The Role of Anger in Screening for Prognostic Risk Factors and Subgrouping Low Back Pain Patients

MISSING IMAGE

Material Information

Title:
The Role of Anger in Screening for Prognostic Risk Factors and Subgrouping Low Back Pain Patients
Physical Description:
1 online resource (71 p.)
Language:
english
Creator:
Nisenzon,Anne N
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Psychology, Clinical and Health Psychology
Committee Chair:
Robinson, Michael E
Committee Members:
Pereira, Deidre B
Bowers, Dawn
George, Steven E

Subjects

Subjects / Keywords:
anger -- back -- emotion -- pain -- risk
Clinical and Health Psychology -- Dissertations, Academic -- UF
Genre:
Psychology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
Low back pain (LBP) is a highly common and costly pain condition that often becomes chronic if not properly addressed. Psychosocial symptoms have been shown to complicate LBP, subsequently necessitating more interdisciplinary treatment methods. The present study specifically investigated the role of anger experience and expression in predicting pain intensity and disability in LBP. This study also examined whether anger contributed to other psychosocial prognostic indicators in identifying treatment outcome risk groups using an empirical statistical approach. One-hundred and three LBP patients receiving physical therapy completed a series of psychosocial questionnaires at their baseline visit and after one month of treatment. They were also asked to complete the Subgroups for Targeted Treatment (STarT) Screening Tool to assess for Low, Medium, or High risk group classification. Outcome variables were pain intensity, performance-based disability, and patient-reported disability. Analyses revealed that the three STarT risk groups differed on anger in addition to other psychosocial measures, indicating corresponding levels of negative affect. General psychosocial distress also predicted disability post-treatment, but did not have a strong effect on pain. Empirical subgrouping procedures revealed two patient groups divided on overall psychosocial distress. Subsequent analyses revealed that the group with less psychosocial symptomatology reported less pain and disability at follow-up. There was also a strong treatment effect, indicating both groups benefitted from therapy. The findings from this study suggest that anger is part of generalized negative affect in LBP rather than a unique predictor of treatment outcome. Methodological strengths and limitations are explored to guide future research on the role of anger and negative affect in LBP screening procedures and, ultimately, treatment outcome.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility:
by Anne N Nisenzon.
Thesis:
Thesis (Ph.D.)--University of Florida, 2011.
Local:
Adviser: Robinson, Michael E.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2012-08-31

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2011
System ID:
UFE0043181:00001


This item is only available as the following downloads:


Full Text

PAGE 1

THE R OLE OF ANGER IN SCREENING FO R PROGNOSTIC RISK FACTORS AND SUBGROUPING LOW BACK PAIN PATIENTS By ANNE NOELLE NISENZON A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2011 1

PAGE 2

2011 Anne Noelle Nisenzon 2

PAGE 3

To m y family, for their encouragement, support, and love 3

PAGE 4

ACKNOWL EDGMENTS I sincerely thank my mentor, Dr. Michael Robinson, for his patience, encouragement, wisdom, and support throughout this project, as well as in my inte llectual career. I would also like to thank my Defense Committee, Drs. Dawn Bo wers, Deidre Pereira, and Steven George, for their time in reviewing this study and providing valuable suggestions. Special thanks are extended to my colleagues in the Center for Pa in Research and Behavioral Health for their endless support and enthusiasm. I am particul arly grateful to Laura Wandner, Calia Torres, Jason Beneciuk, and Dr. Steven Ge orge for assisting me in the completion of this project. Additionally, I would lik e to thank the skilled physical therapis ts at all of our recruitment centers who have made this project possible. I must al so express my appreciation to the many patients that gave us their time and energy to participat e in this study. Finally, I thank my friends and family for their support and guidance in reaching this milestone and beyond. 4

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........6 ABSTRACT.....................................................................................................................................7 CHAPTER 1 INTRODUCTION................................................................................................................. ...9 Commonly Addressed Negative Emotions in Pain................................................................10 The Construct of Anger..........................................................................................................12 The Impact of Anger in Pain..................................................................................................13 Etiology of Anger in Pain...................................................................................................... .14 Subgrouping of Low Back Pain Patients................................................................................18 The Potential Role of Anger in Patient Subgrouping.............................................................20 Study Rationale.......................................................................................................................21 Specific Aims and Hypotheses...............................................................................................23 2 METHODS...................................................................................................................... .......25 Participants.............................................................................................................................25 Measures.................................................................................................................................25 Procedures..................................................................................................................... ..........31 Statistical Analyses........................................................................................................... ......32 3 RESULTS...................................................................................................................... .........36 Aim 1: Comparison of STarT Risk Groups on Anger Variables............................................36 Aim 2: The Contribution of Anger Variables to Treatment Outcome....................................38 Aim 3: Empirical Grouping of Low Back Pain Patients........................................................39 Exploratory Aim: Examining Treatment Ou tcome in Relation to Patient Subgroup.............41 4 DISCUSSION................................................................................................................... ......49 APPENDIX A FLOW DIAGRAM OF STUDY DESIGN.............................................................................60 B STarT MEASURE................................................................................................................ ..61 LIST OF REFERENCES...............................................................................................................62 BIOGRAPHICAL SKETCH.........................................................................................................71 5

PAGE 6

LIST OF TABLES Table page 3-1 Descriptive Data........................................................................................................... ......43 3-2 Descriptive Data for STarT Risk Groups...........................................................................44 3-3 Psychosocial Characteristics of STarT Risk Groups.........................................................44 3-4 Correlations between Psychosocial Predictors and Outcome Variables............................45 3-5 The Effect of Anger and Other Psychosocial Variables on PII.........................................45 3-6 The Effect of Anger and Other Psychosocial Variables on RMDQ..................................46 3-7 Descriptive Data for the Two-Gr oup Cluster Solution of LBP Patients............................47 3-8 Psychosocial Characteristic of Cluster Division of LBP Patients.....................................47 3-9 Treatment Outcome Differences Between the Clusters.....................................................48 3-10 Treatment Outcomes by Cluster Me mbership and Evaluation Period...............................48 6

PAGE 7

Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THE ROLE OF ANGER IN SCREENING FO R PROGNOSTIC RISK FACTORS AND SUBGROUPING LOW BACK PAIN PATIENTS By Anne Noelle Nisenzon August 2011 Chair: Michael Robinson Major: Psychology Low back pain (LBP) is a highly common and costly pain condition that often becomes chronic if not properly addressed. Psychosocial symptoms have been shown to complicate LBP, subsequently necessitating more interdiscipl inary treatment methods. The present study specifically investigated the role of anger experience and expressi on in predicting pain intensity and disability in LBP. This study also examined whether anger contributed to other psychosocial prognostic indicators in identifyi ng treatment outcome risk groups using an empirical statistical approach. One-hundred and three LBP patients receiving physical therapy completed a series of psychosocial questionnaires at thei r baseline visit and after one m onth of treatment. They were also asked to complete the Subgroups for Targeted Treatment (STarT) Screening Tool to assess for Low, Medium, or High risk group classifi cation. Outcome variables were pain intensity, performance-based disability, and patient-reported disability. Analyses revealed that the th ree STarT risk groups differed on anger in addition to other psychosocial measures, indicating corresponding levels of negative affect. General psychosocial distress also predicted disabil ity post-treatment, but did not have a strong effect on pain. Empirical subgrouping procedures revealed two patient groups di vided on overall psychosocial 7

PAGE 8

8 distress. Subsequent analyses revealed that the group with less psychosocial symptomatology reported less pain and disability at followup. There was also a strong treatment effect, indicating both groups benefitted from therapy. The findings from this study suggest that anger is part of generalized negative affect in LBP rather than a unique predictor of treatm ent outcome. Methodological strengths and limitations are explored to guide future research on the role of anger a nd negative affect in LBP screening procedures and, ultimately, treatment outcome.

PAGE 9

CHAP TER 1 INTRODUCTION Pain is a multidimensional experience, imp acting individuals physically, mentally, and emotionally. Although historically conceived of as a purely sensor y experience, research in the past 40 years has shown that psychosocial stress and disability accompany pain, thereby leading to interpersonal and intrapersonal problems far beyond the scope of the actual pain condition. Melzack & Casey (1968) were the first to describe the motivational-affective dimension of pain, which entails how one reacts emotiona lly to acute or chronic sensati ons of pain. Since this time, the definition of pain has changed to accommoda te the sensory and emotional experience associated with actual or pot ential tissue damage (Internatio nal Association for the Study of Pain, 1994). Therefore, a major fo cus of the psychologica l contribution to the pain literature has been the assessment and treatment of emotional aspect s of pain and their role in patient quality of life. Given that pain is a nearly ubiquitous c ondition, with approximat ely 25 million people suffering from acute pain and 50 million with chronic pain in the United States, the cost of pain is high. Direct costs of pain, such as dia gnostic assessments and physical and pharmacological treatments, and indirect costs of pain, including lost time from wo rk and disability, are estimated to be $125 billion per year (T urk & Melzack, 2001). Specifica lly, the primary cause of jobrelated disability is back pain, costing Americans upwards of $50 billion annually (Strine & Hootman, 2007). Often present with no known underl ying pathology, back pain is difficult to treat and leads to chronic pain problems in 6080% of cases consulting primary care (Hill, et al., 2008). Thus, primary prevention efforts aimed at identifying prognostic indicators prior to the development of chronic pain syndromes are cruc ial in helping control these costs. Current research on contributing factors to chronic pain development a nd maintenance points towards the 9

PAGE 10

presence of com orbid negative affect, namely depression, anxiety, and, more recently, fearavoidance beliefs. These commonly addresse d psychosocial prognostic indicators will be discussed briefly to elucidate cu rrent methods used in primary car e and physical therapy settings to help guide treatment in low back pain patients. Commonly Addressed Negative Emotions in Pain Depression is the most frequently noted como rbid emotional component of pain, with as many as 30-60% of chronic pain patients in clin ic-based samples reporting depressive symptoms (Banks & Kerns, 1996). Depression in pain has b een associated with reports of greater pain intensity (Williams, Jacka, Pasco, Berk, & Dodd, 2006), overall nega tive mood (Feldman, Downey, & Schaffer-Neitz, 1999), pain-related disability (Tan, Jensen, Thornby, & Sloan, 2008), and poorer treatment outcome (Bair, Robinson, Katon, & Kroenke, 2003). Depression has also been shown to be uniquely associated with pain intensity in low back pain (LBP) patients after controlling for ot her negative emotions, lending some support that negative emotions may have differential effects in the pa in experience (George, Wittmer, Fillingim, & Robinson, 2006). In terms of the temporal relationship between depression and pain, there is evidence to suggest a bidirectiona l effect, such that individual s with premorbid depression may be more sensitive to painful stimuli, and those with a chronic pain condition often experience significant life interference and disability leading to depres sed mood (Tan, et al., 2008). Anxiety is another common negative emotion in pain, and has been related to decreased pain threshold and tolerance in experimental settings (Janssen, Spinhoven, & Arntz, 2004) as well as greater disability in clin ical settings (Tan et al., 2008). Recently, the anxiety literature in chronic pain has focused on the role of anxiety sensitivity and fear-avoidance beliefs in the development and maintenance of chronic pain sy ndromes. Anxiety sensitivity, or heightened physiological arousal to stimuli that may be perceived as threatening, has been shown to 10

PAGE 11

contribute to fear of pain (M artin, McGrath, Brown, & Katz, 2007; W att, Stewart, Lefaivre, & Uman, 2006). Subsequently, fear of experiencing pain following in jury can lead to maladaptive coping styles and further avoida nce of physical activities. Ac cording to the Fear Avoidance Model of Exaggerated Pain Perception (FAMEPP), initially conceived by Lethem and colleagues (1983), pain patients may develop a disuse syndrome through avoi dance of activity, leading to a vicious circle of greater disability, increased pa in severity, and negative affect. This model is quickly gaining support in the physical therapy li terature, which has shown that fear-avoidance beliefs accounted for nearly 20% of the variance in pain-relate d disability in chronic LBP patients (George, Wittmer, et al., 2006). The similar concept of pain catastrophizing, or ruminative thoughts that one cannot to lerate painful situations, has also been linked to increased pain severity, though to a le sser degree (George, Dannecker, & Robinson, 2006). Thus, the presence of anxiety, particularly fear-avoidan ce beliefs and related c ognitive coping styles, is considered a key predictor of an unfavorable prognosis and, therefore, remains an important target in successful physical therapy treatment approaches (George & Zeppieri, 2009). When identifying psychosocial risk factors for the progression of acu te pain to chronic pain cases, comorbid depressi on and anxiety are often named as the main predictors. Furthermore, these were two of the main psycholo gical constructs used by Hill et al. (2008) to classify pain patients into treatment subgroups ranging from those manageable by primary care (i.e., few psychosocial correlates) to those requiring multidisciplinary care. However, there is a growing body of literature on the role of ange r in pain, and how both the experience and expression of anger can impact ones pain pres entation and chronicity. Despite the relative dearth of empirical research on a nger in pain relative to other negative emotions, Okifuji, Turk, & Curran (1999) found that 69% of chronic pain patients endorsed some form of anger. Given 11

PAGE 12

this finding, a closer exam ination of what defines anger as a cons truct, how it affects pain, and, ultimately, how it impacts treatment choices is warranted. The Construct of Anger Anger is a negatively-valenced emotion usually involving a sense of injustice and a desire to have the injustice resolved (Fernandez & Tu rk, 1995). It is comprised of both a cognitive appraisal of the situation and an action tendency, together which de fine the intensity and type of anger ultimately expressed. Like other emotions, th ere is a wide range of anger intensities, from the relatively mild (e.g., frustr ation and annoyance) to the mo re severe cases (e.g., rage), depending on the degree of wrongdoing perceive d by the angry individual and the individuals reactive tendency (Ortony, Clore, & Collins, 1988) The overarching construct of anger may be further branched into the conceptually similar, yet behaviorally disparate constructs of hostility and aggression. Hostility has been defined as an enduring tendency towards making cognitive appraisals of malicious intent of others (G reenwood, Thurston, Rumble Waters, & Keefe, 2003), while aggression is characterized by the action tendency of outward expression (e.g., fighting or damaging property due to angry feelings) (Baron, 1977). Additionally, there is a third concept related to anger labeled passi ve aggression, in which one ex presses anger covertly through interpersonal victimization. These concepts are all inter-r elated and are often used interchangeably in the literatu re, although others argue that th ey capture different affective qualities, and therefore should be assessed as different cons tructs (Fernandez & Turk, 1995). Within the construct of anger, the clinical literature addresses the distinction between state and trait anger, describi ng ones anger experience over ti me, and between anger-in and anger-out, which refers to anger self-regulation styles. First addresse d by Spielberger et al. (1983), state anger is a transitory feeling of anger that arises due to a specific occurrence, whereas trait anger refers to more stable pers onality characteristics. Trait anger is often 12

