|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 TRAUMA THERAPISTS QUALITY OF LIFE : THE IMPACT OF INDIVIDUAL AND WORKPLACE FACTORS ON COMPASSI ON FATIGUE AND COMPASSION SATISFACTION By LINDSAY G. LEONARD 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 2008
2 Lindsay G. Leonard
3 To my family, for your love, support, and encourag ement; to my best friend and the love of my life, Shane, for everything that you are and ever ything you have brought to my life; and also to Sondra Smith-Adcock, for your mentorship and guidance, and belief in my capabilities.
4 ACKNOWLEDGMENTS I would first like to thank m y supervisor y committee chair, Dr. Sondra Smith-Adcock. Sondra helped me at every step of the way and also provided emotional support and encouragement whenever I got stuck. Words ca nnot fully describe how much Sondras insight and support meant to me over the last 4 years. I would also like to express gratitude towa rd the other members of my supervisory committee for their expertise, guida nce, and support. Each of th ese three people offered a unique perspective that ultimately helped me to crea te a methodologically sound and important study. I thank Dr. West-Olatunji for opening my eyes to ne w ways of thinking as a clinician, researcher, and professor. I thank Dr. Shaw for her interest in my topic, he r expertise in the academic world, and her warm demeanor. I thank Dr. Griffin for his shared knowledge in the field of crisis and trauma and his mentorship. Additional thanks go to Dr. Jocelyn Lee for her expertise in statistical analysis, her guidan ce through the methodological process, and her efficiency. Last, but certainly not least, I want to thank my family for their emotional and financial support over my many years of study I especially thank my parents Brad and Lora for their ongoing encouragement which helped me reach my academic goals. I thank my brother Adam for all the theoretical discussions and collaborative insight. I thank my sister Stephanie for her creativity and her inspir ing words. I am also thankful for my best friend, my love, Shane McKim, for his support and unwavering belief in me.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4LIST OF TABLES................................................................................................................. ..........8LIST OF TERMS.............................................................................................................................9ABSTRACT...................................................................................................................................11 CHAP TER 1 INTRODUCTION..................................................................................................................12Scope of the Problem........................................................................................................... ...17Theoretical Framework.......................................................................................................... .19Trauma Theory................................................................................................................20Empathy and Countertransference.................................................................................. 21Constructivist Self-Development Theory........................................................................ 22Integrated Theoretical Framework.................................................................................. 22Need for the Study..................................................................................................................24Research Problem...................................................................................................................25Research Questions............................................................................................................. ....262 LITERATURE REVIEW.......................................................................................................27Negative Effects of Trauma on the Therapist......................................................................... 30Vicarious Traumatization................................................................................................30Burnout............................................................................................................................33Compassion Fatigue........................................................................................................ 35Interrelationships among Po sitive and Negative Effects........................................................ 44Workplace Factors.............................................................................................................. ....45Recommendations for Therapists and the Workplace............................................................ 49Need for the Study..................................................................................................................503 METHODOLOGY................................................................................................................. 52Research Design and Relevant Variables...............................................................................53Research Hypotheses............................................................................................................ ..54Population and Sample.......................................................................................................... .55Sampling Procedure................................................................................................................56Instrumentation................................................................................................................ .......57Professional Quality of Life Scale (ProQOL)................................................................. 57Psychologists Burnout Inventory (PBI)......................................................................... 59Stressful Life Experiences Screening-Short Form.......................................................... 60Data Analysis..........................................................................................................................60
6 4 RESULTS...............................................................................................................................63Demographics.........................................................................................................................63Measurement Reliabilities...................................................................................................... 65Correlational Analyses......................................................................................................... ...66Regression Analyses............................................................................................................ ...67Hypothesis 1: Relationships among wor kplace variables and compassion fatigue........ 68Hypothesis 2: Relationships among workplace variables and compassion satisfaction...................................................................................................................69Hypothesis 3: Relationships among indi vidual and workpl ace variables and compassion fatigue....................................................................................................... 69Hypothesis 4: Relationships among indi vidual and workpl ace variables and compassion satisfaction............................................................................................... 70Summary.................................................................................................................................71DISCUSSION................................................................................................................................77Overview of the Study............................................................................................................77Discussion of Descri ptive Statistics........................................................................................ 77Discussion of Correlational Rela tionships Among Variables................................................ 82Discussion of Hypotheses....................................................................................................... 85Relationships among Control, Overinvol vement, Negative Clientele, Workplace Support, amount of Secondary Exposure, and Compassion Fatigue........................... 85Relationships among Control, Overinvol vement, Negative Clientele, Workplace Support, amount of Secondary Exposur e, and Compassion Satisfaction....................88Relationships among Individual and Wor kplace Variables and Compassion Fatigue.... 90Relationships among Individual and Workplace Variables and Compassion Satisfaction...................................................................................................................92Theoretical Implications....................................................................................................... ..93Practical Implications......................................................................................................... ....95Research Implications and Future Directions......................................................................... 99Limitations.................................................................................................................... ........101Summary...............................................................................................................................104APPENDIX A PROFESSIONAL QUALITY OF LIFE SCALE................................................................. 105B PSYCHOLOGISTS BURNOUT INVENTORY................................................................ 107C STRESSFUL LIFE EXPERIENCESSHORT FORM...................................................... 108D DEMOGRAPHIC QUESTIONNAIRE................................................................................ 110E INFORMED CONSENT......................................................................................................112
7 LIST OF REFERENCES.............................................................................................................114BIOGRAPHICAL SKETCH.......................................................................................................120
8 LIST OF TABLES Table page 4-1 Descriptive Statistics for Ca tegorical Individual V ariables............................................... 73 4-2 Descriptive Statistics for Continuous Individu al and Workplace Variables..................... 74 4-3 Measurement Reliabilities for ProQOL, PBI, and SLES................................................... 75 4-4 Pearson Product Moment Correlation s between Independent and Dependent Variables ............................................................................................................................76
9 LIST OF TERMS Burnout A state of physical, emotional and mental exhaustion caused by longterm involvement in emotionally demanding situations. Compassion fatigue The beha viors and emotions resulting from a close, empathic connection with a traumatized or suffering person. Compassion fatigue involves PTSD-like symptoms in the therapist including emotional di stress, cognitive shifts, and relational disturbances. Compassion satisfaction The pleasure derived from effectiv ely helping others through psychotherapeutic work. Compassion stress The emotional energy from the empathic response and the desire to relieve the suffering of the client. Control Therapists control over their wo rk activities and decisions. Control is a factor asso ciated with burnout. Counseling psychologists Professional therapists with a license in Psychology who have completed a doctoral degree from a Counseling Psychology program. Countertransference The emotional response a therapist has to a client related to personal issues and internal conflicts. Empathy A cognitive-affective state that involves both intellectually understanding the perspectiv e of another person and experiencing the emotions of another person. Empathic response When the therapist uses empathic understanding to put him or herself into the phenomenological world of the client, experiencing the suffering and other emotions experienced by the client. Negative clientele Therapists experi ence with aggressive, dangerous, or threatening client behaviors as well as clients with severe psychopathology and/or high le thality. Experience with negative clientele is a factor associated with burnout. Overinvolvement Therapists feelings of being too involved (e.g., overresponsibility, excessive emotiona l involvement) in the care of their clients. Overinvolvement is a factor associated with burnout and is the opposite of disengagement.
10 Professional counselors Professiona l therapists with a license to practice in Professional Counseling who have completed a masters, specialist, or doctoral degree in a Counseling or Counselor Education program. Psychologists Professional therapists with a license to prac tice in Psychology who have completed a masters or doctoral degree in Counseling or Clinical Psychology. Social support Assistance provided by significant others involving emotional support, information, social companionship and instrumental support. Social workers Professional therapists with a license to practic e Clinical Social Work who have completed at least a masters degree in a Social Work program. Trauma When a person experienced or witnessed an event or events that involved actual or threatened death or serious injury or a threat to the physical integrity of self or others and is beyond ones ability to cope with current resources. Trauma therapist A therapist who spends a significant amount (at least 25% of caseload and 15-20% of total hour s) of their client contact hours working with traumatized clients. Vicarious traumatization Changes in the th erapists worldview, psychological needs, belief systems, and cognitions as a result of empathic engagement with the clients trauma material. Vicarious posttraumatic growth Positive psychological changes that individuals may experience as a result of an intimate, empathic relationship with someone who is struggling with highly stressful or traumatic life circumstances. Vicarious posttraumatic growth involves changes in beliefs about se lf, others, and the world. Workplace support Ones professional p eer group providing various forms of social support (e.g., emotional, instrumental, informational).
11 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 TRAUMA THERAPISTS QUALITY OF LIFE : THE IMPACT OF INDIVIDUAL AND WORKPLACE FACTORS ON COMPASSION FATIGUE AND COMPASSION SATISFACTION By Lindsay G. Leonard May 2008 Chair: Sondra Smith-Adcock Major: Mental Health Counseling Therapists working with trauma survivors can experience both psychological costs and benefits as a result of this work. The ps ychological cost of compassion fatigue involves symptoms of traumatic distress, cognitive shifts, and relational disturbances, whereas the psychological benefit of compassion satisfaction consists of pleas ure derived from effectively helping clients in distress. In this study, rela tionships between individu al factors (e.g., gender, ethnicity, personal trauma history, years of clinical experience, trauma training) and workplace factors (e.g., perceived control, overinvolveme nt, negative clientele, workplace support, amount of secondary exposure) were examined to determ ine their impact on levels of compassion fatigue and compassion satisfaction. Participants, who included 98 trauma therapists, completed an online survey consisting of the Professional Quality of Life Scale, the Psychologists Burnout Inventory, the Stressful Life Experiences-Short Form and a demographics questionnaire. Multiple regression analyses revealed that co ntrol, overinvolvement, and secondary exposure were significantly related to compassion fatigue Control and personal trauma history were significantly related to compa ssion satisfaction, accounting fo r substantial amounts of the variance of each dependent variable (i.e., compassi on fatigue and compassion sa tisfaction).
12 CHAPTER 1 INTRODUCTION Som etimes after a session, I w ill be traumatizedI will feel overwhelmed, and I can remember a particular situation with a sexuall y abused person where II just didnt want to hear any more of her stories about what actually happened. She seemed to want to continue to tell me those over and over and I remember feeling almost contaminated, you know, like I was abused (Be ll, 1998; as quoted in Bell, Kulkarni, & Dalton, 2003). Therapists working with clients coping with traumatic experiences provide an invaluable service to the clients and comm unities they serve. Working w ith this population requires the therapist to be extremely present, empathic, and skilled. It is highly emotionally-charged work that can be both draining and rewarding. What ar e the specific psychological costs of this type of caring? What are the psychol ogical benefits? What factors in fluence how this type of work affects the professional? Ho w do professional counselors who work with clients who have endured trauma maximize the benefits and minimize the costs? The potential psychological costs of caring for mental health professionals are often referred to in the literature as compassion fatig ue or secondary traumatic stress (Figley, 1995; Monroe et al, 1995), vicarious traumatiza tion ( McCann & Pearlman, 1990; Pearlman & Saakvitne, 1995), and burnout ( Pines, 1993 ). The potential psychological benefits are labeled in the literature as compassion satisfaction (Stamm, 2002) or vi carious posttraumatic growth (Arnold, Calhoun, Tedeschi, & Cann, 2005) Various factors influence therapists experience of positive or negative outcomes, including personal or individual factors, self-care strategies, workplace aspects, client character istics, and social support (Figle y, 2002). All of these factors impact the therapeutic relationship and the therap ists experience of positive or negative personal outcomes. Many of these factors are specific to the individual professional. Workplace factors also have a strong impact on therapists pers onal psychological response to trauma work. Workplace factors, in contrast to individual factors, often can be altered on a systemic level to
13 positively affect a large number of professionals. Therefore, studies that consider individual as well as workplace factors are needed. Of the terms used to describe the psychol ogical costs of caring, the term compassion fatigue, previously referred to as secondary traumatic stress, cons ists of behaviors and emotions that often follow from helping a significant othe r who has experienced a traumatic event (Figley, 1993). Given the close connection that therapists have with their c lients, they can be especially vulnerable to compassion fatigue. Compassion fatigue usually develops quickly following secondary exposure to trauma and includes re-exper iencing of the clients traumatic event as it was recounted to them, avoidance or numbing of emotions, and physiological arousal. In comparison, burnout, which is defined as physical emotional and mental exhaustion caused by longterm involvement in an emotionally demanding environment, often co-occurs with compassion fatigue (Pines and Aronson, 1988; Trippany, Kress, & Wilcoxon, 2004). Therapists with high levels of both burnout and compassion fatigue are at the greatest risk for negative outcomes, such as depression and PTSD (Stamm, 2005). Why, more specifically, are ment al health professionals work ing with trauma susceptible to negative consequences such as compassi on fatigue and burnout? Much of these negative consequences, especially compassion fatigue, have been attributed to an essential foundation for forming a relationship with clients, empathy. Em pathy as a cognitive-affective state leaves the therapist vulnerable to being traumatized by th e clients suffering thr ough their close connection (Figley, 1995). Following from this empathic c onnection, therapists, part icularly those with a personal trauma history, may experience trauma tic countertransference, which may include overidentification with the traumatized client and/or avoidance reactions (e.g., denial, minimization, disengagement) (Saakvitne, 1995; Wilson & Lindy, 1994). Depending upon how
14 these countertransference issues are dealt wi th, positive (e.g., compassion satisfaction) or negative (e.g., compassion fatigue) outcomes may occur. Considering the consistent documentation of negative outcomes in the literature for therapists working with trauma, why would anyone want to do this work? The psychological benefits, or the energizing psyc hological rewards, that come fr om working with clients who are traumatized have been discussed more recently in the literature, represente d by several concepts. Terms used in the literature include compassi on satisfaction (Stamm, 2002; 2005), posttraumatic growth (Tedeschi & Calhoun, 1996), stress-rel ated growth (Park, Cohen, & Murch, 1996), perceived benefits (McMillen & Fisher, 1998), and thriving (Abraido-Lanza, Guier, & Colon, 1998). Concepts referring to positive psychological outcomes are linked with theories of both resilience and posttraumatic growth. Resilience, which is the ability to maintain a stable equilibrium in the face of adverse or traumatic experiences, is differentiated from recovery, which is equilibrium preceded by a period of decreased psychological functioning (Bonanno, 2004). Recent resilience literatur e purports that resilience is more common than has been described in previous litera ture (Bonanno, 2004). Thus, resilience or psychological growth should also be common in therapists working in highly stressful and tr aumatic environments. Resilience can occur through multiple pathways, including through individual traits such as hardiness as well as the use of repressive c oping behaviors and positive emotions and laughter (Bonanno, 2004). Hardiness refers to a personality in which one is committed to a life purpose, belief that one has control over their surrounding s, and the belief that one can grow from adversity (Kobasa, Maddi, & Kahn, 1982). Re silience theory provides a framework through which positive psychological outcomes for trau ma therapists can be understood.
