttime pain intensity, and the independent variables were morning depressed mood, morning anxious mood, morning pain intensity, the appraisals of Catastrophizing, control over pain, Self-Efficacy for pain, and the C S Q coping strategies of Distraction, Reinterpreting Pain Sensation, Ignoring Pain, and Praying and Hoping. For the analyses that examined the relationship between daily processes and psychosocial and functional variables, the dependent variables were the coefficients from the regression equations for the dependent variables in the daily diary analyses, and the independent variables were the nine subscales of the pre-diary monitoring MPI (i.e., Affective Distress, Life Control, Support, General Activity Level, Interference, Pain Severity, Punishing Responses, Distracting Responses, and Solicitous Responses). I expected to find that (a) people with higher levels of negative mood would experience higher levels of catastrophic thinking and use more Praying and Hoping coping strategies, whereas people with lower levels of negative mood would experience higher levels of Self-Efficacy and control appraisals, and use more Distraction and Ignoring Pain coping strategies; (b) people with higher levels of 47 pain intensity would experience higher levels of catastrophic thinking and use more Praying and Hoping, Distraction, and Ignoring Pain coping strategies, whereas people with lower levels of pain would experience higher levels of Self-Efficacy and control appraisals; and (c) people who generally perceived themselves to be adapting well to their pain from both a psychosocial and functional perspective would experience lower levels of negative mood and pain intensity on a daily basis. Average daily pain was the measure of pain intensity used in the analyses of the interaction effects between pain intensity and the three pain appraisals, and pain intensity and the four coping strategies. Because this analyses was exploratory, no specific predictions were made concerning the findings. Data were analyzed using Hierarchical Linear Modeling (HLM; Bryk & Raudenbush, 1992). The advantage of H L M is twofold. First, H L M allows you to examine data with an hierarchical structure. For this study it meant that relationships identified in the analyses of the daily data (first level of the hierarchy) could be examined as a function of psychosocial and functional variables (second level of the hierarchy). The second advantage of H L M is that it partitions the variance explained at each level of the hierarchy into both within-and between-participant. 48 Hypotheses The Relationships Between Daily Nighttime Depressed Mood, Nighttime Anxious Mood, Nighttime Pain Intensity, Appraisals, and Coping Strategy Use Hypothesis 1. Daily appraisals of Catastrophizing, control, and Self-Efficacy, and the daily use of Praying and Hoping, Ignoring Pain, and Distraction coping strategies will demonstrate a significant linear relationship with participants' average scores for daily nighttime depressed mood, nighttime anxious mood, and nighttime pain intensity, after controlling for the morning covariate. Specifically: (a) Daily nighttime depressed mood will be positively associated with Catastrophizing and Praying and Hoping, and negatively associated with perceived control over pain, Self-Efficacy, Distraction, and Ignoring Pain, when controlling for morning depressed mood. (b) Daily nighttime anxious mood will be positively associated with Catastrophizing and Praying and Hoping, and negatively associated with perceived control over pain, Self-Efficacy, Distraction, and Ignoring Pain, when controlling for morning anxious mood. (c) Daily nighttime pain intensity will be positively associated with Catastrophizing, Praying and Hoping, Ignoring Pain, and Distraction, 49 and negatively associated with perceived control over pain and Self-Efficacy, when controlling for daily morning pain intensity. The Relationships Between the MPI Subscales and Daily Responses to Pain Hypotheses 2. The average scores across participants for daily nighttime depressed mood, nighttime anxious mood, and nighttime pain intensity will be a function of psychosocial and functional variables (the Time 1 MPI subscales) when controlling for the corresponding morning score, the three appraisals, and the coping strategies entered into the model. Specifically: (a) Daily nighttime depressed mood will be negatively associated with Life Control, Support, General Activity Level, and Solicitous Responses, and positively associated with Interference, Affective Distress, Pain Severity, and Punishing and Distracting Responses. (b) Daily nighttime anxious mood will be negatively associated with Life Control, Support, General Activity Level, and Solicitous Responses, and positively associated with Interference, Affective Distress, Pain Severity, and Punishing and Distracting Responses. (c) Daily nighttime pain intensity will be negatively associated with Life Control, Support, and General Activity Level, and positively associated with Interference, Affective Distress, Pain Severity, and Punishing, Solicitous, and Distracting Responses. 50 Additional Exploratory Analyses Exploratory Analysis 1: No hypotheses were generated for the C S Q Reinterpreting Pain Sensation coping subscale as results from previous research have been quite inconsistent. However, because this coping strategy has been examined in numerous other studies, it was included in this analysis for exploratory purposes. Exploratory Analysis 2: As well as examining whether the average daily scores across participants on the dependent variables were a related to the MPI subscales, I also examined whether these subscales changed the magnitude of the daily relationships between the dependent and independent variables. For example, if a linear relationship was found between morning pain intensity and nighttime pain intensity, the results would indicate whether any of the MPI subscales made this relationship stronger or weaker. Because these questions have not been investigated previously in the research literature, this analysis was exploratory. Exploratory Analysis 3: Because there was no empirical or theoretical basis for predicting specific directions for the potential interaction effects between pain intensity and pain appraisals and pain intensity and coping strategy use, these effects were examined in an exploratory manner only. 51 C H A P T E R T H R E E Method Design On each day of the 30-day-diary period, participants completed questionnaires concerning Self-Efficacy for pain, anxious and depressed mood, and pain intensity within one hour of getting up in the morning; and coping strategy use, anxious and depressed mood, Catastrophizing, control, and average and current pain intensity within one hour of going to bed. Measures of the psychosocial and functional variables were completed pre-diary (range 1 to 3 days) and post-diary (range 1 to 4 days) completion. Participants Women with chronic low back pain were recruited following announcements about the research project in local newspapers and on radio stations. Participants met the following criteria: (a) low back pain experienced for at least 6 months, (b) low back pain experienced on a daily basis, (c) not attending a multidisciplinary pain clinic at the time of the study (i.e., a health care delivery facility staffed by physicians of different specialties and other non-physician health care providers who specialize in the diagnosis and management of patients with chronic pain; International Association for the Study of Pain, 1990), (d) the ability to read, speak, and write English, (e) able to attend a personal interview, and (f) completion of a written consent to participate in the study. In addition 52 participants were required to have a spouse or partner living with them (married or common-law relationship) to ensure that they could complete Section II of the MPI, which involves questions concerning the spouse or partner's responses to the participant's pain (i.e., the MPI interpersonal subscales). One hundred and thirty two women responded to the advertisements. Thirty-two of these women did not meet the study criteria as they either did not have a spouse or partner (n = 20), or they were unable to meet for a personal interview due to their living location (n = 12). This resulted in 100 women being accepted into the study. Twelve participants who started the study dropped out: Six did not return phone calls and the reasons for their non-completion are unknown; one said she had completed all of the diaries but did not return them; two said their pain had completely stopped within one week of starting the diaries; two felt too distressed by other events in their lives to continue; and one said she was too busy. Thus, 88 participants were included in the data analyses. Eighty-one percent of these participants completed all 30 days of diaries, 14% completed between 25 to 29 days, and 5% completed between 14 to 24 days. A full description of the demographic data is provided in Appendix B. The mean age of the women was 46.83 years (SD - 11.90) with a range of 23 to 80 years, and 78% had some college or university fraining. Fifty-five percent were employed in some capacity outside of the home, and 50% had a yearly family income of under $60,000. Ninety-four percent of the women were Caucasian. 53 The mean years since pain began was 16.69 (SD = 12.78), and ranged from 6 months to 47 years, and mean years since pain was experienced on a daily basis was 10.75 (SD = 10.46) with a range of 3 months to 40 years. Forty-one percent of participants' back pain was the result of an accident, although only 6% were involved in litigation for their pain. Fifty-five percent of participants had seen two or more health care professionals about their back pain within the last 6 months, and only 14% had not seen anyone. Finally, 73% of participants were taking medication for back pain. Measures The measures were the Coping Strategy Questionnaire (CSQ, Rosentiel & Keefe, 1983; Swartzman et al., 1994), Arthritis Self-Efficacy for Pain Scale (Lorig et al., 1989), State-Trait Personality Inventory (STPI) Anxiety and Depression subscales (C. D. Spielberger, personal communication, November 8, 1995, February 28, 1997), Numerical Rating Scales of current and daily average pain intensity and perceived control over pain, and the Mulitidimensional Pain Inventory (MPI, Kerns et al., 1985). The McGi l l Pain Questionnaire (MPQ, Melzack, 1975) was used for descriptive purposes to measure pain quality. Demographic information (Appendix B) was gathered and included age, number of children living at home, ethnicity, education, family income, employment status, pain duration, and any past (i.e., witliin the last 6 months) or current treatment including medication intake. 54 Coping Strategy Questionnaire The C S Q (Rosentiel & Keefe, 1983) is the most widely used measure of pain coping strategies (Jensen et al., 1991), and is designed to assess cognitive and behavioural responses. The CSQ items were based on a review of relevant laboratory and clinical studies (Rosentiel & Keefe, 1983). The subscales used to measure coping in this study were those derived from a factor analysis of the individual test items by Swartzman, Gwadry, Shipiro, and Teasell, (1994), which was based on a sample of 126 chronic pain patients suffering from whiplash who were attending a Physical Medicine and Rehabilitation Clinic. This analysis resulted in a 32-item questionnaire and five subscales which are quite similar to those found in the original Rosentiel and Keefe (1983) study. The Catastrophizing (six items) and Reinterpreting Pain Sensations (six items) subscales were exactly the same as Rosentiel and Keefe's. The Praying and Hoping subscale (four items) was comprised of four of the original six Rosentiel and Keefe items. Ignoring Pain Sensations (eight items) consisted of items from Rosentiel and Keefe's corresponding subscale as well as items from their Coping Self-Statements subscale. Finally, Distraction (eight items) was comprised of items from Rosentiel and Keefe's Diverting Attention and Increasing Activity subscales. Respondents are asked to rate on a 7 point scale (0 = never do that and 6 = always do that) how frequently they engage in the different activities listed in 55 the CSQ when they experience pain. However, the instructions were changed slightly for this study to reflect the use of the questionnaire as a daily measure. Instead of "For each activity, I want you to indicate using the scale below, how much you engage in that activity when you feel pain" the instructions were "For each activity, I want you to indicate, using the scale below, how much you have engaged in that activity from the time you got up this morning until now, when you felt pain." The score for each subscale is the mean of that subscale, which can range from 0 to 6, with higher scores indicating more frequent engagement in the coping strategy. Internal consistency of the subscales were within acceptable limits (Swartzman et al., 1994). Catastrophizing Appraisals The Catastrophizing subscale of the C S Q (Rosentiel & Keefe, 1983; Swartzman et al., 1994) is one of the most frequently used measures of catastrophic appraisals (Jensen et al., 1991). Although it is administered as part of the entire CSQ, it has been used individually to measure catastrophic thiriking (e.g., Gi l et al., 1990; Keefe et al., 1989). The Catastrophizing subscale contains six items describing catastrophic thoughts concerning current pain experience (e.g., "It is terrible and I feel it is never going to get better"). The scores on the items are summed and the mean scores range from of 0 to 6, with higher scores indicating more frequent catastrophic thoughts. In order to reflect the fact that this subscale was used as a 56 measure of appraisal, the instructions were changed somewhat for this study. Instead of being asked to respond to whether they engaged in the activity (as per the instructions for the entire CSQ), respondents were instructed: "Below are a list of thoughts people have reported thinking when they feel pain. For each thought I want you to indicate, using the scale below, how often you thought that way from the time you got up this morning until now when you felt pain?" The anchors ranged from 0 = never thought that to 6 = always thought that. Test-retest stability for this subscale has been shown to be moderate to strong over intervals ranging from 24-hours to 6 months (Keefe et al., 1990; Main & Waddell, 1991). Perceived Control Over Pain Appraisal The control subscale of the C S Q (Rosentiel & Keefe, 1983) is one of the most frequently used measures of perceived control over pain (e.g., Affleck et al., 1992; Hill , 1993; Reesor & Craig, 1988). This is a 7-point numerical rating scale ranging from 0 = no control to 6 = complete control. Respondents are asked to rate how much control they feel they have over pain, based on all the things they do to cope, or deal with it. Due to the daily recording required by this study, the original instructions were changed slightly. Instead of being asked "Based on all the things you do to cope, or deal with your pain, on an average day, how much control do you feel you have over it?" participants were asked "Based on all the things you have done today to cope, or deal with your pain, how much control do you feel you have had over you pain today?" Test-retest stability has been shown 57 to be moderate to strong over intervals ranging from 24 hours to 6 months (Keefe et al., 1990; Main & Waddell, 1991). Self-Efficacy Appraisals The Arthritis Self-Efficacy Scale (ASE; Lorig et al., 1989; O'Leary et al., 1988; Regan et al., 1988) is a 20-item self-report questionnaire that was developed from a pool of items generated from rheumatologists and arthrritis patients. The A S E consists of three factors: (a) Self-Efficacy for Pain, (b) Self-Efficacy for Function, and (c) Self-Efficacy for controlling symptoms other than arthritis. Respondents are asked to rate their level of confidence about their ability to perform certain tasks or to decrease pain on a 10 to 100 point scale, where 10 = very uncertain and 100 = very certain. The A S E has been shown to demonstrate good internal consistency, as well as test-retest stability over periods ranging from 2 to 29 days (Lorig et al., 1989; O'Leary et al., 1988; Regan et al., 1988). The Self-Efficacy for Pain subscale was used for this study, which has been validated by Anderson, Dowds, Pelletz, Edwards, and Peeters-Asdourian (1995) on a sample of chronic back pain patients. This subscale contains five items, resulting in a range of mean scores from 10 to 100, with higher scores indicating higher levels of self-efficacy The only change made to the original questionnaire was that the term arthritis was replaced by the term back pain. 58 State-Trait Personality Inventory - Anxiety and Depression Subscales Anxious and depressed mood were measured by the state anxiety (STPI-SA) and depression (STPI-SD) subscales of the State-Trait Personality Inventory (STPI; C. D. Spielberger, personal communication November 8, 1995, February 28, 1997). Because this study was interested in looking at daily anxious and depressed mood responses rather than anxiety and depression proneness, only the state anxiety and depression scales were used. The STPI-SA scale was based on the State-Trait Anxiety Inventory-State scale (STAI; C. D. Spielberger, personal communication, November 8, 1995, February 28, 1997; Spielberger, 1983), which is one of the most widely used measures of state anxiety (Cleeland & Syrjala, 1992). More than 600 high school and college students, approximately 600 neuropsychiatric medical and surgical patients, and 200 prisoners were used for the test construction, standardization, and validation of the original STAI. The questionnaire was then revised in order to drop or replace ambiguous or difficult to comprehend items, and re-tested on 5,000 more respondents. The STAI-S has demonstrated good internal validity, although the test-retest stability is low (Spielberger, 1992). This low stability is to be expected as the STAI-S scale is designed to be sensitive to fluctuations in anxiety resulting from situational influences (Spielberger, 1992). The 10 strongest items for the STAI-S were used to construct the STPI-SA. The STPI-SA correlates 59 highly with the parent scale, and has demonstrated high internal consistency (C. D. Spielberger, personal communication, November 8, 1995, February 28, 1997). The STPI- SD was developed to measure state depression, and the scale was formed by choosing 20 items from the four most widely used measures of depression; the Beck Depression Inventory (BDI; Beck et al, 1961); Zung Self-Rating Depression Scale (Zung, 1965); Center for Epidemiologic Studies Depression Scale (CES-D; Radloff, 1977); and the Multiple Affect Adjective Checklist ( M A A C ; Zuckerman & Lubin, 1985). Each of these measures have demonstrated good stability and validity over numerous studies (see Ritterband, 1995 for a comprehensive review), and the items chosen for the STPI-SD were stated, as nearly as possible, as they appeared in the original scale. The 20 items were factor analyzed, and the best set of 10 items were selected for the STPI-SD scale. The STPI-D scale was administered to 251 university students (170 females and 81 males) recruited from undergraduate psychology courses at the University of South Florida (C. D. Spielberger, personal communication, November 8, 1995, February 28, 1997). Participants also completed the STPI-SD, BDI, Zung, and CES-D. The STPI-SD correlated .70 or higher with the BDI, Zung, and C E S - D , and demonstrated high internal consistency. The STPI-SA and the STPI-SD scales each consist of 10-items, and respondents are asked to rate how they feel on these items right now, that is. at this 60 moment on a 4-point scale ranging from 1 = not at all to 4 = very much so. Scores are summed (a number of items have reverse scoring), resulting in a range of scores of 10 to 40, with higher scores indicating higher levels of depressed or anxious mood. Pain Intensity Participants were asked to rate their pain intensity on numerical rating scales (NRS), 0 = no pain and 10 = pain as bad as it could be. Participants were asked to complete three pain intensity measures: One was completed in the morning and asked about current pain intensity, and two were completed at bedtime and asked about current pain intensity and average pain intensity for the day. The validity of NRSs has been well documented (Jensen & Karoly, 1992). NRSs have demonstrated significant, positive relationships with other measures of pain intensity (Jensen et al., 1986; Jensen et al., 1989; Kremer et al., 1981), sensitivity to changes due to treatment (Kaplan et al., 1983; Keefe et al., 1981), and a high rate of compliance with the measurement task (Jensen et al., 1986). They have also been shown to demonstrate moderate test-retest stability over a period of 7 days (Jensen & McFarland, 1993). Multidimensional Pain Inventory The MPI (Kerns et al., 1985) is a self-report questionnaire designed to measure psychosocial and functional variables important to the experience of 61 chronic pain. The original development of the MPI was based on a sample of 120 chronic pain patients attending pain management programs at two Medical Centres. Eighty-one percent of the sample was male, and mean pain duration was 10.2 years. The MPI contains 12 subscales and one composite scale that are divided into three parts. Corrfirmatory factor analysis identified five subscales for Part I (22 items) designed to measure perceptions of (a) Pain Severity, (b) perceived Interference of pain on vocational, marital/family, and recreational/social functioning, (c) Support from significant others, (d) Life-Control, and (e) Affective Distress. For this study, the item concerning current pain intensity was dropped from the Pain Severity subscale as it would overlap too highly with the nighttime pain intensity dependent variable, and the item concerning rating of control over pain was dropped from the Life Control subscale as it would overlap too highly with the independent variable of perceived control over pain. These adjustments do not affect the psychometric stability of the subscales (Rudy, 1989). Part II assesses perceptions of the frequency of Punishing, Solicitous, and Distracting responses of the spouse or partner to the respondent's demonstrations and complaints of pain. The items for Part II were derived from interviews with patients' significant others. Exploratory factor analysis was undertaken on the responses of a subsample of 95 patients who reported living with a spouse or 62 partner, resulting in the retention of 14 items and the identification of the three subscales. Part III assesses participation in four types of common activities, for example household chores and socializing with family and friends. The General Activity Level (GAL) scale is a composite scale based on these four activity subscales, and was the measure of activity used in this study. The items for Part III were derived from two sources: published activity lists, and activity goals established by patients as part of their treatment at one of the V A Medical Centres. Factor analysis resulted in the retention of 18 items and the identification of the four activity subscales. The number of items loading on each of the MPI subscales ranges from 3 to 11, and 18 items comprise the G A L scale. Scores for the subscales and the G A L scale range from 0 to 6, and the subscale score is the mean of the items, with higher scores indicating more frequent endorsement of the response. A l l subscales of the MPI have demonstrated adequate levels of internal consistency and stability over a 2-week test period (Kerns wt al., 1985). Kerns et al. (1985) examined the validity of the MPI by submitting the 12 MPI subscales and several other standardized measures of pain relevant constructs (e.g., State-Trait Anxiety Inventory, Spielberger, 1983; McGi l l Pain Questionnaire, Melzack, 1975) to exploratory factor analysis. The MPI subscales loaded 63 significantly on factors comprised of these conceptually related standardized measures. In this study, the MPI was administered pre- and post-diary monitoring in order to assess the stability of the measure for this population, although only the Time 1 MPI was used in the main analyses. Potential differences between Time 1 and Time 2 MPI subscale scores were examined through doubly multivariate repeated measures analysis of variance ( M A N O V A ) . Table 1 lists the means, standard deviations, and the Cronbach's alphas for the Part I and Part II MPI subscales and the composite G A L scale. A significant difference was found for the overall time effect, F(9,79) = 910.52, p< .01, but the univariate analysis of the individual subscales revealed that there was a significant difference only for G A L , F ( l , 87) = 7.04, p = .009. A n examination of the means for this variable indicated that from Time 1 to Time 2 participants' general levels of activity decreased by a score of 0.17. However, the G A L scale is comprised of four activity subscales for a total of 18 items based on a 0 to 6 numerical rating scale. It would take only a one unit decrease in one of the items to create the type of difference noted above, which would be seen as very little change if one were treating a person in a chronic pain management program. Consequently, although the Time 1 to Time 2 difference on the G A L scale is statistically significant, the difference is not really meaningful from a clinical perspective. 