PAGE 13

considered to be synonymous with hostility in th e literature, particular ly if accom panied by antagonistic behavior (Fernand ez & Turk, 1995). When anger is experienced, individuals may attempt to regulate their emotions either thr ough suppression, commonly la beled anger-in, or outward expression, or anger-out. In terms of th e role of anger in the pain experience, the literature is dominated by how tr ait anger and anger-in versus ange r-out interplay to impact pain intensity, disability, and treatment outcome. The Impact of Anger in Pain As anger has received relatively little attention in the literature in comparison to the role of other negative emotions in pain, it is important to emphasize the unique contribution of this emotion to suffering in the pain experience. In general, the concept of suffering in pain has been used as the overarching term to descri be the depression, anxi ety, and anger patients experience as part of the motivational-affective component. Furthermore, some studies suggest that these separate negative emo tions should be considered one cons truct as they tend to load as one factor in principle component analys es (Hirsh, Waxenberg, Atchison, Gremillion, & Robinson, 2006) or have highly correlated values (Gaskin, Greene, Robinson, & Geisser, 1992). Also, from a theoretical standpoint, it has been suggested in the psychodynamic orientation that depression is actually anger turn ed inwards. While this theo ry has not been empirically validated, there is some evidence in the pain literature to suggest that anger-in accounts for a significant proportion of variance in depression in chronic pain pa tients, thereby demonstrating some link between these constr ucts in pain (Tschannen, Duckro, Margolis, & Tomazic, 1992; Wade, Price, Hamer, Schwartz, & Hart, 1990). Despite the interconnectivity of negative emotions frequently found in chronic pain, there is sufficient evidence to suggest that anger imp acts the pain patient when controlling for other emotions (Bruehl, Chung, & Burns, 2006). For in stance, Bruehl et al., (2002) found that chronic 13

PAGE 14

LBP patients with increased anger-out reported gr eater pain intensity during an experim ental ischemic pain procedure even af ter controlling for de pression, despite the two emotions being significantly correlated. Anothe r study involving negative emo tion induction paradigms during an experimental pain procedure found that high anger expression was unrelated to general emotional expressivity, but was related to in creased physiological arousal and pain (Burns, Quartana, & Bruehl, 2007). Importantly, anger seem s to be particularly relevant to low back pain. Burns (2006) found that ch ronic LBP patients reported great er pain intensity, greater blood pressure reactivity, and slower recovery time during an anger induction than a sadness induction, even when statistically controlli ng for the effects of other negativ e affects. Thus, although it is undeniable that anger co-occurs with other nega tive emotions, they may still be differentially related to pain, and, therefore, may have di fferent implications for pain treatment. Etiology of Anger in Pain There are several theories as to why there is such a high prevalence of anger in pain, which describe neurobiological cognitive, and psychosocial mechanisms behind their association. The common thread of these theories deals more w ith the way anger is expressed than with the intensity or frequency of ones anger. For instance, the ironic processes model of anger suppression and pain, proposed by Burns, Quarta na, and Bruehl (2007), suggests that anger suppression leads to greater pain intensity. This theory is based on Wegners ironic process theory of mental control (Wegner, 1994), which essentially states that any conscious attempt to suppress unwanted thoughts will cause the individua l to devote more attention to the suppressed thought through automatic cognitive monitoring processes. Thus in the context of pain, suppression of experiential anger causes the automatic monitoring system to focus in on more instances of anger and frustration related to the pain. This theo ry has been empirically supported in several studies showing that individuals who were asked to consciously suppress their anger 14

PAGE 15

during an anger-induction proce dure reported greater pain severity (Quartana & Burns, 2007) and greater systolic blood pressure reactivity (Burns, et al., 2007) to an acute pain condition than those who were allowed to express their anger. A neurobiological approach to explaining the connection between anger and pain is the endogenous opioid dysfunction theory, which suggest s that a more expres sive style of anger regulation is associated with el evated pain intensity (Bruehl, Chung, & Burns, 2006). According to this model, the relationship between anger-out and pain sensitivity is mediated by impaired overlapping opioid inhibitory systems which serv e to modulate pain and emotional regulation. Furthermore, this theory postu lates that high levels of a nger-out trigger endogenous opioid release for those with dysfunc tional systems, thereby making anger expression temporarily adaptive in emotional and physical well-being. However, over time, frequent anger expression may lead to a stressful, unsupportive environment for the individual, further establishing his or her opioid debt and subsequent difficulty in ma naging pain. There is substantial empirical evidence for this theory from studies that reveal that patients endorsing high anger expressivity do not experience increased pain severity duri ng an experimental pain procedure following a naloxone (i.e., opioid block) injection as co mpared to a saline in jection, indicating a nonresponsive endogenous opioid system (Brueh l, Burns, Chung, & Quartana, 2008; Bruehl, Chung, Burns, & Biridepalli, 2003). There is also research s howing the relationship between elevated anger-out and increased pain sensitivity in acute clini cal pain (Bruehl, Chung, Donahue, & Burns, 2006; Voulgari, et al., 1991) and in chronic low back pain (Bruehl, et al., 2002; Kerns, Rosenberg, & Jacob, 1994; Lombar do, Tan, Jensen, & Anderson, 2005). Seemingly, the ironic proce ss theory of anger suppressi on in pain and the endogenous opioid dysfunction theory contradict on whether elevated anger-in or anger-out is more likely to 15

PAGE 16

incite or aggravate pain conditions. However, th ey both describe the deleterious effects of the inappropriate regulation of high trait anger on pain. Following this pa ttern, the state-trait matching hypothesis of anger expres sion states that those with hi gh anger-out regulation style can experience a reduction in anger arousal and resulting nega tive physiological effects through behavioral expression of the emotion (Engebret son, Matthews, & Scheier, 1989). Furthermore, this theory posits that a trait x situation mismatch such as suppr ession of anger in high anger-out individuals or, conversely, forced anger expression in anger-in individuals, would lead to increased pain sensitivity in both acute and chronic pain cases. The state-trait matching hypothesis is mostly associated w ith the cardiac literature, which has shown that individuals with matching trait anger and an expressive anger regulation style experien ce improved post-stress blood pressure recovery (Faber & Burns, 1996) whereas the tendency to suppress anger is associated with carotid stiffening (Anderson, Metter, Hougaku, & Najjar, 2006). Similar findings have been demonstrated in the pain lite rature, showing analgesic effects of verbal and behavioral anger expression in chronic pain patients with high anger-o ut tendencies (Burns, Kubilus, & Bruehl, 2003) and increased pain se nsitivity when anger was suppressed (Burns, et al., 2007). Interestingly, the emotion of anger, as opposed to other negative emotions, is singularly evoked in this line of research, perhaps due to the greater degree of physiological reactivity accompanying anger. Thus, the tre nd of research on the relationship between anger and pain seems to be less focused on the presence or absence of this emotion, but on how it is managed by the pain patient. Finally, there is some speculation as to the impact of anger expression on patient-provider relationships. Studies have shown that pain pa tients with high anger-out appear hostile to physical and occupational therapists, thereby ev oking negative emotional responses from these 16

PAGE 17

care providers and generally alienating these wo rking alliances (Sm ith & Zimny, 1988). Burns, Higdon, Mullen, Lansky, & Wei (1999) found that therapists reported their poorest working relationships with patients who were both depr essed and expressed anger. Furthermore, poor patient-provider relationships have been shown to be predictive of worse compliance with treatment, which may ultimately affect treatment outcome (Sluijs, Kok, & van der Zee, 1993). This effect seems to be particularly pertinent to men, as studies on low back pain show that men who score high on anger-out measures are more li kely to have higher levels of pain-related disability, more difficulties in establishing th erapeutic relationships, and poorer functional outcomes when compared to women (Burns, Johnson, Devine, Mahoney, & Pawl, 1998; Greenwood, et al., 2003). Thus, although suppression of anger has been show n in several studies to be related to increased pain severity, outwa rd anger expression may also indirectly impact ones pain condition through crea ting poorer working alliances w ith health care providers. Given the plentiful empirical and theoreti cal evidence supporting the role of anger and other negative emotions in the development and maintenance of pain conditions, it is important to examine how they impact different aspects of the pain condition, namely pain intensity and disability. Disability is often a difficult construct to define, as it is composed of several factors, including performance-based functional limitations, psychological issues that compound limitations, and societal influences that may main tain or reject disabil ity (Bair, et al., 2003; Gesztelyi & Bereczki, 2006; Truchon, 2001). Thus, a thorough assessment of pain and disability, both performance-ba sed and self-reported ratings, become vitally important in understanding how patients a nd their providers view thei r pain and prognosis. Among common pain conditions, low back pain is often the most targeted syndrome for disability and treatment outcome research due to its high diagnos tic frequency (an estimated 65% 17

PAGE 18

1-year p revalence rate in the U.S.) and resulting exorbitant direct and i ndirect costs (Walker, 2000). Additionally, although only 3-10% of acute b ack pain patients go on to develop a chronic pain condition, these individuals consume appr oximately 75-80% of the recourses, further highlighting the importance of early screen ing measures (Nachemson, Waddell, & Norlund, 2000). Current methods used to elucidate risk factors for poor treatment outcome in pain patients, specifically those with LBP, will be discussed as well as identified prognostic indicators. Most importantly, the potential contributing role of anger in this line of research will be explored. Subgrouping of Low Back Pain Patients The concept of patient subgrouping has become an extensively researched area for treating patients with low back pain. Patient subgrouping involves the implementation of evidence-based treatment classification systems that identify patients who ma y benefit from more targeted approaches designed for their specific constella tion of signs and symptoms (Delitto, Erhard, & Bowling, 1995). Prior to this method, the trad itional medical model emphasized use of pathoanatomical symptoms to formulate treatm ent; however, due to the heterogeneity of LBP cases and contributing and maintaining factors, this method was not shown to be effective (Fritz, Cleland, & Childs, 2007). In physical therapy se ttings, the development of multivariate clinical prediction rules (CPR) has signifi cantly helped guide identificat ion of treatment subgroups for LBP patients (Delitto, et al., 1995; Fritz & George 2000). CPRs identify several factors that predict pain treatment response, such as pain duration, pain location, and spinal mobility. They are then used to classify LBP patients into one of 4 treatment categories: manipulation, stabilization, specific exercises, or traction. CPRs have been found to be useful in treatment matching procedures and, ultimately, in predicti ng treatment outcome (Fritz, Delitto, & Erhard, 2003). 18

PAGE 19

However, a m ajor limitation to currently publis hed CPRs involving interventions is that they do not incorporate a thorough psychosocial assessment during the developmental phase, and are not conclusive in terms of the potential treatment approaches available for these patients (Beneciuk, Bishop, & George, 2009). For example, studies have shown that inclusion of a cognitive-behavioral intervention for fear-avoidan ce beliefs related to LBP has been helpful in treatment outcome (George, Fritz, Bialosky, & Donald, 2003). Furthermore, there is a large body of empirical research using psychological and affective predic tors to distinguish subgroups of pain patients. For example, the Minnesota Multiphasic Personality Inventory (MMPI-2) has been repeatedly used to identify psychological and affective factors that group chronic pain patients according to pain coping styles (Riley & Robinson, 1998), risk for disability (Gatchel, Mayer, & Eddington, 2006) and surgical treatm ent outcome (Riley, Robinson, Geisser, Wittmer, & Smith, 1995). These studies have provided substantial evidence fo r the usefulness of including psychosocial factors, especially those assessing negativ e emotions, when attempting to empirically define subgroups of patients for further treatment plans. There have been a variety of screening tool s developed to subgroup pa in patients in order to guide treatment (Dionne, et al., 2005; Duijts, Kant, Landeweerd, & Swaen, 2006; Truchon & Cote, 2005; Westman, Linton, Ohrvik, Wahlen, & Leppert, 2008). However, these measures vary substantially in terms of methodological quality, conceptual pur poses, and the normative samples pain conditions (Hill, et al., 2008). To date, there has been only one attempt at creating a comprehensive, yet brief screening tool for prognostic indicators in no n-specific LBP patients that includes both physical and psychosocial scales. The Subgroups for Targeted Treatment Back Screening Tool (STarT), developed by Hill and colleagues (Hill, et al., 2008) is a brief measure designed to be used in primary care se ttings to delineate subgroups of nonspecific back 19

PAGE 20

pain patients according to their risk of deve loping a more chronic pain condition. Through a non-systematic literature search a nd a clinical advisory panel of primary care pain specialists, these authors identified several negative prognos tic indicators for continued pain, including the psychosocial constructs of fear-avoidance, a nxiety, pain catastrophizi ng, and depression. To discriminate between low-risk patients (e.g., those who would be nefit from basic primary care pain management) and high-risk patients (e.g., those who required both physical and cognitivebehavioral treatments for a favorable outcome), predictive validity was calculated using univariate receiver operating charac teristic (ROC) curve statistical analyses. While the STarT tool has shown promise in identifying risk s ubgroups of LBP patients for further treatmentmatching (Hill, et al., 2008), there are several important caveats of this measure. First, the subgroups of LBP patients were not derived usin g multivariate procedures, thereby not taking into account the interrelations hips of the variables used in the models. Secondly, the psychosocial subscale of the sc reening measure may have been under-specified, given that the construct of anger or other poten tially important factors were not included. Thus, the STarT tool for screening prognostic indicato rs in LBP patients may requi re further development via validation studies to ascertai n its clinical util ity in different patient settings. The Potential Role of Anger in Patient Subgrouping In accordance with the limited empirical litera ture focused solely on anger in acute and chronic pain, there have also been just a few st udies examining this emotions role in patient subgrouping in terms of treatment response. Kinder, Curtiss, & Kalichman (1986) found a unique effect of trait anger as a suppressor variab le for male chronic pain patients, such that those who endorsed high trait anger had lo wer hysteria (Hy), depression (D), and hypochrondriasis (Hs) scale scores on the MMPI-2 than those who were lower on trait anger. Following theories supporting the role of suppressed anger in pa in exacerbation, these authors 20