15 Tedeschi & Calhoun (1996; 2004) use the term posttraumatic growth (PTG) because it focuses more on major crises rather than lowe r level stress and emphasi zes the role of the process of coping with the trauma as the vehicl e for growth. When PTG occurs, it involves an increased appreciation for life in general, more meaningful relati onships, an increased sense of personal strength, changed life prio rities, and a richer existential or spiritual life (Tedeschi & Calhoun, 1996). This idea has recently been exte nded to describe positive consequences for therapists working with traumatized c lients (Calhoun, Tedeschi, & Cann, 2005). Compassion satisfaction has been described more generally as the pleasure derived from effectively helping others through psychotherapeutic work (Stamm, 2005). Stamm (2002) coined this term as a positive parallel to the negative effects related to compassion fatigue. Compassion satisfaction is similar to other terms used to describe positive outcomes associated with stress and trauma such as posttraumatic grow th. However, the concept can be differentiated by its more specific focus on helpers (e.g., c ounselors, crisis responders). Compassion satisfaction has also been described as a sense of efficacy in ones work and ones ability to make a positive impact on the world (Stamm 2002). Burnout and compassion fatigue, especially in combination, have been found to result in less efficacy a nd compassion satisfaction (Stamm, 2002). Previous researchers have begun to identify related factors for negative outcomes, such as compassion fatigue. However, little research has been done to examine factors related to positive outcomes, such as compassion satisfaction. Several factors have been shown to impact the therapists experience of trauma work a nd negative psychological outcomes. Individual factors commonly associated with compassion fatigue include the therapists personal experience with trauma and amount of expos ure to trauma clients (Schaube n & Frazier, 1995; Pearlman &
16 MacIan, 1995). Therapists persona l trauma history in particular has been most consistently shown to be associated with compassion fatigue (Kassam-Adams, 1999; Pearlman & MacIan, 1995). Other individual factors, such as gender and years of experience have had more inconsistent findings across studies Ethnicity has not been system atically investigated as an individual factor associated with compassion fatigue most likely due to the lack of heterogeneity of ethnic background in samples collected. The ethnic background of the therapist may significantly impact their experience of compa ssion fatigue due to cu ltural influences on expression of traumatic responses and definition of trauma (Young, 2001). Gender has been shown to be associated with compassion fatigue in some studies (Kassam-Adams, 1995; Meyers & Cornille, 2002; Wee & Myers, 2002), but not in others (e.g., Pearlman & MacIan, 1995). Years of clinical experience as a related factor for compassion fatigue and other negative psyc hological outcomes also had mi xed findings in prior research, with some researchers finding a relationship (e.g., Cunningham, 2003), whereas others did not (e.g., Kassam-Adams, 1999). Individual factors, incl uding gender, years of clinical experience, and personal trauma history, need to be re-eva luated to determine their relationship with compassion fatigue. Individual factors such as ethni city and specialized trauma training need to be evaluated to determine their asso ciation with compassion fatigue. When workplace factors have been examined, lack of availability of supervision or consultation and other forms of workplace suppo rt, lack of perceived control over work activities, working with more highly disturbed clientele with difficult client behaviors, overinvolvement with clients, and a larger number of hours spent with trau ma clients have been associated with negative ther apist outcomes (Ackerley et al., 1988; Kassam-Adams, 1995;
17 Pearlman & Maclan, 1995; Rupert & Morgan, 2005). Amount of exposure to traumatized clients, defined as a heavier caseload of trau ma survivors with more weekly hours of client contact, was found to be related to compassion fatigue in many studies (e.g., Brady et. al, 1999; Chrestman, 1999; Kassam-Adams, 1999; Schauben & Frazier, 1995), but a few studies found no relationship (e.g., Cunningham, 2003; Landry, 2001). The majority of the literature concerning workplace factors and negative therapist outcomes involved examinations of burnout, rather than compassion fatigue. Therefore, a study is needed to determine the impact of workplace factors on both the negative and positive effects of trauma work (i.e., compassion fatigue and compassion satisfaction), rather than burnout. The current study examined relationships betw een individual factors and workplace factors associated with compassion fatigue and compassi on satisfaction. The res earcher in the present study investigated how individual factors of persona l trauma history, years of clinical experience, gender, ethnicity, and trauma training along w ith workplace factors of sense of control, overinvolvement, negative clientele, workpl ace support, and amount of secondary exposure impacted therapists levels of compassion fatig ue and compassion satisfaction. The practical significance of this study is to inform agencies and counseling centers on ways they can set up the work environment to minimize compassion fa tigue and maximize compassion satisfaction for therapists. Scope of the Problem A growing num ber of mental he alth counselors are increasingl y working with more clients who have experienced some form of trauma (Trippany, White Kress, & Wilcoxon, 2004). Common types of trauma experienced by client s and encountered vicar iously by therapists include childhood sexual abuse, sexual assault, do mestic violence, natural disasters, and school violence (Trippany, White Kress, & Wilcoxon, 2004). Recent widespread traumatic events, such
18 as the terrorist attacks of 9/11, Hurricane Katr ina, the War on Terror, and the Virginia Tech shootings impact the number of people affected by traumatic events. Additionally, the media allows a greater number of people to be exposed to traumatic images through the news broadcasts depicting the traumatic events. Thus, more people may be traumatized or retraumatized by the viewing of these media imag es. Because therapists are treating more traumatized clients, the topic of compassion fati gue is becoming more relevant and crucial to address. Statistics that report the widespread nature of trauma signify the various types of trauma work as well as the high incidence of this work. It has been estimated that 1 in 6 women and 1 in 10 men will be victims of childhood sexual abuse (Trippany, White Kress, & Wilcoxon, 2004). One in five women (21 percent) reported being raped or physically or sexually assaulted in her lifetime (Commonwealth Fund, 1999). Nearly one-third of American women (31 pe rcent) report being physically or sexually abused by a husband or boyfriend at some point in their lives (Commonwealth Fund, 1999). According to a recent survey, 20 percent of public schools have experienced one or more violent crimes such as sexual assault, robbery, and aggravated assault (US Department of Education and US Department of Justice, 2003). Disasters may include natural (e.g., earthquak es, tornadoes) or human made (e.g., war, criminal violence, public health disaster) and include traumatic reactions with similar, yet somewhat unique traumatic features. As therapists are increasingly exposed to treati ng traumatized clients, it is assumed that the incidence of compassion fatigue may increase as well. As a result of secondary exposure to these traumatic events, the prevalence of ex treme distress in therapists who work with traumatized clients has been increasing (M eldrum, King, & Spooner, 2002; Wee & Myers, 2002). One study found that 27 percen t of therapists experienced ex treme distress as a result of their work (Meldrum, King, & Spooner, 2002). Another study reveled that 64.7 percent of
19 trauma workers exhibited some degree of posttr aumatic stress symptoms (Wee & Myers, 2002). Thus, therapists, especially those exposed to tr auma through their work wi th clients, are at a greater risk for developing negative psychologi cal outcomes such as compassion fatigue. Compassion fatigue not onl y negatively impacts therapists, but also the clients they serve and the organizations in which they work. Compassion fatigue, as mentioned previously, involves symptoms similar to th at of posttraumatic stress disorder (PTSD). Therapists who experience compassion fatigue may show sympto ms such as re-experiencing the clients traumatized event, avoidance behaviors, di minished affect, hypervigilance, difficulty concentrating, physiological reactiv ity, and irritability (Figley, 1995). These symptoms may lead to a diminished capacity to provide effective ps ychotherapy to their clients. Thus, client care may suffer. Additionally, the th erapists personal relationships and home life may suffer as a result of these symptoms. There are also significant negative impacts on the organization or workplace. First, high therapist turnover is a pred ictable outcome of compassi on fatigue and other negative consequences (e.g., burnout) (Sexton, 1999). Hi gh therapist turnover impacts the workplace through greater need for training new workers an d difficulty in creating a cohesive workplace community. Other outcomes for the workplace include a loss of energy, commitment, and optimism, which has a deleterious effect on th e emotional culture of the work environment (Sexton, 1999). Thus, there is reason to be concerned for the psychological health and wellbeing of therapists working with traumatized clie nts and a necessity to find effective ways to help buffer them from stress and other negative consequences. Theoretical Framework There exists no single unifying theory for the constructs of compassion fatigue or com passion satisfaction. Although Figley (1995; 2002) proposed a m odel to describe the process
20 of developing compassion fatigue, a single underlying theoretical framework has not been discussed in literature on compa ssion fatigue nor compassion satisfa ction. In this section, this researcher will build a theoretical framework th at encompasses these relevant constructs. Trauma Theory The construct of com passion fatigue, also referre d to as secondary traumatic stress, shares terminology associated with the symptomology of posttraumatic stress disorder (PTSD). Thus, the psycho-physiological theory of trauma res ponse will be reviewed as it relates to the phenomenon that therapists experience vicarious ly through being closely connected with their clients trauma material. PTSD involves three major categories of symptoms: (1) reexperiencing the event in sensory form; (2) avoidance of reminders of the trauma; (3) chronic hyperarousal in the autonomic nervous system (Rothschild, 2000). These symptoms also occur with compassion fatigue. During the experience of a traumatic event, th e limbic system in the brain activates the amygdala, which signals the hypothalamus, mobilizi ng the body to fight, flee, or freeze. The normal response to threat is halted by the produc tion of cortisol. However, in persons with PTSD there is a deficiency in cortisol producti on which makes it more difficult to halt the bodys alarm response (Rothschild, 2000). Thus, peopl e with PTSD are more easily physiologically aroused and have greater difficulty returning to a resting state. Memory is also affected by a traumatic e xperience. During a traumatic event, the hippocampus is suppressed, which is normally fu nctioning in the processing and storage of events. The hippocampus is responsible for d eclarative or narrative memory which puts the memory in context and proper time and space. During a trauma, the amygdala is solely responsible for processing the event. The amyg dala is responsible for implicit or sensory memory and is linked with highly charged emo tions. Thus, traumatic memory is usually
21 experienced as flashbacks that have no acco mpanying narrative and are highly emotional (Rothschild, 2000). This explanation of the phys iological response to trauma may shed light on therapists experience of compassion fatigue. A major difference is that the traumatic event occurs in the clients recounting of the trauma tic event, which the therapist is empathically connected with. Empathy and Countertransference Em pathy has been defined as the action of understanding, being aware of, being sensitive to, and vicariously experiencing the feelings, thoughts, and experience of another (MerriamWebsters Collegiate Dictionary, 1998). Empath ic connection involves putting oneself into the phenomenological world of the client (Wils on & Lindy, 1994). Empathic connection is a foundational skill for all therapists and is necessa ry to build rapport and establish a safe and trusting environment in which healing can occur. However, in working with traumatic material, therapists may have difficulty maintaining e nough emotional distance to protect them from vicariously experiencing the nega tive effects of the trauma. The experience of compassion fatigue necessitates an empathic connection with a traumatized individual. The c ognitive-affective theory of em pathy describes empathy as a vicarious experiencing of another persons private world through both cognitively understanding and feeling the other persons lived experience (Rogers, 1959; Duan & Hill, 1996). According to this theory, there are two types of empathy: c ognitive and affective (Gla dstein, 1983). Cognitive empathy refers to intellectually taking the perspective of another person, whereas affective empathy denotes experiencing the emotions of the other person (Gladstein, 1983). Affective empathy is a central construct in Figley s (1995, 2002) model of compassion fatigue. Countertransference, originally posited as part of psychodynamic theory, is a widely accepted phenomenon that can occur in the therapeu tic context. Countertransference refers to
22 the activation of the th erapists unresolved issues through wo rking with a clients material (McCann & Pearlman, 1990). Wilson and Lindy (1994) identified two types of defensive countertransference reactions co mmon to trauma therapists: a voidance reactions and overidentification reactions. Avoi dance reactions include denial, minimization, distortion, detachment and disengagement from the empath ic connection. Over-identification involves idealization, enmeshment, and exce ssive advocacy for the client. Constructivist Self-Development Theory Constructivist Self-Developm ent Theory (CSDT) is a useful theory for understanding the cognitive changes that occur for therapists e xperiencing trauma vicariously through therapeutic work with clients. CSDT asserts that human beings construct thei r own realities through development and use of cognitive schemas (McCann & Pearlman, 1990). These schemas include assumptions about self, others and the wo rld which allow people to make sense of their experiences. Therapists may experience disruptions in these basic schemas of self, others, and the world through their work with trauma survivors (McCann & Pearlman, 1990). CSDT outlines five components of the se lf that are affected through vicarious traumatization: (1) frame of reference; (2) self-capacities; (3) ego resources; (4) psychological needs; and (5) cognitive schemas, memory, and perception (Pearlman & Saakvitne, 1995). These changes in the therap ists core beliefs can lead to feelings of helplessness and loss of control which can chal lenge the counselors identity (Pearlman & Saakvitne, 1995). Integrated Theoretical Framework No single unifying theory exists to encom p ass the constructs of compassion fatigue, compassion satisfaction, and how the work environment impacts the occurrence of these phenomena. A physiological theory of trauma was explicated to account for the physiological