64 To further assess the stability of the MPI subscales and the G A L scale, the correlations between the scores at Time 1 and Time 2 for each MPI subscale were examined and were as follows: Affective Distress, r = .39; Life Control, r = .43; Table 1 Means and Standard Deviations for the MPI Subscales for Time 1 and Time 2. and Cronbach's alphas for Time 1(N = 88) Time 1 Time 2 M SD a M SD Affective Distress 3.06 1.20 .81 2.80 1.33 Life Control 3.78 1.34 .81 3.93 1.15 Support 3.80 1.66 .87 3.78 1.62 General Activity L e v e l a 3.10 0.77 .78 2.93 0.75 Interference 3.35 1.36 .92 3.25 1.48 Pain Severity 3.50 1.10 .82 3.32 1.21 Distracting Responses 2.02 1.49 .74 2.03 1.55 Punishing Responses 1.68 1.32 .80 1.50 1.34 Solicitous Responses 2.90 1.50 .80 3.07 1.54 Note. The higher the score, the greater the attribution a significant at p < .01 Support r = .86; General Activity Level, r = .69; Interference, r = .83; Pain Severity, r = .71; Distracting Responses, r = .78; Punishing Responses, r = .76; and 65 Solicitous Responses, r = .78. Although all of the correlations were significant at the p < .01 level, the magnitude of the relationships indicate that the subscales are not stable for Affective Distress and Life Control, and are only moderately stable for the remaining subscales. The internal reliability of the Time 1 MPI for this population was examined through Cronbach's alphas. As can be seen in Table 1, the alphas for the subscales are within acceptable limits, indicating that this measure had acceptable internal consistency. McGi l l Pain Questionnaire The M P Q (Melzack, 1975) was ao'ministered pre-diary monitoring for descriptive purposes. The M P Q was designed to assess the sensory, affective, and evaluative dimensions of the pain experience (Melzack & Torgeson, 1971). The measure consists of three major dimensions and 16 subclasses. The three dimensions are (a) words that describe the sensory qualities of pain in terms of temporal, spatial, thermal, and other characteristics (subclasses 1-10, sensory subscale); (b) words that describe affective qualities in terms of fear, tension, and autonomic characteristics (subclasses 11-15, affective subscale); and (c) evaluative words that describe the subjective overall intensity of the entire pain experience (subclass 16, evaluative subscale). Each subclass consists of words that were considered by most respondents to be qualitatively similar. The first word in a subclass is the least intense and the last is the most intense. The scores 66 are summed for each of the three subscales, and range from 0 to 42 for the sensory subscale, 0 to 14 for the affective subscale, and 0 to 5 for the evaluative subscale, with higher scores indicating more choices of "intense" words on that subscale. Studies of the validity of the 3-factor framework of the M P Q are numerous (see Reading, 1989, for a comprehensive review). Test-retest studies have yielded evidence to suggest that there is consistency in an individual's scores on the subscales across a one week time period (Fox & Melzack, 1976; Love, Leboef, & Crisp, 1989; Melzack, 1975). For study participants, the mean for the M P Q sensory subscale was 15.79 and the standard deviation was 7.79; for the affective subscale it was 2.74 (M) and 2.56 (SD), and for the evaluative subscale it was 2.35 (M) and 1.53 (SD). The internal consistency of the sensory and affective subscales of the M P Q for this population were examined through Cronbach's alphas. No alpha could be computed for the evaluative subscale as participants could only pick one word in this subscale. The alpha for the sensory subscale was .73, which is within acceptable limits, but for the affective subscale it was .68, which is low, indicating that this subscale did not have good internal consistency for this population. Examples of questions from each of the multi-item measures are given in Appendix I. 67 Procedure Participants were initially screened over the phone to make sure they met the inclusion criteria, understood the purpose of the study, and knew what would be required of them. This was followed by a personal interview where the nature of the study was reviewed, and participants signed the consent form (Appendix C), and completed the demographic questionnaire, the MPQ, and the first of two administrations of the MPI. Participants were given a daily diary package that consisted of a concertina file holder with 30 morning and 30 nighttime sets of questionnaires arranged in a mornmg-mghttime sequence. The order of the questionnaires for the morning was Self-Efficacy for Pain, anxious and depressed mood scales, and current pain intensity. For the nighttime questionnaires, the order was the CSQ coping strategies, control, average pain, current pain intensity, CSQ Catastrophizing subscale, and the anxious and depressed mood scales. Each 7 days of dairies were separated by a stamped, addressed envelope. My phone number was pasted to the file holder, and a pen and adhesive labels were included in the package. During the personal interview, participants were trained in how to keep the diaries. As practice, they completed a nighttime diary based on their back pain experience the day before the appointment, and a morning diary for that day. Instructions for keeping the 30 days of diaries included (a) completing the morning diary within one hour of getting up and the nighttime one within one hour of going 68 to bed, (b) to miss the recording period entirely if they forgot or were unable to fill in a diary within that time period, (c) to complete the diary based on their experience for that day only, (d) to seal the diary with an adhesive label after completing it, and (e) to mail in their responses at the end of each block of 7 days. Participants were called at the end of each week to remind them to mail in their diaries and to problem solve any recording difficulties. At the end of the recording period participants completed the second MPI, and mailed it back with the last set of diaries. Participants were phoned the day they were due to finish the study to remind them to complete the second MPI, to thank them for their involvement, to inquire about their experience of being in the study, and to ask them whether they would prefer to be notified about the results of the study through group information meetings or a written summary. No incentives for participation were offered except for receiving the results of the study. Data Analysis Missing Data HLM (Bryk & Raudenbush, 1992, 1987) can handle missing data for the independent variables at the Level-1 analyses, but cannot handle any missing data at the Level-2 analyses. For the daily variables the following approach was taken to missing data. For variables that consisted of single numerical analogue scales (control over pain and pain intensity), the participant's mean was substituted for missing items if she had recorded 80% or more of the overall data points for that 69 item. If she had recorded less than this, the data point was coded as missing data. No score for the day was derived for the multi-item variables (i.e., Self-Efficacy, C S Q subscales, anxious and depressed mood) if these subscales had more than one item missing. However, the mean of the subscale was substituted if the participant had scores for that subscale on 80% or more of the recording days. This resulted in complete daily data being available for 96 to 100% of respondents, depending on the scale. There was no missing data for the Time 1 or Time 2 MPI. Preliminary Analyses Differences between completers and non-completers on demographic variables and the Time 1 MPI and M P Q were examined via analysis of variance and Chi-square. For those participants that completed the study, means, standard deviations, and Cronbach's alphas were computed for days 1, 15, and 30 as this is a more efficient way of describing the data over time than presenting all 30 days of data. The data were checked in three stages. First, outliers were examined for within- (i.e., variation in individual daily recording) and between-participant (differences between individuals) data. Although there were numerous between-participant outliers (classified as 3 standard deviations from the mean), no one participant was alone in having extreme scores. These outliers were not deleted from the analyses. 70 Second, to determine whether the assumptions for regression were met, box-plots, PP plots of the residuals, and scatter plots were examined for both the dependent and independent variables. With the exception of Catastrophizing and Reinterpreting Pain Sensation, the data appeared to meet the assumptions. Catastrophizing and Reinterpreting Pain Sensation data appeared skewed and non-normal. Various transformations were attempted, but were unsuccessful in creating a more normal distribution. Consequently, they were left unchanged. Third, the assumption of homogeneity of variance was examined through Chi-square analyses, and was not met for any of the models. However, H L M , like Analysis of Variance, is rather robust against the violation of the assumption of equal variances (Bryk & Raudenbush, 1992). What these findings indicate is that some women were quite variable in their responses to pain, whereas others demonstrated relatively consistent responses. Analyses of Diary Data: H L M Overview H L M was used to analyze the data as it provides a means for examining data that have an hierarchical structure. For the Level-1 of the hierarchy, I examined relationships between the three daily dependent variables (nighttime depressed mood, mghttime anxious mood, and nighttime pain intensity) and the daily independent variables (pain appraisals and coping strategy use) for each participant, and also assessed whether participants differed from each other on these relationships. For the Level-2 of the hierarchy, I examined whether these 71 differences were influenced by psychosocial and functional variables important to the experience of chronic pain. H L M also partitions the variance in the dependent Ovariables into variance due to individual variation in daily recording (within-participant) and variation due to differences between individuals (between-participant) at both the Level-1 and Level-2 analyses of the hierarchy. A l l of the independent variables at Level-1 and Level-2 were centered on the grand mean (i.e., centering). This computation involves subtracting the grand mean of a variable from each individual's score on that variable. This does not alter the estimated values for the regression slopes (i.e., the linear relationship between a dependent and independent variable), but does change the value of the intercepts. With centering, the intercept becomes the estimate of what the score on the dependent variable would be for a participant who scored at the mean on all of the independent variables; that is, the average score for the average participant. The analyses at Level-1 addressed the question of whether daily pain appraisals and coping strategy use were related to same night depressed mood, anxious mood, or pain intensity. A separate H L M model was examined for each of the three dependent variables. Because H L M examines these relationships for each individual (N = 88), there were 88 separate estimates of the average score on the dependent variable (the intercepts), and 88 separate estimates for each of the relationships between the dependent and the independent variables (i.e., the slopes). Figure 1 is a graphic representation of the Level-1 analyses. At Level-1 72 Figure 1 Graphic Representation of H L M Level-1 Analysis Within-Participant Analysis LEVEL-1 INTERCEPT Within-Participant Analysis = Average score on the dependent variable for a participant who scores at the mean on every independent variable - based on every participant's average score (N = 88) SLOPES -- HYPOTHESIS 1 STEP 1 ? = Which independent variables demonstrate a significant linear relationship with a participant's average score on the dependent variable -based on the relations for each independent variable for every participant Within- and Between-Participant Analysis STEP 2 VARIANCE EXPLAINED Between-Participant Analysis ? = How much of the variance in the intercept is attributable to within-participant and how much to between-participant - and are these significant RELIABILITY STEP 3 ? = Can you reliably distinguish between participants on their intercepts and slopes 73 H L M tests: (a) whether there is a significant linear relationship between each independent variable and participants' average scores on the dependent variable (Step 1; Hypothesis 1), (b) how much of the variability in the average scores on the dependent variable is attributable to variation in individual daily recording (within-participant) or to differences among participants (between-participant), and whether these variations are larger than would be expected by chance alone (Step 2), and (c) the extent to which you can reliably distinguish between participants on their average scores on the dependent variable and on their relationships between the dependent and independent variables (Step 3). There are two main types of analyses at Level-2. The first analysis addressed the question of whether psychosocial and functional variables (the MPI subscales) predict participants' levels of nighttime depressed mood, anxious mood, or pain intensity after taking into account their morning mood or pain, appraisals, and coping strategy use (the Level-1 intercepts). This is called the analysis of the Level-1 intercepts. H L M examines this first question for each individual, resulting in 88 separate regression equations. The coefficients for the Level-1 intercepts now become the dependent variables, and the MPI subscales are the independent variables. The Level-2 analyses are graphically displayed in Figure 2. H L M examines whether (a) the Level-2 intercepts (the average score on the Level-lintercept adjusted for the MPI subscales) can be predicted by any of the Level-2 independent variables (Stepl, Hypothesis 2), (b) how much of the variance in Figure 2 H L M L E V E L - 2 A N A L Y S E S F O R T H E INTERCEPTS 74 Within-Participant Analyses LEVEL-2 INTERCEPT Within-Participant Analyses Within- and Between Participant Analyses Average score on Level-1 intercept for a participant who scores at the mean on every Level-1 and Level-2 independent variable -based every participant's Level-1 intercept (N = 88) SLOPES -- HYPOTHESIS 2 STEP1 Which Level-2 independent variables demonstrate a significant linear relationship with participants' Level-2 intercepts - based on relationships for each independent variable for each participant VARIANCE EXPLAINED Between-Participant Analyses STEP 2 ? = How much of the variance in the intercept is attributable to within-participant and how much to between-participant - and are these significant STEP 3 ? = Can you reliably distinguish between participants on their Level-2 intercepts 75 theLevel-2 intercept can be attributed to within- and between-participants, and whether these variances are larger than would be predicted by chance (Step 2), and (c) the extent to which you can reliably distinguish between participants on their Level-2 intercepts (Step 3). The second analyses at Level-2 concerns the question of whether the MPI subscales influence the relationships between the dependent variable and independent variables (the slopes) identified at Level-1. H L M examines whether any of the Level-2 independent variables weaken or strengthen these linear relationships. This is called the Level-2 analyses of the slopes. Because this has not been investigated in the research literature before, the analyses of the slopes was exploratory in this study. Steps in H L M Model Building Null model. The first model in the analysis is called a null model as it does not include any independent variables from either Level-1 or Level-2 data (e.g., appraisals, coping strategies, and MPI subscales). The null model is used to partition the variance in the dependent variable into within-participant and between-participant component. If all of the variance were attributable to onlywithin-participant variation, the examination of between-participant differences would be unnecessary. However, if there is variation to be explained between participants, the null model provides a baseline from which to calculate the amount of variance explained by subsequent models. A null model was 76 computed for each of the dependent variables (nighttime depressed mood, anxious mood, and pain intensity). The equation for the null model is presented in Appendix D. Level-1 models. At this level, the daily relationships between the dependent and independent variables were examined. Prior to exaniining the full Level-1 models, preliminary analyses were conducted where each appraisal and coping strategy was added separately to the null model in order to test whether its slope was significant. This tested whether an appraisal or coping strategy was related to the dependent variable. This was done for two reasons. First, because the coping strategy of Reinterpreting Pain Sensation was only being examined in an exploratory manner, I planned to drop it from the model if it showed no significant relationship with the dependent variable. Second, if an independent variable demonstrated a significant relationship with a dependent variable but the relationships did not vary between individuals, its slope was constrained to be equal (i.e., fixed). A significant slope and non-significant variance suggests that although there is a linear relationship between the dependent and independent variable for most participants, one can not differentiate between participants on that basis. Consequently, by fixing the slopes, H L M assumes that this linear relationship is the same for every participant. Each independent variable was also paired with every other independent variable, and these pairs were examined in separate Level-1 models to check for potential suppression effects. 77 A Level-1 model was computed separately for each dependent variable using the following steps. First, each dependent variable's corresponding morning score was added to the null model as a covariate. Second, the model was then extended to include the appraisal variables of Self-Efficacy, Catastrophizing, and perceived control over pain. This provided an estimate of the proportion of both the within- and between-participant variance of the dependent variable that was explained by the pain appraisals, over and above that explained by the morning covariate. Third, this model was again extended to include the coping strategy variables. This allowed for the examination of the amount of variance in the dependent variables that was explained by the addition of the coping strategies, above and beyond that explained by the morning covariate and the three pain appraisals. The equations for the Level-1 model are presented in Appendix D. Level-2 models. This level of analysis examined whether psychosocial and functional variables important to the experience of chronic pain were associated with the daily pain processes identified in the Level-1 analyses. The first part of the Level-2 analyses examined whether the MPI subscales predicted participant's scores on nighttime depressed mood, nighttime anxious mood, or mghttime pain intensity that had been adjusted for the corresponding morning score, the appraisals, and the coping strategies (analyses of the Level-1 intercepts). The second part of the Level-2 analyses examined whether the MPI subscales altered the strength of the linear relationships between the three dependent variables and 78 the appraisals and coping strategies (analyses of the Level-1 slopes). The MPI subscales were divided into two groups: functional subscales and interpersonal subscales, although each subscale in the grouping was entered into the analyses. The functional subscales are Pain Severity (PS), Interference (I), Life Control (LC), Affective Distress (AD), Support (S), and General Activity Level (GAL), and the MPI interpersonal subscales are Distracting Responses (DR), Solicitous Responses (SR), and Punishing Responses (PR). Two Level-2 models were examined for each dependent variable; one for the functional subscales and one for the interpersonal subscales. Each model examined both the Level-1 intercepts and slopes. The equations for the Level-2 analyses are presented in Appendix D. Interaction effects The final step of the analysis was to examine, via H L M , the interaction effects (see Cohen & Cohen, 1983) between average daily pain and daily appraisals, and average daily pain and daily coping strategy use. The question being addressed here was "Does pain over the day moderate the relationships between pain appraisals or coping strategy use and nighttime depressed mood, nighttime anxious mood, or nighttime pain intensity?" This was an exploratory analysis and did not address any specific hypotheses. 79 C H A P T E R F O U R Results In this chapter, I report the differences between completers versus non-completers, descriptive statistics, and the main analyses. Completers Versus Non-Completers Comparability To assess whether those who dropped out of the study (n = 12) were different from those who completed it (n = 88), demographic variables were examined as well as Time 1 MPI and M P Q scores. A one-way analysis of variance ( A N O V A ) indicated no significant differences between completers and non-completers on age, F ( l , 96) = 2.27, p = .13; the number of years since their pain began, F ( l , 98) = 1.93, g = . 17; or the number of years years since pain was experienced daily, F ( l , 95) = 2.06, g = . 15. Means and standard deviations for all three variables are presented in Table A in Appendix E . The remaining demographic data were nominal and were analyzed using Chi-square analyses. The number of cells with an expected frequency of less than five was too high for a Chi-square analysis for employment status, household income, cause of the pain, number of health care professionals involved, and treatment for back pain in the last 6 months, therefore they were collapsed into smaller categories. No significant differences were found between completers and non-completers on these variables (Table B in Appendix E). Data on medication, litigation, ethnicity, and educational level could not be examined through Chi-80 square analysis as the collapsed groupings were still too small. A full description of the demographic variables are provided in Appendix B. A multivariate analysis of variance ( M A N O V A ) indicated a significant difference between completers and non-completers on the nine MPI subscales, F(9, 90)= 2.90, p < .01. Univariate analyses of the individual subscales revealed that non-completers compared to completers were more affectively distressed, F(l,99) = 6.84, p < .01, felt less control over their lives, F(l,99) = 6.85,_p_ <.01, and felt more supported in their pain by their spouse or partner, F(l,99) = 6.64, p_ < .01, but the two groups did not differ significantly on the other six subscales. Table C in Appendix E provides the means for the two groups on all nine MPI subscales. The results of the M A N O V A revealed no significant differences between the Completers and Non-completers on pain quality measured by the M P Q Sensory, Affective, and Evaluative subscales, F(3, 95) = 1.17, p. > .50. Descriptive Statistics Means, Standard Deviations and Cronbach's Alphas for the Daily Variables Table 2 shows the means, standard deviations, and Cronbach's alphas for days 1, 15, and 30 of the daily variables. The data were examined in this three day format as it was felt that this would give a clearer and more concise picture of participants' responses over time than just examining aggregated means or presenting descriptive statistics for all 30 days. A doubly M A N O V A indicated a significant overall time effect, F(14, 54) = 375.51, p. < .01, although univariate Table 2 Means, Standard Deviations, and Cronbach's Alphas for the Daily Variables for Days 1, 15, and 30 Day l Day 15 Day 30 Variable n a M SD n M SD a n M SD a Appraisals Self-efficacy 87 59.10 16.10 .75 86 60.05 20.11 .85 73 60.52 19.54 .84 Catastrophizing 86 0.98 1.27 .90 86 0.72 1.08 .89 72 0.63 1.06 .92 Control ^ 82 3.16 1.44 — 83 3.41 1.55 — 71 3.28 1.44 — Coping strategies Reinterpreting Pain 87 1.08 1.20 .86 86 0.72 1.18 .92 72 0.66 1.19 .93 Praying and Hoping 87 1.48 1.55 .80 87 1.32 1.70 .87 73 1.19 1.53 .89 Distraction 87 2.44 1.28 .83 86 2.38 1.53 .89 73 2.24 1.52 .89 Ignoring Pain 87 3.61 1.20 .87 87 3.25 1.61 .94 73 3.24 1.72 .94 Mood Anxiety A M 87 18.16 5.31 .86 88 16.91 5.46 .90 73 17.11 5.26 .87 Anxiety PM 85 17.37 4.44 .78 85 16.99 5.43 .89 72 16.97 5.00 .85 Depression A M 87 17.70 4.86 .82 86 17.28 5.46 .89 73 17.33 5.05 .86 Depression PM 85 17.87 4.63 .79 85 17.35 4.91 .84 72 17.21 4.88 .83 Pain b A M Pain 85 4.44 2.32 — 85 4.42 2.41 — 72 4.56 2.41 — PM Pain 87 4.70 2.57 — 86 4.40 2.60 — 73 4.64 2.64 — Average Pain 87 5.08 2.14 — 86 4.67 2.11 — 73 4.79 2.24 — Note. The higher the score, the greater the attribution. Dashes indicate data not applicable. a Missing cases due to incomplete monitoring.b No Alpha is given as the measure is a single, numerical analogue scale. 82 analysis of the individual variables revealed a significant difference only on Catastrophizing, F(2, 87) = 5.67, p < .01, with the mean score decreasing over time. However, the largest decrease was only .35 of a unit (day 1 to day 30). Because this subscale is based on a 7-point scale, this decrease is quite small, and from a clinical perspective, not very meaningful. Correlations For the Daily Variables and the Time 1 MPI Tables 3 and 4 present the correlations among all of the continuous variables for day 1 of data collection. Although H L M bases its analysis on a variance-covariance matrix, the correlations are presented for interpretive purposes. Correlations were also examined for days 15 and 30, and are presented in Tables A - D in Appendix F. Because the correlation matrix is so large, it has been divided into two tables. Table 3 lists the correlations for the daily and demographic variables, whereas Table 4 lists the correlations for the Time 1 MPI subscales with these same variables. The correlations for all of the variables for all 3 days are quite similar in magnitude and direction. There are some differences in whether a correlation is significant, but when this occurs, the significant correlation is quite small and not greatly different from the non-significant correlation. The following discussion concerns the correlations for all 3 days, and is based on the significant correlations only. CO 0 0 Table 3 Pearson-Product Moment Correlation Matrix of Demographic and Daily Variables for Day 1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 1 Age 2 Years Pain Began .26* 3 Years Pain Daily .19 .68* 4 Anxiety A M -.10 .02 .04 5 Anxiety PM -.08 -.10 -.14 .67** 6 Depression AM -.09 -.04 .02 .75** .48** 7 Depression PM -.06 -.11 . -.10 .47** .70** .62** 8 Average Pain .20 .19 .19 .23* .13 .10 .09 9 A M Pain .06 .24* .17 .31** .13 25** .00 .57** 10 PMPain .03 .14 .10 .14 .18 .11 .10 .75** .48** 11 Catastrophizing .12 .04 .10 .34** .31** .35** 39** .30** .04 .19 12 Control -.01 .21 .18 -.21 -.25* -.20 -.18 -.14 -.23* -.03 -.16 13 Self-efficacy -.17 -.04 -.15 -.34** -.31** -.38** -.24* -.36** -.35** -.32** -.27* .38** 14 Distraction -.02 .15 .05 .00 .03 -.08 -.04 .21* .12 .27** .12 .34** .03 15 Ignoring Pain -.10 .06 .07 -.03 -.11 -.09 -.17 .16 .03 .12 -.07 29** .22* .15 16 Praying and Hoping .13 .05 .10 .14 .09 .14 .04 .17 .04 .18 .42** .18 .01 .38** .25* 17 Reinterpreting Pain Sensation -.03 -.07 . -.09 .11 .04 .14 .07 .07 -.07 .09 .18 .13 .06 .34** .38** .34** Note. Due to missing data, minimum n = 79 *p_<.05 **p_<.01 Table 4 Pearson-Product Moment Correlation Matrix of the Time 1 MPI, Demographic, and Daily Variables for Day 1 18 19 20 21 22 23 24 25 26 1 Age -.14 .14 -.00 .17 -.24* .07 -.03 .01 .08 2 Years Pain Began -.14 .02 .31** .31** -.08 -.02 .04 .20 .17 3 Years Pain Daily .01 .08 .25* .35** -.06 -.02 .05 .04 -.14 4 Anxiety A M .45** -.42** .26* .21 -.01 -.18 .08 .20 -.10 5 Anxiety PM .36** -.13** .16 .12 -.04 -.14 .17 .26* .00 6 Depression AM .42** _ 46** .26* .14 -.13 -.30** -.09 .09 -.20 7 Depression PM .46** -.42** .12 .13 .02 -.21 .03 .25* -.08 8 Average Pain .07 -.20 .36** .60** .03 .20 22* .07 .28** 9 A M Pain .05 -.25** .27* .45** -.06 -.07 .02 .14 .03 10 PM Pain .12 -.13 .41** .58** -.14 .12 .14 .09 .22* 11 Catastrophizing .17 -.13 .22* .22* -.05 .04 .16 .00 .09 12 Control -.17 .33** .02 .04 .13 -.01 .10 .12 .13 13 Self-efficacy -.25* .25* . 