PAGE 21

inte rpreted these results to indicate that those le ss willing to express thei r anger tend to exhibit the neurotic triad profile, whereas thos e who acknowledge their anger appear less psychoneurotic. It may be noted that the neurotic tria d personality profile has not been found to be particularly predictive of poor treatment outcome in pain patients when compared to other high negative affect profiles (Rile y, et al., 1995). Thus, more re search must be done to uncover the functional relevance of ange r experience and expression in th e scope of ones psychosocial presentation, and how it impacts pain and disability in patients. Despite the dearth of empi rical studies examining the role of anger in patient subgrouping, there is sufficient evid ence to suggest that it is a wo rthwhile investigation. Trait anger has been associated with perceived disabi lity and negatively rela ted to activity levels, independent of other negative emotions (Kerns, et al., 1994; Okifuji, et al., 1999). Furthermore, Burns et al., (1998) demonstrated that male LB P patients with high ange r-out levels showed lower improvement on lifting capacity following tr eatment than patients who did not endorse anger. However, another study performed by this group of researchers found that suppression of hostility leads to greater pain severity and activity interference (Burns, Johnson, Mahoney, Devine, & Pawl, 1996). These findi ngs suggest that anger in its various forms has clinical as well as theoretical importance in the development of functional difficulties for pain patients, and therefore, may be informative in trea tment-matching classification systems. Study Rationale The current study, which aimed to identify the added utility of anger in screening and subgrouping LBP patients according to their prog nostic indicators, is important for several reasons. First, this study is one of the very few examining anger in a wide range of LBP patients, and how it may contribute to treatment outcome. As previously stated, low back pain is a leading cause of disability, and, as of now, it is unclear what modi fiable physical and 21

PAGE 22

psychosocial factors play a role in the m aintena nce of this condition. Although anger is a unique part of the negative affect spectrum pain pati ents endorse, found both an ecdotally in clinical settings and in empirical rese arch, it has never before been included in patient subgrouping studies. Furthermore, the current research on wh ether expressive or ex periential anger is the main factor related to pain is mixed; this study aimed to clarify this issu e in terms of identifying risk factors for continued disability. Another novel aspect of this study is the e xploration of the impact of anger on treatment in a physical therapy setting. Presently, there is abundant evidence for the role of fear-avoidance beliefs in the development of chronic pain fo llowing injury (George, Dannecker, et al., 2006; Vlaeyen & Linton, 2000), but there ha ve not been any studies on anger in this respect. This is rather surprising given the resear ch showing a negative impact of expressive anger on working alliances between pain patients and health care pr oviders (Burns et al., 1999), and the importance of the patient-provider interactio ns in adherence to physical th erapy exercises (Sluijs, et al., 1993). As physical therapy has been shown to significantly improve low back pain, it is important to consider all psychos ocial factors shown to affect pain, particularly when screening for treatment-matching and creating new treatment systems. Finally, the current study is uni que in that we have used empi rical rather than theoretical methods to subgroup pain patients when examining the utility of the screening tool. The STarT measure, which was used in the present study to initially define risk groups, describes three subgroups of patients that are delineated based on heuristic procedures. However, many subgrouping studies emphasize the us e of hierarchical cluster analys es to identify mathematically different groups of patients based on the exam ined variables, thereby providing more sound evidence for the division of pain patients in to specific groups (Burns, Kubilus, Bruehl, & 22

PAGE 23

Harden, 2001; Riley & Robinson, 1998). Thus, the current study aim ed to improve upon the methods described by Hill et al. (2008) to validat e the STarT tool in two ways: creating a more specified model of psychosocial variables in LBP with the inclusion of a nger, and using more a rigorous methodology to define patient risk subgroups. Specific Aims and Hypotheses The specific aims for the current study were as follows: 1) To examine whether the three risk groups of LBP patients, as defi ned by the STarT tool (i.e., low risk, moderate risk, high risk for poor treatment outco me and subsequently sustained pain conditions), differed on a measure of ange r in a similar pattern to other psychosocial constructs used in the STarT measure (Hill et al., 2008). Sub-aim: To explore whether the 3 risk groups of LBP patients differed on anger regulation styles. 2) To examine whether anger contributed to predicting risk for poor treatment outcome in LBP patients after 1 month of physical therapy treatment. 3) To examine whether the addition of an anger measure to the current psychological constructs used in the STarT tool affected empirical subgrouping of patients in terms of identifying negative prognostic indicators. Exploratory sub-aim: To explore whether empirically derived cl usters of low back patients could be used to predict their 1 m onth physical therapy treatment outcome. Hypotheses based on the above aims were: 1) We predicted that those who were classi fied as being high risk on the STarT tool would have higher anger scores than the lo w and moderate subgroups of LBP patients. 23

PAGE 24

24 Sub-aim: We predicted that patients in the high risk subgroup would be more likely to score high on trait anger and low on anger-out measures, indicating di ffering anger regulation styles from their anger experience. 2) We predicted that patients endorsing hi gh trait anger would show less improvement on the physical examination and grea ter self-reported disa bility and pain leve ls at the 1 month follow-up appointment than those with low trait anger. 3) We predicted that the empirical subgroupi ng of LBP patients with the included anger measure would be similar to that achieved by Hill et al. (2008). Thus, we predicted that there would be 3 clusters of LBP patients, showing lo w, moderate, and high risk for continued pain following treatment, divided by levels of physic al and psychosocial prognostic indicators Exploratory sub-aim: We hypot hesized that the clusters of LBP patients created during their baseline measurement would be useful in predicting treatment outcome at their 1-month follow-up physical therapy appointment. Thus, if we were to identify separate clusters of patients based on the prevalence of negative pro gnostic indicators includi ng anger, we predicted that those in the higher risk group would show le ss improvement and greater disability and pain than those with low risk af ter one month of treatment.

PAGE 25

CHAP TER 2 METHODS Participants Patients with a current diagnosis of LBP who were referred for physical therapy services were recruited for the present study. Patients were recruited from severa l outpatient physical therapy clinics throughout Nort h and North-Central Florida, including the Orthopaedics and Sports Medicine Institute (OSMI) and Shands Re hab Center at Magnolia Parke in Gainesville, Florida, and several Brooks Rehabi litation Centers in Jacksonville, Florida. Inclusion criteria for this study were: (1) adults aged 18 years or older, (2) a referral to outpatient physical therapy for LBP with or without radiating symptoms of any duration, and (3) the abi lity to read and speak English fluently. Exclusion criteria for the study were: (1) physical or psychological disorders related to metastatic disease, vis ceral disease, or fract ure, and (2) osteoporosis. All participants were required to provide signed informed cons ent prior to enrolling in the study. Financial compensation served as an incentive for participation. Measures The following self-report psychological questi onnaires, pain assessments, and clinicianadministered physical impairment measures were give n to all participants at their initial physical therapy appointment. These measures were selected because they assess factors often associated with negative pain treatment outcomes. It may be noted that this study is part of a larger study examining the validation of the STarT tool in physical therapy settings, which warranted the administration of additional physical tests and psychological measures during the evaluation. However, as they are not pertinent to the curren t investigation and related hypotheses, they will not be described in detail here, but may be viewed in Appendix A. 25

PAGE 26

Demographic and Clinical Ch aracteristics Questionnaire A questionnaire was administ ered eliciting information pertaining to the participants age, gender, race, years of education, marital status, and employment status. Additionally, clinical characteristics of the patients pain condition were obtained, including duration and history of pain symptoms and surgical procedures. Subgroups for Targeted Treatment (STarT) Back Screening Tool The STarT is a 9-item measure used to scre en for back pain prognostic indicators with the intention of ultimately aiding treatment decisi on-making in primary care settings. An Overall score is achieved by summing items 1-9, and a Psychosocial subscale is derived by summing the scores of the bothersomeness, fear, catastr ophizing, anxiety, and depression items (Items 1, 4, 7, 8, and 9) (Appendix B). Both the overall scor e and the psychosocial subscale are used to subdivide LBP patients into low, moderate, and high risk groups, based on their likelihood for a favorable treatment outcome in primary care. The authors described the Low risk group (Overall tool score 0-3) as t hose with few negative prognostic i ndicators who would benefit from standard primary care practices (e.g., analgesia, education). The Moderate risk group (Overall tool score >3, but Psychosocial subscale score <4) tended to endor se higher levels of physical prognostic indicators but minimal presence of psychos ocial factors, and woul d, therefore, benefit from physiotherapy. Finally, the High ri sk group (Psychosoc ial subscale score 4) was found to benefit the most from comb ined intensive physical and psyc hological approaches. The full measure is presented in Appendix B. The STarT demonstrated adequate test-retest reliability for both the overall tool score (kappa= 0.73) and the psychosocial subscale (k appa= 0.69). Cronbachs alpha ranged from 0.74 to 0.79, indicating adequate intern al consistency. The measure also demonstrated adequate 26

PAGE 27

predictive validity, with 78.4% of patients identif ied as being high risk for having poor disability outcome at their 6-month follow-up appointment. Furthermore, discriminant validity values for the measure were considered outstanding for identifying risk for disability (AUC= 0.92). However, it may be noted that the utility of the STarT tool may be limited by the lack of systematic review in item selection, thereby pote ntially necessitating furthe r specification of the model. Pain Intensity Numerical Rating Scale (NRS) Patients were asked to rate their pain intens ity using a numerical rating scale, anchored from 0 (No pain sensation) to 10 (Worst pain sensation imaginable). In addition to current pain intensity, patients were aske d to provide their best and wors t pain intensity levels over the past 24 hours. Numerical rating sc ales for measuring pain intensity have been shown to have high convergent validity, to be sensitive to treatments, and ar e easy to administer (Jensen & Karoly, 2001). Pain Bothersomeness Patients were asked to rate the bothersomene ss, or unpleasantness, of their LBP with the question, In the last week, how bothersome has your low back pain been? using one of the following responses: Not at all, Slightly, M oderately, Very Much, or Extremely. A single question assessing pain bothersomeness has been shown to be significantly related to pain intensity, and had 80% sensitivity in identifying patients with work-related absence and health care consultations in a 6-m onth period (Dunn & Croft, 2005). State-Trait Anger Expression Inventory (STAXI) The STAXI is a 44-item questionnaire designe d to measure several dimensions of anger experience and expression. Participants were as ked to indicate their agreement with statements 27

PAGE 28

expressing present feelings of anger (e.g., I am burned up.), endorsem ent of angry traits (e.g., I am a hotheaded person.), and actions taken when angry (e.g., I strike out at whatever infuriates me.) on a 4-point Likert scale. Six separate dimensions may be derived from the STAXI: State Anger (STAXI-S) (i.e ., intensity of anger at the time of testing); Trait Anger (STAXI-T) (i.e., dispositional anger); Anger-i n (AX/IN) (i.e., tendency to suppress angry feelings); Anger-out (AX/OUT) (i.e., tendency to express angry feelings); Anger Control (AX/CON) (i.e., frequency of attempts to cont rol anger); and Anger E xpression (AN/EX) (i.e., general index of ange r expression). For the purposes of this study, we examined th e Trait Anger, Anger Control, and Anger-in and Anger-out scales, as previous research indicates that these di mensions have a significant and independent association with pain intensity and chronicity (Bruehl, Chung, & Burns, 2006; Fernandez & Turk, 1995). The STAXI boasts high internal consistency, with Cronbach alpha scores of State-Trait Anger scales ranging from = .70 to .93, and from .73 to .80 for Anger Expression scales. Additionally, co nstruct validity studies indicate a strong relationship between STAXI-Trait and measures of hostility, su ch as the MMPI Hostility subscale ( r= .59) (Spielberger, 1988). Fear Avoidance Beliefs Questionnaire (FABQ) The FABQ is a 16-item questionnaire designe d to assess fear-avoidance beliefs regarding physical activity and work in pain patients, part icularly those with low back pain conditions. Patients were asked to rank endorsement of beliefs (e.g., Physical activity may harm my back) on a scale of 0 (Completely disagr ee) to 6 (Completely agree). Items on the FABQ are divided into Work and Physical Activity subscales to as sess fear-avoidance beliefs in these two domains separately. For the current study, the FABQ Work subscale was chosen for analyses due to its 28

PAGE 29

dem onstrated association with cu rrent and future disability in patients with LBP (Fritz, George, & Delitto, 2001; Waddell, et al., 1993). The FABQ has demonstrated high levels of internal consistency (Cronbachs alpha= 0.88 ) and test-retest reliability ( r= 0.95) (Waddell, Newton, Henderson, Somerville, & Main, 1993). Spielberger State-Trait An xiety Inventory (STAI) The STAI is a 40-item measure assessing both transient and long-standing anxiety in adults. It is composed of two 20-item scales, one examining st ate anxiety, or how anxious the respondent is in the given moment, and one for trait anxiety, or ones dispositional anxiety levels. For the purposes of this study, only trait anxiety was measured, as it is more closely related to disability following pain (H adjistavropoulos, Asmundson, & Kowalyk, 2004). Participants were asked to indi cate their agreement with various statements (e.g., I worry too much over something that really doesnt matte r.) on a 4-point Likert scale ranging from 1 (Almost Never) to 4 (Almost Always). The STAI-Trait has demonstrated high internal consistency (Cronbachs alpha= .90) and has been used extensively in research and clinical practice. Patient Health Questionnaire9 (PHQ-9) The PHQ-9 is a 9-item measure used to measure depression in medical settings. Participants were asked to rate the frequency with which they experience each of the 9 DSM-IV criteria for clinical depression on a 4-point Likert scale, ranging from 0 (Not at all) to 3 (Nearly everyday). The PHQ-9 has demonstrat ed strong test-rete st reliability ( r= 0.84) and internal consistency (Cronbachs alpha = .89) when used in primary care settings, and is considered a valid measure for depression in clinical sa mples (Kroenke, Spitzer, & Williams, 2001). 29