23 and behavioral symptoms associated with comp assion fatigue. Physiological changes create a vicious cycle with changes in th e cognitive system that is explained by the CSDT model. The unique exposure of the therapist to the clients trauma material is explained by the process of empathic connection and countertransference. A theoretical and causal model for compassion fatigue was proposed by Figley (1995) and subsequently revised (Figley, 2002). The model rests on the assumption that empathy is an essential ingredient in the therapeutic relations hip and is necessary to effectively work with clients (Figley, 2002). However, empathic engagement can also have costs, especially for therapists working with traumatized clients because it leaves the therapist vulnerable to indirectly experiencing the clients trauma. Figleys (1995) compassion fatigue model c ontains ten key variables involved in the occurrence of compassion fatigue. Fi rst, the therapist must have th e ability to empathize with the client and the empathic concern to motivate them to help others in need. Next, exposure to the client involves experiencing the emotional pain of the client through direct exposure. The empathic response occurs when the therapist uses empathic understanding to put him or herself into the phenomenological world of the client, experiencing the suffering and other emotions experienced by the clie nt (Figley, 2002). The therapist may then experience compassion st ress, which is the residue of emotional energy (p.1437) from the empathic response and the desire to relieve the suffering of the client (Figley, 2002). When the ther apist experiences compassion stress, it can develop into compassion fatigue unless protective factors are present. One of these protective factors is sense of achievement, which is the therapists sense of sa tisfaction with his or her efforts to assist the client. Sense of achievement is consistent w ith the concept of compassion satisfaction which
24 thus fits within Figleys (1995) model. A nother protective factor is disengagement. Disengagement involves distanci ng oneself emotionally from the clients pain between sessions and letting go of thoughts and emotions concerning the client in order to live their life (Figley, 2002). The concept of disengagement in its oppos ite form is overinvolvement, which is related to the inability to disengage from the trauma of clients. There are also several factor s that can exacerbate the condi tion of compassion stress to develop into compassion fatigue. First, prolonged exposure to traumatized clients involves the therapists ongoing sense of res ponsibility for relieving the suffe ring of the client over a period of time. Prolonged exposure, or amount of seconda ry exposure to traumatized clients, has been consistently associated with compassion fati gue in the literature (Kassam-Adams, 1995; Pearlman & Maclan, 1995). Second, traumatic reco llections may occur, or memories associated with the clients trauma material that elicits an emotional reaction for the therapist. Third, disruptions in the therapists life (e.g., illness, li fe changes) that would normally be manageable but may compound compassion stress to lead into compassion fatigue (Figley, 2002). Need for the Study The occurrence of com passion fatigue and other negative psychological outcomes for trauma therapists have been well-documente d in the literature (e.g., Maslach, 1993; McCann & Pearlman, 1990; Salston & Figley, 2003). Howeve r, most research has focused on individual characteristics (e.g., gender, amount of experience personal trauma history) associated with these concepts rather than on the characterist ics of the workplace (Bell, Kulkarni, & Dalton, 2003), with inconsistent findings Though research on the construct of burnout has begun to focus on factors in the workplace that are asso ciated with burnout (e.g., Ackerley et al., 1988; Rupert & Morgan, 2005), more research is needed on the organizational correlates of compassion fatigue and compassion satisfacti on. Numerous studies have recommended
25 interventions at the organi zational level (e.g., Sexton, 1999; Collins and Long, 2003), however it is unknown to what extent these interventions occur in different work settings or how much they help. Although factors associated with negative psycho logical outcomes for therapists have been discussed in the literature, there has been li ttle focus on factors associated with positive psychological outcomes for therapists (e.g., comp assion satisfaction). A broader understanding of contextual factors related to both compassion fatigue and co mpassion satisfaction is needed (Salston & Figley, 2003). In the current study, this researcher examined the relationship between individual factors (i.e., personal trauma history, years of clinical experience, gender, ethnicity, trauma training) and workplace factors (i.e., control, overinvolvement, negative clientele, workplace support, amount of secondary exposur e) and compassion fatigue and compassion satisfaction. Results demonstrated the extent to which workplace factors in addition to individual factors account fo r the variance in compassion fatigue as well as compassion satisfaction. Research Problem The m ain problems that was addressed in th e present study involved assessing the impact of individual and workplace factors on the expe rience of negative psychological outcomes (i.e., compassion fatigue) and positive outcomes (i.e., compassion satisfaction) for therapists who work with traumatized clients. To explore this problem, five workplace factors known to be associated with burnout were assessed: (1) contro l, (2) overinvolvement, (3) negative clientele, (4) workplace support, (5) amount of secondary exposure (Ackerley et al., 1988). Individual factors of gender, ethnicity, pe rsonal trauma history, years of experience, and trauma training were also assessed to better account for variance in compassion fatigue. The population of interest was Professional Couns elors, Psychologists, and Social Workers working with
26 traumatized client populations. The outcome of the study was intended to address the question of how individual factors and wor kplace factors were related to le vels of compassion satisfaction and compassion fatigue. Research Questions The following research questions w ere proposed in this study: 1. What is the relationship among workplace fact ors (e.g., control, overinvolvement, negative clientele, workplace support, amount of secondary exposure) and compassion fatigue? 2. What is the relationship among workplace fact ors (e.g., control, overinvolvement, negative clientele, workplace support, amount of sec ondary exposure) and compassion satisfaction? 3. Does the addition of workplace factors (e.g., co ntrol, overinvolvement, negative clientele, workplace support, amount of secondary exposure) significantly improve the variance accounted for by individual factors (e.g., gender, ethnicity, personal trau ma history, years of experience, trauma training) in compa ssion fatigue and compassion satisfaction?
27 CHAPTER 2 LITERATURE REVIEW Therapis ts, particularly those working with clients healing fr om traumatic events, have the potential for rewarding, growth-enha ncing experiences as well as negative consequences such as burnout and psychological distress as a result of their therapeutic wo rk (Maslach & Leiter, 1997). It is understandable that trauma therapists have the potential to experience extreme stress due to the high demands of their work, including empathic strain from repeated exposure to clients traumatic material and emotional pain. Cush way and Tyler (1996) have observed consistent levels of moderate to significant distress among mental health practitioners. Numerous other studies have found similarly high levels of trauma tic stress in therapists (e.g., Brady et al., 1999; Kassam-Adams, 1999; Chrestman, 1999). Overa ll distress is mediated by both risk and resiliency factors including the therapists personal characteristics, characte ristics of the client and the trauma, the therapists attempts to cope, and the environment in which the therapy takes place (Dutton & Rubenstein, 1995; Figley, 1995; Ch restman, 1999). These risk and resiliency factors impact the therapeutic relationship and the therapists consequent experience of positive or negative personal and professional outcomes. Trauma counselors are trained to work with clients in the aftermath of traumatic events (Dutton & Rubinstein, 1995). A traumatic stress or is defined diagnosti cally as when a person experienced, witnessed, or was confronted with an event or events that involved actual or threatened death or serious injury or a threat to the physical integrity of se lf or others (American Psychiatric Association, p.427). Each persons experience of tr auma is different. Trauma therapists may experience psyc hological distress through effects of their exposure to clients traumatic material (Dutton & Rubenstein, 1995).
28 Why are therapists who work with traumatized clients vulnerable to negative consequences? First, therapists display empathy as a major resource in helping their clients. Empathy helps the therapist understand and emotiona lly connect with the c lients experience, but it also leaves the therapist vulne rable to being traumatized by the client trauma material as a result of this intimate connecti on (Figley, 1995). The concept of empathy has been defined in a variety of ways throughout years of theoretic al writings on the topic (Duan & Hill, 1996). Rogers (1959) defined empathy as a process of feeling as if one actually were that other person. The cognitive-affective theory of empathy desc ribes empathy as a vicarious experiencing of another persons private world through both cognitively understanding and feeling the other persons lived experience (Rogers, 1959; Duan & Hill, 1996). A cognitive-affective state of empathy can be broken down and described in two parts (Gladstein, 1983). First, cognitive empathy refers to intellectua lly taking the perspective of another person. Second, affective empathy denot es experiencing the emotions of the other person (Gladstein, 1983). Affective empathy wa s focused on in Figleys (1995, 2002) model of compassion fatigue as one of the central constructs. Aspects of working with trauma survivors in therapy that make the trauma therapist susceptible to negative outcomes include increased awareness of the potential for trauma in their own life, the difficulty of working through client s powerful emotions an d mistrust in therapy, clients reenactments of earlier relationships projected onto the therapist, and the helplessness that ensues when therapists witness the rec ounting of the clients traumatic experiences (Pearlman & Saakvitne, 1995). Therapists own history of traumatic events and unresolved issues related to these experiences may incr ease their risk for negative consequences. Therapeutic countertransference, es pecially as it relates to the therapists unresolved traumas, has
29 been related to the risk of negative outcomes for therapists Saakvitne (1995) defines countertransference as the aff ective, ideational, physical re sponses a therapist has to a clientand the therapists conscious and unconsci ous defenses against affects, intrapsychic conflicts and associatio ns around the former (p.23). W ilson and Lindy (1994) identified two types of defensive countertransference reactio ns common to trauma therapists: avoidance reactions and over-identification reactions. Av oidance reactions includ e denial, minimization, distortion, detachment and disengagement from th e empathic connection. Over-identification involves idealization, enmeshment, and excessive advocacy for the client. Countertransference reactions can be useful tools to increase unde rstanding of the client and can be effectively processed through supervision and consultation (Pearlman & Saakvitne, 1995). Depending upon how countertransference issues ar e dealt with, the ther apist can experience either positive or negative consequences both personally and in the therap eutic interaction. Therapeutic work can also be extremely rewa rding and contribute to therapists personal growth and spiritual tr ansformation. Therapists have repo rted enjoying creativity, strength, and witnessing the resilience of su rvivors which led to a sense of spiritual growth and deeper intimate relationships with significant others (Schauben and Frazier, 1995; Pearlman & Saakvitne, 1995; Arnold, Calhoun, Tedeschi, & Cann, 2005). The literature reviewed contains studies predominantly concerning the negative effects for trau ma counselors, however recent literature has begun to explore the po sitive effects for counselors. There are many explanatory factors for ther apists positive and negative psychological outcomes including individual, workplace, and cl ient factors (Figley, 1995). The following literature review is organized as follows: review of the negative and positive effects of trauma on the therapist, associated factors, theoretical foundations, interrelations hips between positive and
30 negative effects, and a discussion of workplace factors that impact therapists psychological outcomes. Negative Effects of Trauma on the Therapist Various term s have been described in th e literature on the negative effects of psychotherapeutic work with traumatized client s. It is important to review, define and differentiate how these various te rms clarify constructs related to how working with traumatized clients negatively impacts therapists. The fo llowing terms have been used to describe the negative effects for counselors : vicarious traumatization (McC ann & Pearlman, 1990; Pearlman & Saakvitne, 1995); traumatic countertransfer ence (Herman, 1992); burnout (Pines, 1993); and compassion fatigue or secondary traumatic stress disorder (Figley, 1995). These terms are often used interchangeably to describe the variety of potential negati ve effects for counselors working with trauma victims. The te rms vicarious traumatization a nd burnout will be reviewed and differentiated from compassion fatigue, which w ill also be defined and discussed. Relevant literature for each of these constructs will also be reviewed in the following section. These three terms related to negative therapist outcomes we re chosen for this literature review because current theoretical and empirical li terature frequently utilizes these terms as compared to other terminology previously used to describe ne gative effects for the trauma therapist. Vicarious Traumatization Vicarious traum atization has been defined as t he transformation in th e inner experience of the therapistas a result of empathic engagement with the clients trauma material (McCann & Pearlman, 1990, p.145). In the process of providi ng therapy for clients exposed to trauma, the traumatic material can begin to affect the th erapists worldview, emotional and psychological needs, belief systems, and cognitions (Pearlman & Saakvitne, 1995). This process is viewed as normal and inevitable when therapists work with clients healing from traumatic events.