39** . 4i** .28** -.20 -.06 -.04 -.14 14 Distraction -.05 -.04 .16 .17 .06 .15 .40** .08 .30** 15 Ignoring Pain -.09 .09 -.16 .09 .12 -.12 .08 .07 .04 16 Praying and Hoping -.07 .02 .01 .11 -.12 -.08 .13 .05 -.07 17 Reinterpreting Pain Sensation .00 -.08 -.07 -.09 .07 .04 .12 -.09 .12 18 MPI Affective Distress -.70** .31** .31** .07 -.24* -.01 39** -.14 19 MPI Life Control -.19 -.25* .00 .31** .06 -.28** .20 20 MPI Interference .52** -.12 .24* .18 .04 .34** 21 MPI Pain Severity -.13 .09 -.01 .20 .10 22 MPI General Activity Level .07 .22* -.00 .13 23 MPI Support .46** -.53** .70** 24 MPI Distracting Responses -.11 .63** 25 MPI Punishing Responses -.27* 26 MPI Solicitous Responses Note. Due to missing data, minimum n = 79, *p<.05 **p<.01 85 With regards to the demographic variables, although the correlations are low, older women reported longer pain duration and less general activity. Years since pain began has a strong positive correlation with years since pain was experienced daily, moderate positive correlations with the MPI subscales of Interference and Pain Severity, and low positive correlations with pain intensity. Years since pain was experienced daily has low positive correlations with the MPI subscales of Interference and Pain Severity. With regards to the daily variables, the correlations for appraisals, pain, and mood are in the predicted direction, although are low to moderate in magnitude. Self-efficacy and control are negatively correlated with anxious and depressed mood, and pain, whereas pain and Catastrophizing are positively correlated. The four coping strategies demonstrate few significant relationships with the daily variables, and the results are inconsistent across the three days. With regards to the functional variables of the MPI, Affective Distress has moderate positive correlations with anxious and depressed mood, and low negative correlations with Self-Efficacy and control. Life Control has moderate negative correlations with anxious and depressed mood, and low positive correlations with control, Self-Efficacy, and pain intensity. Pain Severity has moderate to high positive correlations with pain intensity, and moderate negative correlations with Self-Efficacy. Finally, General Activity Level has a low negative correlation with 86 Self-Efficacy. The interpersonal variables of the MPI (Solicitous, Distracting, and Purnshing Responses) have weak relationships with the other variables. Comparison With Select Clinical Pain Populations on the MPI and M P Q Because chronic pain sufferers from the general community have rarely been included in pain research, it is useful to place these participants' scores on the psychosocial and functional variables in the context of results from clinical pain populations. Therefore, in Tables E - H in Appendix F I present data from select clinical studies. The studies for the MPI are taken from the MPI manual (Rudy, 1989), and are based on a sample of heterogeneous chronic pain patients and a sample of lower back pain patients. Unfortunately, no comparisons could be made concerning demographic data as this information was not provided in the manual. The analyses of the differences between this study's participants and those of the selected pain populations were conducted via two-tailed t-tests. Because the risk of a Type I error was seen as more important than a Type II error, the p values for the analyses were set at < .001. Study participants had significant lower levels of Affective Distress, Pain Interference, Pain Severity, and Support, and higher levels of Life Control and General Activity Level. They were also significantly different on Solicitous Responses, but not on Distracting and Punishing Responses. This suggests that even though pain patients may not perceive their spouses to respond to their pain in a more pumshing or distracting way than non-clinical pain sufferers, they may perceive them to respond more to their pain behaviour. Comparisons were also made with two studies of chronic low back pain patients on pain quality (MPQ subscales). One data set was comprised of 40 females recruited from a comprehensive back pain assessment centre (Reesor & Craig, 1988), and the second data set was comprised of 32 patients attending a pain treatment centre (Keefe & Dolan, 1986). Compared to the other two groups, study participants were somewhat older, and had experienced back pain substantially longer. The results from the t-tests showed no significant differences between the study participants and the other two groups on any of the M P Q subscales. Although based on limited comparisons, these findings suggest that even though this study's community sample appeared to be similar to certain clinical pain populations on measures of pain quality (i.e., MPQ), they were less functionally and psychologically impaired, and may not have received as much reinforcement for their pain behaviour from their spouses. Main Analyses Overview To test the various models implied by the hypotheses, I first of all considered the daily relationships between the three pain appraisals, the four coping strategies, and nightly depressed mood, anxious mood, and pain intensity. 88 Next, I determined the relative importance of psychosocial and functional variables (the MPI subscales) in predicting the Level-1 results. Finally, I examined the interaction effects between daily pain appraisals or coping strategies and average daily pain. Null Models Table 5 shows the mean scores across participants for nighttime depressed mood, nighttime anxious mood, and nighttime pain intensity, which were the results of testing three null models (i.e., no independent variables included). The analysis also estimated the within- and between-participant standard deviations of the means, and tested their statistical significance. A l l three between-participant standard deviations were statistically significant (all gs < .001). The .95 confidence interval was 10.56 to 24.56 for nighttime depressed mood, 9.67 to 25.03 for nighttime anxious mood, and 0.28 to 8.78 for nighttime pain intensity. These results showed a substantial amount of variation between participants on all three dependent variables, indicating that further analyses to determine the sources of these differences were warranted. For nighttime depressed mood, the estimated within-participant variance was 10.72 and the between-participant variance wasl2.28. Expressed as percentages, these variance components indicated that 47% of the variance in nighttime depressed mood was attributed to variation in participant's own daily 89 Table 5 H L M Results for the Null Models for Nighttime Depressed Mood. Nighttime Anxious Mood, and Nighttime Pain Intensity (N = 88) Nighttime Nighttime Nighttime Depressed Anxious Pain Mood Mood Intensity Mean of Individuals' Average Ratings 8 17.56 17.35 4.53 Standard Deviation (Within) a 3.27 3.31 1.46 Standard Deviation (Between)a 3.50 3.85 2.13 Model Statistics Variance: Within Participants 10.72 10.98 2.14 Between Participants 12.28 14.80 4.53 Components of Variation as Percentages Within Participants 47 43 32 Between Participants 53 57 68 Reliability Estimates .97 .98 .98 Significant at p < .001 recording (within-participant), and 53% of the variance to differences between participants' mean scores. For nighttime anxious mood, 43% of the variance in cores was within-participant and 57% between-participant. Finally, for nighttime pain intensity, 32% of the variance was within-participant and 68% between-participant. 90 The reliability estimates are indicators of our ability to distinguish between participants on their mean scores on the dependent variables. A l l three had very high reliability, indicating that we can reliably distinguish between individuals in their mean scores on the dependent variables. In summary, the results of the analyses of the null models indicated that there was substantial variation within- and between-participants on all three dependent variables. However, participants' varied more when compared with each other than they did in their individual daily recordings. Explanation of these variances was attempted in the Level-1 and Level-2 analyses. For each of the dependent variables, I discuss the Level-1 results and then the Level-2 results before proceeding on to the results for the next dependent variable. Nighttime Depressed Mood Level-1 analyses. It was predicted that higher levels of nighttime depressed mood would be associated with lower levels of perceived control over pain, Self-Efficacy, Distraction, and Ignoring Pain, and higher levels of Catastrophizing and Praying and Hoping strategies. Table 6 shows the results of the analyses, the reliability estimates, and the within- and between-participant explained variances. The slopes for Catastrophizing and control were fixed in order to avoid the loss of participants who scored zero on these variables for the recording days. The slopes for Self-Efficacy also were fixed as they did not vary 91 Table 6 Level-1 H L M Analyses of the Effects of Morning Depressed Mood, the Three Pain Appraisals, and Distraction, Ignoring Pain, and Praying and Hoping on Nighttime Depressed Mood (N = 88) Fixed Effects Random Effects Unstandardized SE Standardized Variance Coefficient Coefficient Intercept 17.526** 0.226 3.994** Depressed mood - am 0.373** 0.032 1.719 0.049** Self-efficacy 0.015* 0.006 0.289 fixed Catastrophizing 1.444** 0.093 1.662 fixed Control -0.424** 0.064 -0.627 fixed Distraction -0.382** 0.091 -0.571 fixed Ignoring Pain -0.227** 0.074 -0.346 fixed Praying and Hoping -0.068 0.091 -0.111 fixed Level-1 (within variance) 6.857 Reliability of Estimates Intercept .830 Depression - am .530 Percent Variance Explained Morning Depressed Mood Within-participant 23 Between-participant 69 Morning Depressed Mood and Appraisals : Within-participant 35 Between-participant 69 Morning Depressed Mood, Appraisals, and Coping Strategies: Within-participant 36 Between-participant 67 Dashes indicate data are not applicable. *p_<.05 **p_<01 92 between participants in the preliminary analysis. Even though the slopes and variance for Praying and Hoping were not significant in the preliminary analysis this coping strategy was kept in the analyses in order to adhere to the original hypotheses concerning the model, but its slopes were fixed. The retention of Praying and Hoping and the fixing the slopes did not substantially change the other coefficients. Remterpreting Pain Sensation was dropped from the model as it was not significant at either level in the preliminary analysis. Even after taking into account morning depressed mood, Catastrophizing had a significant positive effect on nighttime depressed mood, whereas control had a significant negative effect, with Catastrophizing being the strongest predictor. The unstandardized coefficients indicate that, on average, for every unit increase in the Catastrophizing score, the nighttime depressed mood score would increase by 1.44, and for every unit increase in the control score, the nighttime depressed mood score would decrease by 0.42. Self-Efficacy was positively related to nighttime depressed mood, which was in the opposite direction to what was predicted. Although this may indicate that higher levels of Self-Efficacy during the day are associated with higher levels of depression at night, there was some indication that this result was due to a suppression effect between the Self-Efficacy and control variables. First, based on days 1, 15, and 30 of the diary monitoring, the correlations between Self-Efficacy and nighttime depressed mood ranged from -.24 to -.50 (Tables A and C in 93 Appendix F), whereas the correlations between the Self-Efficacy and control variables were somewhat higher (rs ranging from .35 to .65). Second, as part of the preliminary analysis for this model, each pain appraisal was added separately to the null model. For Self-Efficacy, results indicated a significant negative relationship with nighttime depressed mood, even after taking into account morning depressed mood (b = -0.02, t = 2.51, g < .01). This significant relationship disappeared when the model was expanded to include the control variable. These results suggest there may be a significant negative relationship between Self-Efficacy and nighttime depressed mood that is being suppressed in this analysis by the control variable. With regards to the coping strategies, both Distraction and Ignoring Pain were significantly and negatively associated with nighttime depressed mood. Based on the unstandardized coefficients, for every unit increase in the Distraction or Ignoring Pain scores, the mghttime depressed mood score would decrease by .38 or .23, respectively. Praying and Hoping was not significant. Table 6 shows the estimates of the random effects, and the significance of these effects across participants. The variance of the intercept was significant, indicating that there were significant differences between participants on their adjusted mghttime depressed mood scores. Reliability estimates for the model were high, suggesting we can reliably distinguish between participants in their adjusted nighttime depressed mood mean. 94 The partitioning of the variance in the null model (Table 5) indicated that 47% of the variance in nighttime depressed mood was within participants, and 53% was between participants. The results for this Level-1 analysis indicates that morning depressed mood explained 23% of the within-participant variance, the three pain appraisals (after controlling for morning mood) explained 12%, and the three coping strategies (after controlling for morning mood and the three appraisals) explained 1%. For the between-participant variance, 69% percent of the variance was attributable to morning depressed mood. Because all of the slopes for the pain appraisals and coping strategies were fixed, these variables could not explain any of the between-participant variance. However, the coping strategies actually decreased the explained between-participant variance by 2%, which may be due to the inconsistency among participants on coping strategy use, or the inter-relationship between the appraisals and coping strategies. Together, morning depressed mood, the three pain appraisals, and the three coping strategies explained 36% of the within-participant variance and 67% of the between-participant variance, which is 53% of the total variance. However, almost all of this explained variance was attributable to morning depressed mood and the three pain appraisals. Forty-seven percent of the total variance was left to be explained (64% within-participant and 33% between-participant). Further explanation of the variance was attempted in the Level 2 analysis. 95 In summary, Hypothesis la was partly supported as, after controlling for morning depressed mood, participants who experienced higher levels of depression at night catastrophized more, felt less control over their pain, and used fewer Distraction and Ignoring Pain strategies during the day. However, in contrast to the predictions of Hypothesis la, Self-Efficacy was associated with higher levels of nighttime depression, and there was no relationship found for Praying and Hoping. Level-2 analysis. The main thrust of this analyses concerned the examination of the relationships between the Level-2 independent variables and the Level-1 intercept (i.e., the analyses of the intercepts). That is, did psychosocial and functional variables important to the experience of chronic pain (Time 1 MPI subscales) predict participants' average daily mghttime depressed mood scores, adjusted for morning depressed mood, the appraisals, and coping strategies. Specifically, for the MPI functional subscales, it was predicted that mghttime depressed mood would be negatively associated with Life Control, Support, and General Activity Level, and positively associated with Interference, Affective Distress, and Pain Severity. For the MPI interpersonal subscales it was predicted that nighttime depressed mood would be positively associated with Punishing and Distracting Responses, and negatively associated with Solicitous Responses. 96 The exploratory part of this analyses examined whether the Level-2 independent variables changed the magnitude of the Level-1 relationships between the independent variables and nighttime depressed mood (i.e., the analyses of the slopes). The morning depressed mood slope was the only slope examined in this analyses as it was the only slope not fixed in the Level-1 model. Table 7 shows the results of the Level 2 analyses for the MPI functional subscales. For efficiency sake, only those Level-2 effects that were statistically significant are presented in Table 12, and the complete results are presented in Table A in Appendix G. With regard to the analyses of the intercepts, the Support subscale had a significant and negative effect on the average depressed mood score across participants, whereas Affective Distress had a significant positive effect. This indicates that, on average, after taking into account the other MPI functional subscales, the more supportive a participant perceived her spouse to be concerning her pain, the less depressed she felt at night, and the more emotionally distressed she was in general, the more depressed she was at night. There were no significant effects for the remaining MPI functional subscales, or for the morning depressed mood slope. The MPI subscales did not explain any of the witiun-participant variance for nighttime depressed mood, but did explain 9% of the between-participant variance. This indicates that the MPI functional subscales contributed 97 Table 7 Level-2 H L M Results for Nighttime Depressed Mood. The Relationships Between the MPI Functional Subscales and the Level-1 Intercept and Morning Depressed Mood Slope (N = 88) Fixed Effects Random Effects Unstandardized Coefficient SE Variance Intercept 17.511** 0.217 3.617** Depressed mood - am 0.370** 0.033 0.051** Self-efficacy 0.014* 0.007 fixed Catastrophizing 1.436** 0.093 fixed Control -0.423** 0.064 fixed Distraction -0.371** 0.092 fixed Ignoring Pain -0.213** 0.075 fixed Praying and Hoping -0.092 0.090 fixed Level-1 (within variance) 6.856 Effects of MPI-level variables Support on intercept -0.350* 0.153 Affective Distress on intercept 0.570* 0.269 Reliability of the Estimates Level-1 Variance Explained Intercept .818 Within-participant 0% Depressed Mood - am Between-participant 9% slope .538 Note. Only statistically significant Level 2 effects are presented. See Table A in Appendix G for complete results. Dashes indicate data are not applicable *P<.05 **p_<01 98 more to differences between people on depressed mood than to individual day-to-day variation. The reliability estimate for the Level-2 nighttime depressed mood model remains high, indicating that we can reliably distinguish between participants in their nighttime depressed mood mean adjusted for the MPI functional subscales In summary, Hypothesis 2a was partially supported as lower levels of average nighttime depressed mood were associated with lower levels of emotional distress and higher levels of spousal support, after taking into account morning depressed mood, the appraisals, and the three coping strategies. Hypothesis 2a was not supported for Interference, Pain Severity, Life Control, or General Activity Level as these showed no significant relationships with average nighttime depressed mood. In the exploratory analyses, the MPI functional subscales did not change the magnitude of the relationship between the morning and nighttime depressed mood slope. Table 8 shows the results of the analyses for nighttime depressed mood and the MPI interpersonal subscales. As in the previous Level-2 analysis, only the Level-1 intercept and morning depressed mood slope were regressed on the MPI interpersonal subscales, and only the significant Level-2 effects are presented in the table. Table B in Appendix G gives full details of the Level-2 effects. With regard to the analyses of the intercepts, there was a significant positive relationship between Punishing Responses and the average nighttime 99 Table 8 Level-2 H L M Results for Nighttime Depressed Mood. The Relationships Between the MPI Interpersonal Subscales and the Level-1 Intercept and Morning Depressed Mood Slope (N = 88) Fixed Effects Random Effects Unstandardized Coefficient SE Variance Intercept 17.533** 0.221 3.779** Depressed mood - am 0.370** 0.033 0.051** Self-efficacy 0.015* 0.006 fixed Catastrophizing 1.438** 0.093 fixed Control -0.428** 0.065 fixed Distraction -0.388** 0.092 fixed Ignoring Pain -0.225** 0.074 fixed Praying and Hoping -0.074 0.091 fixed Level-1 (within variance) 6.856 Effects of MPI-level variables Punishing Responses on intercept 0.479* 0.172 Reliability of the Estimates Level-1 Variance Explained Intercept .823 Depressed mood - am slope .538 Within-participant 0% Between-participant 3% Note. Only significant Level 2 effects are given. See Table B in Appendix G for complete results. Dashed indicate data are not applicable *P<.05 **p<.01 100 depressed mood score across participants. This indicates that on average, after taking into account Solicitous and Distracting Responses, the more participants perceived their spouses or partners to be acting in a punishing way to their pain, the higher their levels of nighttime depressed mood. There were no significant effects for Distracting or Solicitous Responses, or for the morning depressed mood slope. The MPI interpersonal subscales did not explain any of the within-participant variance for nighttime depressed mood, but did explain 3% of the between-participant variance. This indicates that the MPI interpersonal subscales contributed more to differences between people on mghttime depressed mood than to individual day-to-day variation. The reliability estimate for nighttime depressed mood remained high, and indicates that we can reliably differentiate between participants on their nighttime depressed mood mean after adjusting for the MPI interpersonal subscales. In summary, Hypothesis 2a was partly supported as people who perceived their spouses' responses to be punishing experienced higher levels of depression at night, after taking into account morning depressed mood, the appraisals, and the three coping strategies. Hypothesis 2a was not supported for Solicitous or Distracting Responses as no relationships were found between these predictors and nighttime depressed mood. In the exploratory analysis, the MPI interpersonal subscales did not change the magnitude of the relationship between the morning and nighttime depressed mood slope. 101 Nighttime Anxious Mood Level-1 analyses. It was predicted that higher levels of nighttime anxious mood would be associated with higher levels of Catastrophizing and Praying and Hoping, and lower levels of perceived control over pain, Self-Efficacy, Ignoring Pain, and Distraction when controlling for morning anxious mood. Table 9 shows the results of the analyses, the reliability estimates, and the explained within- and between-participant variances. Reinterpreting Pain Sensation was not significant in the preliminary analysis and was dropped from the model. Both the Catastrophizing and control slopes were fixed in order to avoid loss of participants. The slopes for Distraction, Ignoring Pain, and Praying and Hoping were fixed as their relationships with nighttime anxious mood did not vary across participants in the preliminary analysis. With regard to the appraisals, after controlling for morning anxious mood, Catastrophizing had a significant positive effect on nighttime anxious mood, whereas control had a significant negative effect, with Catastrophizing being the strongest predictor. Based on the unstandardized coefficients, on average, for every unit increase in the Catastrophizing score, the predicted nighttime anxious mood score would increase by 1.12, and for every unit increase in the control score, the anxious mood score would decrease by .39. 