PAGE 30

Roland-Morris Disability Quest ionnaire (RMDQ) The RMDQ is a 24-item measure used to asse ss functional status of patients with LBP (Roland & Morris, 1983). Patien ts were asked to select st atements describing different functional limitations associated with LBP that apply to their current phy sical abilities (e.g. I sleep less because of the pain in my back.). The RMDQ has been shown to have good criterionbased construct and discriminant validity (B aldwin, Butler, Johnson, & Cote, 2007), as well as well-established intern al consistency (Cronbachs alpha = 0.84 to 0.93) (Roland & Fairbank, 2000). The RDMQ was chosen over other disability measures, such as the Modified Oswestry Disability Questionnaire (Fai rbank, Couper, Davies, & O'Br ien, 1980; Fritz & Irrgang, 2001) because it has been shown to be a more sensitive measure in populations with lower levels of disability (Roland & Fairbank, 2000). As LBP pa tients were recruited fr om outpatient physical therapy centers, they are more likely to have low to moderate levels of physical disability. Physical Impairment Index In addition to the above psychosocial measur es and pain assessment tools, a Physical Impairment Index (PII) was used to establish a clinician-guided measurement of physical impairment in the LBP patients. The PII cons ists of 7 physical examination tests routinely implemented in a physical therapy examination for patients with low back pain. Each test is scored as being either positive or negative based on published cut-off values (Waddell, Somerville, Henderson, & Newton, 1992). The overall PII score ranges from 0-7, with higher scores indicating greater levels of physical impa irment. Adequate reliability has been reported for individual items of the PII and convergent va lidity has been supported via correlations with disability in patients with chr onic low back pain (Waddell, et al., 1992) and acute low back pain 30

PAGE 31

(Fritz & Piva, 2003). The 7 physical functioning te sts included in the P II are: (1) flexion range of motion, (2) extension range of motion, (3) lateral flexion range of motion, (4) straight left raise range of motion, (5) spinal tenderness, (6) bilatera l active straight leg raise, and (7) active sit-up. All baseline physical examinations we re performed by trained, licensed physical therapists. Procedures All research procedures took place within the confines of one of the physical therapy clinics employed in this study. Licensed physical therapists we re requested to ask potential participants whether they woul d like to take part in the st udy at the time of their initial evaluation. The licensed physical therapists th en obtained informed consent from all study participants via signed documentation on an informed consent form, which includes information about the study procedures, duration of the study, and possible risks and benefits of participation. Following informed consent procedures, particip ants were asked to complete a packet of intake questionnaires and undergo a routine phys ical examination, the details of which are outlined above. Completion of all included m easures and the physical examination was, on average, 2 hours in duration. Patients were then provided with a 4-week treatment plan by their physical therapist based on their individual sy mptoms. Importantly, treatments were not standardized in this study, and there were no experimental contro ls or randomization procedures included. The physical therapist involved in th e treatment of an indi vidual study participant determined the appropriate inte rventions to be administered based on his or her professional opinion, which is typical in physical therapy pr actice. Essentially, treatments were not standardized and a control group was not employed in this study so as to not compromise the standard of care for the LBP patients, following ethical practice guidelines. 31

PAGE 32

After four weeks, patients were asked to return to their clinic f or an abbreviated follow-up assessment. Adherence to the patients individualized treatment plans was not necessary for reevaluation in this study. Specifical ly, patients were re-administered the Physical Impairment Index, Patient Satisfaction meas urement, ODQ, RMDQ, PCS, FA BQ, STAXI, and pain intensity and bothersomeness evaluations (Appendix A). All data collected was stored in a secure electronic database. Notably, this study only inte nded to address measurements taken during the baseline assessment and at the 4-week follow-up a ppointment. Participants were also asked to attend an additional 6-month follow-up appointment, although data collected from this session is beyond the scope and purpose of the current study. Statistical Analyses 1) Power analyses. Power analyses were conduc ted to determine the number of participants needed to detect a sizable effect when using th e STAXI subscales as the key outcome variables. It may be noted that, due to the different anal yses required to test th e hypotheses of this study, several methods were used to es timate the needed sample size. The final sample size was based on the most conservative value achieved in order to maximally power the study. Kerns, Rosenberg, & Jacobs (1994) f ound a unique effect size of R2= .41 and power over 0.95 when examining the impact of anger-in on pain interference in chronic pain patients. A similar effect size (R2= .39) was achieved by Nicholson, Gramling, Ong, & Buenaver (2003) when examining differences between anger levels on the STAXI in headache patients vers us healthy controls, after controlling for depression and anxiety. Th ese findings suggest that with a power set at 0.80 and an alpha value of 0.05, a total sample size of 54 subjects would be sufficient to obtain a similar effect size for group differences. It further shows the general sensitivity of the STAXI measures to group and treatment effects. Howe ver, to determine patient subgroups (Specific Aim 3), a sample size of N= 80 to 100 was needed to allow for cluster analyses to be built with 9 32

PAGE 33

variables (5 psychological variables represented in the STarT tool and 4 anger subscale scores), with no less than the suggested 5-10 subjects pe r variable. Thus, recruitment of 80-100 fulfilled statistical requirements needed to perform a cluster analysis, as well as adequately power the study based on previous work using the STAXI to examine the relationship between anger and pain. 2) Descriptive statistics. Means and standard deviations were cal culated for continuous demographic and clinical variables (e.g., age, years of education, and pain duration), and frequency tables were establis hed for categorical demographi c and clinical variables (e.g., gender, race, marital status and employment status). 3) To examine whether the heterogeneous samp le of LBP patients differed on measures of anger (Specific Aim 1), we first divided our sample into 3 groups: Low risk, Moderate risk, and High Risk, based on their STarT screening tool scores. One-way analyses of variance (ANOVAs) were then used to compare groups on their anger experience a nd regulation styles, as measured by the trait anger (STAXI-T), anger-in (AX/IN), and anger-out (AX/OUT), and anger control (AX/CON) subscales of the STAXI. Additionally, scores of other mood measures that have been shown to impact pain prognosis, namely the PHQ-9, STAI-T, and FABQWork Scale (FABQ-W), were added as covari ates to investigate whether a nger was a unique contributor to outcome risk grouping, separate from depressi on, trait anxiety, and fear-avoidance beliefs, respectively. 4) To examine whether anger uniquely contri buted to treatment outcome in LBP patients after treatment (Specific Aim 2), a multiple regr ession analysis was conducted with the anger subscale scores (STAXI-T, AX/IN, AX/OUT, an d AX/CON) and other mood measure scores (PHQ-9, STAI-T, and FABQ-W) as the independen t variables, and the patients Physical 33

PAGE 34

Im pairment Index (PII) score at their one month follow-up appointment as the dependent variable. Additionally, multiple regressions us ing these dependent variables were conducted with patients self-reported disability levels on the RMDQ as the independent variable in order to further understand the relationship between anger and patient perceived disability levels. Parallel analyses were also conducted with pa tients self-reported pa in intensity as the independent variable. Regression diagnostics were performed to assess for multicollinearity between the predictor variables on all multiple regression analyses. 5) To examine whether the addition of anger to the current psychosocial constructs used in the development of the STarT tool affects empi rical subgrouping of patients (Specific Aim 3), a hierarchical agglomerative clus ter analysis (Wards method, s quared Euclidian distance) was employed. The 5 psychosocial items used in the STarT tool (bot hersomeness, pain catastrophizing, fear-avoidance beliefs, trait anxiety, and depre ssion) were represented by the STarT Bothersomeness item, total PCS score, total FABQ-Work Scale score, STAI-Trait score, and PHQ-9 Total Score, respectively. From th e STarT measure, only the Bothersomeness item could be entered as an independe nt variable in the cluster anal ysis, because the rest of the individual psychosocial items on the STarT ar e not continuous variables. However, the additional psychosocial indicators used in STarT were derived from similar measures as the ones employed in this study, and therefore, may be acc urately captured by their total scores (Hill, et al., 2008). Additionally, the 4 anger subscales of the STAXI (STAXI-T, AX/IN, AX/OUT, and AX/CON) were entered into the cl uster analysis, totaling 9 variab les. The number of clusters retained were determined through assessing the ch ange in agglomerative coefficients at any given step, with substa ntial increases denoting cluster fo rmation (Hair, Anderson, Tatham, & Black, 1995). Following cluster analyses, cl assification accuracy was evaluated using a 34

PAGE 35

discrim inant function analysis. We also performe d a chi-square test to compare the categorical composition of the patient subgroups achieved by H ill et al. (2008) with that derived from our cluster analysis. 6) To examine whether the empirically-derived subgroups of LBP patients were useful in predicting treatment outcome at their follow-up evaluation (Explorator y Sub-aim), Repeated Measures ANOVAs (rmANOVAs) were conducted. The within-subjects variable in this analysis was time (i.e., baseline measurement vs. one mont h follow-up) and the between-subjects variable was group (i.e., patient cluster placement). Separa te rmANOVAs were conducted for the (1) PII, (2) RMDQ, and (3) pain intensity level as the dependent variable, indi cating performance-based impairment, patient-perceived physical impairment, and pain intensity. 35

PAGE 36

CHAP TER 3 RESULTS A total of 106 LBP patients were recruited for the study, with 103 LBP patients fully completing the baseline protocol, and 87 part icipants (84.47%) fully completing the 4-week follow-up protocol, either through mail-in ( N= 13) or on-site ( N= 74) evaluations. There were no differences found between methods of follow-up on outcome variables of pain intensity or patient-rated disability or on self-re port psychosocial questionnaires at =.05. Patients included in the study provided informed consent prior to pa rticipation. Demographic variables as well as clinical information (e.g., pain duration and inte nsity) are presented in Ta ble 3-1. Fifty percent of participants reported having pain for 90 days or less, 11.8% of participants reported having pain for 91-180 days, and 38.2% indicated being in pain for 181 days or more. Fifty-six participants (54.4%) reported havi ng a prior history of LBP, and 17 (16.5%) participants reported having a history of surgery for LBP. The mean number of physical therapy sessions completed by the four-week follow-up evaluation was M= 6.82, SD= 2.73, and ranged from 1 session to 12 sessions. There were no significant differen ces found in terms of outcome variables (e.g., physical impairment, patient-rated disability, or pain intensity) or risk outcome grouping according to the STarT measure between physical therapy recruitment sites at = .05. Normality assumptions were tested and met for all data; ther efore, parametric procedures were used in all subsequent analyses. Aim 1: Comparison of STarT Ri sk Groups on Anger Variables The division of participants according to the STarT measure is presented in Table 3-2, and psychosocial characteristic s of each STarT group are presented in Table 3-3. One-way ANOVAs and chi-square tests did not reveal signif icant differences in demographic variables or 36

PAGE 37

in duration of pain sym ptoms between risk groups. However, there was a significant relationship between the presence of low back pain history and STarT risk grouping, su ch that those who fell in the High Risk group were more likely to re port having prior episodes of low back pain, 2 (2)= 8.36, p<.05. Correlations between all the psychosoci al predictors and outcome variables are presented in Table 3-4. Correlations between the psychosocial variables ranged from .01 to .60, with the strongest relationship existing between anger-out and trait ange r, similar to values reported by Spielberger (1988). One-way ANOVAs using planned comparisons revealed a difference in trait anger between groups [ F (1,102)= 10.55, p< .01]. Individual contrasts s how a significant difference between all three groups, such that the Low Ri sk group endorsed less anger than the Medium Risk group [ t (99)= 2.53, p<.05], and the Medium Risk group e ndorsed less anger than those identified as High Risk [ t (99)= 2.23, p<.05]. However, when controlling for other mood variables (depression, a nxiety, and fear-avoidance beliefs as measured by the PHQ-9, STAITrait, and FABQ-W respectively) trait anger was no longer signifi cantly different between risk groups [ F (2,95)= 2.03, p> .05]. Trait anxiety held the only si gnificant relationship in the model [ F (1,95)= 4.33, p<.05], indicating an influence of anxiet y on trait anger within risk groups, although the two mood variables ar e only modestly correlated ( r = .32). In terms of anger regulation styles (Anger-i n versus Anger-out), pl anned comparisons did not reveal a significant differe nce between risk groups at =.05. Notably, there was a trend towards significance for a relationship between risk groups an d anger-in when only anger measures were examined [ F (1,101)= 3.88, p=.05], such that the Medium and High Risk groups combined showed a greater tendency towards an internalized anger expression style than the Low Risk group [ t (99)= 1.60, p=.11]. However, similar to trait a nger, when controlling for trait 37

PAGE 38

anxiety, the effect of anger-in was no longer close to significa nce. There was a m oderate correlation between anger-i n and trait anxiety ( r= .47). Anger Control (AX/Con) was the only anger va riable that significantly differed between groups [ F (1,101)= 6.86, p<.01], and then remained significant when controlling for other psychosocial variables, although th e effect size was notably low [ F (2,94)= 3.26, p<.05, p 2= .07]. Sidak-adjusted pairwise comparisons reve aled that individuals in the Low Risk group indicated greater ange r control abilities ( M= 26.13, SD= 4.49) than those in the High Risk group ( M= 22.93, SD= 5.15). However, it may be noted that tra it anxiety showed a stronger association to risk grouping when included in the model, F (1,94)= 29.23, p<.01, p 2= .24 There was also a moderate correlation between ange r control and trait anxiety ( r= .48). Aim 2: The Contribution of Anger Variables to Treatment Outcome Regression diagnostics were performed on a ll multiple regression analyses conducted, and did not indicate excessive multicollinearity betw een independent variables. For all regression models, assumptions of homoscedasticity and linea rity of standardized residuals were met, and independence of residuals was confirmed through th e Durbin-Watson statistic. Furthermore, the averages of variance inflating factors (VIF) were not substantially greater than one, and were all under 10, and tolerance was well above 0.2, sugge sting the absence of excessive collinearity according to published guidelines (Field, 2005). To assess the unique contribution of anger to treatment outcome across all participants, a hierarchical regression model was used, with the anger variables (STAXI-T, AX/IN, AX/OUT, and AX/CON) entered in the first step, and other psychosocial variables known to impact treatment outcome entered in the second step for all three outcome variables. Results indicate that neither anger variables alone, nor anger in addition to other psychosocial symptoms predicted pain intensity ratings at their 4-week follow-up evaluation, yielding an insignificant 38