31 Although inevitable, if vicarious traumatization is not addressed and worked through (e.g., in supervision, consultation, or pers onal therapy), serious negative e ffects can occur in the therapist both personally and professionally (Pearlman & Saakvitne, 1995). Vicarious traumatization stems from a theo retical framework in constructivist selfdevelopment theory (CSDT), which is a deve lopmental interpersonal theory (Pearlman & Saakvitne, 1995). The underlying philosophical pos ition of this theory is the constructivist notion that individuals cons truct their own realities. A main assumption of this theory is that the meaning of the traumatic event relates to the su rvivors experience of the event. A second assumption is that the individuals early developmen t is primary in their wa y of interacting with self and others. CSDT also assumes that survivors symptoms are adaptive strategies developed for psychological survival in th e face of traumatic events (Pearlman & Saakvitne, 1995). Five areas of the self are expected to be affected by trauma, reflecting both experiential and cognitive realms (Saakvitne, Tennen, & Afflec k, 1998). First, the individuals frame of reference is disrupted, includi ng the way of understanding self and the world and sense of spirituality. Second, the survivors self-capacity to recognize, tolerate, and integr ate affect are disrupted. Third, ego resources are affected, which disrupt ones ab ility to meet their psychological needs in mature ways. This category also includes cognitive and social skills. Fourth, disruptions occur in cognitive schemas in areas of trust, safety, control, esteem, and intimacy. Finally, perceptual and memory system s are affected, including neuro-chemical and sensory changes (Saakvitne, Tennen, & Affleck, 1998). This theoretical position has been used to describe the effects of compassion fatigue as well as posttraumatic growth that occur for the therapist working with traumatized clients (Saakvitne, Tennen, & Affleck, 1998).
32 Schauben & Frazier (1995) explored the ps ychological effects of female counselors working with sexual violence survivors. Counsel ors who had more survivors of sexual violence in their caseload reported greater PTSD symptoms higher levels of vicar ious traumatization, and more disrupted beliefs. Counselors reporte d using active coping st rategies (e.g., seeking instrumental and social support) that were more adaptiv e than coping by behavioral disengagement. Pearlman & Maclan (1995) examined vicarious traumatization with male and female selfidentified trauma counselors. The researchers found that newer therapists, especially those not receiving supervision, and therap ists with a personal trauma history experienced the most negative outcomes. The most frequent disruptio ns experienced by therapists were in selfintimacy and other-esteem. Participants with more time devoted to trauma work had fewer disruptions in self-trust. Lee (1995) explored vicarious traumatization in marriage and family therapists (Jenkins & Baird, 2002). These therapists reported a mean of 63% of clients with a di agnosis of PTSD. It was found that the more hours spent with exposure to clients traumatic ma terial, the greater the intrusion scores. Thus, more secondary exposure to trauma through therapeutic work was associated with intrusive thought s and images related to the c lients traumatic material. To summarize, studies related to vicarious traumatization have found that therapist who worked primarily with trauma clients experience d various levels of vi carious traumatization. Factors that differentiated therap ists who experienced higher leve ls of vicarious traumatization include experience with persona l trauma, lack of experience w ith trauma clients, lack of supervision, a heavier caseload of trauma clients, and a grea ter number of hours spent with trauma clients. Methodological limitations in these studies include mostly female, Caucasian
33 samples and low response rates using a survey design methodology. Additionally, vicarious traumatization was measured through various questionnaires, not a psychometrically sound, multi-item measure of this construct. Burnout The term burnout was coined by Freudenberger (1974). Currently, th ere is no standard definition of burnout (Edwards et al., 2000). Pines and Arons on (1988) define burnout as a state of physical, emotional and mental exhaustion caused by longterm involvement in emotionally demanding situations (p.9). Burnout is differentiated from vicarious traumatization and compassion fatigue as being a more gradual process that results from long term emotional exhaustion (Figley, 1995). Burnout is not limite d to those who treat traumatized populations, whereas the other terms (e.g., vica rious traumatization, compassion fa tigue) are specific to work with these populations. Skovholt (2001) has identified two types of burnout. Meaning burnout occurs when therapists feel that the calling of caring for others through foster ing their emotional healing and development no longer gives sufficient meaning to th e therapists life. Meaning burnout can also occur when the practitioner no longe r feels that their work is help ful to clients. Meaning burnout can occur due to workplace factors such as cas e and paperwork overload, client relapses, or agency dysfunctions that prevent the therapist from practicing in the way they prefer. Caring burnout refers to a decreased ability to professiona lly attach with clients as a result of depleted energy from previous attachments and separations from past clients (Skovholt, 2001). This cycle of attachment and separation can be psychologi cally draining for the practitioner. Also, continuous contact with clients w ho seem unappreciative can lead he lpers to view future clients in more negative terms, which negatively affect s the therapeutic relationship and the therapists overall well-being (Corey, 1989).
34 Maslach (2001) identified a model consisting of three main aspects of burnout: (1) emotional exhaustion; (2 ) depersonalization; (3) dec lines in sense of personal accomplishment (Maslach, 2001). Emotional exhaus tion is the defining f eature of burnout. The depersonalization dimension involves cynicism and interpersonal distancing. The reduced personal accomplishment dimension refers to a se nse of incompetence and lack of achievement and productivity in ones work (Maslach, 2001). Kahill (1988) reviewed empirical research on burnout symptoms and identified five main categories of burnout symptoms. The first category is physical symptoms consisting of fatigue, headaches, sleep disturbance, and other soma tic problems. The second category, emotional symptoms, includes anxiety, depression, irritabili ty, and guilt. Third, behavioral symptoms occur including pessimism, aggre ssion, callousness, and substance abuse problems. Fourth is work-related symptoms involving poor work perfor mance, absenteeism, ta rdiness, low morale, and resigning from work. The final categor y is interpersonal symptoms, consisting of communication problems, social withdrawal, loss of positive feelings toward ones clients leading to depersonalized and intellectualized therapeutic interactions. Numerous studies have been conducted on the topic of work-related burnout. Several of these studies will be reviewed in this section. Etzion (1984) explored the relationship between burnout and social support. Suppor t was found to be significantly negatively correlated with burnout. Social support in the ther apists personal life was found to be more effective in burnout prevention for females, whereas social support at work was found to be more beneficial for males. Women were found to have higher levels of burnout than men in this sample. A study by Ackerley et al. (1988) sampled doctoral-level psychologists and found that over one third of their sample experienced high levels of burnout Factors related to burnout
35 included therapists who were younger, low-inco me, with little experience in personal psychotherapy who experienced a lack of c ontrol in their work environment and felt overcommitted to their clients. The study was limited to a sample of doctoral-level psychologists. Raquepaw and Miller (1989) examined levels of burnout among psychologists and social workers. They found that agency therapists ha d higher levels of burnout. They also found that therapists caseloads did not predict burnout, but th eir level of satisfactio n with their caseloads was predictive. The sample size in this study, however, was smalle r than in comparable studies on the topic. In summary, the lack of a universal definition has been a problem in effectively studying burnout in mental-health workers. Similar to vicarious traumatizati on, studies found those therapists who were younger and less experien ced had higher levels of burnout. Additional explanatory factors were a sense of lack of control, overcommitment to clients, and work in agency settings. Gender differences also were found, with women experiencing higher levels of burnout. Social support was found to be an important preventive factor for burnout. Methodological limitations of these studies include low response rates and primary reliance on self-report. Compassion Fatigue Com passion fatigue, previously referred to as secondary traumatic stre ss, has been defined by Figley (1995) as the natural consequent behaviours and emotions resulting from knowing about a traumatizing event experienced by a signif icant other the stress resulting from helping a traumatized or suffering person (p.7). This term has its roots in the term compassion, which is defined as a feeling of deep sympathy and sorrow for another who is stricken by suffering or
36 misfortune, accompanied by a strong desire to alle viate the pain or remove its cause (Websters Encyclopedia, 1989, p.299). Compassion fatigue is viewed as an acute r eaction of PTSD-like symptoms as a result of empathic strain and exposure of helpers to the experiences of thei r clients (Figley, 1995; Pearlman & Saakvitne, 1995). Other symptoms include feelings of helplessness, confusion, isolation from supporters and experiences of intrusive, avoidance and arousal symptoms. Recovery, similar to onset, is usually faster than recovery from work-related burnout (Figley, 1995). Compassion fatigue, or secondary trau matic stress reactions, similar to vicarious traumatization, is considered inevitable to so me extent with trauma work (Herman, 1992). Figley (1995) proposed and s ubsequently revised (Figley, 2 002) a theoretical model of compassion fatigue. The model assumes that empa thy is necessary to effectively work with clients (Figley, 2002). Although essential, empathic engagement can also have costs, especially for therapists working with traumatized clients. Figleys (1995) mode l contains ten factors related to the occurrence of co mpassion fatigue. First, the therapist must have the ability to empathize with the client and empathic concern fo r the client. Next, exposure to the client must occur, which involves experienci ng the emotional pain of the client. Next, the empathic response occurs in which the therapist uses empathic understanding to put him or herself into psychological world of the client, experiencing th e emotions that are felt by the client (Figley, 2002). The therapist may then experience compassion stress, which is described as the residue of emotional energy (p.1437) from the empathic res ponse and the desire to relieve the suffering of the client (Figley, 2002). Comp assion stress can develop in to compassion fatigue unless protective factors are present. One of these f actors is sense of achievement, which is the
37 therapists sense of satisfaction with his or her efforts to help the client. Another protective factor is disengagement, which involves distancing oneself emotionally from the clients pain in order to live their life (Figley, 2002). There are also several factors that can exacerbate the condition of compassion stress to lead it to develop into compassion fatigue. First, prolonged exposure, which involves the therapists feelings of responsibility for relieving the sufferi ng of the client over a period of time, is a risk factor. Second, traumatic recolle ctions, or emotional memories associated with the clients trauma material, can exacerbate compassion stress Third, disruptions in the therapists life (e.g., illness, life changes) that can compound with compassion stress to develop into compassion fatigue (Figley, 2002). Figley (1995) conducted a meta-analysis of current compassion fatigue literature and categorized it into three main ar eas: (1) indicators of psychological distress; (2) cognitive shifts; (3) relational disturbances. The first area involves distressing emotions (e.g., depression, anxiety, fear, anger, shame), flas hbacks or nightmares, numbing of affect and avoidance, somatic complaints (e.g., sleep problems, headaches), addi ction or compulsive be haviors, physiological arousal (e.g., hypervigilance), a nd impairment in day to day functioning. The second area, cognitive shifts, involves changes in core beliefs rela ted to trust, safety, power, and personal control which arise as a result of exposure to clients traumatic experiences. Herman (1992) introduced the concept of witness guilt, in which therapists experience guilt for enjoying life while their clients have experienced such horrible traumatic events. Third, relational disturbances include the therapis ts personal relationships sufferi ng as a result of alterations in trust and intimacy (Clark & Gior o, 1998). Relationships with c lients are also affected through either overidentification or detach ment, negatively affecting the therapeutic alliance. Therapists
38 may also emotionally detach from their support sy stems and relationships w ith significant others (Dutton & Rubinstein, 1995). Landry (2001) and Wee and Myers (2002) studi ed compassion fatigue in mental health workers responding to the Oklahoma City bombi ng in 1995 (Creamer & Liddle, 2005). Landry (2001) found that compassion fatigue was related to empathy but not to ye ars of experience with trauma survivors or social support involvemen t. Wee and Myers (2002) showed men having higher compassion fatigue symptoms. Numerous studies have found a relationship between compassion fatigue and therapist personal histor y of trauma (Kassam-Adams, 1995; Pearlman & MacIan, 1995), whereas others ha ve not (Schauben & Frazier, 1995). Reasons for these mixed findings include lack of a standard way to meas ure personal trauma history. Suggestions have been made for this construct to be measured on a scale rather than as a dichotomous variable (Creamer & Liddle, 2005). Several studies have found a relationship between heavier caseloads of trauma survivors and compassion fatigue (Brady et al., 1999; Chrestman, 1999; Schauben & Frazier, 1995). Positive correlations have been found between compassion fatigue and reduced longevity of career, larger caseloads, increased contact with clients, and lo ng working hours (Beaton & Murphy, 1995). Gender of the therapists was associated with compassion fatigue in some studies (KassamAdams, 1995; Meyers & Cornille, 2002; Wee & Myers, 2002), but not in others (Pearlman & MacIan, 1995). Two studies found that women reported more compassion fatigue symptoms than men (Kassam-Adams, 1995; Meyers & Co rnille, 2002) and one found that men reported more symptoms (Wee & Myers, 2002). These studies have often had predominantly female and
39 Caucasian samples. Years of therapeutic experience had mixed findings, with some studies finding a relationship (Cunningham, 2003), whereas others did not (Kassam-Adams, 1999). Methodological limitations to expl ain these contradictory findings include lack of diversity in samples (i.e., mostly mainly female Caucasian samples) and reliance on different measures of trauma history and exposure to trauma clients. All studies utilized a survey methodology with varying return rates and are therefore limited due to unknown factors associ ated with those who did not respond. The most consistent findings in these studies were the relationship between both amount of trauma clients in the caseload a nd personal trauma histor y of the therapist and compassion fatigue. Although related terms for negative ps ychological outcomes are often used interchangeably in the literature, there are some significant differences between the terms. Vicarious traumatization is ofte n used to refer to the negati ve cognitive effects, whereas compassion fatigue encompasses cognitive, emotional, somatic, and behavioral outcomes. Burnout is differentiated from compassion fatigue as being a more gradual process that results from long term emotional exhaustion (Figley, 1995) Burnout is not limited to those who treat traumatized populations, whereas the other terms are limited to work with these populations. Finally, traumatic countertransferen ce can be differentiated from other terms as warning signs or risk factors for compassion fatigue rather than symptomology of compassion fatigue. In conclusion, the research on negative effects of therapeutic work w ith traumatized clients has found significant percentages of practitioners experiencing various negative consequences, although these percentages vary from study to stu dy. Relationships between factors of personal trauma history and amount of exposure to traumatized clients, and social support have been found consistently across constructs. In many of the studies reviewed, the primary method of