102 Table 9 Level-1 H L M Analysis of the Effects of Morning Anxious Mood, the Three Pain Appraisals, and Distraction, Ignoring Pain, and Praying and Hoping on Nighttime Anxious Mood (N = 88) Fixed Effects Random Effects Unstandardized Coefficient SE Standardized Coefficient Variance Intercept 17.254** 0.272 5.712** Anxious mood - am 0.341** 0.031 1.709 0.002** Self-efficacy -0.000 0.009 0.017 0.044* Catastrophizing 1.121** 0.100 1.271 fixed Control -0.389** 0.070 -0.563 fixed Distraction -0.482** 0.101 -0.735 fixed Ignoring Pain -0.138 0.081 -0.244 fixed Praying and Hoping 0.152 0.102 0.258 fixed Level-1 (within variance) 7.791 Reliability of Estimates Intercept Anxious mood - am Self-efficacy .691 .487 .268 Percent Variance Explained Morning Anxious Mood: Within-participant 18 Between-participant 63 Morning Anxious Mood and Appraisals: Within-participant 28 Between-participant 68 Morning Anxious Mood, Appraisals, and Coping Strategies: Within-participant 29 Between-participant 61 Dashes indicate data are not applicable. *p_ < .05 **p_ <.01 103 Self-Efficacy was not significantly related to nighttime anxious mood. However, this result may reflect a suppression effect between the Self-Efficacy and control variable because the correlations between Self-Efficacy and nighttime anxious mood for days 1, 15, and 30 (see Tables A and C in Appendix F) were moderate (range -.21 to -.42), indicating that there could be a negative relationship between these two variables. In addition, when Self-Efficacy was added separately to the null model in the preliminary analysis, the result for Self-Efficacy was significant, b = -0.03, t = -3.68, g < .01, even after controlling for morning anxious mood. This significant relationship disappeared when the model was expanded to include the control variable. With regard to the coping strategies, Distraction had a significant negative relationship with nighttime anxious mood. Based on the unstandardized coefficients, for every unit decrease in the Distraction score, the nighttime anxious score would increase by .48. Praying and Hoping and Ignoring Pain were not significantly associated with nighttime anxious mood. Table 9 also shows the estimates of the random effects and the significance of these effects across participants. The variance of the intercept was significant, indicating that there were significant differences between participants on their nighttime anxious mood scores after adjusting for morning anxious mood, the three pain appraisals, and the three coping strategies. The reliability estimate 104 for the adjusted mean is still moderately high, indicating that we can distinguish between participants in their adjusted mghttime anxious mood. Participants differed significantly from each other on Self-Efficacy even though there was not a significant relationship between Self-Efficacy and nighttime anxious mood across participants. This difference between findings at the two levels can be interpreted in two ways. Either the non-significant findings for the slopes is the result of a suppression effect between the Self-Efficacy and control variable, or that, although on average across individuals, there is no relationship between Self-Efficacy and nighttime anxious mood, individuals do vary in that relationship. That is, for some individuals, lower scores on Self-Efficacy are associated with higher scores on nighttime anxious mood. However, the reliability estimate for Self-Efficacy was low (.27) indicating that it is difficult to distinguish between participants on that basis. The partitioning of the variance in the null model (Table 5) indicated that 43% of the variance in nighttime anxious mood was within-participant and 57% was between-participant. The results of this Level-1 analysis indicated that morning anxious mood explained 18% of the within-participant variance, the three pain appraisals (controlling for morning anxiety) explained 10%, and coping strategy use (controlling for morning anxiety and the three appraisals) explained 1%. Sixty-three percent of the between-participant variance was attributable to morning anxious mood, and 5% to the three pain appraisals. Because all of the 105 slopes for the coping strategies were fixed, these variables could not explain any of the between-participant variance. However, these variables decreased the explained between-participant variance by 7% which may be due to the fact that participants were inconsistent in their coping strategy use, or to the inter-relationship between the appraisals and coping strategies. Together morning anxious mood, the three pain appraisals, and the three coping strategies explain 29% of the within-participant variance and 61% of the between participant variance, which is 48% of the total variance. However, almost all of this variance is attributable to morning anxious mood and the three pain appraisals. Fifty-two percent of the total variance was left to be explained (71% within and 39% between). Further explanation of the variance was attempted in the Level 2 analysis. In summary, Hypothesis lb was partly supported as, after controlling for morning anxious mood, higher levels of nighttime anxious mood were associated with increased Catastrophizing, decreased perceived control over pain, and less frequent use of Distraction as a coping strategy. However, Hypothesis lb was not supported for Self-Efficacy, Praying and Hoping or Ignoring Pain as no relationships were found between these variables and nighttime anxious mood. Level-2 analyses. This level of analyses examined the relationships between the Level-2 independent variables and the Level-1 intercepts (analyses of the intercepts). That is, did psychosocial and functional variables (Time 1 MPI 106 subscales) predict participant's average mghttime anxious mood scores, adjusted for morning anxious mood, the appraisals, and coping strategy use. Specifically, for the MPI functional subscales, it was predicted that nighttime anxious mood would be negatively associated with Life Control, Support, and General Activity Level, and positively associated with Interference, Affective Distress, and Pain Severity. For the MPI interpersonal subscales it was predicted that nighttime anxious mood would be positively associated with Punishing and Distracting Responses, and negatively associated with Solicitous Responses. The exploratory part of the analyses examined whether the Level-2 independent variables changed the magnitude of the relationships between the Level-1 independent variables and nighttime anxious mood (the analyses of the slopes). This analysis only included the morning anxious mood slope, as apart from Self-Efficacy, all the other slopes for the independent variables had been fixed. Self-Efficacy was allowed to be random in the Level-2 model in order to match the Level-1 model, but it was not regressed on the predictors as its slope was not significant at Level-1. Table 10 shows the results of the Level-2 analyses for nighttime anxious mood and the MPI functional subscales. Only those Level-2 effects that were statistically significant are presented, and complete results are presented in Table A in Appendix G. With regard to the analyses of the intercepts, Interference and Affective Distress were significantly and positively related to the average 107 Table 10 Level-2 H L M Results for Nighttime Anxious Mood. The Relationships Between the MPI Functional Subscales and the Level-1 Intercept and Morning Anxiety Slope (N = 88) Fixed Effects Random Effects Unstandardized Coefficient SE Variance Intercept 17.236** 0.254 4.745** Anxious Mood - am 0.333** 0.031 0.040** Self-efficacy 0.001 0.009 0.002** Catastrophizing 1.100** 0.100 fixed Control -0.399** 0.069 fixed Distraction -0.490* 0.101 fixed Ignoring Pain -0.105 0.082 fixed Praying and Hoping 0.127 0.100 fixed Level-1 (within variance) 7.785** Effects of MPI-level variables Interference on intercept 0.647* 0.266 Affective Distress on intercept 0.700* 0.316 Support on intercept -0.381* 0.192 Reliability of the Estimates Level-1 Variance Explained Intercept .663 Within-participant 0% Anxious Mood- am Between-participant 17% slope .466 Self-Efficacy .280 Note. Only statistically significant Level 2 effects are presented. See Table A in Appendix G for complete results. Dashes indicate data are not applicable. *p_<.05 **p_<01 108 mghttime anxious mood score across participants. This indicates that, on average, after taking into account all of the MPI functional subscales, higher levels of nighttime anxious mood were reported by participants who perceived pain to be interfering in their lives in general, or by those who were more emotionally distressed overall. No relationships were found for the other five MPI functional subscales, or for the morning anxious mood slope. The MPI functional subscales did not explain any of the within-participant variance for nighttime depressed mood, but did explain 17% of the between-participant variance. These results indicate that these subscales contributed more to differences between people on anxious mood than to individual daily variation. The reliability estimates for the Level-2 model was only moderate, indicating that although we can distinguish between participants in their adjusted nighttime anxious mood mean for the Level-2 model, we should exercise some caution when doing so. In summary, Hypotheses 2b was partially supported as higher levels of anxious mood were associated with higher levels of interference of pain on daily life and overall emotional distress, and lower levels of spousal support, after taking into account morning anxious mood, the appraisals, and the three coping strategies. Hypothesis 2b was not supported for Pain Severity, Life Control, or General Activity Level as no significant relationships were found between these variables and nighttime anxious mood. In the exploratory analyses, the MPI 109 functional subscales did not change the magnitude of the relationship between the morning and nighttime anxious mood slope. Table 11 shows the results of the Level-2 analyses for nighttime anxious mood and the MPI interpersonal subscales. As in the previous Level-2 analyses, only the intercept and morning anxious mood slope were regressed on the MPI interpersonal subscales, and only the significant Level-2 effects are presented. Table B in Appendix G gives full details of the Level-2 effects. With regard to the analyses of the intercepts, increased Punishing Responses were associated with higher average nighttime anxious mood scores across participants. This indicates that, on average, after taking into account Solicitous and Distracting Responses, the more participants perceived their spouses or partners to be acting in a punishing way to their pain, the higher their level of mghttime anxious mood. There were no significant relationships for the Solicitous or Distracting Responses subscales, or for the morning anxious mood slope. The MPI interpersonal subscales did not explain any of the within-participant variance for nighttime depressed mood, but did explain 5% of the between-participant variance. These results suggest that these subscales contribute more to differences between people on anxious mood than to individual variation across time. The reliability estimate for nighttime anxious mood is moderate, indicating that although we can differentiate between participants in their 110 Table 11 Level-2 H L M Results for the Nighttime Anxious Mood. The Relationships Between the MPI Interpersonal Subscales and the Level-1 Intercept and Morning Anxious Mood Slope (N = 88) Fixed Effects Random Effects Unstandardized Coefficient SE Variance Intercept 17.276** 0.270 5.612** Anxious Mood - am 0.339** 0.031 0.043** Self-efficacy -0.000 0.009 0.002** Catastrophizing 1.111* 0.100 fixed Control -0.395** 0.070 fixed Distraction -0.496** 0.102 fixed Ignoring Pain -0.139 0.081 fixed Praying and Hoping 0.161 0.102 fixed Level-1 (within variance) 7.784 Effects of MPI-level variables Punishing Responses on intercept 0.647** 0.210 Reliability of the Estimates Level-1 Variance Explained Intercept .689 Within-participant 0% Anxious Mood - am Between-participant 5% slope .483 Self-Efficacy .266 Note. Only significant Level 2 effects are given. See Table B in Appendix G for complete results. Dashed indicate data are not applicable *P<.05 **p<.01 Ill nighttime anxious mood mean adjusted for the MPI interpersonal subscales, we should do this with some caution. In summary, Hypothesis 2b was partially supported as participants who reported more punishing types of responses from their spouses or partners reported experiencing higher levels of anxiety at night, after taking into account morning anxious mood, the appraisals, and the three coping strategies. Hypothesis 2b was not supported for Distracting or Solicitous Responses as no significant relationships were found between these variables and nighttime anxious mood. In the exploratory analysis, the MPI interpersonal subscales did not change the magnitude of the relationship between the morning and nighttime anxious mood slope. Nighttime Pain Intensity Level-1 analyses. It was predicted that higher levels of nighttime pain intensity would be associated with higher levels of Catastrophizing, Praying and Hoping, Ignoring Pain, and Distraction, and lower levels of perceived control over pain, Self-Efficacy, and Distraction, after controlling for morning pain intensity. The control and Catastrophizing slopes were again fixed. Because the slopes for Reinterpreting Pain Sensation were significant in the preliminary analysis, it was entered into the model. Although in the preliminary analysis the linear relationships between nighttime pain intensity and the four coping strategies were 112 significant across participants, participants did not vary in these relationships. Consequently, the slopes for all four coping strategies were fixed. Table 12 presents the results for the H L M analyses for nighttime pain intensity. Catastrophizing and control were significantly related to mghttime pain intensity, with control being the strongest predictor. Based on the unstandardized coefficients, on average, for every unit increase in the Catastrophizing score, the predicted mghttime pain intensity score would increase by .49, and for every unit increase in the control score the pain score would decrease by .42. There was no significant linear relationship between Self-Efficacy and nighttime pain intensity across individuals. This may be the result of a suppression effect between the Self-Efficacy and control variables because the correlations between Self-Efficacy and nighttime pain intensity ranged from -.32 to -.44 for days 1, 15, and 30 of the diary monitoring (Table A and C in Appendix F). In addition, a significant negative relationship between the Self-Efficacy and nighttime pain intensity variables was identified, even after controlling for morning pain intensity, when Self-Efficacy was entered separately into the model in the preliminary analysis, b = -0.014, t = -3.24, g = .002. This significant relationship disappeared when the model was expanded to include the control variable. Although Self-efficacy did not demonstrate a significant linear relationship with mghttime pain intensity, participants did vary significantly on these relationships. This finding can be interpreted in two ways. Either the results 113 Table 12 Level-1 H L M Analysis of the Effects of Morning Pain, the Three Pain Appraisals, and the Four Coping Strategies for Nighttime Pain Intensity (N = 88) Fixed Effects Random Effects Unstandardized SE Standardized Variance Coefficient Coefficient Intercept 4.570** 0.177 2.581** Morning pain 0.248** 0.029 0.586 0.036** Self-efficacy 0.005 0.004 0.100 0.000** Catastrophizing 0.489** 0.046 0.563 fixed Control -0.422** 0.031 -0.624 fixed Distraction -0.021 0.047 -0.032 fixed Ignoring Pain -0.036 0.038 -0.055 fixed Praying and Hoping 0.272** 0.050 0.446 fixed Reinterpreting Pain Sensation 0.067 0.057 0.080 fixed Level-1 (within-variance) 1.528 Reliability of Estimates Intercept 0.807 Morning Pain 0.436 Self-efficacy 0.294 Percent Variance Explained Morning Pain Intensity: Within-participant Between-participant Morning Pain Intensity and Appraisals: Within-participant 28 Between-participant 45 Morning Pain Intensity, Appraisals and Coping Strategies: Within-participant 29 Between-participant 43 Dashes indicate data are not applicable. *P< .05 **p<01 114 are due to a suppression effect between the Self-Efficacy and control variables, or even though on average across individuals, there is no relationship between Self-Efficacy and nighttime pain intensity, some individuals do demonstrate a relationship. However, even if the later explanation was the case, the reliability estimate for Self-Efficacy is so low that it is not possible to reliably distinguish between people based on those relationships. Praying and Hoping was the only coping strategy that demonstrated a significant relationship with nighttime pain intensity. Based on the unstandardized coefficients, on average, for every unit increase in the Praying and Hoping score, nighttime pain intensity would increase by .27. Table 12 shows the estimates of the random effects, and the significance of these effects across participants. The variance of the intercept was significant, indicating that, on average, there were significant differences between participants on nighttime pain intensity after adjusting for morning pain intensity, the three pain appraisals, and the four coping strategies. The reliability estimate for the adjusted mean is quite high, and indicates that we can reliably distinguish between participants in their adjusted nighttime pain intensity mean. The partitioning of the variance in the null model (Table 5) indicated that 32% of the variance in nighttime pain intensity was within-participant and 68% was between participant. The results of this Level-1 analysis indicate that morning pain intensity explained 11% of the within-participant variance, the three pain 115 appraisals (controlling for morning pain) explained 17%, and the four coping strategies (controlling for morning pain and the appraisals) explained 1%. Forty-five percent of the between-participant variance was explained by morning pain intensity, and none by the pain appraisals. Even though the slopes for all of the coping strategies were fixed, they decreased the between-participant variance by 2%, which may be due to the inconsistency among participants in their coping strategy use, or to the inter-relationships between the appraisals and the coping strategies. Together, the three pain appraisals and the four coping strategies accounted for 29% of the within-participant variance and 43% of the between-participant variance, which is 38% of the total variance. However, almost all of this variance was attributable to morning pain intensity and the three pain appraisals. Sixty-two percent of the total variance was still left to be explained (71% within and 57% between). Further explanation of the variance for nighttime pain intensity was attempted in the Level-2 analysis. In summary, Hypothesis lc was partially supported as higher levels of pain at night were associated with higher levels of catastrophic thinking or the use of Praying and Hoping strategies, and lower levels of perceived control over pain, after controlling for morning pain intensity,. Hypothesis lc was not supported for Self-Efficacy, Distraction, or Ignoring Pain as no significant relationships were found between these variables and nighttime pain intensity. 116 Level-2 analysis. The analyses of the intercepts examined whether the Level-2 independent variables predicted the Level-1 intercept. The question being addressed was whether psychosocial and functional variables (Time 1 MPI subscales) predicted participant's average mghttime pain intensity scores adjusted for morning pain intensity, the appraisals, and coping strategy use. Specifically, for the MPI functional subscales, it was predicted that nighttime pain intensity would be negatively associated with Life Control, Support, and General Activity Level and positively associated with Interference, Affective Distress, and Pain Severity. For the MPI interpersonal subscales it was predicted that nighttime pain intensity would be positively associated with Punishing, Distracting, and Solicitous Responses. Exploratory analyses of the slopes examined whether the Level-2 independent variables changed the relationships between the independent variables and mghttime pain intensity identified at Level-1. This involved only the analyses of the, morning pain slope as, apart from Self-Efficacy, all of the other Level-1 independent variable slopes had been fixed. Self-Efficacy was allowed to be random to match the Level-1 model, but was not included in the slope analyses as the reliability of its variance was too low. Table 13 presents only the significant Level-2 effects for the MPI functional subscales. The complete results can be found in Table A in Appendix G. With regard to the analyses of the intercepts, there was a significant 117 Table 13 Level-2 H L M Results for the Nighttime Pain Intensity. The Relationships Between the MPI Functional Subscales and the Level-1 Intercept and Morning Pain SlopeLN = 88) Fixed Effects Random Effects Unstandardized Coefficient SE Variance Intercept 4.563** 0.153 1.872** Pain Intensity - am 0.246** 0.030 0.036** Self-efficacy 0.007 0.004 0.000** Catastrophizing 0.471** 0.046 fixed Control -0.425** 0.031 fixed Distraction -0.033 0.047 fixed Ignoring Pain -0.020 0.038 fixed Praying and Hoping 0.243** 0.049 fixed Reinterpreting Pain Sensation 0.081 0.057 fixed Level-1 (within variance) 1.530 Effects of MPI-level variables Pain Severity on intercept 0.858** 0.176 Affective Distress on morning pain slope 0.069* 0.031 Reliability of the Estimates Level-1 Variance Explained Intercept .765 Pain Intensity - am slope .439 Self-Efficacy .269 Within-participant 0% Between-participant 27% Note. Only statistically significant Level 2 effects are presented. See Table A in Appendix G for complete results. Dashes indicate data are not applicable. *p_<.05 **p_<.01 118 relationship between Pain Severity and the average score across participants on nighttime pain intensity. This indicates that the more severe participants' perceived their pain to be in general, the more pain they experienced on a nightly daily basis. No relationships were found for Solicitous or Distracting Responses. With regard to the analyses of the slopes, there was a significant positive relationship between Affective Distress and the morning pain intensity slope. This indicates that the relationship between morning pain and nighttime pain intensity was stronger for participants who perceived their pain to be more severe in general. However, the reliability estimate for the morning pain slope was low, indicating that it is difficult to reliably distinguish between participants based on this relationship. The MPI functional subscales did not explain any of within-participant variance for mghttime pain intensity, but did explain 27% of the between-participant variance. The reliability estimates for the nighttime pain intensity Level-2 model remained quite high, indicating that we can distinguish between participants in their adjusted Level-2 mghttime pain intensity mean. In summary, Hypotheses 2c was partially supported as participants reported higher levels of pain if they perceived their pain to be more severe overall, after taking into account their morning pain intensity, the appraisals, and the four coping strategies. Hypothesis 2c was not supported for Interference, Support, Affective Distress, Life Control, or General Activity Level as no 119 significant relationships were found between these variables and nighttime pain intensity. Table 14 shows the results of the analyses for mghttime pain intensity and the MPI interpersonal subscales. Once again, only the Level-1 intercept and morning pain slope were regressed on the Level-2 independent variables, and Self-Efficacy was allowed to be random to match the Level-1 model. The significant Level-2 effects are presented in Table 14, and the full results are presented in Table B in Appendix G. With regard to the analyses of the intercepts, Punishing Responses demonstrated a significant positive relationship with the average score across participants for nighttime pain intensity. This indicates that, on average, after taking into account Solicitous and Distracting Responses, the more a participant perceived her spouse or partner to be acting in a punishing way to her pain, the higher her level of nightly pain. There were no relationships between Solicitous or With regard to the analyses of the slopes, there was a significant negative relationship between Punishing Responses and the morning pain slope. This indicates that participants who felt that their spouses acted in a punishing way to their pain demonstrated a less consistent relationship between morning and nighttime pain. However, the reliability estimates for the morning pain slope was somewhat low, indicating that it may be difficult to reliably differentiate between participants based on these relationships. 120 Table 14 Level-2 H L M Results for the Nighttime Pain Intensity. The Relationships Between the MPI Interpersonal Subscales and the Level-1 Intercepts and Slopes (N = 88) Fixed Effects Random Effects Unstandardized Coefficient SE Variance Intercept 4.578** 0.171 2.415** Pain Intensity - am 0.248** 0.029 0.033** Self-efficacy 0.005 0.004 0.000** Catastrophizing 0.485** 0.046 fixed Control -0.426** 0.031 fixed Distraction -0.035 0.048 fixed Ignoring Pain -0.033 0.038 fixed Praying and Hoping 0.269** 0.050 fixed Reinterpreting Pain Sensation 0.067 0.056 fixed Level-1 (within variance) 1.525** Effects of MPI-level variables Punishing Responses on intercept 0.361** 0.136 Punishing Responses on AM pain slope -0.040* 0.019 Reliability of the Estimates Level-1 Variance Explained Intercept .799 Within-participant 0% Pain Intensity - am .421 Between-participant 6% Self-Efficacy .305 Note. Only significant Level 2 effects are given. See Table B in Appendix G for complete results. Dashed indicate data are not applicable. *p_<.05 **p_<.01 121 The MPI interpersonal subscales did not explain any of the within-participant variance, but did explain 6% of the between-participant variance. These findings suggest that these subscales contribute more to differences between individuals on nighttime pain intensity than to individual variation across time. Distracting Responses and nighttime pain intensity. The reliability estimates for mghttime pain intensity remained high, indicating that we can differentiate between participants in their mghttime pain intensity mean adjusted for the MPI interpersonal subscales. In summary, Hypothesis 2c was partially supported as participants who perceived their spouses to respond to their pain in a punishing way reported higher levels of pain at night, after taking into account morning pain intensity, the appraisals, and the four coping strategies. Hypothesis 2c was not supported for Solicitous or Distracting Responses as no significant relationships were found between these variables and nighttime pain intensity. Exploratory Analyses of the Interaction Effects A l l of the independent variables were standardized to facilitate interpretation of the interactions (Cohen & Cohen, 1983). The three appraisals were entered into the H L M models together to exarnine their interactions with average pain intensity for each of the dependent variables. The rationale for this was that since it is possible for a person to engage in all three appraisals at the same time (i.e., some level of Self-Efficacy, Catastrophizing, and control), their 122 relationship to each other needed to be considered in the regression equation. The interaction effects for each of the four coping strategies were examined separately based on the assumption that is was possible for a participant to engage in only one coping strategy at a time. Each model contained the morning covariate for the dependent variable, daily average pain, either the three pain appraisals or one of the four coping strategies, and the multiplicative term for either average pain times each appraisal, or average pain times each coping strategy (interaction terms). Because the independent variables were standardized, the unstandardized beta weights from the H L M analyses were used to plot the significant interaction terms for high and low values of average pain, and for the post-hoc analyses of the simple slopes in order to determine the significance of these values. The figures for the significant interactions are given in Appendix H . There were no significant interaction effects for any of the pain appraisals and average pain intensity for nighttime depressed mood, nighttime anxious mood, or nighttime pain intensity. However, two interaction effects were identified for average pain intensity and nighttime depressed mood and coping. First, there was a significant negative effect for average pain and Ignoring Pain, b = -0.26, t = 3.43, P < .001. The analyses of the simple slopes indicated that this interaction effect was only significant for people with high levels of average pain, b = 0.44, t = 6.65, P < .001. This indicated that the more frequent use of Ignoring Pain as a coping 123 strategy was associated with lower levels of nighttrme depressed mood for people with high levels of pain during the day. Second, there was a significant positive effect for average pain and Praying and Hoping, b = 0.20, t = 2.31, p < .02. The simple slopes analyses showed this effect was only significant for people with high levels of average pain, b = 0.21, t = 2.82, g < .005, although this relationship was somewhat weak. For participants with high levels of average pain, the more frequent use of Praying and Hoping was associated with higher levels of nighttime depressed mood. At low levels of average pain, there was a weak relationship, but in the opposite direction. Two interaction effects were identified for nighttime anxious mood. First, there was a significant positive effect for average pain and Praying and Hoping, b = 0.23, t = 2.54, p < .01, although this was a somewhat weak relationship. The simple slopes analysis indicated that this interaction was only significant for people with high levels of average pain, b = 0.28, t = 3.48, g < .001. When pain levels were high, the more frequent use of the Praying and Hoping strategy was associated with higher levels of nighttime anxious mood. Second, there was a significant negative interaction between average pain and Distraction, b = -0.21, t = 2.60, g < .01. This interaction was only significant for people with low levels of pain, b = -0.33, t = 4.58, g < .001. When average pain levels during the day were low, frequent use of the Distraction strategy was associated with lower levels of 124 nighttime anxious mood. The relationship was the same at high levels of pain, but weaker and non-significant. One interaction effect was identified for mghttime pain intensity. There was a significant positive effect for average pain and Reinterpreting Pain Sensation, b = 0.10, t = 2.85, p < .005. The post-hoc analyses indicated that this relationship was only significant for people with high levels of average pain, b = 0.13, t = 3.69, p < .001, although this relationship was weak. When pain levels are high, the more frequent use of the Reinterpreting Pain Sensation strategy was associated with higher levels of pain intensity at night. Whereas, at low levels of pain there was a weak relationship in the opposite direction. 