PAGE 39

overall m odel, F (7,75)=.85, p>.05, R2=.07. The model used to predict performance-based disability, as measured by scores on the Physic al Impairment Index (PII), was significant only after the addition of other mood variables [ F (7,63)= 2.30, p<.05], shown in Table 3-5. Specifically, fear-avoidance beliefs held the greatest weight in the model, t (63)= 3.46, p<.001. Non-anger psychosocial variables explained an additional 17.2% of the variance in PII scores above anger variables alone, yielding a complete model that accounted for 20.4% of the variance. Similarly, anger alone did not si gnificantly predict patient-rat ed disability as measured by the RMDQ, although a trend for anger-in emerged, t (73)= 1.82, p=.07. However, the other psychosocial variables held a stronger relati onship to disability when added, yielding a significant full model, F (7,73)= 3.27, p<.01. The work items on the FABQ also held the greatest influence in this model, t (73)= 3.08, p<.01. As shown in Table 3-6, the complete model explained 23.9% of the variance, which was 16.5% greater than when the anger variables were entered alone. Aim 3: Empirical Grouping of Low Back Pain Patients Hierarchical agglomerative cluster analys es (Wards method, Euclidian distance) were used to examine subgrouping of patients based on psychosocial risk factors, and most closely resulted in a 2 group division. Patient cluster characteristics are presented in Table 3-7. Independent-samples t-tests and Pearsons chi-square tests revealed that the two clusters did not differ significantly on clinical variables (e.g., pa in symptom duration, surgical history, prior episodes of back pain) or on dem ographic variables, although it may be noted that analyses for race and employment were underpowered and may not have detected a meaningful effect. As shown in Table 3-8, the groups appeared to divi de across all psychosocial variables, creating a group in which patients endorsed greater psychosocial symptoms ( N = 17) and a group with lower psychosocial symptom endorsement ( N =81). Of note, scores on the AX/CON indicate Anger 39

PAGE 40

Control, in which higher scores are related to m ore adaptive management of angry emotions. Discriminant function analysis revealed that 96.9% of crossvalidated grouped cases were correctly classified. An examina tion of the factor loadings for eac h variable reveal that scores on the FABQ-W (.58), PHQ-9 (.56), and PCS (.55) were the most relevant in differentiating the clusters, whereas STAI-T (.28), AX/IN (.26), STAXI-T (.19), bothersomeness (.18), AX/OUT (.10), and AX/CON (-.05) held s ubstantially less influence. Independent samples t-tests were conducte d to examine the relationship between the empirically-derived subgroups and outcome meas ures of patient-rate d disability (RMDQ), performance-based impairment (PII), and pain intens ity at baseline. Results, shown in Table 3-9, revealed significant differences in all three measures between groups. The groups differed on the RMDQ [ t (96)= 4.36, p<.001], such that the group with lower psychosocial symptom endorsement indicated lower disabi lity scores than those with hi gher psychosocial risk factors. Similarly, the lower psychosocial symptom gr oup had lower baseline scores on the PII, indicating less performance-based impairment, t (95)= 2.34, p<.05. Finally, those with lower psychosocial symptoms reported lower pain intens ity than those with higher psychosocial risk factors, t (96)= 2.98, p<.01. Chi-square tests were performed to examine the relationship between the division of our sample according to scores on the STarT measure and the two-group solution derived from the cluster analysis. Results indicate d a highly significant relationship [ 2(2)= 12.69, p < 0.01] suggesting related categorical composition of pa tient subgroups, despite the addition of anger variables in the cluster analysis. Given that the STarT measure di vides patients into three ri sk groups depending on severity and variety of psychosocial vari ables, the three cluster solution was also examined. Results 40

PAGE 41

yielded groups based upon high ( N= 17), m oderate ( N= 37), and low ( N= 44) endorsement of psychosocial variables, similar to the STarT gr oup division, with 87.8% of cross-validated cases correctly classified. Interesti ngly, FABQ-W scores was the only psychosocial variable that did not follow the division, as the lowest group repo rted higher fear-avoidance beliefs than the moderate group. One-way ANOVAs and chi-square an alyses did not reveal significant clinical (e.g., duration of pain symptoms, surgical history, prior episodes of back pain) or demographic differences between the three groups. In co mparing the three groups on treatment outcome variables, a one-way ANOVA revealed significant differences in RMDQ [ F (2,97)= 13.51, p<.01] and pain [ F (2,97)= 8.16, p<.01]. Sidak-adjusted multiple comparisons indicated significant differences between all three cluste rs on the RMDQ, such that those in the high psychosocial risk group reported the greatest disab ility, and those in the moderate group reported higher disability than those in the low group. However, for pain intensity, there were only significant differences between the high ( M= 6.59, SD= 1.43) and low ( M= 4.59, SD= 1.94) psychosocial risk factor groups, a nd between the low and moderate ( M= 5.67, SD= 1.87) psychosocial symptom endorsement groups. Thus, the moderate and high psychosocial risk groups did not differ significantly on pain. Scor es on the Physical Impairment Index were not significantly different between groups at =.05. In comparing the two-cluster vers us the three-cluster solution, it appears that the data more closely follows a two group division, as evidence d by the excellent classification accuracy and significant relationships to all three outcome vari ables. Thus, the two cluster solution was used for the remainder of the analyses. Exploratory Aim: Examining Treatment Ou tcome in Relation to Patient Subgroup In examining the relationship between the two-group cluster soluti on of LBP patients and scores on the Physical Impairment Index, results indicated a signifi cant effect of testing occasion 41

PAGE 42

on PII scores [ F (1,67)= 18.72, p<.001, p 2= .22] such that both groups had lower performancebased physical impairment scores at 4 weeks. There was a trend towa rds significance for the main effect of group [ F (1,67)= 3.82, p=.06, p 2= .05], indicating a tendency for patients with lower psychosocial risk factors to have lower physical impairment at both time points. There was not a significant interaction effect of time and group membership on impairment. In terms of patient-rated disabi lity, results revealed a signif icant effect of time and group. Thus, after 4 weeks of treatment, both subgroups of patients reported lower disability [ F (1,77)= 39.14, p<.001, p 2= .34]. Also, those who fell in the low psychosocial symptom endorsement group had lower scores overall on the RMDQ than those with greater psychosocial symptoms [ F (1,77)= 17.07, p<.01, p 2= .18]. Results also yielded a tr end towards significance for the interaction of time and group [ F (1,77)= 3.45, p=.07, p 2= .04], such that those in the higher psychosocial risk group showed a tendency toward s greater decreases in RMDQ scores at the 4week follow-up evaluation than thos e in the lower risk group. Similarly, there was a significant effect for ti me and group for pain intensity, such that both groups experienced less pain af ter four weeks of treatment [ F (1,79)= 64.61, p<.001, p 2= .45], and those with less psychosocial symp toms reported less pain overall [ F (1,79)= 4.05, p<.05, p 2= .05]. Results did not show significant interactional effects of group membership and testing occasion on patients pain ratings. Means and standard deviations of outcome variables for the two groups over both testing times are presented in Table 3-10. 42

PAGE 43

Table 3-1. D escriptive Data Demographic/ Clinical Variable N Age, in years M(SD) 10340.48 (13.84) Education 7-12 years 12-16 years 16+ years 22 67 14 21.4% 65.0% 13.6% Sex Male Female 40 63 38.8% 61.2% Race White Black/ African-American Asian Pacific Islander More than one race 76 18 4 3 2 73.8% 17.5% 3.9% 2.9% 1.9% Marital Status Single Partnered/Married Divorced Widowed 34 59 9 1 33.0% 57.3% 8.7% 1.0% Employment Status Employed Unemployed Retired 66 31 6 64.1% 30.1% 5.8% Recruiting Clinic OSMI Magnolia Parke Brooks Rehabilitation Centers 27 41 35 26.2% 39.8% 33.9% Average Pain Intensity Rating M (SD) 1035.41 (1.96) Duration of Pain Symptoms, in days M (SD) 102508.05 (1101.68) 43

PAGE 44

Table 3-2. D escriptive Data for STarT Risk Groups STarT Risk Group Low Risk ( N =39) Medium Risk ( N =34) High Risk ( N =30) Age, in years M (SD) 39.26 (14.28) 42.74 (12.96) 39.50 (13.65) Education 7-12 years 12-16 years 16+ years 9 24 6 4 23 7 9 20 1 Sex Male Female 18 21 12 22 10 20 Race White Black/ African-American Asian Pacific Islander More than one race 29 9 1 0 0 24 6 1 2 1 23 3 2 1 1 Employment Status Employed Unemployed Retired 25 13 1 25 5 4 16 13 1 Duration of Pain Symptoms, in days M (SD) 495.34 (883.23) 497.76 (1331.32) 535.80 (1101.70) History of Low Back Pain Yes No 15 24 19 18 22 8 Table 3-3. Psychosocial Characteristics of STarT Risk Groups STarT Group M (SD) Psychosocial Symptom Low Risk Group Moderate Risk Group High Risk Group FABQ-W 10.64 (10.38) 12.50 (10.53) 15.60 (12.04) PHQ-9 3.92 (4.26) 6.79 (5.10) 12.23 (6.43) STAI-Trait 33.49 (7.71)35.76 (10.26)38.67 (10.06) STAXI-Trait 13.90 (3.80)14.74 (3.60)17.37 (5.51) AX/IN 13.87 (4.24)14.62 (4.29)16.03 (4.88) AX/OUT 12.82 (3.17)12.74 (2.51)13.67 (3.38) AX/CON 26.13 (4.49)25.03 (5.25)22.93 (5.15) FABQ-W: Fear Avoidance Beliefs QuestionnaireWork Scale, PHQ-9: Patient Health Questionnaire, STAI-T: Spielberger State-Trait Anxiety InventoryTrait Scale, STAXI-Trait: State-Trait Anger Expression InventoryTrait Subscale, AX/IN: Anger-in Subscale, AX/OUT: Anger-out Subscale, AX/CON: Anger Control Subscale 44

PAGE 45

Table 3-4. C orrelations between Psychosoc ial Predictors and Outcome Variables Pain PII RMDQ FABQW PHQ-9 STAITrait STAXITrait AX/ IN AX/ OUT AX/ CON Pain 1.0 PII .23a 1.0 RMDQ .60 b .41b 1.0 FABQ-W .32 b .19 .42 b 1.0 PHQ-9 .39 b .26 b .50 b .32 b 1.0 STAI-T .15 b .12 .25 a .20 a .46 b 1.0 STAXI-T .06 .01 .10 .04 .32 b .31 b 1.0 AX/IN .01 .06 .13 .07 .35 b .47 b .52 b 1.0 AX/OUT .02 .07 .1 1 .01 .14 .34 b .60 b .31 b 1.0 AX/CON -.26 b -.12 -.29 b -.01 -.17 -.48 b -.28 b -.12 -.45 b 1.0 PII: Physical Impairment Index, RMDQ: Roland-Morris Disability Questionnaire a= p< .05 b= p< .01 Table 3-5. The Effect of Anger and Other Psychosocial Variables on PII Model: PII Score B SE B Step 1: Anger variables onlya STAXI-Trait .04 .07 .09 AX/IN .06 .06 .17 AX/OUT -.03 .10 -.05 AX/CON -.05 .05 -.13 Step 2: Anger variables plus mood variablesb STAXI-Trait -.04 .06 -.09 AX/IN .02 .06 .06 AX/OUT -.02 .09 -.03 AX/CON -.03 .05 -.09 FABQWork Scale .07 .02 .44* STAI-T -.01 .03 -.05 PHQ-9 .01 .04 .04 a: Model accounted for 3.2% of the variance in PII, F (4,66)= 0.55, p= .70. b: Model accounted for 20.4% of the variance in PII, F (7,63)= 2.30, p< .05. p< .05 45

PAGE 46

Table 3-6. T he Effect of Anger and Ot her Psychosocial Variables on RMDQ Model: RMDQ Score B SE B Step 1: Anger variables onlya STAXI-Trait .03 .20 .02 AX/IN .31 .17 .25+ AX/OUT -.34 .28 -.19 AX/CON -.17 .14 -.15 Step 2: Anger variables plus mood variablesb STAXI-Trait -.04 .18 -.03 AX/IN .28 .18 .22 AX/OUT -.18 .26 -.10 AX/CON -.19 .16 -.17 FABQWork Scale .18 .06 .34* STAI-T -.11 .09 -.18 PHQ-9 .15 .11 .17 a: Model accounted for 7.3% of the variance in RMDQ, F (4,76)= 1.51, p= .21. b: Model accounted for 23.9% of the variance in RMDQ, F(7,73)= 3.27, p< .01. p< .05 + p< .10 46

PAGE 47

Table 3-7. D escriptive Data for the Tw o-Group Cluster Solution of LBP Patients Cluster Solution Demographic Low Psychosocial Symptom ( N =81) High Psychosocial Symptom ( N =17) Age, in years 39.84 (14.30) 42.00 (11.36) Education 7-12 years 12-16 years 16+ years 15 53 13 4 12 1 Sex Male Female 32 49 6 11 Race White Black/ African-American Asian Pacific Islander More than one race 61 15 2 2 1 13 0 2 1 1 Employment Status Employed Unemployed Retired 52 24 5 11 5 1 Duration of Pain Symptoms, in days M (SD) 399.41 (776.97) 950.58 (2026.82) History of Low Back Pain Yes No 43 38 11 6 Table 3-8. Psychosocial Characteristic of Cluster Division of LBP Patients Cluster Solution M (SD) Psychosocial Symptom Low Psychosocial Symptoms Cluster ( N =81) High Psychosocial Symptoms Cluster ( N =17) Bothersomeness 2.55 (0.88) 3.21 (0.66) PCS 12.43 (8.65) 32.00 (10.24) FABQ-W 13.54 (5.55) 17.46 (4.82) PHQ-9 5.69 (5.27) 12.46 (5.74) STAI-Trait 33.18 (6.43) 45.58 (10.98) STAXI-Trait 14.31 (3.68) 18.13 (5.46) AX/IN 13.55 (3.61) 18.50 (4.80) AX/OUT 12.65 (2.93) 14.42 (3.11) AX/CON 25.39 (4.75) 22.54 (5.64) 47

PAGE 48

Table 3-9. T reatment Outcome Differences Between the Clusters Low Psychosocial Cluster ( N =81) High Psychosocial Cluster ( N =17) Outcome Variable M (SD) M (SD) t Outcome (BL) p Outcome (BL) PII 3.55 (1.80) 4.65 (1.50) 2.34 .02 RMDQ 10.06 (5.58) 16.35 (4.50) 4.36 <.001 Pain Intensity (NRS 0-10) 5.08 (1.98) 6.59 (1.43) 2.98 .004 Table 3-10. Treatment Outcomes by Cluste r Membership and Evaluation Period Outcome Variable Testing Occasion PII RMDQ Pain Intensity (0-10) Low Psychosocial High Psychosocial Low Psychosocial High Psychosocial Low Psychosocial High Psychosocial Baseline 3.55 1.92 4.21 1.25 9.58 5.46 16.67 4.72 4.99 1.90 6.51 1.51 4-weeks 2.27 1.70 3.43 1.60 6.47 5.53 10.93 5.86 3.20 2.29 3.80 2.41 48