40 investigation has been through su rveys with low response. Lack of a standard definition for burnout and lack of a standard instrument for measuring vicarious traumatization may have led to mixed findings related to these constructs. Review of Positive Effects Whereas there has been a plentiful amount of literature concerning th e negative effects of psychotherapy with traumatized clients, there is a paucity of research to date on the positive effects of this type of work. Positive consequences have b een mentioned in the literature somewhat tangentially in the c ontext of a more thorough covera ge of negative consequences (Arnold, Calhoun, Tedeschi, & Cann, 2005). Th is overemphasis on negative consequences shapes the reality that trauma work is primar ily negative and hazardous. A paradigm shift is occurring in psychotherapy through the postmodern movement, which has focused attention on client strengths rather than pat hologies. Similarly, a paradigm sh ift seems to be occurring in the literature related to the conseque nces for therapists working w ith trauma populations. In the following section, a review of the literature on positive conse quences of therapeutic work, including trauma work, will be presented along with an introduction to the constructs of vicarious posttraumatic growth and compassion satisfaction. A useful framework for understanding the possi bility of positive psychological outcomes for trauma therapists is resilience theory. Resili ence refers to the ability to maintain a stable psychological equilibrium in the face of adverse or traumatic experiences. It is differentiated from recovery, which is equilibrium preced ed by a period of decreased psychological functioning (Bonanno, 2004). Recent re silience literature supports the notion that resilience is more common than has been described in previous literature (Bonanno, 2004). Thus, resilience or psychological growth should also be possible, even likely, in therapists working with
41 traumatized clients. Resilience can occur thr ough multiple pathways. Resilience can occur due to individual traits such as hardiness, which is a personality profile consisting of commitment to a life purpose, belief that one has control over their surroundings and belief that one can grow from adversity (Kobasa, Maddi, & Kahn, 1982). Re silience can also occu r through repressive coping, which involves the tendency to avoid pain ful memories and emotions. Finally, the expression of positive emotions and laughter in th e face of adversity are related to resilience (Bonanno, 2004). Given the common experience of and the multiple pathways to resilience, it is necessary to explore associated factor s to increase growth and resilience fo r trauma therapists. Herman (1992) reported that therapists d eepen their sense of integrity through trauma work and appreciate life more fully. She also stated that these therapists take life more seriously, have a greater understanding of others, have deep er and more intimate relationships and feels inspired by their clients courage. Pearlman and Saakvitne (1995) discu ssed the increased sense of connection with others that grows from the intimacy of working with clients healing from trauma. Other rewards include spiritual growth, increased respect for the resilience of the human spirit, and learning from witne ssing the strength exhibited by c lients (Pearlman and Saakvitne, 1995). Schauben and Frazier (1995), in their survey of female counselors working with sexual violence survivors, found that 45% of counselors reported enjoyable aspects of their work, such as creativity, strength, and witnessing the resilien ce of survivors. They also reported joy in seeing clients grow from their adversity and expe rienced growth and change in themselves as a result of their work with survivors. Fi nally, counselors reported feeling good about the importance of their work. These early results were obtained through qualitative methods. This study contributed to the understa nding of the positive psychologi cal outcomes for therapists.
42 However, a broader understanding of the phenome non of positive growth through trauma work is still needed, particularly through quantitative methods. Stamm (1998) purported that not all trauma workers experience distress and negative consequences. He surmised that some therapis ts may have protective mechanisms that enhance and maintain their well-being. King et al. (1998) stated that hard iness and social support are key protective mechanisms (Collins & Long, 2003). Hardin ess, as previously discussed, refers to a personality profile consisting of control, commitment, and viewing change as a challenge (King et al., 1998). Another study focused on vicarious traumatizat ion in female psychotherapists found that practitioners who tr eated more survivor s of sexual abuse had more existentially and spiritually satisfying lives (Brady et al., 1999). The authors surmised that work with survivors of sexual abuse may force therapists to challenge constructs of meaning and their faith in a way that may have promoted growth. Exposure to trauma mate rial may temporarily produce a spiritual crisis for therapists but often results in a stronger a nd healthier sens e of spiritual well-being (Brady, Guy, Poelstra, & Brokaw, 1999). It is also possibl e that people who are more spiritual are drawn to trauma work due to the focus on spiritual issu es that often arises for trauma survivors. Limitations of this study are that the sample consisted of female psychologists primarily in private practice settings (58%). The construct of posttraumatic growth has recently been extended to include trauma therapists experience of this phenomenon vicariously. The construct of posttraumatic growth was developed by Tedeschi and Calhoun (1996) and involves three basic ca tegories: changes in self-perception, interpers onal relationships, and philosophy of life. The conceptual model of posttraumatic growth for clients who directly experience a traumatic event includes an increased
43 appreciation for life in general, more meaningful relationships, an increased sense of personal strength, changed life priorities, and a richer existent ial or spiritual life (Tedeschi & Calhoun, 1996). Although positive consequences of psychotherapy for therapists has been mentioned tangentially in the literature on negative outcomes, one study by Arnold, Calhoun, Tedeschi, and Cann (2005) attempted to qualitatively explore po sitive consequences of therapeutic work and the concept of vicarious posttr aumatic growth for therapists. These researchers used a convenience sample of licensed psychotherapists from various disciplines All 21 therapists interviewed mentioned some positive consequen ces of their therapeutic work. The most frequently reported positive consequence was ob serving clients posttr aumatic growth, with contributed to their own growth and development. They also reported trait-oriented changes in themselves, such as increased sensitivity, compassion, insight, tolerance, and empathy. Many therapists (76%) said that trauma work imp acted their spiritual gr owth and 48% reported a deepened appreciation of the human spirit. A lthough these therapists also mentioned negative consequences from their work, all of them also mentioned many positive outcomes. Because a small sample was used with a qualitative method, th e findings in this study cannot be generalized to other populations of therapists. Additional st udies are needed using quantitative methods for exploring vicarious posttraumatic growth in therapists. Compassion satisfaction has been described as the enjoyment derived from effectively helping others through psychotherapeutic work (Stamm, 2005). Stamm (2002) developed this term while revising the Compassi on Fatigue Self Test (Figley, 1995; Figley & Stamm, 1996) to include positive aspects that pa rallel the negative effects of compassion fatigue. Compassion satisfaction has also been described as a sense of efficacy in ones work (Stamm, 2002). Stamm
44 (2002) additionally described compassion satisfac tion as the happiness that comes from being able to make the world a closer reflection of what one wants it to be (Stamm, 2002). Although little empirical research has been conducted on compassion satisfaction, some preliminary related factors have been purported. Compassion sa tisfaction is proposed to be related to the protectiv e factors of hardiness and good social support (King et al., 1998). Stamm (2002) surmised that positive collegial support in the workplace could enhance compassion satisfaction and decrease compassion fatigue and other negative consequences through enhancing therapists sense of competency. Preliminary research on compassion satisfaction indicated that burnout and compassion fatigue, especia lly in combination, result in less of a sense of efficacy and lower levels of co mpassion satisfaction (Stamm, 2002). Interrelationships among Positive and Negative Effects Recent stud ies examined the interrelationships between compassion fatigue, compassion satisfaction, and burnout (Conrad & Kellar-Guenther, 2006). Compassion satisfaction refers to ones sense of enjoyment and fulf illment in their therapeutic wo rk. Conrad and Kellar-Guenther (2006) surveyed child protective se rvice workers to determine thei r levels of compassion fatigue, compassion satisfaction, and burnout. The research ers found that approximately half of their participants had high (i.e., raw sc ore of 36-40) or very high (i.e., raw score of 41-115) levels of compassion fatigue. Additionally, 70% of participants had good potential for compassion satisfaction (i.e., raw score of 8299), which seemed to moderate th e effects of burnout, of which significantly lower levels were found. Participan ts with high compassion satisfaction had lower levels of both compassion fatigue and burnout. More research is needed to determine the interrelationships between the positive effects (compassion satisfaction) and the negative eff ects (compassion fatigue) of psychotherapeutic work with traumatized populations. Additionally, to add to the li terature on positive effects of
45 psychotherapeutic work, there needs to be more information on how to enhance positive outcomes as well as prevent negative outcomes. Workplace Factors Overall psychological outcom es are mediated by both risk and protective factors including the therapists personal characteristics, characteristics of the client and the trauma, the therapists attempts to cope, and aspects of the work envi ronment in which the therapy takes place (Dutton & Rubenstein, 1995; Figley, 1995; Chrestman, 1999). The interactions between these factors influence the therapeutic relationship. The relationship between workplace factors and positive and negative outcomes for counselors will be disc ussed in more detail in this section. There is a significant impact, es pecially in agency settings, on organizations in which therapists are affected by negative consequences of therapeutic work. First, high-turnover is a predictable outcome of burnout and other negative consequences. Th ere is likely to be a loss of energy, commitment, and optimism, which has a deleterious effect on the work environment (Sexton, 1999). The more negative countertran sference effects experienced by therapists, the greater the possibility that issues of clients will be played out within the organization. Given that negative consequences such as vicarious traumatization and compassion fa tigue are considered inevitable outcomes of trauma work, this raises issues of occupational he alth and safety within the workplace (Sexton, 1999). Maslach and Leiter (1997) take a strong work-climate view with respect to the causes of burnout and other negative effects on the counselor They see burnout as an interaction between the person and his or her work environment. Si x sources of burnout include work overload, lack of control, insufficient rewards, breakdown of community, unfairness, and significant value conflicts (Maslach & Leiter, 1997) First, work overload refers to long work hours and seeing more clients than one can successfully manage. Second, a sense of lack of control, or having no
46 voice concerning policies in the workplace or trea tment of clients, also seems to negatively impact therapists. Third and fourth, insufficien t rewards, such as low wages for counselors in many settings, may bring about feelings of unfairn ess. Fifth, counselors often feel alienated from one another, leading to a sense of br eakdown of community and social support in the workplace. Sixth, value conflicts may occur betw een the professional goals of the practitioner and policies in the setting in which he or she works (Maslach & Leiter, 1997). Additional work-related factors related to burnou t are lack of challe nge on the job, role ambiguity, low professional self-esteem, difficulties in providing services to clients, and negative attitudes toward the profession (Solderfeldt, Solderfeldt, & Wa rg, 1995). Personal factors that interact with these work-related factors include chronic minor hassles in daily living, family income, attitudes toward the profession, years of experience, and low education (Solderfeldt, Solderfeldt, & Warg, 1995). The studies reviewed previous ly have yielded numerous inconsistent results concerning individual associated factors due to flaws in methodology and meas urement. Factors related to the workplace that influence therapist outcomes have been somewhat more consistent in the literature. Private practice therapists have b een consistently found to have lower levels of burnout than agency therapists (Ackerley et al., 1988; Hellman & Morrison, 1987; Vredenburgh, Carlozzi, & Stein, 1999; Rupert & Morgan, 2005). This difference has been attributed to several factors, including independent practitioners having greater control over their work, less paperwork, and less disturbed clie nts (Rupert & Morgan, 2005). Vredenburgh, Carlozzi, & Stei n (1999) surveyed 521 psychol ogists (52.1% response rate) and also found that private practitioners had the lowest levels of emoti onal exhaustion and higher levels of personal accomplishment (i.e., lower le vels of burnout). Additionally, they found
47 positive correlations between client load and sense of personal accomplishment and nonsignificant correlations between client load and emotional exhaustion or depersonalization. They also found that males had higher levels of depersonalization than females. Hellman & Morrison (1987) found that therapists in private practice expe rienced less personal depletion but more stress. They also found that therapists with more disturbed clients re ported more stress. Methodological flaws may account for the differe nces in findings, including survey methods with low response rates, small samples, and sa mples limited to primarily female, Caucasian, doctoral-level psychologists. Rupert & Morgan (2005) conducted a large nati onal survey of 571 doctoral psychologists in solo private practice, group private practice, and agency settings. Overall, the researchers found similar burnout rates as those found by Acke rley and colleagues (1988). Solo and group practitioners had lower overall levels of burnout, including dimens ions of emotional exhaustion and sense of personal accomplishment. The re searchers used ANCOVA to measure differences with age as a covariate, since agency worker s tended to be younger and less experienced which has been found to be related to burnout. Solo and group private practitioners reported a greater sense of control over their work, less negative client behavior, more direct pay clients, and more overinvolvement with clients. Level of exha ustion was positively related to number of hours worked, paperwork hours, percentage of managed care clients, negative client behaviors, and overinvolvement with clients. Ackerley, Burnell, Holder, and Kurdek (1988) conducted a study examining various correlates of burnout, including demographics, obj ective work characteristics, types of work activities, types of client issu es, and factors within the therap y setting. The researchers found that various subcomponents of burnout were posi tively related to younger age of the therapist,
48 agency work, and work with couples and as a clin ical supervisor. A fina l category of predictors of burnout examined in this study was specific asp ects of the work setting itself. Four factors associated with the work setting were included in the development of the Psychologists Burnout Inventory (PBI), which is a 15 item survey on a 7point Likert scale. The four subscales were first conceptually derived and were upheld through f actor analysis procedures. The four subscales of the PBI are nega tive clientele (e.g., hi gh lethality, psychopathology, resistance, noncompliance), perceived control (e.g., initiative, co ntrol over work activi ties), workplace support (e.g., sharing responsibilities, ab ility to consult), a nd overinvolvement (e.g., overly responsible for clients, taking work home). Ackerley and colleagues (1988) found that all four subcomponents of the PBI were related to variou s subscales of the Maslach Burnout Inventory (MBI; Maslach & Jackson, 1986). To summarize findings concerning workplace f actors and negative therapist outcomes, the work setting has been consistently found to have an impact on burnout for therapists. Specifically, independent practitioners have lowe r levels of burnout. Factors that have been related to lower levels of bur nout include a greater sense of control over their work, less paperwork, less disturbed clients, and a more manageable caseload (e.g., severity of cases and number of clients seen). Although there have been several studies c onducted on workplace factors associated with burnout, there have been no such empirical studies on workplace factors associated with compassion fatigue or compassion satisfaction. There has been recent literature surrounding recommended changes in the workplace for prev enting compassion fatigue. However, these recommendations have not been empirically explored. These recommendations will be discussed in the following section.