125 C H A P T E R 5 Discussion The purpose of this study was to examine the relationships between daily pain appraisals (Catastrophizing, Self-Efficacy, and perceived control over pain) and coping strategy use (Distraction, Ignoring Pain, Praying and Hoping, and Reinterpreting Pain Sensation) and nighttime depressed mood, anxious mood, and pain intensity for women with chronic low back pain who were not attending a specialized chronic pain treatment program. These relationships were examined at two levels. The first level of analyses examined whether pain appraisals and coping strategy use during the day predicted levels of nighttime negative mood and pain. This analysis was based on 30 days of monitoring for each participant. The second level of analyses examined whether these daily processes could be predicted by psychosocial or functional variables important to the experience of chronic pain. This analysis was based on the MPI subscales that were completed prior to participants beginning the daily monitoring. I also examined in an exploratory manner whether average pain during the day moderated the relationships between daily appraisals or coping strategy use and nighttime negative mood or pain. First, I discuss the results for three dependent variables and the exploratory analyses of the interaction effects, and then compare these results to research 126 findings based on clinical pain populations. Then, I discuss the lhnitations of this study, and research and clinical implications. Predicting Daily Negative Mood and Pain Appraisals. The results of the within-participant analyses (individual variation across days) indicated that, after controlling for the corresponding morning score, participants who reported higher levels of depressed mood, anxious mood, or pain intensity at night also reported experiencing higher levels of Catastrophizing and lower levels of perceived control over pain during the day. These results were as predicted, and are consistent with previous research findings based on clinical pain populations (Flor & Turk, 1988; Hill , 1993; Keefe et al., 1989). Average daily pain did not moderate any of the relationships between the appraisals and mood or pain. Catastrophizing was the strongest predictor of nighttime depressed and anxious mood, whereas control was the strongest predictor of nighttime pain intensity. This difference in findings may be explained in two ways. First, Dufton (1989) found that even though people experiencing chronic pain made cognitive errors (a negatively distorted belief about oneself or one's situation) about the emotional difficulties associated with living with pain and the pain intensity, they tended to make more cognitive errors about their emotions. Thus cognitive errors such as catastrophizing would be more strongly associated with negative affect than pain intensity. Second, the measure for control was related specifically to 127 perceived control over pain, which has been found to be highly correlated with perceived ability to decrease pain intensity, r = .97, g < .0001 (Keefe et al., 1997). Participants may have been responding to the control measure largely in terms of their ability to decrease pain. Consequently, although perceived control is related to negative affect, the nature of the measure would lead it to be more strongly associated with pain intensity. The results of the within-participant analyses for the Self-Efficacy appraisal were unexpected for all three dependent variables, and contrary to previous research findings (Lorig & Holman, 1993; Lorig et al., 1989; O'Leary et al., 1988; Regan et al., 1988). When the morning covariate, the appraisals, and the coping strategies appropriate to the model were entered into the H L M analyses, higher levels of Self-Efficacy were associated with higher levels of mghttime depressed mood, and there were no significant relationships with nighttime anxious mood or nighttime pain intensity. There was some evidence that these results may have been due to a suppression effect between the Self-Efficacy and control variable, which can decrease the magnitude of the relationship between a predictor and outcome variable, or change the direction of that relationship (Cohen & Cohen, 1983). After controlling for the corresponding morning score, the within-participant variance explained by the three pain appraisals was reasonably high for each of the dependent variables. These results indicate that appraisals made a 128 substantive contribution to levels of negative mood and pain intensity, although the strength of that contribution varied according to the outcome being measured. The results of the between-participant analyses indicated that a large proportion of the variability among participants on the dependent variables could be explained by the corresponding morning score: 69% for nighttime depressed mood, 63% for nighttime anxious mood, and 45% for nighttime pain intensity. However, further examination of whether the Catastrophizing and control appraisals added to the explained between-participant variance for all of the dependent variables was limited by the strategy of "fixing" the slopes (assuming the relationships between these predictors and the dependent variables were the same for all participants). Although the slopes for Self-Efficacy were not fixed in any of the models, it did not explain any additional between-participant variance for nighttime depressed mood or pain intensity, but did increase the explained variance for nighttime anxious mood by 5%. However, the reliability estimate for this variance was quite low, indicating that, although the differences were statistically significant, one should be cautious about differentiating between participants based on their relationships between Self-Efficacy and nighttime anxious mood. Coping strategy use. The results of the within-participant analyses and the interaction analyses for the coping strategies indicated that different coping strategies were related to different components of the pain experience, and varied 129 according to different pain levels. Individuals who reported lower levels of depressed mood at night used more Distraction coping strategies during the day. The CSQ Distraction strategies involve the deliberate engagement in pleasurable activities. This positive outcome for Distraction may be explained by the cognitive-behavioural (CBT) theory of depression (Beck et al., 1961; Lewinsohn, 1974). CBT postulates that depression is both a function of cognitive errors and the reduction in general and pleasurable activities. Thus, Distraction may mitigate depression by increasing a person's participation in activities that are personally rewarding. Lower levels of nighttime depressed mood were associated with higher levels of Ignoring Pain, although this relationship held only for those participants who reported higher levels of average pain during the day. This result may be due to the fact that many of the Ignoring Pain strategies could be interpreted as an attempt to deliberately view the pain as a challenge (e.g., "I tell myself I can't let pain stand in the way of what I have to do."). Lazarus and Folkman (1984) state that this type of cognitive strategy is accompanied by emotions such as eagerness and excitement, which are incompatible with feelings of depression. It may be that, for some participants, this type of strategy is most useful when pain is particularly demanding or challenging. Although Praying and Hoping demonstrated no relationship with nighttime depressed mood across participants in the main analyses, participants who reported 130 higher levels of average pain for the day engaged more frequently in Praying and Hoping. These types of strategies focus on relying on some powerful other (e.g., God) rather than personal action to take away the pain. These results suggest that for some participants, higher levels of pain lead to a decrease in a sense of personal agency regarding the pain. A lack of personal agency has been shown to be associated with negative mood in clinical pain populations (Pellino & Oberst, 1992; Tommey et al., 1993). No relationships were found for Reinterpreting Pain Sensation and nighttime depressed mood. Individuals who reported lower levels of anxiety at night used more Distraction coping strategies during the day. It may be that engaging in pleasurable activities generates a more relaxed state of mind, which mitigates the feelings of anxiety. On average, across participants, there was no relationship between the Praying and Hoping coping strategy and nighttime anxious mood. However, for participants who experienced higher levels of average pain, increased use of Praying and Hoping was associated with increased anxiety at night. It may be that relying on powerful others is an indication of a belief that a person can do nothing about the pain themselves, which leads to increased anxiety about the pain experience. No relationships were found for Reinterpreting Pain Sensation and nighttime anxious mood. Individuals who reported higher levels of nightly pain intensity engaged more frequently in Praying and Hoping coping strategies during the day. 131 Reinterpreting Pain Sensation was unrelated to nighttime pain intensity across participants. However, for participants reporting higher levels of average pain, the increased use of Reinterpreting Pain Sensation was associated with increased mghttime pain. The reason for this difference may be due to the fact that Reinterpreting Pain Sensation was the strategy used the least by participants. Therefore, even though for most participants it did not demonstrate a relationship with nighttime negative mood or nighttime pain intensity, those participants with high levels of pain during the day used Reinterpreting Pain Sensation, but it was not effective in reducing their nighttime pain. After controlling for the corresponding morning score and the three pain appraisals, the coping strategies only added an additional 1% of the explained within-participant variance for mghttime depressed mood and nighttime pain intensity, and decreased the explained variance for nighttime anxious mood by 1%. Because the slopes for all of the coping strategies were fixed, they could not explain any additional between-participant variance. However, it may be because the relationships between the coping strategies and the three dependent variables was inconsistent across participants, the coping strategies decreased the explained between-participant variance by 2% for mghttime depressed mood and nighttime pain intensity, and by 7% for mghttime anxious mood. The relationships found between increased use of Distraction and decreased depression and anxiety, increased use of Ignoring Pain and decreased anxiety, and 132 increased use of Praying and Hoping and increased nightly pain were as predicted, and similar to those found in previous research based on clinical pain populations (Geisser, Robinson, & Henson, 1994; Jensen & Karoly, 1991; Keefe & Williams, 1990). I had also predicted that Distraction and Ignoring Pain would be positively associated with nighttime pain intensity. However, these relationships were non-significant and may be due to differences in the populations studied. Previous research had been based on clinical pain populations, whereas this study was based on a community sample. Although based on only anecdotal information, some participants reported that they found having to monitor their pain daily interfered with their ability to use distraction or ignoring pain strategies, which were strategies they generally found to be quite helpful. Therefore, it is possible that non-clinical pain populations differ from clinical pain populations in that Distraction and Ignoring Pain are actually useful strategies for managing pain intensity. The design of this study may have suppressed the positive benefits of these strategies and rendered them ineffective. In summary, the findings that pain appraisals, compared to coping strategy use, were stronger predictors of negative mood and pain are similar to findings based on clinical pain populations (Jensen, Turner, & Romano, 1994; Keefe et al., 1997; Spinhoven & Lissen, 1991). There are at least two possible explanations of why appraisals are stronger predictors of adaptational outcomes than coping strategy use. First, chronic pain, by its very definition, is difficult, intractable, 133 unrelenting, and always unpleasant (International Association for the Study of Pain, 1986). It is, therefore, possible that there is a limit to the effectiveness of coping strategies, leaving people to rely more on changing their appraisals. Second, participants varied a great deal in coping strategy use. The choice of preferred coping strategies appears to be quite individualized, and may be influenced by a number of factors not measured in this study. These results do not imply that coping strategies are not helpful in pain management. Rather, it is probably more realistic to say that that different strategies will be more effective than others for some people at some times, but not necessarily for all people at all times. The Role of Psychosocial and Functional Variables on Daily Responses to Pain The question addressed at this level of analyses was whether, after taking into account morning mood or pain, the appraisals, and coping strategy use, psychosocial or functional variables important to the chronic pain experience predicted participants' daily negative mood or pain, or the relationships between morning and nighttime scores on the dependent variables. The psychosocial and functional components were measured by the MPI subscales. These subscales were divided into functional (Interference, Life Control, Support, General Activity Level, Affective Distress, and Pain Severity) and interpersonal (Punishing, Distracting, and Solicitous Responses) groupings. 134 The results for the wimin-participant level of analyses for the MPI functional subscales indicated that participants reported less nightly depression and anxiety if they felt their spouse or partner customarily acted in a supportive manner in response to their pain, or if they themselves felt less emotionally distressed overall. Participants also reported higher levels of nightly anxiety if they perceived pain to be interfering more in their lives. Finally, participants experienced higher levels of nightly pain if they perceived their pain to be generally severe. These results are similar to studies based on clinical pain patients (Flor & Turk, 1988; Kerns, et al., 1985; Turk et al., 1995). The lack of significant findings between affective distress and nighttime pain intensity was unexpected, and the reasons for this finding are unclear. There were no relationships between Interference and nighttime depressed mood or nighttime pain intensity. These latter findings may be due to the fact that, compared to clinical pain patients, study participants reported lower levels of interference of pain on their lives in general (data taken from the MPI manual, Rudy, 1989). That is, they appear to have limited the impact of pain on their lives despite pain intensity, and thus there seems to have been far less of a relationship between pain intensity and Interference than found with clinical pain patients. There was also a more consistent relationship between morning and nighttime pain intensity for those women who experienced higher levels of overall emotional distress. That is, an emotionally distressed participant was more likely 135 than a non-emotionally distressed participant to have higher levels of pain at night if she started out with higher levels of pain in the morning. There were no relationships found for Life Control or General Activity Level, which was different from what was predicted. The non-significant results for Life Control may be due to the fact that, in this study, this subscale did not demonstrate stability across the 30 day monitoring period. Although there were no significant differences between the means for the Time 1 and Time 2 Life Control scores, the correlation between these two measures was low. This lack of stability may be due to how the questions for this subscale are phrased. Most of the questions for the MPI subscales are not time limited. However, the questions for Life Control ask respondents to base their responses on their experiences "During the past week." This relatively short time period makes the subscale more of a time-variant measure, and thus weakens its strength as a more stable, time-invariant predictor of adaptation, at least for this sample. The difference in findings for General Activity Level in this study, compared to previous studies (Geisser, Robinson, & Hensen, 1994; Regan et al., 1988), may be due to the fact that this study was based on a community sample. A comparison between the MPI functional subscales for study participants and clinical pain patient studies indicated that study participants demonstrated far higher levels of general activity, and lower levels of variability around those activity levels than clinical populations. Although based on comparisons with 136 only a limited number of clinical studies, these results suggest that participants in this study, compared to clinical pain patients, maintained more consistent and higher levels of activity over time. Consequently, this lack of variability may have decreased the utility of the General Activity Level subscale as a predictor of mood or pain for study participants. The MPI functional subscales did not explain any of the within-participant variance on any of the dependent variables. However, they did explain 9% of the between-participant variance for nighttime depressed mood, 17% for mghttime anxious mood, and 27% for nighttime pain intensity. These results suggest that the reason why people differ in their levels of negative mood and pain is partly attributable to psychosocial and functional variables important to the experience of chronic pain, although the strength of these relationships vary according to the outcome being measured. The results for the within-participant analyses for the MPI interpersonal subscales indicated that punishing spousal responses are more predictive of negative mood and pain than solicitous or distracting spousal responses. Although no relationships were found for Solicitous or Distracting Responses, participants who perceived their spouses as customarily responded to their pain in a punishing way, experienced more depression, anxiety, and pain on a nightly basis. They were also more likely than participants who did not feel punished for their pain by their spouses to have less consistency between morning and nighttime levels of 137 pain. This later finding suggests that the more participants perceived their spouses to be reacting negatively to their pain during the day, the higher the level of pain they experienced at night. The MPI interpersonal subscales did not explain within-participant variation in nighttime depressed or anxious mood, but did account for a relatively small proportion of the explained within-participant variance for mghttime pain intensity (5%). At the between-participant level, these subscales explained 5% of the variance for nighttime depressed mood, 2% for mghttime anxious mood, and 6% for nighttime pain intensity. These results suggest that differences between people in daily negative mood and pain can be partly explained by their perceptions of how their spouses or partners customarily respond to their pain, and that most of this difference is attributable to how punishing they feel their spouses are. The results for Punishing Responses are consistent with findings based on clinical pain populations (Kerns, Southwick et al., 1991; Schwartz et al., 1996). The non-significant findings for Solicitous Responses are consistent with those found by Burns et al. (1991) and Schwartz et al. (1996), but different from other studies that have found positive associations between solicitous or distracting spousal behaviours and pain intensity (Flor et al., 1987; Kerns et al., 1990; Lousberg et al., 1992; Romano et al., 1992). These contradictory results may be due to an as yet unidentified complex pattern of interrelationships that interact with Solicitous and Distracting Responses and differentially predict responses to 138 pain (Schwartz et al., 1996). For example, Jensen (1996) found that solicitous spousal responses predicted increased pain behaviour if they were given immediately following an expression of pain. Conclusions This study sought to address a number of methodological problems in previous chronic pain research. Data were collected via daily diaries, which typically yields stronger predictive validity by decreasing recall error and increasing reports of the prevalence of coping responses. A broad range of psychosocial and functional variables important to the experience of chronic pain were examined, reflecting the multidimensional nature of the phenomena of pain. Participants were recruited from the general community rather than from specialized pain clinic. Appraisals and coping strategy use were examined in the same analysis. This made the design closer to Lazarus and Folkman's (1984) model of stress and coping, and accounted for any overlapping variance between the appraisals and coping strategies. Finally, the morning score for the dependent variable was controlled for in each of the analyses. Thus, the results indicated the relationship between daily appraisals and coping strategy use and the change between morning and nighttime negative mood and pain. There were four major findings in this study. First, pain appraisals were more predictive of nighttime negative mood and pain than coping strategy use, with Catastrophizing the strongest predictor of depressed and anxious mood, and 139 perceived control over pain the strongest predictor of pain intensity. Second, overall emotional distress predicted higher levels of depressed and anxious mood on a daily basis. Third, the more a participant perceived her pain to be interfering in her life in general, the more anxious she was on a daily basis. Fourth, perceived punishing spousal responses predicted negative mood and pain more than solicitous or distracting spousal responses . These results are similar to findings based on clinical pain populations, even though study participants, compared to clinical pain populations, were functioning at a higher level and were less affectively distressed. Thus, although definitive statements can not be made until these results are replicated, they suggest that some of the same processes identified in clinical pain patients may apply to low back pain sufferers in the community who are similar to study participants. Exploratory analyses indicated that the level of pain during the day moderated the relationships between the coping strategies and the dependent variables, but not the relationships between the appraisals and the dependent variables. This result is probably due to the fact that participants were more inconsistent in coping strategy use than appraisals. Limitations of the Study This study had several limitations. A principle limitation is the demographic characteristics of the participants. First, only women with a spouse or partner were included in the study. Second, as in most clinical pain studies, 140 participants were mostly Caucasian, middle-aged, and middle class. Third, the women were all from a metropolitan area. Fourth, the study was limited to people with low back pain. It is possible that the results may have been somewhat different if participants were dissimilar to participants involved in this study, or if they suffered from other types of chronic pain. A second limitation is that the analyses in this study were correlational, and the daily variables of catastrophizing, control, the four coping strategies, and nighttime mood and pain were measured at the same point in time. Therefore, causality can not be inferred from these relationships. For example, even though increased use of Distraction was associated with lower levels of mghttime anxious mood, it does not necessarily indicate that this coping strategy causes anxiety to decrease. It may be just as plausible that high levels of anxiety make it harder for people to use distraction as a coping strategy. A third limitation is that the 12 women who dropped out of the study were more affectively distressed and felt less control over their lives than those women who completed the study. This may limit the generalizability of the results to community pain sufferers who are similar to study participants. A fourth limitation concerns the choice of the outcome variables. Clinicians in pain treatment programs focus not only on interventions that target negative adaptational responses to pain, but also on interventions that can have a beneficial effect on positive outcomes, regardless of their effects on reducing pain 141 or emotional distress. Because this study focused only on negative moods and pain intensity, the potential beneficial effects of appraisals and coping strategies on more positive mood states were not identified. A fifth limitation concerns the design of the study. Although daily monitoring allows investigators to measure closer to "real time" the variables of interest, participants in this study were asked to recall their appraisals and coping strategy use over approximately a 9 hour time period. Thus, there may have been some recall error or bias, or situationally specific confounds. A sixth limitation concerns the reactive effects of diary monitoring. A potential drawback of this type of data collection is that it causes participants to attend to their pain more than they otherwise might. This heightened attention might potentially change a person's experience of the pain, compromising the validity of the diaries as a measurement tool. However, there is some indication that may not be of as much concern as previously thought, at least for pain and mood. First, Cruise, Broderick, Porter, Kaell, and Stone (1996) addressed this issue in their study of 35 rheumatoid arthritis patients. These patients completed diaries of pain and mood seven times a day for one week. The authors found no significant time effects for any of the measures. They suggest that chronic pain patients may be so accustomed to their pain that attending to it exerts little, if any, influence in their experience of pain or mood. The results of the Cruise et al. study are supported by von Baeyer (1994) who failed to find reactive effects for 142 patients who rated their chronic low back pain once a day for 8 days. Second, participants were asked at the end of the study whether they thought their participation had in any way changed the way they thought about their pain or the types of coping strategies they engaged in. A number of participants said that even though some of the strategies were new to them, they did not want to try them out during the study as they thought this might bias the results. However, some participants did report that the daily monitoring made it more difficult for them to ignore the pain or distract themselves from it, which were strategies they normally found quite helpful. Therefore, reactive effects may have been a problem in this study in regards to these two coping strategies. A seventh limitation relates to the use of H L M for the analyses. First, H L M is a relatively new method of statistical analyses, and research on its robustness for violations of its assumptions has been limited. Second, the strategy of fixing the slopes for all of the coping strategies and most of the appraisals produced average effects for these predictors, which potentially masked important individual differences in the daily processes. Third, even though one can test the overall significant difference in variance between two models, H L M does not provide a test of whether there is significant change at the within- and between-participant variance levels. Finally, data was collected via self-report questionnaires, which may have been influenced by person or situational variables not measured in this study. 