PAGE 49

CHAP TER 4 DISCUSSION The current study aimed to provide an in-depth examination of the presentation and impact of anger in low back pain, particularly as it pertains to physical therapy treatment outcome. Although anger is infrequently me ntioned in the literature on psychosocial determinants of pain and disability, it is a commonly noted trait of LBP patients in various healthcare settings. It has also been shown to lead to poor pa tient-provider relationships and subsequent unfavorable treatment outcome (Sluijs, et al., 1993). In order to bridge the gap between research and practice, the current study examined th e multidimensional construct of anger in addition to other psychos ocial factors used to identify risk for poor prognosis in low back pain. As physical thera py treatment has been moving towards the use of more targeted approaches based on patient classi fication systems, creating a well-specified measure to identify risk factors is becoming more important to the field. To this end, the Subgroups for Targeted Treatment Back Screening Tool (STarT) was devel oped to help clinicians identify LBP patients who may need a multidisciplinary approach to address psychological factors, and has shown promising predictive validity. The present study atte mpted to contribute to this line of literature, by adding anger as a possible progn ostic indicator among others included in the STarT measure, thereby further specifying the model. Furthermore, this study us ed an empirical approach in identifying patient subgroups based on psychosocial risk fact ors, rather than heuristic procedures. As initially predicted, LBP patients did differ on trait anger, such that those in higher risk groups, as classified by the STar T measure, endorsed more ange r than those with lower risk profiles. However, other psychosocial variables held stronger relationships to risk grouping after being included in the model, thereby minimizi ng the impact of anger on group classification. 49

PAGE 50

Additionally, anger regulation styl e, specifically anger-in, had only a m arginal effect on risk grouping when measured alone, although the full psychosocial model did show significant differences between the Low Risk group and the hi gher risk STarT groups. These results suggest that anger may have contributed to the overall co nstruct of psychosocial distress, as opposed to uniquely impacting STarT risk classification. This finding s upports the substa ntial body of literature suggesting that genera l negative affect is related to somatic symptom magnification, and subsequently, increased pain and disabili ty (Hirsh, et al., 2006; Watson & Pennebaker, 1989). Importantly, this line of literature does not negate th e effect of any individual psychosocial or mood construct. Rather, it s uggests that the experien ce and expression of suffering in LBP is multifactorial a nd cannot be decomposed easily. The current study did find a si gnificant, albeit small, effect of patient risk grouping on anger control, as patients in the Low Risk group reported greater attempts to control anger than those in the higher risk groups. While this find ing was not originally hypothesized, it is not unexpected given the context of the study question. Patients in our sample scored in the 35th percentile, on average, in trait anger, indi cating anger levels comparable to the general population. As our sample did not endorse particularly high levels of anger, there may have been a floor effect, thereby limiting expl oration of anger regulation styl es in pain. The dimension of anger control as measured by the STAXI, on th e other hand, focuses more on how situational anger is managed when it is present (e.g., I contro l my temper., I calm down faster than most people.). Thus, a significant relationship betwee n low anger control abi lities and risk for poor treatment outcome may be more substantiated in th e context of situational anger to pain. This outcome is more in line with th e state-trait matching hypothesis of anger in pain, which suggests that those who unsuccessfully try to control their anger tend to experience more pain. Use of the 50

PAGE 51

STAXI-2, an updated version of the STAXI that in cludes separate scales of Anger Control-In and Anger Control-Out, in subsequent studies m a y help elucidate the nature of anger regulation style in response to pain (Spielberger, 1999). This updated meas ure may need to be validated in the chronic pain population, however, as it has not yet been implemented in published pain studies. In addition to the inclusion of more sensit ive measures of anger control, the methodology used to access anger in LBP pa tients may be improved in this study. Several studies have advised against the use of self -report measures of anger experi ence and expression due to strong covariation with general negativ e affect, thereby clouding variance attributable to anger alone (Burns, Quartana, & Bruehl, 2008; Quartana, Bounds, Yoon, Goodin, & Burns, 2010). In order to obtain a more accurate account of the unique c ontribution of anger to pain and disability, a recent study conducted by Quartana, et al. (2010) used numerical ra ting scales to assess selfreported anger suppression during anger and pa in induction tasks. These authors found a positive association between anger suppression a nd pain intensity when controlling for both positive and negative emotions. Other studies us ing anger induction procedures have also found a strong unique effect of anger suppression on pa in severity and other physiological responses (Burns, et al., 2007; Burns, Quartana, & Bruehl 2009; Quartana & Burns, 2007). However, a different set of studies have c oncluded that both anger and sadn ess induction results in increased pain perception during experimental pain tasks (Rainville, Bao, & Chretien, 2005; van Middendorp, Lumley, Jacobs, Bijlsma, & Geenen 2010). Thus, including an anger induction procedure in the present study may have yielded a clearer picture on the unique relationship between anger management and risk profiles in LBP patients, or if one ev en exists outside of generalized negative affect. 51

PAGE 52

The strong relationship of tra it anxiety to risk group member ship was an interesting and somewhat surprising finding in the study. The pr esence of anxiety in chronic pain is often conceptualized as fear of pain (George, Dannecker, et al., 200 6), or is subsumed under other psychosocial dimensions, such as anxiety sens itivity, fear-avoidance beliefs, and general emotional distress (Carleton, Abrams, Kac hur, & Asmundson, 2009; Ryan, Gray, Newton, & Granat, 2010). However, few examine the contribution of dispositional anxiety to treatment outcome. Interestingly, recent studies have show n support for the influence of trait anxiety on the development of chronic low back pain, sepa rate from fear-avoidance beliefs (Newcomer, Shelerud, Vickers Douglas, Larson, & Crawford, 2010). Further research must be done to explore this novel phenomenon, as it may help shape screening procedures for negative prognostic indicators in physic al therapy settings. Our study indicated a signifi cant relationship between high tr ait anger and greater patientrated and performance-based disabi lity at the 4-week follow-up ev aluation, but did not detect a unique effect above and beyond ot her included psychosocial factors. Furthermore, results revealed that these added psychosocial constructs especially fear avoidance beliefs, were the main factors predicting patient-rated disability an d performance-based impair ment in our sample. As previously noted, anger experience and expres sion may have been difficult to access due to methodological constraints, and may therefore have been subsumed under the construct of overall negative affect. This proposed explanat ion falls in line with the notion that higher general psychosocial distress is related to risk for poor treatment outcome, which was supported in this study. Another issue to consider is that half of our sa mple reported having low back pain for less than 3 months and two-thirds of our samp le reported having pain for 6 months or less, indicating a relatively short dura tion of symptoms. While the cu rrent study did not indicate a 52

PAGE 53

relationship between pain durat ion and psychosocial factors, LBP episode duration has been significantly associated with poor treatm ent outcome in other studies (Dunn, Jordan, & Croft, 2010; Mallen, Peat, Thomas, Dunn, & Croft, 2007). Our study did show a significant positive association between low back pain history and ou tcome risk, indicating that general pain course and history may be influencing our results. Given that predic tors of pain and disability prognoses may be different across settings (D unn, et al., 2010), additional research on how clinical pain presentations can variably c onfound the impact of psychosocial symptoms is warranted. The current studys findings suppor t previous research indicati ng the importance of fear avoidance thoughts and behaviors in developing and maintaining disability in low back pain (George, Wittmer, et al., 2006; Vlaeyen, Kole-S nijders, Boeren, & van Eek, 1995; Waddell, et al., 1993). However, it is intere sting that fear avoidance was the only significant psychosocial predictor of physical impairment and self-reported disab ility among those included in the model. Furthermore, the complete regression model, with anger, anxiety, depression, and fear avoidance, accounted for less than 8% in reported pain. As screening measures such as the STarT are heavily reliant on psychos ocial symptoms to assess risk for poor treatment outcome, it is important to have a firm understanding of the magnitude of their effect. Several studies have emphasized caution in interpreting the relationshi p between psychosocial distress and disability due to low or moderate effect sizes (Estlande r, Takala, & Viikari-J untura, 1998; Gesztelyi & Bereczki, 2006; Schiphorst Preuper, et al., 2008), or the absence of significant findings when controlling for other factors (Dunn, et al., 2010). In light of the mi xed findings in this line of literature, it would be interesti ng and worthwhile to explore th e apparently dynamic combination of factors that impact pain and disability. 53

PAGE 54

Although we had originally hypoth esized that L BP patients woul d divide into three groups based on psychosocial prognostic indicators, our results more closely followed a two group division, separating patients in to higher and lower psychosoc ial symptom endorsement. To compare to the STarT division, a three group soluti on was also analyzed. While patients were well-balanced between the high, moderate, and lo w groups, this categoriza tion yielded slightly lower classification results and held weaker relationships to the out come variables. On the other hand, the group sizes found in the two cluster solution were notably mismatched, as the group with lower scores on the psychosocial measures ha d more than four times the number of patients than those who endorsed greater symptomatology. This finding further supports the conclusion that the study sample was generally low on psyc hosocial distress, there by restricting possible statistical effects. Importantly, however, by identifying the few of those with significant distress, the two-cluster solution may help single out pa tients who would benefit from multidisciplinary treatment options. Overall, the finding that empirical subgrouping techniques yielded two groups, as opposed to the three patient groups id entified by the STarT measure, highlighted the relevance of empirical division of data over heuristic procedures. However, replication studies should be conducted with different LBP patient samp les to confirm generalizability of results. Given the pattern of results prec eding the cluster analysis, it is not surprising that the anger variables did not carry as much weight in dividi ng the groups as some of the other psychosocial variables, namely pain catastrophizing, fear avoi dance, and depression. Furthermore, the finding that the cluster analysis division and the STar T measure grouping of our sample were highly related suggests similar group composition, regardless of the inclusion of anger. While it is very possible that anger simply does not play a role in patient subgrouping, it may be that anger was one of the components within the general cons truct of emotional distress, as suggested by 54

PAGE 55

Gaskin et al. (1992). Additionally, there is resear ch to suggest that, in the context of clinical pain, anger is not as clearly differentiated from other psychosocia l constructs when using selfreport measures. As noted previously, studies examining chronic pain and anger using selfreport measures included experiment al pain (Bruehl, et al., 2003) or mood induction procedures (Burns, 2006) to understand the relationship between pain and anger. In clinical pain, research has shown that the lines between negative affect constructs tend to get blurred and may have state-dependent qualities that in fluence patient ratings (Burns, et al., 2008; Gaskin, et al., 1992). Thus, it is speculated that unders tanding how anger uniquely contri butes to risk for poor outcome in LBP may simply require a differe nt experimental model. Our exploratory analysis reve aled a significant effect of group membership on outcome after four weeks of treatment, such that those with higher psychosocial endorsement were more likely to have higher self-rated pain and disabili ty in their follow-up evaluation. This finding provides some support that risk grouping according to the magnitude of psychosocial symptomatology is associated with treatment outcome, and may, therefore, be useful in guiding treatment. Notably, the effect of testing occasion was more powerful across all outcome variables, indicating that patients, regardless of their risk group, reported less pain and disability after engaging in physical therapy. Furthermore, there was a slight tendency for those with higher psychosocial risk f actors to improve more over four we eks of treatment, indicating that the standard, non-targeted physical therapy course was sufficient in addressing their disability without added psychosocial intervention. Thus, these results appear to corroborate the notion that psychosocial factors may be a factor, but not necessarily th e defining feature, of a poorer prognosis. However, given that our sample did not report pathognomonic symptoms of 55

PAGE 56

em otional distress overall, it is currently unclear whether these fi ndings are generalizable to the broader LBP population. Interestingly, our results only showed a tende ncy towards significance for the influence of patient grouping on performance-based impairment, as measured by the Physical Impairment Index. While it may be that the potential effect was underpowered, a broader examination of the data shows that performance-based impairment generally has a weaker association to risk grouping than patient-rated disab ility. This finding has been s upported in the l iterature, as a study comparing the two forms of impairment as sessment found a stronger relationship between psychological variables and self-report disability measures than with performance-based measures (Schiphorst Preuper, et al., 2008). Furthermore, validation studies of the PII show rather weak relationships to psychosocial symptoms of depression, fear-avoidance and catastrophizing, ranging from .05 to .28 (Fritz & Piva, 2003). Also, othe r studies have shown significant differences between se lf-report and performance-based measures, and have concluded that assessing disabilities using multiple methods allows for a broader perspective of patient disability and functiona l capacity (Brouwer, et al., 2005; Lee, Simmonds, Novy, & Jones, 2001). The findings from the current study support the us e of multiple, varied out come measures when assessing for negative prognostic indicators in treatment, as focusing on one modality may inflate or, conversely, minimize potential effects. Many of the general limitations of the curr ent study have already been mentioned, including the methodological cons traints of using self-report psychosocial questionnaires as opposed to painand moodinduction procedures to access these constructs. Additionally the anger measure selection could have been improve d, as the updated versi on of the STAXI better delineates anger control styl es. The comparison of the STarT measure grouping and the 56

PAGE 57

em pirical grouping used in the current study should be interpreted with caution, as some of the psychosocial measures used in the development of the STarT were different from those included in the cluster analyses. However, it may be noted that the measures included in the cluster analyses show high correlations with those used to develop the STarT, indicating that similar constructs were examined. Another limitation of the current study is that patients were not evaluated for past or ongoing psychological trea tment or psychiatric diagnoses prior to their enrollment. This limitation is especially notable given the low levels of psychosocial symptoms endorsed by LBP patients in the sample. Understa nding patients psychiatric histories may have clarified this finding, as well as how psychiatric treatment may cont ribute to risk grouping. Although the scores on psychosocial measures were notably low across all patients, there is no apparent reason to suspect that our sample of LBP physical therapy patients were clinically or demographically different from other physical therapy clinics. Furthermore, the mean scores on the RMDQ and on pain intensity were not differen t from other studies i nvestigating disability and pain in LBP recruited from physical reha bilitation centers (Cairns, Foster, Wright, & Pennington, 2003; Schiphorst Preuper, et al., 2007). Of note, the STarT measure was originally created to assess for prognostic indicators in primary care setti ngs, and had not been used in physical therapy clinics prior to th is investigation. To this end, an associated study is currently exploring the external, convergen t, and predictive valid ity of the STarT measure in physical therapy settings, which may help clarify differences in patient risk profiles between primary care and PT settings. Nevertheless, the current st udy did reveal that pati ent subgroups differed on pain, disability, and performance-based impairment in the expected direction, demonstrating the impact of psychosocial symp tomatology on outcome risk. 57