49 Recommendations for Therapists and the Workplace Recommendations in the literature for fostering positive outcomes and mitigating negative effects include the need for mentor and peer support, involvement in diverse activities, and creating health-promoting work environments (e .g., Skovholt, 2001). More extensive research has been conducted on therapist self-care, which could be built in to the workplace to promote positive outcomes for counselors. Proactive st rategies recommended for therapist self-care include being able to anticipate emotional dist ress, maintaining personal awareness, developing a plan of action if problems arise, maintaining a support network, participating in organizational memberships and meetings, and seeking mental health consultation (Morrissette, 2001). Several additional strategies for mitigating negative effects for therapists have been suggested in the litera ture. Stamm and Pearce (1995) proposes positive collegial support systems in the workplace to provide structural and functional social support (Collins & Long, 2003, p.422). Figley (1989) recommends making time for pleasure, staying involved in professional activities for support, and sett ing realistic goals and boundaries. Cerney (1995) encourages seeking regular supervision or consultation to process painful client material and personal emotions or cognitions that may be perceived as overwhelming. Other recommendations include seeking social support, journaling, progressive relaxation, physical act ivity, diet, spiritual strength, involvement in activities of interest, and achieving balance in life (Salston & Figley, 2003). It is recommended to ha ve crisis intervention available and offered continuously on a voluntary basis (Salston & Figley, 2003). The work environment should create a space of acceptance in which therapists do not feel ashamed for having negative reactions (Sexton, 1999). Additional sugg estions for the work environment include clinical supervision and consultation, ca se conferences, peer process groups, personal psychotherapy made available without penalty, trau ma therapy training,
50 professional development and regular organiza tional team meetings (Sexton, 1999, p.399). The work environment should be an atmosphere of acceptance, and non-judgmental support (Neuman & Gamble, 1995). Collins and Long (2003) also arti culate the necessity of a work culture that values and supports its staff. Additionally, th ese authors proposed self-awareness programs, clinical supervision, and confidential counseling services available. Munroe et al. (1995) proposed a team treatment model for preventing compassion fatigue. These researchers proposed the formation of trea tment teams to meet the psychological needs of staff. Three main tenets underlie the model: (1) the acceptance of the reality of secondary trauma, (2) therapist responses are natural and valuable processes, (3) each team member can be an observer as to how other te am members are responding to sec ondary trauma. They suggest that this community can serve to absorb the trau matic experience of an individual by diffusing its effects among many people and demonstrating that the survivors feelings are understood (p.215). Teams are recommended to meet at least two and a half hours per week to cover administrative and clinical issues. Need for the Study W hile its existence has been well-documente d, most research on compassion fatigue has focused on individual correlates (e.g., gender, years of experience, personal trauma history) rather than characteristics of the organization (Bell, Kulkarni, & Dalton, 2003). Research on the construct of burnout has begun to focus on factors in the workplace associat ed with burnout (e.g., Ackerley et al., 1988; Rupert & Morgan, 2005), but more resear ch is needed on workplace factors related to compassion fatigue and comp assion satisfaction. Numerous studies have recommended interventions to prevent compassion fatigue in the workplace (e.g., Sexton, 1999; Collins and Long (2003), however no empirical data supports these recommendations.
51 A broader understanding of contextual factors related to compassion fatigue and compassion satisfaction is needed (Salston & Figley, 2003). Measures of control, overinvolvement, negative clientele, workplace s upport, and amount of exposure will be assessed to determine relationships between compa ssion fatigue and workplace factors. As previously stated in this literature review, there exists a gap in the research concerning positive effects of psychotherapeutic work and fact ors associated with these positive outcomes. The present study attempted to address this gap by including a measure of compassion satisfaction to determine this construc ts relationship with workplace factors. To begin to create a model of workplace f actors and compassion fatigue and compassion satisfaction, individual factors need to be accounted for. Individual fact ors found to relate to compassion fatigue are personal trauma history and years of experience (e.g., Bell, Kulkarni, & Dalton, 2003). Additional individual factors with more mixed findings need to also be accounted for. These individual factors are gend er, ethnicity, and trauma training. This study will contribute to the literature by elucidating the impact of workplace factors as well as individual factor s on the psychological outcomes (i.e., compassion fatigue and compassion satisfaction) for therapists working with trauma clients. These results should have important implications for organizations by de monstrating aspects of the workplace, both structural and functional, that ca n be altered so as to help pr event therapists from developing compassion fatigue and to encourage thei r experience of compassion satisfaction.
52 CHAPTER 3 METHODOLOGY The purpose of this stud y was to examine the relationships among workplace variables (e.g., control, overinvolvement, negative client ele, workplace support, secondary exposure), individual variables (e.g., gender, ethnicity, personal trauma histor y, years of clinic al experience, trauma training) and levels of compassion fatig ue and compassion satisfaction for Professional Counselors, Psychologists, and So cial Workers working with trau matized clients and/or clients in crisis. Workplace variables we re analyzed to determine if th ere are any explanatory properties for compassion fatigue and compassion satisfacti on, while including the i ndividual variables of therapists gender, ethnicity, personal trauma hist ory, years of clinical experience, and trauma training experience in the model. The study spec ifically explored (a) the relationships among levels of control, overinvolvement, negativ e clientele, workplace support, and amount of secondary exposure and compassion fatigue, (b) the relationships among levels of control, overinvolvement, negative clientele, workpl ace support, amount of secondary exposure and compassion satisfaction, and (c) changes in va riance of compassion fatigue and compassion satisfaction when workplace variables (e.g., c ontrol, overinvolvement, negative clientele, workplace support, amount of secondary exposure) are added to individual variables (gender, ethnicity, personal trauma history, years of clinical experience, trauma training). Levels of compassion fatigue and compassion satisfaction were measured using the Professional Quality of Life Scale (ProQOL; St amm, 2005; see Appendix A). The Professional Quality of Life Scale consists of 30 questions measured on a 6-point scale intended to measure three core components of therap ists quality of life in thei r professional work: compassion fatigue, compassion satisfaction, and burnout. Wo rkplace variables were measured using the Psychologists Burnout inventory (PBI; see Appe ndix B), which was developed by Ackerley et.
53 al (1988). The Psychologists Burnout Inventory is a 29-item scale measured on a 7-point scale intended to assess for the presence of four wo rkplace factors related to burnout: Control, Overinvolvement, Support, and Negative Clientele. A fifth workplace variable, therapists amount of secondary exposure to traumatized clients, was measured through asking the following question: How many hours per week are sp ent working directly with clients who have been traumatized? Gender and ethnicity were measured in a demographics section. The individual variable of therapists personal trauma history was measured by the Stressful Life ExperiencesShort Form (Stamm, 1997, see Appendi x C). Years of clin ical experience was measured through an open question asking how ma ny years the therapist ha s been doing clinical work including internship and practicum experi ences. Finally, trauma training was assessed by a closed question asking if the participant has had any speciali zed training in working with traumatized populations. This chapter contains a description of th e methodology utilized in the collection and analysis of data in this study. Included are a description of the research design and relevant variables, research hypotheses population, sample, sampling procedures, instrumentation, data collection procedures, and proposed data analysis procedures. Research Design and Relevant Variables A survey design utilizing correlational and com parative methods was used in this study. The dependent variables are compassion fati gue and compassion satisfaction, which were measured by the Professional Quality of Life Scale (Stamm, 2005). The ProQOL scale yields three sub-score scales for each of the three components of professional quality of life: Compassion fatigue, Compassion satisfaction, and Burnout. The Burnout subscale was not used in this study since it is not one of the variables being investigated.