143 These potential confounds can not be ruled out, although there seems no obvious advantage to respondents deliberately completing the questionnaires with a self-presentational bias (cf. Jensen, 1996). Research Implications As far as I am aware, this is the first study to examine the relationships between pain appraisals and coping strategy use and negative mood and pain in a non-clinical pain population. Thus, there is a need to see if these results can be replicated in similar and dissimilar non-clinical pain populations. There is also a need to determine the effects of gender on the pattern of findings obtained in this study. The analyses of gender differences in response to pain is complex, and involves not just an analysis of male and female mean differences on study variables, but also the consideration of factors such as the moderating effects of situational contextual factors that may influence these responses, such as socioeconomic status, power imbalances, and different familial demands and responsibilities. Coping strategy use was not consistent among participants in this study. Therefore, more research is needed into what factors influence people to choose certain pain coping strategies over others, and what determines the frequency of the use of these strategies. Although researchers in stress and coping have identified that certain factors such as neuroticism (Bolger & Zuckerman, 1995) and familial demands (Moen & Dempster-McClain, 1987) moderate coping 144 choices, to my knowledge there has been no research in this area in the chronic pain field. This type of information could help clinicians design treatment programs that enhance beneficial coping choices. One of the questions participants were asked after they finished their diary monitoring was whether they felt there were certain factors important to their experience of pain that had not been considered in the study. Many participants replied that they wished they could have recorded their levels of fatigue and anger, as they felt these influenced how they responded to pain on a daily basis. Because the examination of how these two factors relate to peoples' pain experience and adaptational outcomes has been neglected in chronic pain research, further study is warranted. At the pre-study interview, a number of participants talked about the struggle they had experienced in coming to accept their pain problem, and how once they did accept it, they felt they were able to live with the pain more effectively even though their pain levels did not necessarily change. Acceptance of the pain is also an attitude that clinicians try to foster in pain treatment programs. However, questions arise as to what exactly does "acceptance of the pain" mean, what are the indicators of acceptance, how do people who have accepted their pain differ in their adaptation to pain from those who have not accepted it, what factors enhance or impede peoples' ability to arrive at this 145 acceptance, and how can clinicians foster this process? These are important research questions that have been neglected in pain research to date. Finally, the results suggest that coping strategy use varies according to the level of pain intensity. Further research is needed to help us understand the processes involved in these types of interactions. Clinical Implications The findings of this study suggest a number of treatment strategies. First, although cognitive-behavioural pain management programs incorporate at least one or more sessions on changing negative thought patterns, this research suggests that the instruction for all pain management strategies should include a component on how to recognize and change appraisals that may interfere with the client's ability to implement that strategy. For example, if a graduated exercise program is the strategy being taught, then appraisals concerning the ability to engage successfully in that strategy, catastrophic thoughts about increased pain or re-injury, or perceptions regarding short- and long-term benefits should be explored with the client. Through this process, clients can be taught throughout the treatment program how to recognize potentially detrimental appraisals, and how to decrease their influence. Another implication of these findings is that, because we do not understand why people prefer certain coping strategies over others, or how they determine how frequently to use a certain strategy, a parsimonious approach to treatment 146 would be to teach a wide variety of coping strategies. This would provide clients with numerous coping options from which they could chose their individual "favourites." Finally, this research suggests that marital therapy may be an important pain treatment component for clients who perceive their interactions with their spouse or partner to be negative and hostile. These types of interactions may create a feedback cycle, where negative responses produce increased pain intensity (or visa versa), which precipitate further negative responses, which in turn leads to increased pain, and so on. Unless this cycle is broken, this system may perpetuate itself, leading to entrenchment in the patterns of responses and further distress. In summary, the strengths of this study include a strong theoretical basis, a daily diary data collection research design, and the application of a relatively new methodology for studying associations between pain -related appraisals and coping and adjustment to pain. Interesting results emerged which make an important contribution to our understanding of appraisals and coping, their relation to adjustment to chronic pain, and potential treatment implications. Many findings were consistent with previous research even though the participants in this study were not clinical pain patients. Further research is needed to establish whether these findings can be replicated in similar and dissimilar pain populations in order to determine whether the relationships identified in this study are generalizable across different pain populations. 147 R E F E R E N C E S Affleck, G. , Tennen, H . , Pfeiffer, C , & Fifield, J. (1987). Appraisals of control and predictability in adapting to a chronic disease. Journal of Personality and Social Psychology. 53. 237-239. Affleck, G. , Urrows, S., Tennen, H . , & Higgins, P. (1992). Daily coping with pain from rheumatoid arthritis: Patterns and correlates. Pain. 51. 221-230. Anderson, K. O., Dowds, B. N . , Pelletz, R. E . , Edwards, W. T. , & Peeters-Asdourian, C. (1995). Development and validation of a scale to measure self-efficacy beliefs in patients with chronic pain. Pain. 63, 77-84. Bandura, A . (1997). Self-Efficacy: The exercise of control. New York, N Y : W. H . Freeman. Bandura, A . (1986). Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ: Prentice-Hall. Beck, A . T., Ward, C. H . , Mendelson, M . , Mock, J., & Erbaugh, J. (1961). A n inventory for measuring depression. Archives of General Psychiatry, 4, 53-63. Ben-Porath, Y. S., & Tellegen, A . (1990). A place for traits in stress research. Psychological Inquiry, 1, 14-40. Bolger, N. , DeLongis, A . , Kessler, R. C , & Schilling, E . A . (1989). Effects of daily stress on negative mood. Journal of Personality and Social Psychology, 57, 808-818. Bolger, N . , & Zuckerman, A . (1995). A framework for studying personality in the stress process. Journal of Personality and Social Psychology, 69,890-902. Bryk, A . S., & Raudenbush, S. W. (1992). Hierarchical linear models. Newbury Park, C A : Sage. Bryk, A . S., & Raudenbush, S. W. (1987). Application of hierarchical linear models to assessing change. Psychological Bulletin, 101, 147-158. Burns, J. W., Johnson, B. J., Mahoney, N . , Devine, J., & Pawl, R. (1996) Anger management style, hostility and spouse responses: Gender differences in predictors of adjustment among chronic pain patients. Pain, 64, 445-453. 148 Chapman, D. R., Sola, A . E . , & Bonica, J. J. (1979). Illness behavior and depression compared in pain center and private practice patients. Pain, 6, 255-268. Cohen, F. (1987). Measurement of coping. In S. V . Kasl & C. L . Cooper (Eds.), Stress and health: Issues in research methodology (pp. 283-305). New York: John Wiley & Sons. Cohen, J., & Cohen, P. (1983). Applied multiple regression/correlation analysis for the behavioral sciences (second ed.). Hillsdale, N.J.: Lawrence Erlbaum. Costa, P. T., & McCrae, R. R. (1990). Personality: Another "Hidden Factor" in stress research. Psychological Inquiry, 1, 22-24. Council, J. R., Ahern, D. K . , Follick, M . J., & Kline, C. L . (1988). Expectancies and functional impairment in chronic low back pain. Pain, 33, 323-331. Crook, J., Rideout, E . , & Browne, G. (1984). The prevalence of pain complaints in a general population. Pain, 18, 299-314. Cruise, C. E . , Broderick, J., Porter, L . , Kaell, A . , & Stone, A . A. . (1996). Reactive effects of diary self-assessment in chronic pain patients. Pain. 67. 253-258. De Benedittis, G . , Lorenzetti, A . , & Pieri, A . (1990). The role of stressful life events in the onset of chronic primary headache. Pain, 40, 65-75. DeGood, D. E . , & Shutty, M . S. (1992). Assessment of pain beliefs, coping, and self-efficacy. In D. C. Turk & R. Melzack (Eds.), Handbook of pain assessment (pp. 214-234). New York: Guilford Press. DeLongis, A . , Folkman, S., & Lazarus, R. S. (1988). The impact of daily stress on health and mood: Psychological and social resources as mediators. Journal of Personality and Social Psychology, 54, 486-495. DeLongis, A . D. , Hempill, K . J., & Lehman, D. R. (1992). A structured diary methodology for the study of daily events. In F. B. Bryant, J. Edwards, R. S. Tindale, E . J. Posavac, L . Heath, E . Henderson, & Y. Suarez-Balcazar (Eds.), Methodological issues in applied social psychology (pp. 83-109). New York: Plenum. 149 Dolce, J. J., Crocker, M . F., & Doleys, D. M . (1986). Prediction of outcome among chronic pain patients. Behavior Research and Therapy. 24, 313-319. Dolce, J. J., Crocker, M . F., Moletteire, C , & Doleys, D. M . (1986). Exercise quotas, anticipatory concern and self-efficacy expectancies in chronic pain: A preliminary report. Pain, 24, 365-372. Dufton, B. D. (1989). Cognitive failure and chronic pain. International Journal of Psychiatry in Medicine. 19, 291-297. Eich, E . , Reeves, J. L . , Jaeger, B., & Graff-Radford, S. B. (1985). Memory for pain: Relation between past and present pain intensity. Pain, 23, 375-379. Epstein, S. (1986). Does aggregation produce spuriously high estimates of behavior stability? Journal of Personality and Social Psychology, 50, 1199-1210. Epstein, S. (1983). Aggregation and beyond: Some basic issues on the prediction of behavior. Journal of Personality, 51, 360-392. Estlander, A . M . , & Harkapaa, K . (1989). Relationship between coping strategies, disability, and pain levels in patients with chronic low back pain. Scandinavian Journal of Behaviour Therapy. 18, 59-69. Feuerstein, M . , Suit, S., & Houle, M . (1985). Environmental stressors and chronic low back pain: Life events, family and work environment. Pain. 22, 295-307. Flor, H . , & Turk, D. C. (1988). Chronic back pain and rheumatoid arthritis: Predicting pain and disability from cognitive variables. Journal of Behavioral Medicine. 11. 251-265. Flor, H . , Turk, D. C , & Birnaumer, N . (1985). Assessment of stress-related psychophysiological reactions in chronic back pain patients. Journal of Consulting and Clinical Psychology. 53. 354-364. Flor, H . , Turk, D. C , & Rudy, T. E . (1989). Relationship of pain impact and significant other reinforcement of pain behaviors: The mediating role of gender, marital status and marital satisfaction. Pain, 38, 45-50. Folkman, S. (1992). Making the case for coping. In B. N . Carpenter (Ed.), Personal coping (pp. 31-46). Westport, CT: Prager. 150 Geisser, M . E . , Robinson, M . E . , & Henson, C. D. (1994). The Coping Strategies Questionnaire and chronic pain adjustment: A conceptual and empirical reanalysis. The Clinical Journal of Pain. 10, 98-106. Geisser, M . E . , Robinson, M . E . , Keefe, F. J., & Weiner, M . L . (1994). Catastrophizing, depression and the sensory, affective and evaluative aspects of chronic pain. Pain, 59, 79-83. Haaga, D. A . F. (1992). Catastrophizing, confounds, and depression: A comment on Sullivan and D'Eon (1990). Journal of Abnormal Psychology. 101. 206-207. Hadjistavropoulos, H . , & Craig, K . D. (1994). Acute and chronic low back pain: Cognitive, affective, and behavioral dimensions. Journal of Consulting and Clinical Psychology. 62. 341-349. Hill , A . (1993). The use of pain coping strategies by patients with phantom limb pain. Pain, 55, 347-354. Holm, J. E . , Holroyd, K . A . , Hursey, K. G. , & Penzien, M . S. (1986). The role of stress in recurrent tension headache. Headache, 26, 160-167. Jensen, M . P. (1996, August). Validity of self-report and observational measures. Paper presented at the 8 t h World Congress on Pain, Vancouver, B C . Jensen, M . P., & Karoly, P. (1992). Self-report scales and procedures for assessing pain in adults. In D. C. Turk & R. Melzack (Eds.), Handbook of pain assessment (pp. 135-151). New York: Guilford Press. Jensen, M . P., & Karoly, P. (1991). Control beliefs, coping efforts, and adjustment to chronic pain. Journal of Consulting and Clinical Psychology, 59, 431-438. Jensen, M . P., Karoly, P., & Braver, S. (1986). The measurement of clinical pain intensity: A comparison of six methods. Pain, 27, 117-126. Jensen, M . P., Karoly, P., & Huger, R. (1987). The development and preliminary validation of an instrument to assess patients' attitudes toward pain. Journal of Psychosomatic Research, 31, 393-400. Jensen, M . P., & McFarland, C. A . (1993). Increasing the reliability and validity of pain intensity measurement in chronic pain patients. Pain, 55, 195-204. 151 Jensen, M . P., Turner, J. A . , & Romano, J. M . (1994). Correlates of improvement in multidisciplinary treatment of chronic pain. Journal of Consulting and Clinical Psychology. 62. 172-179. Jensen, M . P., Turner, J. A . , & Romano, J. M . (1991). Self-efficacy and outcome expectancies: Relationship to chronic pain coping strategies and adjustment. Pain. 44. 263-270. Jensen, M . P., Turner,. J. A . , Romano, J. M . , & Karoly, P. (1991). Coping with chronic pain: A critical review of the literature. Pain, 47, 249-284. Keefe, J. K. , Affleck, G . , Lefebvre, J. C , Starr, K. , Caldwell, D. S., Tennen, H . (1997). Pain coping strategies and coping efficacy in rheumatoid arthritis: A daily process analysis. Pain, 69, 35-42. Keefe, F. J., Brown, G. K. , Wallston, K. A . , & Caldwell, D. S. (1989). Coping with rheumatoid arthritis pain: Catastrophizing as a maladaptive strategy. Pain. 37. 51-56. Keefe, F. J., Caldwell, D. S., Martinez, S. M . , Beckham, J., & Williams, D. (1991). Analyzing pain in rheumatoid arthritis patients. Pain coping strategies in patients who have had knee replacement surgery. Pain, 46, 153-160. Keefe, F. J., Crisson, B. J., Urban, B. J., & Williams, D. A . (1990). Analyzing chronic low back pain: The relative contribution of pain coping strategies. Pain, 40, 293-302. Keefe, F. J., Caldwell, D. S., Williams, D. A . , Gil , K . M . , Mitchell, D. , Robertson, C , Martinez, S., Nunley, J., Beckham, J. C , Crisson, J. E . , & Helms, M . (1990). Pain coping skills haining in the management of osteoarthritc knee pain: A comparative study. Behavior Therapy, 21, 49-62. Keefe, F. J., Caldwell, D. S., Queen, K. T., Gil , K. M . , Martinez, S., Crisson, J. E . , Ogden, W., & Nunley, J. (1987). Pain coping strategies in osteoarthritis patients. Journal of Consulting and Clinical Psychology. 55, 208-212. Keefe, F. J., & Dolan, E . (1986). Pain behavior and pain coping strategies in low back pain and myofascial pain dysfunction syndrome patients. Pain, 24, 49-56. 152 Keefe, F. J., Schapira, B. , Williams, R. B. , Brown, C , & Surwit, R. S. (1981). EMG-assisted relaxation framing in the management of chronic low back pain. American Journal of Clinical Biofeedback. 4, 93-103. Keefe, F. J., & Williams, D. A . (1990). A comparison of cognitive coping strategies in chronic pain patients in different age groups. Journal of Gerontology. 45, 161-165. Kerns, R. D., Haythornthwaite, J., Southwick, S., & Giller, E . L . (1990). The role of marital interaction in chronic pain and depressive symptom severity. Journal of Psychosomatic Research, 34, 401-408. Kerns, R.D. , Southwick, S., Giller, E . L . , Haythornwaite, J. A . , Kacob, M . C . , & Rosenberg, R. (1991). The relationship between reports of pain related social interactions and expressions of pain and affective distress. Behavior Therapy. 22, 101-111. Kerns, R. D. , Turk, D. C , & Rudy, T. E . (1985). The West-Haven Yale Multidimensional Pain Inventory (WHYMPI). Pain. 23. 345-356. Lazarus, R. S. (1990). Response article. Psychological Inquiry, 1, 41-51. Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping. New York: Springer. Lewinsohn, P. M . (1974). Clinical and theoretical aspects of depression. In K. S. Calhoun, H . E . Adams, & K. M . Mitchel (Eds.), Innovative treatment methods for psychopathology (pp. 201-217). New York: John Wiley & Sons. Linton, S. J., & Melin, L . (1982). The accuracy of remembering chronic pain. Pain. 13. 281-285. Lorig, K . L . , Chastain, R. L . , Ung, E . , Shoor, S., & Holman, H . R. (1989). Development and evaluation of a scale to measure perceived self-efficacy in people with arthritis. Arthritis and Rheumatism. 32. 37-44. Lorig, K. , & Holman, H . (1993). Arthritis self-management studies: A twelve year review. Health Education Quarterly. 20(1). 17-28. Lousberg, R., Schmidt, A . J. M . , & Groenman, N . A . (1992). The relationship between spouse solicitous and pain behavior: Searching for more experimental evidence. Pain, 51. 75-79. 153 Love, A . , Loeboeuf, D. C , & Crisp, T. C. (1989). Chiropractic chronic low back pain sufferers and self-report assessment methods. Part 1. A reliability study of the Visual Analogue Scale, the pain drawing, and the McGil l Pain Questionnaire. Journal of Manipulative and Physiological Therapeutics. 12. 21-25. Lubin, B. (1965). Adjective Check List for the measurement of depression. Archives of General Psychiatry. 12. 57-67. Main, C. J., & Waddell, G. (1991). A comparison of cognitive measures in low back pain: Statistical structure and clinical validity at initial assessment. Pain. 46, 287-298. McCracken, L . M . , & Gross, R. T. (1993). Does anxiety affect coping with chronic pain? The Clinical Journal of Pain, 9, 253-259. McCreary, C , Turner, J., & Dawson, E . (1981). Principal dimensions of the pain experience and physiological disturbance in chronic low back pain patients. Pain. 11. 85-92. Melzack, R. (1975). The McGi l l Pain Questionnaire: Major properties and scoring methods. Pain, 1, 277-299. Melzack, R., & Torgerson, W. S. (1971). On the language of pain. Anesthesiology, 34, 50-59. Moem, P., & Dempster-McClain, D (1987). Employed parents: Role strain, work time, and preference for working less. Journal of Marriage and the Family, 49, 579-590. Nachemson, A . L . (1984). The prevention of chronic back pain - challenge for the 80's. Bulletin of the Hospital Joint Diseases. 44. 1-10. O'Leary, A . , Shoor, S., Lorig, K. , & Holman, H.'R. (1988). A cognitive-behavioral treatment for rheumatoid arthritis. Health Psychology, 7, 527-544. Pellino, T. A . , & Oberst, M . T. (1992). Perception of control of illness in chronic low back pain. Orthopaedic Nursing. 11, 22-26. Ritterband, L . M . (1995). Evaluation of the Beck Depression Inventory's sensitivity and state-trait properties. Unpublished master's thesis, University of South Florida, Tampa, F L . 154 Reading, A . E . (1989). Testing pain mechanisms in persons in pain. In P. D. Wall & R. Melzack (Eds.), The textbook of pain (2 n d ed., pp. 269-280). Edinburgh: Livingstone Churchill. Reading, A . E . (1979). The internal structure of the McGi l l Pain Questionnaire in dysmenorrhea patients. Pain. 7, 353-358. Reesor, K . A . , & Craig, K. D. (1988). Medically incongruent chronic back pain: Physical limitations, suffering, and ineffective coping. Pain, 32, 35-45. Regan, C. A . , Lorig, K. , & Thorensen, C. E . (1988). Arthritis appraisal and ways of coping: Scale development. Arthritis Care and Research, 3, 139-150. Romano, J. M . , Turner, J. A . , Syrjala, K. L . , & Levy, R. L . (1987, August). Coping strategies of chronic pain patients: Relationship to patient characteristics and functioning. Poster session presented at the International Association for the Study of Pain Fifth World Congress on Pain, Hamburg, Germany. Rosentiel, A . K . , & Keefe, F. J. (1983). The use of coping strategies in chronic low back pain patients: Relationship to patient characteristics and current adjustment. Pain, 17. 33-44. Spielberger, C. D. (1992). State-Trait Anxiety Inventory (Form Y). In M . Hersen & A . S. Bellack (Eds.), Dictionary of behavioral assessment techniques (pp. 448-450). New York: Pergamon Press. Spielberger, C D . (1984). The State-Trait Anxiety Inventory: A comprehensive bibliography. Palo Alto, C A : Consulting Psychologists Press. Spielberger, C. D. (1983). Manual for the State-Trait Anxiety Inventory (Revised Ed.). Palo Alto, C A : Consulting Psychologists Press. Spielberger, C. D. , & Sydeman, S. J. (1994). State-Trait Anxiety Inventory and State-Trait Anger Expression Inventory. In M . E . Maruish (Ed.), The use of psychological tests for treatment plarming and assessment (pp. 292-320). Hillsdale, NJ: L E A . Spinhoven, P., & Linssen, A . C. G. (1991). Behavioral treatment of chronic low back pain. I. Relation of coping strategy use to outcome. Pain, 45, 29-34. 155 Spinhoven, P., Ter Kuile, M . M . , Linssen, A . C. G . , & Gazendam, B. (1989). Pain coping strategies in a Dutch population of chronic low back pain patients. Pain. 37. 77-84. Sternbach, R. A . (1986). Pain and "Hassles' in the United States: Findings of the Nuprin Pain Report. Pain, 27. 69-80. Stone, A . A . , & Kennedy-Moore, E . (1992). Assessing situational coping: Conceptual and methodological considerations. In H . S. Friedman (Ed.), Hostility coping and health (pp. 203-214). Washington, D C : American Psychological Association. Stone, A. S., Kennedy-Moore, E . , Newman, M . G. , Greenberg, M . , & Neale, J. M . (1992). Conceptual and methodological issues in current coping assessments. In B. N . Carpenter (Ed.), Personal coping. Theory, research, and application (pp. 16-29). Westport, C T : Prager. Stone, A . A . , & Neale, J. M . (1984). New measures of daily coping: Development and preliminary results. Journal of Personality and Social Psychology. 46. 892-906. Strecher, V . J., DeVellis, B. M . , Becker, M . , & Rosenstock, I. M . (1986). The role of self-efficacy in achieving health behavior change. Health Education Quarterly. 130). 79-91. Sullivan, M . J. L . , & D'Eon, J. J. (1990). Relation between catastrophizing and depression in chronic pain patients. Journal of Abnormal Psychology, 99. 260-263. Suls, J., Wan, C. K. , & Blanchard, E . B. (1994). A multilevel data-analytic approach for evaluation of relationships between daily life stressors and symptomatology: Patients with Irritable Bowel Syndrome. Health Psychology. 