PAGE 58

In term s of clinical implications, the results of the current study woul d be most useful in helping physical therapists understand the benefits and limitations of patient classification based on psychosocial screening procedures. There were a few strong rela tionships found between individual psychosocial variable s and outcome, such as fear-avoidance beliefs, which can help direct the clinician to these specific issues in patients during treatment. However, our results also suggested that an overall construct of psychosocial distress may be impacting risk subgrouping, thereby making targeted psychosocial approaches somewhat difficult. Therefore, psychosocial risk screening tools do appear to be helpful in cl assifying patients, but may fall short in their ability to assist with individualized treatment plans. Regarding the impact of anger on treatment planning, the current study did not pr ovide clear evidence whether the addition of targeted anger management protocols would be relevant to treatment outcome in the physical therapy setting. Thus, more research examining anger in physical thera py settings would help elucidate how it can affect clin ical relationships and, ultimat ely, treatment strategies. To this end, future studies in this area may include obtaini ng ratings of anger and pain immediately following a physical therapy session to better access patients reactive anger to the exercises. Subsequent analyses may then examine how these ratings impact future physical therapy treatment results. Additionally, examining the usefulness of LBP screening measures in a long-term follow-up study may help identify the most pertinent risk factors for poor prognosis and the development of chronic LBP conditions. This investigation will be possible with the current sample, as 6-month follow-up data is currently being collected. In conclusion, the current study yielded some interesting findings regarding the influence of psychosocial factors on treatment outcome in LB P patients. Results suggested that the patient risk groups represented different levels of ps ychosocial distress, although certain constructs, 58

PAGE 59

specifically fear-avoid ance beliefs, were especially influential. While it is evident that anger is part of the negative affect cons truct, the unique contribution of anger experience and expression to treatment outcome is less clear and may re quire additional study, preferably conducted when patients are in the midst of expressing anger. Overall, this study furthe r supported that treatment outcome in low back pain is multifactorial, and is, therefore, difficult to predict. However, it appears that updating and implementing screenin g measures based on psychosocial factors may be able to help guide practice in physical therapy treatment. 59

PAGE 60

APPENDIX A FLOW DIAGRAM OF STUDY DE SIGN Clinician inquires about participation in study Routine demographics obtained Informed consent obtained from clinician Questionnaires completed o STarT Back Tool (9-items) o Pain Catastrophizing Scale (13-items) o Fear-Avoidance Beliefs Qu estionnaire (21-items) o State-Trait Anxiety Inventory (trait portion) (20-items) o State-Trait Anger Expression Inventory (44-items) o Patient Health Questionnaire (depression) (9-items) o Tampa Scale of Kinesiophobia (11-items) o Patient Centered Outcomes (4-items) o Revised Oswestry Disability Questionnaire (10-items) o Roland-Morris Disability Questionnaire o Pain Intensity (Numerical Rating Scale 0-10) Current Best in past week Worst in past week o Bothersomeness In past week Physical Impairment Index completed (7-tests) Additional physical examinat ion procedures completed o Prone instability test o Hip internal rotation o Passive spine mobility testing o Qualitative spine mobility assessment o Centralization / Peripheralization phenomena # of visits determined / was episode of care completed Re-assess: Physical Impairment Index Patient satisfaction with the delivery of PT treatment Re-administer questionnaires o Fear-Avoidance Beliefs Questionnaire o State-Trait Anger Expression Inventory o Patient Health Questionnaire o Pain Catastrophizing Scale o Tampa Scale of Kinesiophobia o Revised Oswestry Disability Questionnaire o Roland-Morris Disability Questionnaire o Pain Intensity (as above) o Bothersomeness (as above) Initial Session Within 1st Week 4-week follow-up This flow diagram indicates all physi cal and psychosocial evaluations that were administered at the patients baseline and 1 month follow-up visits, as part of a larger parent project. Measures that are in bold font indicate those that are pertinent to the current study. 60

PAGE 61

61 APPENDIX B START ME ASURE For this first set of questions, please th ink about your back pain over the past two weeks 1. Overall, how bothersome has your back pain been in the last 2 weeks ? Not at all Slightly Moderately Very much Extremely For each of the following, please cross one box to show whether you agree or disagree with the statement, thinking about the last 2 weeks 2. My back pain has spread down my leg(s) at some point in the last 2 weeks. Agree Disagree 3. I have had pain in the shoulder or neck at some time in the last 2 weeks. Agree Disagree 4. Its really not safe for a person with a co ndition like mine to be physically active. Agree Disagree 5. In the last 2 weeks, I have dressed more slowly than usual because of my back pain. Agree Disagree 6. In the last 2 weeks, I have only walked short distances because of my back pain. Agree Disagree 7. Worrying thoughts have been going through my mind a lot of the time in the last 2 weeks. Agree Disagree 8. I feel that my back pain is terrible and that its never going to get any better Agree Disagree 9. In general in the last 2 weeks, I have not enjoyed all the things I used to enjoy. Agree Disagree

PAGE 62

LIST OF REFERE NCES Anderson, D. E., Metter, E. J., Hougaku, H., & Najjar, S. S. (2006). Suppressed anger is associated with increased carotid arterial stiffness in older adults. Am J Hypertens, 19(11), 1129-1134. Bair, M. J., Robinson, R. L., Katon, W., & Kroenke, K. (2003). Depression and pain comorbidity: a literature review. Arch Intern Med, 163 (20), 2433-2445. Baldwin, M. L., Butler, R. J., Johnson, W. G., & Co te, P. (2007). Self-reported severity measures as predictors of return-to-work outcomes in occupational back pain. J Occup Rehabil, 17(4), 683-700. Banks, S. M., & Kerns, R. D. (1996). Explaining high rates of depression in chronic pain: A diathesis-stress framework. Psychological Bulletin, 119 (1), 95-110. Baron, R. A. (1977). Human Aggression New York: Plenum Press. Beneciuk, J. M., Bishop, M. D., & George, S. Z. (2009). Clinical prediction rules for physical therapy interventions: a systematic review. Phys Ther, 89(2), 114-124. Brouwer, S., Dijkstra, P. U., Stewart, R. E., Goek en, L. N., Groothoff, J. W., & Geertzen, J. H. (2005). Comparing self-report, clinical examination and functional testing in the assessment of work-related limitations in patients with chronic low back pain. Disabil Rehabil, 27 (17), 999-1005. Bruehl, S., Burns, J. W., Chung, O. Y., & Quar tana, P. (2008). Anger management style and emotional reactivity to noxious stimuli among chronic pain pa tients and healthy controls: the role of endogenous opioids. Health Psychol, 27 (2), 204-214. Bruehl, S., Burns, J. W., Chung, O. Y., War d, P., & Johnson, B. (2002). Anger and pain sensitivity in chronic low back pain patie nts and pain-free cont rols: the role of endogenous opioids. Pain, 99 (1-2), 223-233. Bruehl, S., Chung, O. Y., & Burns, J. W. (2006) Anger expression and pain: an overview of findings and possible mechanisms. J Behav Med, 29 (6), 593-606. Bruehl, S., Chung, O. Y., Burns, J. W., & Biridepa lli, S. (2003). The association between anger expression and chronic pain intensity: evidence for partial mediation by endogenous opioid dysfunction. Pain, 106 (3), 317-324. Bruehl, S., Chung, O. Y., Donahue, B. S., & Burn s, J. W. (2006). Anger regulation style, postoperative pain, and relationship to the A118G mu opioid receptor gene polymorphism: a preliminary study. J Behav Med, 29 (2), 161-169. Burns, J. W. (2006). Arousal of negative emo tions and symptom-specific reactivity in chronic low back pain patients. Emotion, 6(2), 309-319. 62

PAGE 63

Burns, J. W ., Higdon, L. J., Mullen, J. T., Lansky, D., & Wei, J. M. (1999). Relationships among patient hostility, anger expre ssion, depression, and the working alliance in a work hardening program. Ann Behav Med, 21(1), 77-82. Burns, J. W., Johnson, B. J., Devine, J., Mahone y, N., & Pawl, R. (1998). Anger management style and the prediction of treatment outco me among male and female chronic pain patients. Behav Res Ther, 36 (11), 1051-1062. 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 (3), 445-453. Burns, J. W., Kubilus, A., & Bruehl, S. (2003). Emotion induction moderates effects of anger management style on acute pain sensitivity. Pain, 106 (1-2), 109-118. Burns, J. W., Kubilus, A., Bruehl, S., & Hard en, R. N. (2001). A fourth empirically derived cluster of chronic pain patients based on th e multidimensional pain inventory: evidence for repression within the dysfunctional group. J Consult Clin Psychol, 69 (4), 663-673. Burns, J. W., Quartana, P. J., & Bruehl, S. (2007 ). Anger management style moderates effects of emotion suppression during initial stress on pain and cardiovascu lar responses during subsequent pain-induction. Ann Behav Med, 34(2), 154-165. Burns, J. W., Quartana, P. J., & Bruehl, S. (2008 ). Anger inhibition and pain: conceptualizations, evidence and new directions. J Behav Med, 31 (3), 259-279. Burns, J. W., Quartana, P. J., & Bruehl, S. (2009 ). Anger management style moderates effects of attention strategy during acute pain inducti on on physiological responses to subsequent mental stress and recovery: a comparison of chronic pain patients and healthy nonpatients. Psychosom Med, 71 (4), 454-462. Cairns, M. C., Foster, N. E., Wright, C. C., & Pennington, D. (2003). Level of distress in a recurrent low back pain populati on referred for physical therapy. Spine (Phila Pa 1976), 28(9), 953-959. Carleton, R. N., Abrams, M. P., Kachur, S. S., & Asmundson, G. J. (2009). Waddell's symptoms as correlates of vulnerabilitie s associated with fear-anxiet y-avoidance models of pain: pain-related anxiety, catastr ophic thinking, perceived disab ility, and treatment outcome. J Occup Rehabil, 19 (4), 364-374. Delitto, A., Erhard, R. E., & Bowling, R. W. (19 95). A treatment-based cl assification approach to low back syndrome: identifying and stag ing patients for conservative treatment. Phys Ther, 75(6), 470-485; discussion 485-479. 63

PAGE 64

Dionne, C. E., Bourbonnais, R., Fremont, P., Rossi gnol, M., Stock, S. R., & Larocque, I. (2005). A clinic al return-to-work rule for patients with back pain. CMAJ, 172 (12), 1559-1567. Duijts, S. F., Kant, I. J., Landeweerd, J. A ., & Swaen, G. M. (2006). Prediction of sickness absence: development of a screening instrument. Occup Environ Med, 63 (8), 564-569. Dunn, K. M., & Croft, P. R. (2005). Classificatio n of low back pain in primary care: using "bothersomeness" to identify the most severe cases. Spine, 30 1887-1892. Dunn, K. M., Jordan, K. P., & Croft, P. R. ( 2010). Contributions of prognostic factors for poor outcome in primary care low back pain patients. Eur J Pain Epub ahead of print. Engebretson, T. O., Matthews, K. A., & Sche ier, M. F. (1989). Relations between anger expression and cardiovascular reactivity: reconciling inc onsistent findings through a matching hypothesis. J Pers Soc Psychol, 57 (3), 513-521. Estlander, A. M., Takala, E. P., & Viikari-Junt ura, E. (1998). Do psychological factors predict changes in musculoskeletal pain? A prospectiv e, two-year follow-up study of a working population. J Occup Environ Med, 40 (5), 445-453. Faber, S. D., & Burns, J. W. (1996). Anger mana gement style, degree of expressed anger, and gender influence cardiovascular recove ry from interpersonal harassment. J Behav Med, 19(1), 31-53. Fairbank, J. C., Couper, J., Davies, J. B., & O'Br ien, J. P. (1980). The Oswestry low back pain disability questionnaire. Physiotherapy, 66(8), 271-273. Feldman, S. I., Downey, G., & Schaffer-Neitz R. (1999). Pain, nega tive mood, and perceived support in chronic pain patient s: A daily diary study of people with reflex sympathetic dystrophy syndrome. Journal of Consulting an d Clinical Psychology, 67 (5), 776-785. Fernandez, E., & Turk, D. C. (1995). The scope a nd significance of anger in the experience of chronic pain. Pain, 61(2), 165-175. Field, A. (2005). Discovering Statistics Using SPSS. (2nd ed.). Los Angeles: SAGE Publications Inc. Fritz, J. M., Cleland, J. A., & Childs, J. D. (2007). Subgrouping patients with low back pain: evolution of a classification approach to physical therapy. J Orthop Sports Phys Ther, 37(6), 290-302. Fritz, J. M., Delitto, A., & Erhard, R. E. (2003) Comparison of classifi cation-based physical therapy with therapy based on clinical prac tice guidelines for patients with acute low back pain: a randomized clinical trial. Spine (Phila Pa 1976), 28 (13), 1363-1371; discussion 1372. 64