54 The independent variables in this study include workplace variables of control, overinvolvement, negative clientele and wor kplace support, which were measured by the Psychologists Burnout Inventory (PBI; Ackerley et. al, 1988) and amount of secondary exposure, which was measured by asking about th e amount of hours per week spent in direct contact with traumatized clients. The other independent variables in this st udy are the individual variables of gender, ethnicit y, personal trauma history, measured by the Stressful Life ExperiencesShort Form (Stamm, 1997), years of clinical experience, measured by a question about the number of years working as a therap ist, and trauma training, measured by a question about specialized training worki ng with traumatized clients. Research Hypotheses The following null hypotheses were tested in this study: H1 There is no varian ce in compassion fatigue e xplained by workplace variables (i.e., control, overinvolvement, negative clientele, workplace support, and amount of secondary exposure). H2 There is no variance in compassion satisfac tion explained by workplace variables (i.e., control, overinvolvement, negative clientele, workplace support, and amount of secondary exposure). H3 There is no change in variance of compassion fatigue when workplace variables are added to a model including individual variables (i.e., gender ethnicity, personal tr auma history, years of clinical experience, trauma training). H4 There is no change in variance of compassi on satisfaction when wo rkplace variables are added to a model including individual variab les (i.e., gender, ethni city, personal trauma history, years of clinical e xperience, trauma training). The expected directions of the association be tween higher levels of compassion fatigue and workplace variables based upon prior research was as follows: lower levels of perceived control, higher levels of overinvolvement, higher levels of negative clientele, lower levels of perceived support, and higher levels of secondary exposur e are expected. Indivi dual variables were expected to relate to higher levels of compassion fatigue as follows: more personal trauma
55 history, less years of clinical e xperience, and no trauma training. The expected direction of the association between higher levels of compassi on satisfaction and workplace variables based on prior research was as follows: higher leve ls of perceived control, lower levels of overinvolvement, lower levels of negative clientele, higher levels of perceived support, and lower levels of secondary exposure are expected. Individu al variables were expected to relate to higher levels of compassion satisf action as follows: less personal trauma history, more years of clinical experience, and having sp ecialized trauma training. The individual variables of gender and ethnicity were exploratory due to mixed findings or lack of explora tion in previous literature. Population and Sample The population to which participants in the present study were in tended to generalize consists of therapists working with clients healin g from traumatic experiences. To estimate this population, professionals from the Internationa l Society for Traumatic Stress Studies were chosen as the population from which to draw a sa mple. The International Society for Traumatic Stress (ISTSS) is an international, multi-discipl inary professional organization dedicated to the development and advancement of the field of tr aumatic stress. The ISTSS was founded in 1985. ISTSS currently has approximately 2,200 member s from a variety of disciplines, including psychiatrists, psychologists, social workers, coun selors, nurses, researchers, administrators, and clergy. A secondary sample was obtained from the A ssociation for Traumatic Stress Specialists (ATSS). This sample was included when sampli ng from the ISTSS did not yield the anticipated number of responses. The ATSS is also an in ternational, multi-disci plinary professional organization focused on providing training, certification, and edu cation for those who provide emotional care to trauma survivors. ATSS wa s founded in 1987 and curr ently has 585 members who are actively involved in cris is intervention, emergency, respons e, or long term treatment of
56 PTSD. Members come from a variety of disciplines including emergency responders, psychologists, psychiatrists, social workers, coun selors, victim assistance staff, and clergy. The main difference between ISTSS and ATSS may be a primary focu s on research (ISTSS) versus practice (ATSS). Sampling Procedure The participants in this study included a repr esentative sam ple of Professional Counselors, Psychologists, and Social Workers who were cont acted from the ISTSS member registry. In order to obtain a representative sample, a list of potential participants was obtained through using the search engine function of the member registr y. The following settings were isolated due to relevance to the desired populati on: public health facility/nonuniversity; privat e health/mental health practice; nongovernmental or ganization; veterans organizati on; university non-educator. From the lists of names obtained through this sear ch, all participants with at least a masters degree were selected, excluding me dical doctors due to the focus of the current study on the population of trauma counselors, soci al workers, and psychologists. To obtain a representative sample from th e ATSS population, the ATSS office manager was contacted to explore potenti al sampling methods. This inve stigator was not granted access to an e-mail list, however it was agreed that a ma ss e-mail be sent out to ATSS members inviting them to participate in the study includ ing a summary and purpose of the study. Of the initial 311 potential participants contacted through the ISTSS, 53 participants were obtained and completed the entire survey, with a response rate of 17%. Eight potential participants partially completed the survey. On ly the data from the 53 participants who fully completed the survey were included in the data analyses. Of the 585 ATSS members contacted through a mass e-mail, 68 participants were obtained who completed the entire survey. The response rate for the ATSS sample was 11.62%. Th irteen members who partially completed the
57 survey were not included in the data analyses. Finally, 23 participants were removed from the sample who indicated that there profession was other than Psyc hologist, Professional Counselor, or Social Worker. A proposal for this study was submitted to the University of Florida Institutional Review Board (UFIRB). A survey including the instru ments selected and a de mographic questionnaire was created on a website. The link to the survey as well as a letter of consent requesting participation was e-mailed to selected particip ants. The survey incl uded (a) the Professional Quality of Life Scale (measures compassion fati gue, compassion satisfaction, and burnout), (b) the Psychologists Burnout Inventory (measures workplace factors of control, overinvolvement, negative clientele, and workplac e support) and one item about the amount of secondary exposure they experience in their work, (c) the Stressf ul Life Experiences Screening Short Form (measures personal trauma history), and (d) a demographics questionnai re, including gender, age, ethnicity, years of clinical experience, and trauma training. Instrumentation The instruments that were used in this study include the Professional Quality of Life Scale (ProQOL; Stamm 2005; see appendix A), the Psyc hologists Burnout Inve ntory (PBI; Ackerley et. al., 1988; see appendix B), and the Stressful Life Experiences Screen ing-Short Form (SLES; Stamm, 1997; see appendix C). Each of th ese instruments are available for public use. Professional Quality of Life Scale (ProQOL) The Professional Quality of Life Scale is th e revised version of the Compassion Fatigue Self Test (Figley, 1995). The CFST was revised due to psychometric problems and the negative connotation associated with the name of the inst rument (Stamm, 2005). The ProQOL is the third revision of the original Compa ssion Fatigue Test (CSF; Figley, 1996). The revision addressed previous issues of separating the constructs of burnout and vicarious trauma and reducing
58 participant strain by shor tening the scale from 66 to 30 items. Items that met both high item-toscale criteria and were most theoretically relevant to the subscale constr uct were retained. The revision was based on over 1000 participants from multiple studies (Stamm, 2005). The ProQOL is measured on a 5-point Likert-t ype scale ranging from never (0) to very often (5). There are three s ubscales: compassion fatigue, comp assion satisfaction, and burnout. A sample item from the Compassion Fatigue subscale is, I feel as though I am experiencing the trauma of someone I have helped In the Compassion Satisfaction subscale, a sample item is the following: I get satisfaction form being able to help people The Burnout subscale consists of items such as, I feel connected to others. Across all three subscales five items are reversescored. Scale distributions for each subscale are gene rally unimodal and symmetric. However, the compassion satisfaction subscale is skewed toward the positive side and the compassion fatigue subscale is skewed toward the side involving littl e disruption (Stamm, 2005) Alpha reliabilities for each subscale are as follows: Compassion Sa tisfaction subscale = .87; Burnout alpha = .72; and Compassion Fatigue alpha = .80. Construct validity has been wellestablished in over 200 article s in peer-reviewed journals (Stamm, 2005). Convergent validity using th e multi-trait multi-met hod analysis (Campbell & Fiske, 1959) demonstrated that the ProQOL does m easure separate constructs. Shared variances between the subscales are generally small. Compassion Satisfaction and Burnout have 5% shared variance. Compassion Satisfaction a nd Compassion Fatigue have only 2% shared variance. Shared variance among Burnout and Co mpassion Fatigue is slightly higher at 21%, reflecting the distress common in both conditions (Stamm, 2005).
59 The ProQOL has been tested on general h ealthcare workers (inc luding counselors and psychologists), child and family workers, and school personnel. Possible scores range from 0 to 50 for each of the three subscales. The average score for the Compassion Satisfaction subscale is 37. The average score for the Burnout subscale is 23. For the Compassion Fatigue subscale, the average score is 13. Psychologists Burnout Inventory (PBI) The PBI is a 15-item survey on a 7-point Likert s cale. This format is identical to that of the Maslach Burnout Inventory (MBI; Maslach & Jackson, 1986). The following four subscales were conceptually derived: control (3 items), workplace support (3 items), types of negative clientele (6 items), and overinvolv ement with clients (3 items). Control refers to the amount of perceived personal control an individual feels they have over their work activities. A sample item is evidenced by the following: I have control over what I do and when I do it during the work day. The Workplace Support subscale measures the amount of perceived support (e.g., emotional, in strumental, and informa tional) an individual receives from coworkers and/or supervisors. The following is a sample item from the Workplace Support subscale: I receive constructive feedback from coworkers or supervisors. Negative Clientele refers to the severity of client issues as well as difficult client behaviors. For example, I work with clients who defe nsively withdraw and withhold, is a sample item. Overinvolvement refers to the professional taki ng on a disproportionate amount of effort and responsibility in the th erapeutic process and being unable to disengage (i.e., taking the work home with them). A sample item is: I find myself feeling responsib le for my clients wellbeing. To test validity of the scale, principal-component factor analysis was performed with varimax rotation (Ackerley et. al, 1988). Four eigenvalues were obtained greater than 1 (2.69,
60 1.93, 1.70, and 1.28), which accounted for 18%, 13%, 11%, and 9% of the total variance, respectively. A scree test indi cated that all four factors shoul d be retained (Ackerley et. al, 1988). Stressful Life Experiences Screening-Short Form The Stressful Life Experiences Screening-Short F orm (Stamm, 1997) was designed to identify traumatic stressors as a clinical tool or for research purposes. The SLES-Short Form provides a list of 20 possible trau matic life experiences. The participant is in structed to enter a zero if they did not experience the event or enter a score between 1 (a little like my experience) and 10 (exactly like my experience). A sample item is I have witnessed or experienced a serious accident or injury. The instrument can also be used on a dichotomous scale in which participants enter 0 if they did not experience the event and 1 if th ey did experience the event. A total score is calculated whether th e scale is used on a 0 to 1 scale or a 0 to 10 scale. The original SLES (Stamm, 1996) was normed on military physical health care workers (n = 15). After data collection and analysis, the in itial items were revised and reduced. The second norm group included mental health care workers (n = 30) and university students (n = 30). Internal consistency was .7 and item-to-scale co rrelations were generally low (e.g., between .0 and .3). Low item-to-scale correlations and lack of changes in overall alpha reliability of the scale indicate a single over all construct is being measured by the scal e (Stamm et. al, 1996). Data Analysis There are a total of 12 variables that were investigated in th is study: 2 variables m easuring compassion fatigue and compassion satisfacti on as measured by the ProQOL, 5 variables measuring workplace factors as measured by the PBI and amount of secondary exposure to trauma clients as measured by percentage of time spent with these cl ients per week, and 5 individual variables measuring gender, ethnicity, pe rsonal trauma history as measured by the
61 SLES, years of clinical experience as measured by years of therapeutic work including during graduate school, and trauma tr aining as measured by whether or not the participant had any specialized training in working with traumati zed clients. The relationships among these variables were analyzed using multiple regression/correlation (MRC) analyses. The first step in data analyses involved obtaining desc riptive statistics for the sample. This was done through calculation of mean s and standard deviations for all continuous variables and frequency distributions for demographic categorical variables. The second step in the analysis was to calculate reliability (Cronbachs alpha) co efficients for the subsca les of each instrument used in the study. The final step in the data analyses was to test the hypotheses presented. The first hypothesis examined the relations hips among workplace variables and compassion fatigue. This hypothesis was tested using a multi ple regression procedure. For this analysis, compassion fatigue was considered the dependent variab le and the workplace variables (e.g., control, overinvolvement, negative clientele, workplace support, and amount of secondary exposure) were considered the inde pendent variables. The second hypothesis concerned the relatio nships among workplace variables and compassion satisfaction. This hypothesis was also tested using a multiple regression procedure. In this regression equation, comp assion satisfaction was considered the dependent variable and workplace variables (e.g., control, overinvolvement, negative clientele, workplace support, and amount of secondary exposure) were considered the independent variables. The third hypothesis examined the change in variance of compassion fatigue after workplace variables were added to a model includ ing individual variables. The hypothesis was to be tested using a stepwise multiple regre ssion procedure. In this regression equation,
62 compassion fatigue was considered the dependent va riable and individual variables (e.g., gender, ethnicity, personal trauma history, years of clinical expe rience, trauma training) were considered the independent variables. Indivi dual variables were to be ente red into the stepwise regression procedure, followed by workplace variables (e.g., c ontrol, overinvolvement, negative clientele, workplace support, and amount of secondary exposur e) to examine change in the variance of compassion fatigue. The fourth hypothesis examined the change in variance of compassi on satisfaction after workplace variables were added to a model includ ing individual variables. The hypothesis was tested using a stepwise multiple regression proc edure. In this regression equation, compassion satisfaction was considered the dependent va riable and individual variables (e.g., gender, ethnicity, personal trauma history, years of clinical expe rience, trauma training) were considered the independent variables. Individual variab les were entered into the stepwise regression procedure, followed by workplace variables (e.g., c ontrol, overinvolvement, negative clientele, workplace support, and amount of secondary exposur e) to examine change in the variance of compassion satisfaction. To test the third and fourth hypotheses, beta coefficients were examined to see if workplace variables significantly improved the va riance associated with both compassion fatigue and compassion satisfaction. For all analyses the Type I error rate was set at 0.05.