13. 103-113. Swartzman, L . C , Gwadry, F. G . , Shapiro, A . P., & Teasell, R. W. (1994). The factor structure of the Coping Strategies Questionnaire. Pain, 57. 311-316. Tabachnick, B. G . , & Fidell, L . S. (1989). Using multivariate statistics (2nd ed.). New York: Harper Collins. Tennen, H . , & Affleck, G. (1996). Daily processes in coping with chronic pain: Methods and analytic strategies. In M . Zeidner & N . Endler (Eds.), Handbook of coping (pp. 151-180). New York: Wiley. 156 Toomey, T. C , Mann, J. D., Abahsian, S. W., Camrike, C. L . M . , & Hernandez, J. T. (1993). Pain locus of control in chronic pain patients and medical clinic patients with and without pain. The Clinical Journal of Pain, 9, 242-247. Turk, D. C , & Holzman, A . D. (1986). Chronic pain: Interfaces among physical, psychological, and social parameters. In A . D. Holzman & D. C. Turk (Eds.), Pain management. A handbook of psychological treatment approaches (pp. 1-9). New York: Pergamon Press. Turk, D. C , Okifuji, A . , & Scharff, L . (1995). Chronic pain and depression: Role of perceived impact and perceived control in different age cohorts. Pain. 61, 93-101. Turk, D. C , & Rudy, T. E . (1992). Classification logic and strategies in chronic pain. In D. C. Turk & R. Melzack (Eds.), Handbook of pain assessment (pp. 409-428). New York: Guilford Press. Turk, D. C , & Rudy, T. E . (1988). Toward an empirically derived taxonomy of chronic pain patients: Integration of psychological assessment data. Journal of Consulting and Clinical Psychology, 56, 233-238. Turk, D. C , Rudy, T. E . , & Salovey, P. (1985). The McGi l l Pain Questionnaire reconsidered: Corifirming the factor structures and examining appropriate uses. Pain. 21. 385-397. Turk, D. C , Wack, J. T., & Kerns, R. D. (1985). A n empirical examination of the "pain-behavior" construct. Journal of Behavioral Medicine, 8, 119-130. Turner, J. A . (1991). Coping and chronic pain. In M . R. Bond, J. E . Charlton, & C. J. Woolf (Eds.), Proceedings of the Vlth World Congress on pain (pp. 219-227). Amsterdam: Elsevier Science. Turner, J. A . , & Clancy, S. (1986). Strategies for coping with chronic low back pain: Relationship to pain and disability. Pain, 24, 355-364. Turner, J. A . , Clancy, S., & Vitaliano, P. P. (1987). Relationship of stress, appraisal and coping, to chronic low back pain. Behavior Research and Therapy, 25, 281-288. 157 Schwartz, L . , Slater, M . A . , & Birchler, G. R. (1996). The role of pain behaviors in the modulation of marital conflict in chronic pain patients. Pain, 65, 227-233. Von baeyer, C. L . (1994). Reactive effects of measurement pain. Clinical Journal of Pain. 10. 18-21. Verbrugge, L . M . (1989). Health diaries - Problems and solutions in study design. Health survey research methods. Proceedings of the 5th. Conference of Health Survey Research Methods, Keystone. Colorado. 171-192. Verbrugge, L . M . (1980). Health diaries. Medical Care, 18. 73-95. Wallston, K. A . , & Wallston, B. S. (1978). Development of Multidimensional Health Locus of Control ( M H L C ) Scales. Health Education Monographs. 6. 160-170. Wethington, E . , & Kessler, R. C. (1993). Neglected methodological issues in employment stress. In B. C. Long & S. E . Kahn (Eds.), Women, work, and coping (pp. 269-295). Montreal: McGill-Queen's University Press. Zung, W. W. K. (1967). A self-rating depression scale. Archives of General Psychiatry, 12, 63-70. A P P E N D I X A Summary Tables of the Literature Review Table A Studies That Have Examined the Relationship Between Pain Appraisals and Adaptation to Chronic Pain For abbreviations see page 168 Study Sample Mean Pain Pain Type Appraisal Appraisal Duration Measure Criterion Measures Results Affleck et al., 1987 92 patients from 9.9 yrs. rheumatology practices 66% females Rheumatoid Control arthritis Control over symptoms, disease course, and treatment Health Care Provider Control over symptoms, illness course IUI POMS Controlling for age, education, HAQ income, employment, illness duration, symptom activity, functional problems, and disease severity: a) perceived control over current symptoms correlated positively with mood for patients with moderate and severe symptoms b) perceived control over disease course correlated negatively with mood and psychosocial adjustment for those with severe symptoms c) perceived control over treatment correlated positively with mood and global adjustment >0 Table A (continued) Study Sample Mean Pain Pain Type Appraisal Appraisal Criterion Results Duration Measure Measures Affleck et al, 1992 Council et al., 1988 75 patients from rheumatology practices 71% females 8.9. yrs. Rheumatoid Control arthritis 40 patients 45.5 mths. attending a pain clinic 50% females 90% low back pain Self-efficacy CSQ -perceived control and ability to decrease pain scales MAPPS DCI a) perceived control Joint pain positively associated with the POMS use of relaxation as a coping NEOPI strategy ATMS b) perceived control negatively associated with Emotional Expression as a coping strategy SIP Self-efficacy correlated Pain positively with motor intensity performance on 7/10 specific Pain physical movements, and behaviour negatively with observed pain Videotape of behaviour during movement movement performance Dolce, Crocker, & Doleys, 1986 63 patients attending a pain management program 54% females 38.1 mths. Mixed Self-efficacy Self-efficacy and concern regarding exercise, work, and medication free coping Medications a) composite of self-efficacy Work status ratings positively associated Exercise with post-treatment work Pain status and exercise, and intensity negatively with post-treatment MMPI medication use BDI b) controlling for depression, MMPI subscales, and pre-treatment pain ratings, self-efficacy for ability to work positively associated with post-treatment work status O Table A (continued) Study Sample Mean Pain Pain Type Appraisal Appraisal Criterion Results Duration Measure Measures Dolce, Crocker, Moletteire, & Doleys, 1986 14 patients 37.2 mths. attending a pain clinic 57% females 86% back pain Self-efficacy Self-efficacy and concern regarding exercise Exercise As self-efficacy beliefs tolerance regarding ability to engage in exercise increased, exercise performance increased and concern regarding exercise decreased over the course of treatment Estlander & Harkapaa, 1989 Flor & Turk, 1988 104 patients from an in-patient a pain program 57% females 70 patients recruited from a pain clinic and community advertisements 80% females 11.3 yrs. Low back Catastrophizing CSQ range 10 13 yrs. Low back n=30 Rheumatoid Arthritis n=40 Catastrophizing PRSS PRCS Disability Patients engaged in questionn- statistically higher levels of aire catastrophizing when Pain experiencing increased pain frequency intensity Pain intensity Pain a) For back pain patients intensity controlling for pain duration MPQ-PRI and X-ray findings, Pain Catastrophizing was positively interference associated with pain intensity, Physician interference activities, and visits physician visits MPI b) For RA patients, AIMS controlling for RA duration, X-ray findings, disease activity and surgeries, Catastrophizing was positively associated with pain, life interference, and disability Table A (continued) Study Sample Mean Pain Pain Type Appraisal Appraisal Criterion Results Duration Measure Measures Geisser, Robinson, Keefe, & Weiner, 1994 Hadjistavropoulos & Craig, 1994 76 patients attending an out-patient pain clinic 47% females 7.8. yrs. 90 patients attending a physiotherapy clinic 50% females 30% low back Catastrophizing CSQ Control > 3 mths. Back Catastrophizing CSQ MPI BDI MPQ PII OLBPDI PES Facial Measures a) Adaptive Copers engaged in less catastrophizing than Dysfunctional patients, even when controlling for depression b) Adaptive Copers demonstrated significantly higher levels of perceived control over pain than Dysfunctional patients Catastrophizing positively associated with Medically Incongruent Back Pain Hill, 1993 90 patients attending an out-patient clinic 100% males 8.4 yrs. Phantom limb Catastrophizing CSQ GHQ Catastrophizing accounted for MPQ 25% of the variance in pain Medication report and 26% of the use variance in psychological Pain triggers distress Jensen & Karoly, 1991 118 patients who had completed an in-patient pain program in the last 7 yrs. 10.96 yrs. Mixed Control SOPA CSQ CES-D Controlling for pain severity: SWLS a) Belief in control over pain HAQ accounted for 11% of the MPI variance in psychological Medication functioning Table A (continued) Study Sample Mean Pain Pain Type Appraisal Appraisal Criterion Results Duration Measure Measures 68% females Professional b) Interaction effects found for service perceived control over pain utilization and activity levels for patients reporting lower levels of pain Jensen et al., 1987 33 patients attending an in-patient pain program 55% females 5.75 yrs. Mixed Control SOPA Hrs. spent working Hrs. spent at chores Uptime Medication use Rest Relaxation Exercise Perceived control over pain positively associated with the use of relaxation at admission, and for males, post-treatment exercise Jensen, Turner, & Romano, 1991 114 patients attending a pain clinic 54% females 6.9 yrs. Mixed Self-efficacy Self-efficacy, immediate outcome and long term outcome expectancies regarding coping Patient rated Controlling for pain severity use of coping and outcome expectancies, strategies expectancies regarding ability to use coping strategies positively associated with reported use of coping strategies Keefe & Williams, 1990 88 patients Not Mixed Control CSQ attending an in- provided patient program SCL-90-R BDI MPQ Perceived control over pain and the belief in the ability to decrease pain negatively Table A (continued) Study Sample Mean Pain Pain Type Appraisal Appraisal Criterion Results Duration Measure Measures 60% females Keefe et al., 1989 223 patients from rheumatology practices 75% females Lorig etal., 1989 Sample 1: 97 patients attending an Arthritis Self-Management course 87% females Sample 2: 144 patients 3.5. yrs Not provided Not provided Rheumatoid Catastrophizing CSQ Arthritis Osteo-arthritis = 56% Rheumatoid arthritis = 15% Other arthritis = 29% OA= 58% Self-efficacy ASES correlated with depression ,but unrelated to indices of general psychological distress AIMS Based on 2 measurement CES-D points, 6 months, apart. Pain Controlling for age, socioeconomic status, gender, disability support status, disease duration, and initial dependent variable scores: Initial catastrophizing scores predictive of depression, pain intensity, and functional impairment at the 6 month point BDI Sample 1: HAQ a) Self-efficacy for physical Pain functioning and for managing intensity arthritis symptoms negatively associated with pain intensity, depression, and disability b) Pre-treatment self-efficacy levels for physical functioning and for managing arthritis symptoms negatively associated with pain intensity, depression, and disability -Table A (continued) Study Sample Mean Pain Pain Type Appraisal Appraisal Criterion Results Duration Measure Measures McCracken & Gross, 1993 attending an Arthritis Self-Management course 80% females 165 patients attending a pain treatment center 50% females 1.8 yrs. Rheumatoid arthritis = 22% Other arthritis = 20% 73% low back Catastrophizing CSQ PASS assessed 4 mths. post treatment Sample 2: a) Self-efficacy appraisals for managing pain, physical functioning, and managing other arthritis symptoms negatively associated with pain intensity, depression, and disability Catastrophizing positively correlated with all of the PASS anxiety scales O'Leary et at, 1988 30 patients attending a pain treatment program 100% females 8 yrs. Rheumatoid Self-efficacy arthritis ASES Pain a) Post-treatment self-intensity efficacy for pain negatively Zung associated with post-treatment PSS pain intensity HAQ b) Post-treatment self-Joint efficacy for physical impairment functioning negatively associated with post-treatment disability c) Post-treatment self-efficacy for managing symptoms negatively Table A (continued) Study Sample Mean Pain Pain Type Appraisal Appraisal Duration Measure Criterion Measures Results associated with post-treatment depression and stress d) Pre- to post-treatment changes in self-efficacy for managing pain unrelated to changes in pain intensity Reesor & Craig, 1988 80 patients attending a back pain clinic 50% females 8.8. yrs. Back Catastrophizing CSQ MPQ BDI OLBPDI PII ISS DDS Non-organic signs Catastrophizing positively associated with Medically Incongruent Low Back Pain Regan et al., 1988 151 patients who were graduates of an Arthritis Self-Management Course or attendees at an Arthritis month lecture 77% females > 10 yrs. Rheumatoid Self-efficacy Arthritis ASES Pain a) Self-efficacy for physical intensity functioning and managing CES-D pain negatively associated MPI activity with pain intensity and scale depression b) Self-efficacy for managing other arthritis symptoms negatively associated with pain intensity and depression, and positively associated with activity level Romano et al., 34 patients 4.6 yrs. Low back 1987 attending an in- and leg Catastrophizing CSQ MPQ SIP a) Catastrophizing positively associated with pain intensity ON ON Table A (continued) Study Sample Mean Pain Pain Type Appraisal Appraisal Criterion Results Duration Measure Measures patient pain program 53% females BDI and depression b) When controlling for depression, catastrophizing positively associated with pain intensity Spinhoven & Linssen, 1991 53 chronic pain patients medically referred or recruited from advertisements 12.7 yrs. Low back Control CSQ Pain Post-treatment to 6 month intensity follow-up increases in Uptime perceived control over pain Medication accounted for 14% of the use variance in decreased pain SCL-90 intensity 57% females Table A (continued) Study Sample Mean Pain Pain Type Appraisal Appraisal Criterion Results Duration Measure Measures Spinhoven et at. 1989 108 patients from an out-patient pain clinic 57% females 13.2 yrs. Low back Control CSQ Pain intensity Uptime Medication use SCL-90 FLI Perceived control over pain positively correlated with uptime Significance levels based on authors criterion of at least g<05 Measures abbreviations: AIMS = Arthritis Impact Measurement Scales. ASES = Arthritis Self-Efficacy Scale. BDI = Beck Depression Inventory. CES-D = Center for Epidemiologic Studies Depression Scale. CSQ = Coping Strategies Questionnaire. DCI = Daily Coping Inventory. DDS = Descriptor Differential Scale. FLI = Functional Limitations Index. GHQ = General health Questionnaire. HAQ = Health Assessment Questionnaire. ISS = Inappropriate Symptom Scale. IUI = Illness Uncertainty Scale. MAPPS = Movement and Pain Prediction Scale. MMPI = Minnesota Multiphasic Personality Inventory. MPI = Multidimensional Pain Inventory. MPQ = McGill Pain Questionnaire. MPQ-PRI = McGill Pain Questionnaire Pain Rating Index. NEOPI = NEO Personality Inventory. OLBPDI = Oswestry Low back Pain Disability Index. PASS = Pain Anxiety Symptoms Scale. PES = Pain Experience Scale. PII = Physical Impairment Index. POMS = Profile of Mood States. PRCS = Pain Related Control Scale. PRSS = Pain related Self-Statements Scale. PSS = Perceived Stress Scale. SCL-90-R = Symptom Checklist 90-Revised. SIP = Sickness Impact Scale. SOPA = Survey of Pain Attitudes. SWLS = Satisfaction With Life Scale. Table B Studies That Have Examined the Relationship Between the CSQ and Adaptation to Chronic Pain For abbreviations see page 175 Study Sample Mean Pain Pain Type CSQ Factors and/or Adaptational Results Duration Subscales Measures Beckham et al., 1991 65 patients attending an out-patient pain clinic 67% females 11.7 yrs. Rheumatoid arthritis Coping Attempts: DA, RPS, CSS, IP PH, IA Pain Control and Rational Thinking: ADP, CN, CAT BDI AIMS HS Controlling for age, gender, disease severity, disability support status: a) Coping Attempts unrelated to any of the adaptational measures b) Pain Control and Rational Thinking negatively related to physical disability, psychological distress, depression, and perceived hassles severity Estlander & Harkapaa, 1989 104 patients attending an in-patient pain program 57% females 11.3 yrs. Low back DA, RPS, CSS, IP, PH, IA, IPB, CAT, ADP, CN Disability a) Mild pain levels associated with the questionnaire use of Increasing Behavioural Activities, Pain Diverting Attention, and Ignoring Pain frequency b) Severe pain associated with the use of Pain intensity Catastrophizing, Increased Pain Behaviour, and Praying and Hoping Geisser, Robinson, 76 patients 7.8. yrs & Henson, 1994 attending an out-patient clinic 47% females 30% Low Conscious Cognitive MPI back Coping: BDI CSS, IP, RPS MPQ Pain Avoidance: PH, DA a) Neither factor related to adaptational measures b) Praying and Hoping positively associated with pain severity, interference, and affective distress c) Ignoring Pain negatively associated with pain intensity ON Table B (continued) Study Sample Mean Pain Pain Type CSQ Factors and/or Adaptational Results Duration Subscales Measures Gross, 1986 50 patients undergoing lumbar laminectomies 44% females > 1 yr. Back Active Coping and Suppression: CSS, CD, IA Loss of Control: CN, CAT Self-Reliance: ADP, PH MHQ Controlling for pre-surgery measures of STAI medical status, somatization, pain Pain intensity, and each criterion variable: interference a) Self-Reliance associated with better Activity levels psychological adjustment pre-surgery, Sleep decreased pain post-surgery, and disturbance increased positive attitude re. surgical Patient rated outcome surgical b) Loss of Control negatively associated outcome with post-surgical pain and positively with patient-rated surgical outcome c) Active Coping and Suppression unrelated to pain intensity or surgical outcome Hill, 1993 90 patients attending an out-patient pain program 100% males 8.4. yrs. Phantom Denial of Pain: limb RPS, DA Cognitive Coping: ADP, CN, CSS, IP Helplessness: IA, PH, CAT GHQ a) Denial of Pain and Cognitive Coping MPQ unrelated to pain intensity or Medication psychological distress use b) Distracting Attention, Increasing Pain triggers Behavioural Activities, and Praying and Hoping positively correlated with pain intensity and psychological distress c) Coping Self-Statements negatively correlated with pain intensity and psychological distress d) Ignoring Pain negatively correlated with pain intensity o Table B (continued) Study Sample Mean Pain Pain Type CSQ Factors and/or Adaptational Results Duration Subscales Measures Jensen & Karoly, 1991 Jensen, Turner, & Romano, 1994 Keefe, Caldwell, Martinez, Beckham, & Williams, 1991 Keefe, Crisson, Urban, & Williams, 1990 118 patients who completed an in-patient pain program within the last 7 yrs. 68% females 94 patients attending an in-patient program 69% females 52 patients, - where 1 to 3 yrs. they were recruited post knee from is not reported surgery 75% females 10.96 yrs. Mixed 62 patients attending a pain program 5.26 yrs. 46% Low back Rheumatoid arthritis 6.5. yrs Low back DA, IA, IP, RPS, PH, CSS. Cognitive Coping: CSS, IP, DA Helplessness: PH, CAT Coping Attempts: DA, IP, IA, RPS, CSS, PH Pain Control and Rational Thinking: CN, ADP, CAT Cognitive Coping and Suppression PRS, CSS, IP CES-D Controlling for pain severity: SWLS a) Ignoring Pain, Increasing HAQ Behavioural Activities, and Coping Self-MPI Statements positively associated with Medication psychological functioning use b) Diverting Attention, Ignoring Pain, Professional and Coping Self-Statements positively service associated with activity level only for utilization those patients reporting low pain levels. BDI a) Cognitive Coping unrelated to post-SIP treatment outcomes Medical b) Decreased Helplessness post-service treatment associated with decreased utilization physician visits post-treatment AIMS Controlling for age, obesity, duration of Observed pain disease, and gender: behaviour a) Pain Control and Rational Thinking Walking negatively associated with pain intensity, speed physical disability, psychological distress, and pain behaviour b) Coping Attempts negatively correlated with walking speed MPQ-PRI Controlling for demographic and medical SCL-90-R status variables: BDI a) Cognitive Coping and Suppression Uptime accounted for 5% of the variance in Table B (continued) Study Sample Mean Pain Pain Type CSQ Factors and/or Adaptational Results Duration Subscales Measures 63% females Keefe et al., 1987 Keefe & Williams, 1990 Parker et al, 1989 51 patients attending a rhuematology clinic 67% females 88 patients attending an in-patient program 60% females 79 patients attending and out-patient clinic 100% males 10.3 yrs. Osteo-arthritis Not provided Mixed Helplessness: CAT, IA, CN, ADP Diverting Attention and Praying: DA, PH Coping Attempts: DA, IP, IA, RPS, CSS, PH Pain Control and Rational Thinking CN, ADP, CAT DA, RPS, CSS, IP, PH, IA, CN, CAT, ADP 11.2 yrs. Rheumatoid Coping Attempts: arthritis DA, IPS, IA RPS, CSS, PH Pain Control and Rational Thinking: Observed pain psychological distress behaviours b) Helplessness positively associated with psychological distress and depression c) Diverting Attention and Praying positively associated with pain intensity ATMS Controlling for age, obesity, duration of Observed pain disease, and gender: behaviours a) Pain Control and Rational Thinking Walking negatively correlated with pain intensity, speed physical disability, psychological distress, and pain behaviours b) Coping Attempts negatively correlated with walking speed SCL-90-R a) Coping Self-Statements and BDI Increasing Behavioural Activities MPQ negatively associated with depression b) Diverting Attention and Increasing Behavioural Activities positively associated with pain severity MPQ-PRI Controlling for education, age, and SCL-90-R disease progression: AHI a) Coping Attempts associated with BDI pain intensity (direction of relationship HS not provided) ATMS b) Pain Control and Rational Thinking Table B (continued) Study Sample Mean Pain Pain Type CSQ Factors and/or Adaptational Results Duration Subscales Measures Romano et al., 1987 34 patients attending an in-patient program 53% females 4.6. yrs. Low back and leg CN, ADP, Cat DA, RPS, CSS, IP, PH, IA, CN, CAT, ADP Pain intensity MPQ SIP BDI negatively associated with pain intensity and psychological distress Ignoring Pain Sensations negatively associated with physical and psychological disability Rosentiel & Keefe, 1983 Spinhoven et al.. 1989 61 patients attending a pain center 69% females 108 patients attending an out-patient clinic 57% females from onset = 6.41 yrs. since continuous = 1.67 yrs. Low back 13.2 vrs. Low back Helplessness: CAT, IA, ADP, CN Diverting Attention and Praying: DA, PH Cognitive Coping and Suppression: RPS, CSS, IP Helplessness: CAT, PH, RPS Active Coping: IP, CSS, IA, DA MHQ Controlling for disability status, pain FCES duration, previous surgeries, and Zung somatization: STAI-S a) Cognitive Coping and Supression Pain intensity positively associated with functional Downtime impairment, but unrelated to pain Interference intensity, depression, or state anxiety b) Helplessness positively associated with depression and state anxiety c) Diverting Attention and Praying positively associated with pain intensity and functional impairment SCL-90 Controlling for disability status, pain FLI duration, previous surgeries, age, and Pain intensity gender: Uptime a) Active Coping accounted for 5% of Medication the variance in depression use b) Perceived Control negatively associated with pain intensity and Table B (continued) Study Sample Mean Pain Pain Type CSQ Factors and/or Adaptational Results Duration Subscales Measures Spinhoven et al. 1991 111 patients attending a pain clinic 62% females 12.2 yrs. Headaches Perceived Control: ADP, CN Helplessness: CAT, PH, RPS Active Coping: IP, CSS, IA, DA Perceived Control: ADP, CN SCL-90 DPQ Pain intensity Pain duration functional limitations, and positively with uptime. c) Helplessness positively associated with pain intensity, anxiety, depression, psychoneuroticism, and functional impairment a) Helplessness accounted for 12% of the variance in psychological distress b) Active Coping accounted for 5% of the variance in daily pain duration c) Perceived Control accounted for 2% of the variance in pain intensity Spinhoven & Linssen, 1991 53 patients attending a pain program 57% females 12.2 yrs. Low back Helplessness: CAT, PH, RPS Active Coping: IP, CSS, IA, DA Perceived Control: ADP, CN SCL-90 a) Helplessness pre-treatment predictice Pain intensity of psychological distress post-treatment, Uptime and accounted for 24% of the variance in Medication depression post-treatment use b) Active Coping accounted for 13% of the variance in uptime post-treatment c) Perceived Control accounted for 14% of the variance in pain intensity post-treatment Sullivan & D'Eon, 125 pts. attending 7.2. yrs. Back and 1990 an out-patient clinic neck 57% females DA, RPS, CSS, IP, PH, IA, CN, CAT, ADP BDI Pain intensity Controlling for age, gender, pain intensity and pain duration, no CSQ subscales were related to depression Table B (continued) Study Sample Mean Pain Pain Type CSQ Factors and/or Adaptational Results Duration Subscales Measures Turner & Clancy, 1986 74 patients attending an out-patient program 36% females 6.8. yrs. Low back Denial of Pain: IP, RPS, ADP Diverting Attention and Praying: D A PH, IA Helplessness: CAT, CN, CSS SIP Controlling for pain duration: BDI a) Denial of Pain positively associated Pain intensity with downtime Downtime b) Diverting Attention and Praying positively associated with pain intensity c) Helplessness positively associated with physical and psychosocial distress and depression Pre- to post-treatment: a) Praying and Hoping negatively associated with pain intensity b) Catastrophizing positively associated with pain intensity and psychological distress Significance levels based on authors criterion of at least p< 05 Coping Strategies Questionnaire subscale abbreviations: ADP = Perceived ability to decrease pain, CAT = Catastrophizing, CD = Cognitive Distraction, CN = Perceived control over pain, CSS = Coping Self-Statements, DA = Diverting Attention, IA = Increasing Behavioural Activities, IPB = Increased Pain Behaviour, RPS = Reinterpreting Pain Sensations. Measures abbreviations: AIMS = Arthritis Impact Measurement Scales. AHI = Arthritis Helplessness Index. BDI = Beck Depression Inventory. CES-D = Center for Epidemiologic Studies Depression Scale. CSQ = Coping Strategies Questionnaire. DPQ = Dutch Personality Questionnaire. FCES = Functional Capacity Evaluation Scale. FLI = Functional Limitations Index. GHQ = General health Questionnaire. HAQ = Health Assessment Questionnaire. HS = Hassles Scale. MHQ = Middlesex Hospital Questionnaire. MPI = Multidimensional Pain Inventory. MPQ = McGill Pain Questionnaire. MPQ-PRI = McGill Pain Questionnaire Pain Rating Index. SCL-90 = Symptom Checklist. SCL-90-R = Symptom Checklist 90-Revised. SIP = Sickness Impact Scale. STAI-S = State-Trait Anxiety Inventory-State Scale. SWLS = Satisfaction With Life Scale. A P P E N D I X B Demographic Questionnaire Summary of the Demographic Information Code# Demographic Information Questionnaire Date: Age: years 1. Ethnic/racial background: (1) Caucasian (4) Hispanic (2) First Nations (5) Asian (3) Black (6) East Indian (7) Other (explain) 2. Education: (1) Less than Grade 12 (4) Bachelors Degree (2) High School Graduate (5) Masters Degree (3) Some College/Umversity (6) Doctoral Degree 3. Approximate Yearly Household Income Before Taxes: (1) Under $19,999 (4) $60,000 to $79,999 (2) $20,000 to $39,999 (5) $80,000 to $99,999 (3) $40,000 to $59,999 (6) Over $100,000 4. Number of Children under 19 Years of Age Living in the Household: (1) None (4) Three (2) One (5) Four or more (3) Two 178 5. Employment Status: (1) Full time (2) Full time with restrictions (3) Full time but on sick leave right now (4) Part time (5) Part time with restrictions (6) Part time, but on sick leave right now (7) No, but not because of pain (8) No, unable to work or unemployed because of pain (9) Homemaker (10) Student 6. How Long Have You Experienced Low Back Pain? (a) Since it began: months (b) On a daily basis: months 7. Under What Circumstances Did Your Pain Begin? (1) _ Accident at work (5) Following surgery (2) At work, but not an accident (6) Following illness (3) Accident at home (7) Pain just began, no reason (4) Motor vehicle accident (8) Other (explain) 8. In the Last 6 Months Have You Had Any of the Following Treatments for Relief of Pain? (1) Nerve blocks (8) Traction (2) TENS (9) Massage therapy (3) Biofeedback (10) Counselling/psychiatric (4) Acupuncture (11) Hypnosis (5) Manipulation (12) Ultrasound (6) Heat therapy (13) Physiotherapy (7) Bed rest (14) Exercise (15) Other (explain) 179 9. What Kind of Medical Specialists are Currently Involved in Your Care? (1) Family Physician (5) Rheumatologist (2) Physiotherapist (6) Chiropractor (3) Acupuncturist (7) Other (4) Naturopath (list) 10. Do you Take Medications for Pain relief? (1) Never (4) 1 to 2 times per day (2) Less than one time per week (5) 3 or 4 times a day (3) Several times a week (6) 5 or more times a day 11. If You Take Medication, When Do You Take It? (1) When needed for pain (2) Regularly, by the clock 12. If You Are Taking Medication For Pain, Which of the Following Do You Take? (1) Analgesics (4) Muscle relaxants (2) NSAIDS (5) Antidepressants (3) Sleeping pills (6) Other (type) 13. Are you currently involved in any kind of litigation concerning your low back pain? (1) Yes (2) No 14. If you answered yes to the above question, who are you litigating against? (1) W C B (3) Private Insurance Company (2) ICBC 180 Table A Means and standard deviations for age, years since pain began, and years since pain experienced daily for completers and non-completers Completers Non-Completers M SD n M SD n A g e a 46.83 11.90 87 41.27 7.42 11 Pain Began 16.69 12.78 88 11.33 10.24 12 Pain Daily b 10.75 10.45 86 6.14 5.29 11 a Missing data due to two participants leaving answers blank b Missing data due to three participants leaving answers blank Table B Frequencies for categorical demographic data a Non-Completers Completers Variable n % n % Cause of Pain Accident at work 12 14 1 8 Accident at home 8 9 1 8 Motor vehicle accident 9 10 0 0 Following surgery 1 1 0 0 Just began 28 32 5 42 Sports injury 22 25 4 33 Other 7 8 1 8 Multiple reasons 1 1 0 0 table continues 181 Table B continued Non-Completers Completers Variable n % n % Children under 19 in house None 58 66 7 58 1 10 11 3 25 2 13 15 1 8 3 6 7 1 8 4 or more 1 1 0 0 ducation Less than grade 12 6 7 0 0 High School graduate 13 15 1 8 Some college/university 42 48 9 75 Bachelors degree 21 24 1 8 Masters degree 5 6 1 8 Doctoral degree 1 1 0 0 mployment status Full-time 23 26 4 33 Full-time with restrictions 4 5 1 8 Full-time but on sick leave 2 2 1 8 Part-time 15 17 1 8 Part-time with restrictions 4 5 0 0 Part-time but on sick leave 1 1 0 0 No, but not due to pain 3 3 0 0 No, due to pain 16 18 1 8 Homemaker 16 18 3 25 Student 1 1 1 8 Retired 3 3 0 0 table continues 182 Table B continued Completers Non-Variable n % n % Ethnic/racial background Caucasian 83 94 11 92 Hispanic 1 1 0 0 Asian 4 5 0 0 East Indian 0 0 1 8 Household income Under $19,000 11 13 2 17 $20-39,999 10 11 1 8 $40-59,999 23 26 3 25 $60-79,999 12 14 2 17 $80-99,999 12 14 1 8 Over $100,000 8 9 1 8 Don't know/decline to answer 12 14 2 17 Litigation status No litigation 82 93 12 100 ICBC 3 3 0 0 Private insurance 1 1 0 0 Multiple companies 1 1 , 0 0 Other 1 1 0 0 Frequency of pain medication Never 24 27 2 17 Less than one time per week 18 20 5 42 Several times per week 21 24 2 17 1-2 times per day 13 15 3 25 3 - 4 times per day 9 10 0 0 5 or more times per day 3 3 0 0 table continues 183 Table B continued Completers Non-Variable n % n % Type of medication for pain N / A 24 27 2 17 Analgesics 19 22 7 58 NSAIDS 9 10 0 0 Sleeping pills 2 2 0 0 Muscle relaxants 1 1 0 0 Antidepressants 1 1 1 8 Multiple medication 31 35 2 17 Other 1 1 0 0 Medication times N / A 24 27 2 17 When needed (prn) 56 64 9 75 By the clock 2 2 1 8 Both prn and by the clock 6 7 0 0 Health professionals involved None 12 14 2 17 Family physician 20 33 3 25 Physiotherapist 3 3 0 0 Naturopath 0 0 1 8 Chiropractor 3 3 1 8 Other 3 3 0 0 Multiple 47 53 5 42 table continues Table B continued Completers Non-Completers Variable n % n % Treatments for pain in the last 6 months None 12 14 2 17 Acupuncture 0 0 1 8 Manipulation 2 2 0 0 Heat therapy 1 1 0 0 Massage 6 7 0 0 Counselling 1 1 0 0 Physiotherapy 2 2 1 8 Exercise 5 6 1 8 Other 1 1 0 0 Multiple 58 66 7 58 a Note. Percentages may not add up to 100 due to rounding A P P E N D I X C Consent Form A P P E N D I X D Equations for the H L M Models: A n example for mghttime depressed mood 188 Null Model The equations for the null model are as follows: Within participants = p 0j + Sij (1) Between-participants POJ = <t>oo+ Uqj (2) Combining equations 1 and 2 yields a single linear model: Yij = d>oo + U 0 j + 8ij (3) where §oo is the grand mean, Uqj is the deviation of each participant's average score from the grand mean (between-participant residual), and ey- is the deviation of each individual's daily score from their own average score (within-participant residual). Level 1 Model The example of an equation for the nighttime depressed mood model for within-participants is as follows: Nighttime depressed mood = p0j + Pij(a.m. depressed mood) + p2j(Self-Efficacy)+ p3j(control) + p4j(Catasfrophizing) + sjj (4) 189 The equations for the between-participants model are: Poj^oo + Uoj (5) PIJ = 4>IO + U , J (6) f32j = 4)20 + U 2 j (7) P3j = <ho + U 3 j (8) P 4 j = <|>40 + U4j (9) Combining equations 4 through 9 yields a single linear model: Nighttime depressed mood = (<p0o + U0j) + (<|>io + Uij)a.m. depressed mood + (<|»2o + U2j)Self-Efficacy + (<t>30 + U3j)control + (<j>40 + U4j)Catastrophizing +Sy (10) This model has both individual level (each day for each participant) residuals, 8y, and group-level (each set of participant's days) residuals, U 0 j , Uij, U 2 j , U 3 j , U 4 j. The intercept is now made up of a fixed effect, <p0o, and a random effect, UOJ, which varies across participants. H L M provides a t-test of the hypothesis that the 190 fixed effect is zero, (i.e., H 0 : <p0o= §) , and a %2 test that the variance is zero (i.e., H 0 : Var (UQJ) = The slope for a.m. depressed mood is made up of a fixed effect, d>i0, and a random effect, Uij. One can think of d>io as the average slope, and Uij as the increment to the slope, either positive or negative, that is unique to each individual. H L M tests whether the average slope is significant (Hypothesis 1), and whether participants vary in their slopes This interpretation is the same for the coefficients of the other three independent variables. A n identical procedure was followed for the addition of the coping strategies to the model, and for the other two dependent variables. Level-2 Equations Continuing with the example for nighttime depressed mood and the three pain appraisals given previously, the equation for the within-participants model is as follows: Nighttime depressed mood = p0j + Pij(a.m. depressed mood) + p2j(Self-Efficacy)+ p3j(control) + p4j(Catastrophizing) + £y (11) 191 The equations for the between-participants model for the MPI functional subscales are: Poj = <t>oo + MPS) + MI) + MLC) + MS) + MGAL) + MAD) +U 0 j (12) Plj = <t>10 + <t»ll(PS) + M O + <t>13(LC) + MS) + (|>15(GAL) + (|)06(AD) -r-Uy (13) P 2 j = d>20 + MPS) + (t)22(I) + MLC) + MS) + MGAL) + MAD) + U 2 j (14) P 3 j = <bo + MPS) + Ml) + MLC) + MS) + MGAL) + MAD) + U 3 j (15) p 4 j = <p40 + 4»4i(PS) + <P42(I) + MLC) + MS) + MGAL) + MAD) + U 4 j (16) APPENDIX E Completers versus Non-Completers Data 193 Table A Means and Standard Deviations for Age. Years Since Pain Began, and Years Since Pain Experienced Daily for Completers and Non-completers Completers Non-completers ~~M SD n M SD n~~ 46.83 11.90 87 41.27 7.42 11 16.69 12.78 88 11.33 10.24 12 10.75 10.45 86 6.14 5.29 11 Age 3 Pain Began Pain Daily b a Missing data are due to two participants leaving answers blank. b Missing data are due to three participants leaving answers blank 194 Table B Results of the Chi-square Analyses for the Categorical Demographic Variables Original Variable Collapsed Groupings n df ^ 2 g (# of categories) Employment 1. Working out-side of the status home 100 1 .06 .80 (11) 2. Not working out-side of the home Household 1. Under $59,999 income 2. Over $60,000 or do not 100 1 .00 1.00 (7) know or refuse to answer Cause of the pain 1. Accident (8) 2. Non-accident 100 1 1.12 .29 # of Health care 1. Two or more professionals 2. One or none 96 1 .38 .54 involved (7) a Treatments in the 1. Two or more last 6 months (10) 2. One or none 100 1 .27 .61 Missing data are due to four participants leaving answers blank. 195 Table C Means and Standard Deviations for Completers and Non-Completers on the Time 1 MPI Subscales Non-Completers (n_= 12) Completers (n = 88) M SD M SD Affective Distress a 4.03 1.29 3.06 1.20 Life Control a 2.73 1.06 3.78 1.34 Support a 2.50 1.46 3.80 1.66 General Activity Level 2.70 0.90 3.10 0.77 Interference 2.99 1.58 3.35 1.36 Pain Severity 3.03 1.52 3.50 1.10 Distracting 1.71 1.08 2.02 1.49 Responses Punishing • Responses 1.79 1.88 1.68 1.32 Solicitous Responses 2.65 1.82 2.90 1.50 Note. The higher the score, the greater the attribution. a significant at g < .01 A P P E N D I X F Correlations for Days 15 and 30, and comparisons with clinical pain studies Table A Pearson-Product Moment Correlation Matrix of Demographic and Daily Variables for Day 15 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 1 Age 2 Years Pain Began .26* 3 Years Pain Daily .19 .68** 4 Anxiety A M -.13 -.12 -.07 5 Anxiety PM -.14 -.17 -.02 .83** 6 Depression A M -.11 -.14 -.06 .88** .73** 7 Depression PM -.12 -.20 .01 .76** .84** .82** 8 Average Pain -.14 .19 .20 .24* .24* .25* .26* 9 A M Pain -.13 .07 .02 44** .39** 44** .38** .62** 10 PMPain -.18 .03 .12 .28* .33** .30** .34** .81** .75** 11 Catastrophizing -.13 -.09 .05 .33** .38** .35** .42** .28** .20 .34** 12 Control .08 .05 .01 -.38** -.36** -.40** -.40** -.29** -.34** -.36** -.28* 13 Self-efficacy -.08 -.03 -.13 -.47** -.42** -.52** -.50** -.41** -.57** -.44** -.33** .65** 14 Distraction .04 .18 .19 -.02 -.01 -.07 -.04 .16 .07 .16 .22* .19 .10 15 Ignoring Pain .04 .14 .02 -.15 -.17 -.17 -.25* -.03 .01 -.08 -.08 .27* .21 .35** 16 Praying and Hoping .04 .07 .22* -.12 .07 -.17 -.03 .07 .08 .12 .31** .09 .07 .41** .21* 17 Reinterpreting Pain Sensation .13 .15 .16 -.07 -.16 -.02 -.07 -.04 .01 .00 .19 .05 .04 .38** .25* .24* *p_<.05 **p_<.01 Note. Due to missing data, minimum n = 79 198 Table B Pearson-Product Moment Correlation Matrix of Time 1 MPI, Demographic, and Daily Variables for Day 15 18 19 20 21 22 23 24 25 26 1 Age .14 .14 -.00 .17 -.24* .07 -.03 .01 .08 2 Years Pain Began .10 .02 .31** .31** -.08 -.02 .04 .20 17 3 Years Pain Daily .01 .08 .25* .35** -.06 -.02 -.05 .04 .04 4 Anxiety A M .33** -.27* .12 .05 -.02 -.36** -.09 .30** -.20 5 Anxiety PM .30** -.19 .17 .07 -.00 -.30** -.03 .22* -.09 6 Depression A M .32** -.32** .03 .01 .03 -.41** -.20 .30** -.30** 7 Depression PM .28* -.25* ' .05 -.06 .06 -.36** -.14 .23* -.15 8 Average Pain .20 -.09 .31** .25* .11 -.11 .02 .20 -.03 9 A M Pain .29** -.18 .23* .14 .13 -.18 .10 .26* -.03 10 PM Pain .17 -.11 .28** .21 .12 -.02 .11 .14 .00 11 Catastrophizing .21 -.08 .26* .06 .09 .02 .08 .02 .00 12' Control -.30** .31** -.12 .04 .05 .20 .12 -.15 .05 13 Self-efficacy -.50** .27* -.22* -.16 .03 .12 .04 -.13 .01 14 Distraction -.06 .11 .23* .07 -.01 .13 .28** -.04 .20 15 Ignoring Pain -.22* .19 -.06 -.03 -.17 .01 .20 .04 .15 16 Praying and Hoping -.05 .20 .18 .07 -.02 .07 .15 -.12 .08 17 Reinterpreting Pain Sensation -.15 .15 .08 .00 .08 .15 .11 .02 .20 18 MPI Affective Distress -.70** .31** .31** .07 -.24* -.01 .39* -.14 19 MPI Life Control -.19 -.25* .00 .31** .06 -.28** .20 20 MPI Interference .52** -.12 .24* .18 .04 .34** 21 MPI Pain Severity -.13 .09 -.01 .20 .10 22 MPI General Activity Level .07 .22* -.00 .13 23 MPI Support .46** -.53** .70** 24 MPI Distracting Responses -.11 .63** 25 MPI Punishing Responses -.27* 26 MPI Solicitous Responses *g<.05 **p<.01 Note. Due to missing data, minimum n = 79 Table C Pearson-Product Moment Correlation Matrix of Demographic and Daily Variables for Day 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 1 Age 2 Years Pain Began .22 3 Years Pain Daily .13 .63** 4 Anxiety AM -.14 -.01 -.02 5 Anxiety PM -.19 -.04 -.01 .88** 6 Depression AM -.01 -.07 -.05 .80** .70** 7 Depression PM -.06 .00 .05 .80** .81** .85** 8 Average Pain -.02 .35** .28* .13 .07 .11 .06 9 AM Pain -.07 .25* .17 .20 .12 .24* .15 .77** 10 PMPain -.05 .27* .26* .07 .08 .07 .07 .87** .83** 11 Catastrophizing -.11 .16 .25* .20 .31** .21 .26* .25* .16 .29* 12 Control -.10 .04 -.05 -.21 -.31 -.29* -.22 -.29* -.24* -.36** -.44** 13 Self-efficacy -.03 -.02 -.05 -.16 -.21 -.35** -.25* -.26* -.40** -.38 -.26* .67** 14 Distraction .06 .17 .09 .15 .18 .11 .07 .11 .22 .14 .17 .09 -.01 15 Ignoring Pain -.03 .08 .09 -.17 -.19 -.13 -.16 -.00 .08 -.08 -.04 . .27* .16 .25* 16 Praying and Hoping .11 .08 .03 .08 .04 .09 .00 .20 .18 .18 .37** -.08 .04 .35** .24* 17 Reinterpreting Pain Sensation .05 -.12 -.08 .03 .05 .09 .04 -.04 .09 -.02 .15 .03 -.02 41** .25* .24* *p_<.05 **p_<.01 Note. Due to missing data, minimum n = 79 200 Table D Pearson-Product Moment Correlation Matrix of Time 1 MPI. Demographic, and Daily Variables for Day 30 18 19 20 21 22 23 24 25 26 1 Age -.12 .14 -.01 .12 -.34** .03 -.10 .01 .04 2 Years Pain Began .13 .04 .39** .33** -.06 .03 .05 .15 .26* 3 Years Pain Daily .02 .12 .25* .34** -.02 -.02 -.08 -.02 .08 4 Anxiety A M .57** -.48** .21 .12 .09 -.18 .19 .28* -.10 5 Anxiety PM 5 5 * * -.50** .22 .05 .08 -.11 .24* .25* -.04 6 Depression A M 5 7 * * -.53** .12 .14 -.01 -.22 .07 ..25* -.14 7 Depression PM .57** -.45** .09 .14 .05 -.12 .15 .29* -.09 8 Average Pain .24* -.23 .47** .81** .03 -.02 .06 .15 .05 9 A M Pain .23 -.28* 4 4 * * .77** -.01 -.05 .07 .23 .08 10 PMPain .23 -.22 4 7 * * .78** -.04 -.03 .04 .16 .04 11 Catastrophizing .20 -.12 .16 .20 -.01 -.13 .09 .21 -.10 12 Control -.28* .48** -.16 -.28* .21 .14 .14 -.10 .22 13 Self-efficacy -.38** 41** -.30** -.46** .32** .04 . .09 -.16 .02 14 Distraction -.02 -.04 .24* .07 .05 .16 .56** .01 .44** 15 Ignoring Pain -.22 .28* -.16 .05 -.07 .02 .06 .04 .16 16 Praying and Hoping -.05 -.00 .12 .05 -.13 -.19 .09 .07 -.10 17 Reinterpreting Pain Sensation -.14 -.05 .10 -.02 -.02 .15 .17 -.11 .20 18 MPI Affective Distress -.40** .27* .30* .18 -.22 .06 .26* -.13 19 MPI Life Control -.23* -.25* -.02 .30* .05 -.29* .19 20 MPI Interference .48** -.12 .17 .15 .05 .31** 21 MPI Pain Severity -.09 .02 -.08 .20 .05 22 MPI General Activity Level .06 .28* .05 .11 23 MPI Support .42** -.59** .67** 24 MPI Distracting Responses -.04 .62** 25 MPI Punishing Responses -.27* 26 MPI Solicitous Responses *P<.05 **p<.01 Note. Due to missing data, minimum n = 79 201 Table E Differences Between Means of Study Participants and a Clinical Sample of a Heterogeneous Pain Population on The MPI Subscales Study Participants Heterogeneous (N = 88) Pain (N = 300) Variables M SD M SD t MPI Subscales Affective Distress 3.06* 1.20 3.78 1.28 4.70 Pain Interference 3.35* 1.36 4.61 1.16 8.60 Life Control 3.78* 1.33 3.19 1.62 3.12 Pain Severity 3.36* 1.08 4.52 1.04 9.12 General Activity Level 3.10* 0.77 2.05 0.98 9.25 Support 3.80* 1.62 4.63 1.42 4.67 Distracting Responses 2.02 1.49 2.42 1.46 2.25 Punishing Responses 1.68 1.32 1.85 1.66 0.88 Solicitous Responses 2.90 1.50 3.57 1.58 1.71 Note. The higher the score, the greater the attribution. D f = 386 * p < .001 (2-tailed t-tests). From "Multiaxial Assessment of Pain. Computer Program User's Manual Version 2.1" by T. E . Rudy, 1989, p. 45. 202 Table F Differences Between Means of Study Participants and a Clinical Sample of a Back Pain Patients on The MPI Subscales Study Participants Back Pain (N = 88) (N = 150) Variables M SD M SD t MPI Subscales Affective Distress 3.06* 1.20 3.81 1.29 4.44 Pain Interference 3.35* 1.36 4.79 0.98 9.45 Life Control 3.78* 1.33 3.09 1.53 3.52 Pain Severity 3.36* 1.08 4.64 1.00 9.25 General Activity Level 3.10* 0.77 1.96 0.91 9.86 Support 3.80* 1.62 4.60 1.43 3.96 Distracting Responses 2.02 1.49 2.47 1.46 2.28 Punishing Responses 1.68 1.32 1.73 1.62 0.24 Solicitous Responses 2.90* 1.50 3.61 1.55 3.45 Note. The higher me score, the greater the attribution. D f = 236 * p < .001 (2-tailed t-tests). From "Multiaxial Assessment of Pain. Computer Program User's Manual Version 2.1" by T. E . Rudy, 1989, p. 47. 203 Table G Differences Between Means of Study Participants and a Clinical Back Pain Sample (Study 1) for the M P Q Subscales Study Participants N = 40 N = 8 6 a Variables M SD M SD ~ t M P Q Subscales: Affective 2.74 2.56 3.10 2.15 0.77 Sensory 15.79 7.79 16.30 6.70 0.36 Evaluative 2.35 1.53 3.10 1.25 2.71 Note. The higher the score, the greater the attribution. a Missing data are due to two participants leaving answers blank. D f = 236. A l l ts non-significant at p < .001. From, "Medically incongruent chronic back pain: Physical limitations, suffering, and ineffective coping." 1988, Pain. 32. pp. 35-45. 204 Table H Differences Between Means of Study Participants and a Clinical Back Pain Sample (Study 2) for the M P Q Subscales Study Participants N = 32 N = 8 6 a Variables M SD M SD t M P Q Subscales: Affective 2.74 2.56 2.70 2.70 0.10 Sensory 15.79 7.79 16.40 6.70 0.39 Evaluative 2.35 1.53 3.00 1.30 1.55 Note. The higher the score, the greater the attribution. a Missing data are due to two participants leaving answers blank. D f = 116. A l l ts non-significant at g < .001. From, "Pain Behavior and pain coping strategies in low back pain and myofascial pain dysfunction syndrome patients," by F. J. Keefe & E . Dolan, 1986, Pain. 24. pp. 49-56. A P P E N D I X G H L M results for MPI functional and interpersonal subscales 206 Table A Level-2 H L M Results of the Three Models. The Relationship Between the MPI Functional Subscales and the Level-1 Intercepts and Morning Covariates Interference Support Activity Level Coefficient SE Coefficient SE Coefficient SE Depressed Mood Intercept 0.330 0.232 -0.350* 0.153 0.131 0.321 Depressed Mood - am slope -0.040 0.0370 -0.006 0.024 -0.001 0.051 Anxious Mood Intercept 0.647* 0.267 -0.381* 0.192 0.104 0.385 Anxious mood - am slope 0.025 0.031 -0.024 0.018 0.025 0.041 Pain Intensity Intercept 0.165 0.163 -0.108 0.107 0.031 0.224 Pain Intensity - am slope -0.034 0.028 0.026 0.017 -0.010 0.038 Pain Severity Life Control Affective Distress Coefficient SE Coefficient SE Coefficient SE Depressed Mood Intercept -0.133 0.252 0.207 0.225 0.570* 0.269 Depressed Mood - am slope 0.050 0.038 0.031 0.032 0.052 0.042 Anxious Mood Intercept -0.075 0.290 0.180 0.261 0.700* 0.316 Anxious mood - am slope 0.013 0.032 0.048 0.026 0.025 0.034 Pain Intensity Intercept 0.857** 0.176 0.067 0.158 -0.181 0.190 Pain Intensity - am slope 0.021 0.030 0.036 0.025 0.069* 0.031 *p_<.05 **p<.01 207 Table B Level-2 H L M Results of the Three Models. The Relationships Between the MPI Interpersonal Subscales and the Level-1 Intercepts and Morning Covariate Slopes Punishing Responses Solicitous Responses Distracting Responses Level 1 Coefficients Coefficient ' SE Coefficient SE Coefficient SE Depressed Mood Intercept 0.479** 0.172 0.034 0.183 0.073 0.179 Depressed Mood - am slope -0.016 0.026 0.013 0.026 -0.016 0.026 Anxious Mood Intercept 0.647** 0.210 0.163 0.217 0.057 0.214 Anxious mood - am slope -0.027 0.022 0.003 0.024 0.002 0.024 Pain Intensity Intercept 0.361** 0.136 0.246 0.138 0.011 0.134 Pain intensity - am slope -0.040* 0.019 0.017 0.020 -0.030 0.020 *g<.05 **p<.01 A P P E N D I X H Tables and Figures of the Significant Interactions 209 Table A H L M Analyses For the Effects For Average Pain and Ignoring Pain for Nighttime Depressed Mood (N = 88) Fixed Effects Unstandardized SE Coefficient Random Effects Variance Intercept Depressed Mood - am Average Pain (AP) Ignoring Pain (IP) A P X I P Level-1 (within-variance) 17.524** 1.880** 0.850** -0.702** -0.256** 0.217 0.152 0.085 0.113 0.075 3.600** 1.107** fixed fixed fixed 7.724 Dashes indicate data are not applicable **p<.001 Figure 1 The Relationship Between Nighttime Depressed Mood and Ignoring Pain for High and Low Levels of Average Pain (AP) -1 0 1 Ignoring Pain (standard scores) 210 Table B H L M Analyses For the Effects For Average Pain and Praying and Hoping for Nighttime Depressed Mood (N = 88) Fixed Effects Unstandardized S Coefficient E Random Effects Variance Intercept Depressed Mood - am Average Pain (AP) Praying and Hoping (Praying and Hoping) A P X Praying and Hoping Level-1 (within-variance) 17.468** 1.914** 0.888** 0.031 0.201* 0.233 0.149 0.087 0.152 0.087 4.204** 1.032** fixed fixed fixed 7.830 Dashes indicate data are not applicable • p<.05 **p<.001 Figure 2 The Relationship Between Nighttime Depressed Mood and Praying and Hoping for High and Low Levels of Average Pain (AP) O O 19 " 18 • <D w w CD M 0) 16 Q 17 -High AP (+1 Sd) Low AP (-1 Sd) -1 0 1 Praying and Hoping (standard scores) 211 Table C H L M Analyses For the Effects For Average Pain and Praying and Hoping for Nighttime Anxious Mood (N = 88) Fixed Effects Random Effects Unstandardized S_ Coefficient E Variance Intercept Anxious Mood - am Average Pain (AP) Praying and Hoping (Praying and Hoping) A P X Praying and Hoping Level-1 (within-variance) 17.243** 1.918** 0.828** 0.219 0.233** 0.264 0.140 0.091 0.164 0.092 5.574** 0.794** f i x e d f i x e d fixed 8.629 Dashes indicate data are not applicable ** p < .001 Figure 3 The Relationship Between Nighttime Anxious Mood and Praying and Hoping for High and Low Levels of Average Pain (AP) O o £ O •H X a < 19 . 18 -17 -16 -'High AP (+1 Sd) Low AP (-1 Sd) -1 0 1 Praying and Hoping (standard scores) 212 Table D H L M Analyses For the Effects For Average Pain and Distraction for Nighttime Anxious Mood (N = 88) Fixed Effects Unstandardized S_ Coefficient E Random Effects Variance Intercept Anxious Mood - am Average Pain (AP) Distraction (Dist) A P X Dist Level-1 (within-variance) 17.269** 1.880** 0.848** -0.852** -0.211* 0.285 0.142 0.090 0.150 0.081 6.600** 0.834** fixed fixed fixed 8.470 Dashes indicate data are not applicable ** g < .001 Figure 4 The Relationship Between Nighttime Anxious Mood and Distraction for High and Low Levels of Average Pain (AP) 19 -O 18 o 0 g 16 15 High AP (+1 Sd) Low AP (-1 Sdi -1 0 1 Distraction (standard scores) 213 Table E H L M Analyses For the Effects For Average Pain and Reinterpreting Pain Sensation for Nighttime Pain Intensity (N = 88) Fixed Effects Random Effects Unstandardized Coefficient SE Variance Intercept 4.500** 0.114 1.057** Pain Intensity- am 0.214** 0.055 0.100** Average Pain (AP) 1.457** 0.037 fixed Reinterpreting Pain Sensation (RPS) 0.023 0.054 fixed A P X R P S -0.096* 0.055 fixed Level-1 (within-variance) 1.199 Dashes indicate data are not applicable *p<.05 **p<.001 214 Figure 5 The Relationship Between Nighttime Anxious Mood and Distraction for High and Low Levels of Average Pain (AP) + J •H a -p g M c •H (TJ 3 -1 -'High AP (+1 Sd) •Low AP (-1 Sd) -1 0 1 Reinterpreting Pain Sensation (standard scores) A P P E N D I X I Examples of Questions for Each of The Multi-Item Measures The Coping Strategy Questionnaire Distraction: I replay in my mind pleasant experiences in the past I do something I enjoy, such as watching T. V . or listening to music Reinterpreting Pain Sensation: I don't think of it as pain but rather as a dull or warm feeling I pretend it is not part of me Ignoring Pain: I tell myself that I can't let pain stand in the way of what I have to do I don't pay any attention to the pain Praying and Hoping: I rely on my faith in God I know someday someone will be here to help me and it will go away for a while Catastrophizing It is terrible and I feel it is never going to get better I feel my life isn't worth living Self-Efficacy for Pain: How certain are you that you can decrease your pain quite a bit How certain are you that you can continue most of your daily activities 217 State-Trait Personality Inventory: Depression Subscale I feel blue I feel hopeful about the future Anxiety Subscale I feel clam I am presently worrying over possible misfortunes Multidimensional Pain Inventory: Interference In general, how much does pain interfere with your day-to-day activities? How much has pain changed your ability to take part in recreational and other social activities? Pain Severity On average, how severe has your pain been in the last week? How much suffering do you experience because of your pain? Support: How attentive is your spouse (significant other) to you because of your pain? How worried is your spouse (significant other) about you because of your pain? 218 Affective Distress Rate your overall mood during the last week During the past week, how irritable have you been? Life Control During the past week, how much do you feel you have been able to deal with your problems? During the past week, how much control do you feel that you have had over your life? Punishing Responses Ignores me Gets angry with me Solicitous Responses Tries to get me to rest Gets me pain medication Distracting Responses Encourages me to work on a hobby Talks to me about something to take my mind off the pain McGi l l Pain Questionnaire: Sensory Subscale Jumping, Flashing, Shooting Pmching, Pressing, Gnawing, Cramping, Crushing Affective Subscale Tiring, Exhausting Fearful, Frightening, Terrifying Evaluative Subscale Annoying, Troublesome, Miserable, Intense, Unbearable 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0053995/manifest

Comment

Related Items