PAGE 65

Fritz, J. M., & George, S (2000). The use of a cl assification approach to identify subgroups of patients with acute low back pain. Interrat er reliability and short-term treatment outcomes. Spine (Phila Pa 1976), 25(1), 106-114. Fritz, J. M., George, S. Z., & De litto, A. (2001). The role of fea r-avoidance beliefs in acute low back pain: relationships with current a nd future disability and work status. Pain, 94(1), 715. Fritz, J. M., & Irrgang, J. J. (2001). A compar ison of a modified Oswestry Low Back Pain Disability Questionnaire and the Quebec Back Pain Disability Scale. Phys Ther, 81(2), 776-788. Fritz, J. M., & Piva, S. R. (2003). Physical impairment index: reliability, validity, and responsiveness in patients with acute low back pain. Spine (Phila Pa 1976), 28 (11), 1189-1194. Gaskin, M. E., Greene, A. F., Robinson, M. E., & Geisser, M. E. (1992). Negative affect and the experience of chronic pain. J Psychosom Res, 36 (8), 707-713. Gatchel, R. J., Mayer, T. G., & Eddington, A. (2006). MMPI disability profile: the least known, most useful screen for psychopathology in chronic occupational spinal disorders. Spine (Phila Pa 1976), 31(25), 2973-2978. George, S. Z., Dannecker, E. A., & Robins on, M. E. (2006). Fear of pain, not pain catastrophizing, predicts acute pain intensity, but neither f actor predicts tolerance or blood pressure reactivity: an experimental investigation in pain-free individuals. Eur J Pain, 10(5), 457-465. George, S. Z., Fritz, J. M., Bialosky, J. E., & Donald, D. A. (2003). The effect of a fearavoidance-based physical thera py intervention for patients with acute low back pain: results of a randomized clinical trial. Spine (Phila Pa 1976), 28 (23), 2551-2560. George, S. Z., Wittmer, V. T., Fillingim, R. B., & Robinson, M. E. (2006). Fear-avoidance beliefs and temporal summation of evoked therma l pain influence self-r eport of disability in patients with chronic low back pain. J Occup Rehabil, 16 (1), 95-108. George, S. Z., & Zeppieri, G. ( 2009). Physical therapy utilization of graded exposure for patients with low back pain. J Orthop Sports Phys Ther, 39(7), 496-505. Gesztelyi, G., & Bereczki, D. (2006). Determinants of disability in everyd ay activities differ in primary and cervicogenic headaches and in low back pain. Psychiatry Clin Neurosci, 60(3), 271-276. Greenwood, K. A., Thurston, R., Rumble, M., Waters, S. J., & Keefe, F. J. (2003). Anger and persistent pain: current stat us and future directions. Pain, 103 (1-2), 1-5. 65

PAGE 66

Hadjistavropoulos, H. D., As mundson, G. J., & Ko walyk, K. M. (2004). Meas ures of anxiety: is there a difference in their ab ility to predict functioning at three-month follow-up among pain patients? Eur J Pain, 8 (1), 1-11. Hair, J. F., Anderson, R. E., Tatham, R. L., & Black, W. C. (1995). Multivariate Data Analysis (4th ed.). Englewood Cliffs, NJ: Prentice Hall. Hill, J. C., Dunn, K. M., Lewis, M., Mullis, R., Main, C. J., Foster, N. E., et al. (2008). A primary care back pain screeni ng tool: identifying patient subg roups for initial treatment. Arthritis Rheum, 59(5), 632-641. Hirsh, A. T., Waxenberg, L. B., Atchison, J. W., Gremillion, H. A., & Robinson, M. E. (2006). Evidence for sex differences in the rela tionships of pain, mood, and disability. J Pain, 7(8), 592-601. Janssen, S. A., Spinhoven, P., & Arntz, A. (2004) The effects of failing to control pain: an experimental investigation. Pain, 107 (3), 227-233. Jensen, M. P., & Karoly, P. (2001). Self-report scal es and procedures for a ssessing pain in adults. In D. C. Turk & R. Melzack (Eds.), Handbook of Pain Assessment (pp. 15-34). New York: Guilford Press. Kerns, R. D., Rosenberg, R., & Jacob, M. C. (1994). Anger expression and chronic pain. J Behav Med, 17(1), 57-67. Kinder, B. N., Curtiss, G., & Kalichman, S. (1986). Anxiety and anger as predictors of MMPI elevations in chronic pain patients. J Pers Assess, 50 (4), 651-661. Kroenke, K., Spitzer, R. L., & Williams, J. B. ( 2001). The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med, 16 (9), 606-613. Lee, C. E., Simmonds, M. J., Novy, D. M., & Jones, S. (2001). Self -reports and clinicianmeasured physical function among patients with low back pain: a comparison. Arch Phys Med Rehabil, 82 (2), 227-231. Lethem, J., Slade, P. D., Troup, J. D., & Bentle y, G. (1983). Outline of a Fear-Avoidance Model of exaggerated pain perception--I. Behav Res Ther, 21 (4), 401-408. Lombardo, E. R., Tan, G., Jensen, M. P., & A nderson, K. O. (2005). Anger management style and associations with self-efficacy and pain in male veterans. J Pain, 6 (11), 765-770. Mallen, C. D., Peat, G., Thomas, E., Dunn, K. M., & Croft, P. R. (2007). Prognostic factors for musculoskeletal pain in primary care: a systematic review. Br J Gen Pract, 57 (541), 655661. 66

PAGE 67

Martin, A. L., McGrath, P. A., Brown, S. C., & Katz J. (2007). Anxiety sens itivity, fear of pain and pain-related disability in childre n and adolescents with chronic pain. Pain Res Manag, 12(4), 267-272. Melzack, R., & Casey, K. L. (1968). Sensory, m otiva tional, and central control determinants of pain. In D. R. Kensalo (Ed.), The Skin Senses (pp. 423-439). Springfield CC: Thomas. Nachemson, A. L., Waddell, G., & Norlund, A. ( 2000). Epidemiology of neck and back pain. In A. L. Nachemson & E. Jonsson (Eds.), Neck and back pain. The scientific evidence of causes, diagnosis, and treatment. (pp. 165-188). Philadelphia: Lippincott Williams & Wilkins. Newcomer, K. L., Shelerud, R. A., Vickers Douglas, K. S., Larson, D. R., & Crawford, B. J. (2010). Anxiety levels, fear-avoida nce beliefs, and di sability levels at baseline and at 1 year among subjects with acute and chronic low back pain. PM R, 2 (6), 514-520. Nicholson, R. A., Gramling, S. E., Ong, J. C., & Buenaver, L. (2003). Differences in anger expression between individuals with a nd without headache after controlling for depression and anxiety. Headache, 43(6), 651-663. Okifuji, A., Turk, D. C., & Curran, S. L. (1999). Anger in chronic pain: in vestigations of anger targets and intensity. J Psychosom Res, 47 (1), 1-12. Ortony, A., Clore, G., & Collins, A. (1988). The Cognitive Structure of Emotions Cambridge, UK: Cambridge University Press. Pain, I. A. f. t. S. o. (1994). Part III: Pain terms: A current List wit definitions and notes on usage. In H. Merskey & N. Bogduk (Eds.), Classification of Chronic Pain: Second Edition (pp. 209-214). Seattle, WA: IASP Press. Quartana, P. J., Bounds, S., Yoon, K. L., Goodi n, B. R., & Burns, J. W. (2010). Anger suppression predicts pain, emotional, and cardi ovascular responses to the cold pressor. Ann Behav Med, 39(3), 211-221. Quartana, P. J., & Burns, J. W. (2007). Painful consequences of anger suppression. Emotion, 7(2), 400-414. Rainville, P., Bao, Q. V., & Chretien, P. (2005) Pain-related emotions modulate experimental pain perception and autonomic responses. Pain, 118(3), 306-318. Riley, J. L., 3rd, & Robinson, M. E. (1998). Valid ity of MMPI-2 profiles in chronic back pain patients: differences in path m odels of coping and somatization. Clin J Pain, 14 (4), 324335. 67

PAGE 68

Riley, J. L., 3rd, Robins on, M. E., Geisser, M. E., Wittmer, V. T., & Smith, A. G. (1995). Relationship between MMPI-2 cluster profiles and surgical outcome in low-back pain patients. J Spinal Disord, 8 (3), 213-219. Roland, M., & Fairbank, J. (2000). The RolandMorris Disability Questionnaire and the Oswestry Disability Questionnaire. Spine (Phila Pa 1976), 25 (24), 3115-3124. Roland, M., & Morris, R. (1983). A study of the natura l history of back pain. Part I: development of a reliable and sensitive measure of disability in low-back pain. Spine, 8 (2), 141-144. Ryan, C. G., Gray, H. G., Newton, M., & Gran at, M. H. (2010). The relationship between psychological distress and free-li ving physical activity in individuals with chronic low back pain. Man Ther, 15 (2), 185-189. Schiphorst Preuper, H. R., Reneman, M. F., Boonstr a, A. M., Dijkstra, P. U., Versteegen, G. J., & Geertzen, J. H. (2007). The relationship between psychosocial distress and disability assessed by the Symptom Checklist-90-Re vised and Roland Morris Disability Questionnaire in patients with chronic low back pain. Spine J, 7 (5), 525-530. Schiphorst Preuper, H. R., Reneman, M. F., Boonstr a, A. M., Dijkstra, P. U., Versteegen, G. J., Geertzen, J. H., et al. (2008). Relationship between psychological factors and performance-based and self-reported di sability in chronic low back pain. Eur Spine J, 17(11), 1448-1456. Sluijs, E. M., Kok, G. J., & van der Zee, J. ( 1993). Correlates of exercise compliance in physical therapy. Phys Ther, 73 (11), 771-782; discussion 783-776. Smith, R. C., & Zimny, G. H. (1988). Physic ians' emotional reactions to patients. Psychosomatics, 29 (4), 392-397. Spielberger, C. D. (1988). State-Trait Anger Expression Inventory: Research Edition Professional Manual Odessa, Florida: Psychological Assessment Resources, Inc. Spielberger, C. D. (1999). Professional manual for the StateTrait Anger Expression Inventory-2 (STAXI-2) Odessa, FL: Psychological Assessment Resources. Spielberger, C. D., Gorsuch, R. L., Lushene, P. R., Vagg, P. R., & Jacobs, A. G. (1983). Manual for the State-Trait Anxiety Inventory (Form Y) Palo Alto: Consulting Psychologists Press, Inc. Strine, T. W., & Hootman, J. M. (2007). US natio nal prevalence and correlates of low back and neck pain among adults. Arthritis Rheum, 57 (4), 656-665. Tan, G., Jensen, M. P., Thornby, J., & Sloan, P. A. (2008). Negative emotions, pain, and functioning Psychological Services, 5 (1), 26-35. 68

PAGE 69

Truchon, M. (2001). Determ inants of chronic disab ility related to low back pain: towards an integrative biopsychosocial model. Disabil Rehabil, 23 (17), 758-767. Truchon, M., & Cote, D. (2005). Predictive validity of the Chronic Pain Coping Inventory in subacute low back pain. Pain, 116 (3), 205-212. Tschannen, T. A., Duckro, P. N., Margolis, R. B. & Tomazic, T. J. (1992). The relationship of anger, depression, and perceived disability among headache patients. Headache, 32 (10), 501-503. Turk, D. C., & Melzack, R. (2001). The measur ement of pain and the assessment of people experiencing pain. In D. C. Turk & R. Melzack (Eds.), Handbook of Pain Assessment (pp. 3-14). New York: Guilford Press. van Middendorp, H., Lumley, M. A., Jacobs, J. W., Bijlsma, J. W., & Geenen, R. (2010). The effects of anger and sadness on clinical pain reports and experimentally-induced pain thresholds in women with and without fibromyalgia. Arthritis Care Res, 62 (10), 13701376. Vlaeyen, J. W., Kole-Snijders, A. M., Boer en, R. G., & van Eek, H. (1995). Fear of movement/(re)injury in chronic low back pain and its relation to behavioral performance. Pain, 62(3), 363-372. Vlaeyen, J. W., & Linton, S. J. (2000). Fear-a voidance and its conse quences in chronic musculoskeletal pain: a state of the art. Pain, 85(3), 317-332. Voulgari, A., Lykouras, L., Papanikola ou, M., Tzonou, A., Danou-Roussaki, A., & Christodoulou, G. (1991). Influence of psychological and c linical factors on postoperative pain and narcotic consumption. Psychother Psychosom, 55(2-4), 191-196. Waddell, G., Newton, M., Henderson, I., Somerville, D., & Main, C. J. (1993). A FearAvoidance Beliefs Questionnaire (FABQ) and the role of fear-avoidance beliefs in chronic low back pain and disability. Pain, 52(2), 157-168. Waddell, G., Somerville, D., Henderson, I., & Newt on, M. (1992). Objective clinical evaluation of physical impairment in chronic low back pain. Spine (Phila Pa 1976), 17 (6), 617-628. Wade, J. B., Price, D. D., Hamer, R. M., Schw artz, S. M., & Hart, R. P. (1990). An emotional component analysis of chronic pain. Pain, 40(3), 303-310. Walker, B. F. (2000). The prevalence of low back pain: a systematic review of the literature from 1966 to 1998. J Spinal Disord, 13 (3), 205-217. Watson, D., & Pennebaker, J. W. (1989). Health complaints, stress, and distress: exploring the central role of negative affectivity. Psychol Rev, 96 (2), 234-254. 69

PAGE 70

70 Watt, M. C., Stewart, S. H., Lefaivre, M. J., & Uman, L. S. (2006). A brief cognitive-behavioral approach to reducing anxi ety sensitivity decreases pain-related anxiety. Cogn Behav Ther, 35(4), 248-256. Wegner, D. M. (1994). Ironic pr ocesses of mental control. Psychol Rev, 101 (1), 34-52. Westman, A., Linton, S. J., Ohrvi k, J., Wahlen, P., & Leppert, J. (2008). Do psychosocial factors predict disability and health at a 3-y ear follow-up for patients with non-acute musculoskeletal pain? A validation of the Orebro Musculoskeletal Pain Screening Questionnaire. Eur J Pain, 12 (5), 641-649. Williams, L. J., Jacka, F. N., Pasco, J. A., Be rk, M., & Dodd, S. (2006). Depression and pain: An overview. Acta Neuropsychiatrica, 18 (2), 79-87.

PAGE 71

BIOGR APHICAL SKETCH Anne Noelle Nisenzon is originally from V oorhees, NJ, and graduated summa cum laude from Boston University with a Bachelor of Arts in psycholo gy. Prior to entering graduate school, she obtained functional neuroimaging research experience at Massachusetts General Hospital in Boston, MA. Subseque ntly, she enrolled in the doctoral program in Clinical and Health Psychology at the Universi ty of Florida, and earned her Master of Science degree in 2008 in neuropsychology. Shortly thereafte r, she joined the Center for Pain Research and Behavioral Health and collaborated on numerous projects examining patient-centered outcomes in pain treatment, the use of placebo in pain, and i nvestigating patient-provider communication using virtual human technology. After completion of a clinical internship at the University of California, San Diego and VA Healthcare System in San Diego, CA, she will commence a postdoctoral fellowship at the San Diego VA Health care System in behavioral medicine. Her clinical interests lie mainly in the area of beha vioral medicine and tertia ry care, namely treating those who are adjusting to chronic pain, injury, or other medical illn esses. Her research interests are primarily in the area of bi opsychosocial treatment techniques for chronic pain in different medical populations. 71