63 CHAPTER 4 RESULTS Results of the data analyses that were conduc ted for the current study are described below. First, sam ple demographics will be descri bed, followed by measurement reliabilities, then Pearson product correlations for each of the variables in the study, and finally the results of regression analyses that were conducted to test each of th e four hypotheses concerning the relationships between compassion fatigue and compassion satisfaction with individual (e.g., gender, ethnicity, years of clin ical experience, pers onal trauma history, trauma training) and workplace variables (e.g., amount of secondary e xposure, control, overinvolvement, workplace support, negative clientele). Demographics The f inal sample that was used for data anal ysis consisted of 98 trauma therapists (73 female, 25 male). The sample also was primarily Caucasian, 91.8% (N = 90), followed by Latino, 3.1% (N = 3), Other, 3.1% (N = 3) and African American, 2.0% (N = 2). Participants were asked to report their profession, which consisted of primarily psychologists, 50.0% (N = 49), followed by social workers, 26.5% (N = 26), and professional counselors, 23.5% (N = 23). Some participants indicated their pr ofession as other (N = 23) and these completed surveys were eliminated from the current analyses because the target population for the current study was trauma ther apists defined as members of the following professions: Professional Counselor Social Worker, Psychologist. Participants who indicated that their profession was other were assumed to be members of other professions who work with traumatized individuals such as emergenc y responders, victim advocates, clergy, or school personnel, and were therefore removed from the sample used for data analysis. Additionally, participants reported their prim ary professional activities as m ostly counseling/therapy and or
64 other clinical activities, 77.6% (N = 76), followed by teaching and/or research activities 11.2% (N = 11), supervisory/administrative 8.2% (N = 8), other 2.0% (N = 2), and case management, 1.0% (N = 1). Participants were also asked to report their clinical work environment. Participants reported their wo rk setting as primarily working with other professionals, 66.3% (N = 65) follo wed by being the only professiona l in their work environment 33.7% (N = 33). With respect to the dependent variable of compassion fatigue, participants reported an average score of 11.03 (SD = 6.13). For the depe ndent variable of compassion satisfaction, participants reported an av erage score of 40.87 (SD = 5.59). Descriptive statistics for the individual independent variables are as follows: The average total number of years of clini cal experience was 18.53 (SD = 9.09) The variable related to trauma training was not included in further an alyses because 99.0% of participants reported having trauma training, with only 1.0% indicating no prof essional training. The mean score for personal trauma histor y was 55.09 (SD = 33.19). Descriptive statistics for workplace independent variables will be described in this section. The average amount of secondary exposure to trauma (measured by amount of time that participants worked with trau ma clients per week) was 16.98 (S D = 11.27). Three participants reported a range of hours (e.g., 15-20 hours) as their response to this item. In order to include this data in the analyses, the mean was taken fr om the range of scores reported by the participant and used in the current analyses. The mean score for perceived control was 3.23 (SD = 2.58). The workplace variable of overinvolvement was 6.98 (SD = 3.47). Participants average score for workplace support was 6.63 (SD = 3.57). Finally, participants reported an average score of
65 18.80 (SD = 6.24) for the workplace variable of negativ e clientele. Sample demographics related to the independent variables are presented in Tables 4-1 and 4-2. Measurement Reliabilities Measurem ent reliabilities for the Professiona l Quality of Life Scale (ProQOL; measuring Compassion Satisfaction and Comp assion Fatigue), the Psychologist s Burnout Inventory (PBI; measuring Control, Workplace Support, Types of Negative Clientele, and Overinvolvement), and Stressful Life Experiences Scale (SLES; measuri ng Personal Trauma History) appear in Table 43. Chronbachs coefficient alpha for the ProQ ol Compassion Satisfact ion subscale = .87 and Compassion Fatigue subscale alpha = .82. This finding is compar able to previous reports by Stamm, (2005) (.87 and .80, respectively). Reliability for the PBI, Psychologists Burnout Inventory, indicated an alpha level of .68 for the Control subscale, .55 for the Workplace Support subscale, .79 for the Types of Negative Clientele subscale, and .64 for the Overinvolvem ent (see Table 3). The alpha level for the overall PBI was .63. There have been no previous reliability analyses conducted on the PBI so there is no existing psychometric information upon which to compare these values. Due to the weaker nature of the alpha levels on the C ontrol, Workplace Support, and Overinvolvement subscales, attempts were made to improve th e internal reliability by removing poorly interrelated items. However, the removal of any si ngle poorly performing item failed to increase the alpha level significantly. Therefor e, the subscales were not modified and were used in their published form for the current analyses. Reliability analyses for the Stressful Li fe Experiences Scale (SLES; Stamm, 1996) revealed a Chronbachs coefficient alpha of .82 fo r the overall score on the SLES. This reliability coefficient is comparable to that found by Stamm (.70). Results of the reliability analyses are found in Table 4-3.
66 Correlational Analyses Pearson Product Moment correlations, using a criterion level of .05 (1-tailed), were computed between the two dependent variable s (i.e., compassion satisfaction and compassion fatigue) and each of the independent variables rela ted to individual (i.e., gender, ethnicity, years of clinical experience, personal trauma histor y, trauma training) and workplace factors (i.e., secondary exposure, control, overinvolvement, workplace support, nega tive clientele) in an attempt to confirm that significant relationships existed among study variables and to determine independent variables to include in the regression analyses. Compassion fatigue was significantly positively correlated with the workplace variables, control (r = 0.34, P .001) and overinvolvement (r = 0.26, P .010), and these relationships were in the expected directions. That is, lowe r reported levels of control and higher levels of overinvolvement were related to higher levels of compassion fatigue. Compassion fatigue was also significantly pos itively correlated with amount of secondary exposure (r = 0.37, P .001). Higher levels of secondary exposure to trauma we re also related to higher levels of compassion fatigue. There was no significant correlation between compassion fatigue and other workplace variables: Types of negative clientele, wor kplace support, personal tr auma history, or the individual variables related to years of clinical experience, ge nder, or ethnicity. Thus, these variables were not included in the subsequent regression equation. See Table 4-4 for a correlation matrix of the independent and depe ndent variables examined in the current study. Compassion satisfaction was si gnificantly negatively correlated with the workplace variables, cont rol (r = -0.47, P .001) and overinvolvement (r = -0.25, P .013). That is, higher reported levels of control and lower leve ls of overinvolvement we re related to higher levels of compassion satisfaction. Compassion satisfaction was also positively correlated with
67 personal trauma history as measured by the SLES (r = .24, P .016) and participants reported years of clinical experience(r = .22, P .027). Personal trauma hi story was not related to compassion satisfaction in the expect ed direction. It was expected that lower levels of personal trauma history would be associat ed with higher levels of compassion satisfaction. However, it was found that higher levels of personal trauma hi story were associated w ith higher levels of compassion satisfaction. Compassion satisfactio n was not significantly correlated with the workplace variables related to types of negativ e clientele, workplace support, amount of secondary exposure, or the indi vidual variables of gender or et hnicity. Thus, these variables were not included in the subse quent regression equation. Also, examination of the correlation matrix revealed that the independent variables were correlated with each other at less than .50, suggesting a minimal risk of multicollinearity and associated Type I error. See Table 4-4 for a correlation matrix of the independent and dependent variables examined in this study. Regression Analyses Tests of skewness and kurtosis were conducted on th e data to assess that it was in line with the norm ality assumptions of multiple regression. Results of these analyses indicate that the assumptions for multivariate normality were met. All skewness and kurtosis estimates for the variables fell within the genera lly accepted values of 2 and Gender and ethnicity were not in cluded in the regression anal yses because there were too few participants (e.g., less than 25) in the diffe rent gender and ethnic gr oups to support their inclusion as individual variable s with meaningful contributions. This decision was based on the statistical assumption of attempting to balance th e complexity of investigation while maintaining meaningful and comprehensible results (e.g., fewer predictors with more meaningful results and
68 lower risk of experimenter-wise errors). Thus, th e removal of these variables served to increase power and reduce the probability of Type II errors (Cohen & Cohen, 1983). Hypothesis 1: Relationships among workpl ace variables and compassion fatigue The first hypothesis concerned the relationshi p between workplace variables (e.g., control, overinvolvem ent, and amount of secondary exposu re) and compassion fatigue. It was expected that therapists with less perceived control over their work activities, more overinvolvement with their clients, and grea ter amount of secondary exposure woul d have higher levels of compassion fatigue. In order to test the first hypothesis, a multiple linear regression analysis was conducted to determine whether therapist workplace variab les were significantly related to compassion fatigue. The scores reported by participants on th e Control and Overinvolvement subscales accounted for significant variation in compassion fatigue scores, F (3, 91) = 12.17, p < .001 ( adjusted R = .263). The standardized beta co efficient for the Control subscale ( = 0.244) was in the positive direction and was significant, t (91) = 2.64, p < .010. The standardized beta coefficient for the Overinvolvement subscale ( = 0.241) was significant and in the positive direction, t (91) = 2.59, p < .011. Lastly, the standardized be ta coefficient for amount of secondary exposure ( = .386) was significant and in the positive direction, t (91) = 4.29, p < .001. The direction of calculated effects indicated that the less a therapist endorsed a sense of control over their workplace, more overinvolvement with clients, and more secondary exposure, the more likely that the therapist was to have hi gher self-reported levels of compassion fatigue. This finding supported the hypothe sis that the workplace variab les, control, overinvolvement, and amount of secondary exposure are related to higher levels of compassion fatigue in therapists. This significant finding was repr esented by a large eff ect size, accounting for approximately 26% of the variance in compassion fatigue.
69 Hypothesis 2: Relationships among workpla ce variables and compassion satisfaction According to the second hypothesis, workplace variables (e.g., control and overinvolvement) would be related to compassion sa tisfaction. It was anticipated that therapists with greater perceived control over their work activities a nd less overinvolvement with their clients would have higher levels of compassion sa tisfaction. Thus, to test the second hypothesis, a multiple linear regression analysis was conducted to determine if therapist workplace variables were significantly related to compassion satisfaction. The scores reported by participants on th e Control and Overinvolvement subscales accounted for significant variation in compassion satisfaction scores, F (4, 93) = 9.92, p < .001 ( adjusted R = .269), as well. The standardized beta coefficient for the Control subscale ( = 0.414) was in the negative direction and was significant, t (93) = -4.58, p < .001. The standardized beta coefficient fo r the Overinvolvement subscale ( = -0.122) was also in the negative direction but was not significant, t (93) = -1.33, p < NS. The direction of calculated effects indicated that the more a therapist endorsed a sense of c ontrol over their workplace, the more likely the therapist was to have higher ratings of compassion satisfaction. This finding supported the hypothesis that the wo rkplace variable of control is related to higher levels of compassion satisfaction in therapists, with the substantial effect size of a pproximately 27% of the variance accounted for. Hypothesis 3: Relationships among individua l and w orkplace variables and compassion fatigue The third hypothesis addressing changes in va riance of compassion fatigue after workplace and individual variables were added to the model was not conducted because no significant correlations were found between compassion fatigue and the individual va riables (e.g., personal trauma history, years of c linical experience).
70 Hypothesis 4: Relationships among individua l and w orkplace variables and compassion satisfaction The fourth and final analysis was designed to test the hypothesis that workplace variables and individual variables were significantly related to therapis ts compassion satisfaction. More specifically, the individual vari ables (e.g., personal trauma hi story and years of clinical experience) were entered into the stepwise regression procedure, followed by workplace variables (e.g., contro l and overinvolvement) to examine the change in the variance of compassion satisfaction. Thus a stepwise linear regression analys is was conducted to determine whether the independent variable s (e.g., individual followed by workplace variables) influenced therapist ratings of the dependent variable (e.g., co mpassion satisfaction). Stepwise multiple regression was utilized to investigate whether individual and workplace variables were related to ther apist reported compassion satis faction. The decision rule was established that a variable must account for a change in R2 of 2% or more and must be significant at the p < .05 level to be included in the final regr ession equation. Individual variables in the initial regression equation included personal trauma history, years of experience, control and overinvolvement. The final equation for compassion satisf action contained only three variables, personal trauma history, years of clinical experiences, and th e control subscale, and yielded a multiple R of .54 (F = 12.54, p = .001) accounting for approximately 26% of the total variance in compassion satisfaction scores. The standardized beta coeffici ent for the indi vidual variable, personal trauma history ( = 0.198) was in the positive direction and was significant, t (94) = 2.22, p < .029. The standardized beta coefficient for the individual variable, participants reported years of clinical experience ( = 0.133) was in the positive direction but was not significant, t (94) = 1.48, p < NS The standardized beta coeffi cient for the Control subscale ( = -0.445) was in the negative direction and was significant, t (118) = -5.08, p < .001.
71 Summary Statis tical evidence from correlational an alyses indicated sign ificant relationships between compassion fatigue and workplace variab les of control, overinvolvement, and amount of secondary exposure. Compassion satisfaction was positively related to workplace factors (i.e., control and overinvolvement) and individual variables (i.e., persona l trauma history and years of clinical experience). All signifi cant correlations were in the expected directions except for the correlation between personal trauma history a nd compassion satisfaction, which was correlated in the opposite direction as was expected (e.g., hi gher levels of personal trauma history were associated with higher levels of compassion satisfaction). Three regression analyses were conducted to test three of the four hypotheses (i.e., Hypotheses 1, 2, and 4) concerning the relationships between indivi dual and workplace variables and the dependent measures of compassion fati gue and compassion satisfaction. Statistical evidence from regression analys es to test Hypothesis 1 (i.e., There is no variance in compassion fatigue explained by workplace variables ) indicated that therapists with less of a sense of control over their workplace, more overinvolvement with their clients, and more secondary exposure to clients trauma had higher levels of compa ssion fatigue, accounting for approximately 26% of the variance. Results from the regression anal yses conducted to test Hypothesis 2 (i.e., There is no variance in compassion satisfaction explained by workplace variables ) indicated that therapists who endorsed more of a sense of control had higher ratings for compassion satisfaction, accounting for approxima tely 27% of the variance. Analyses were not conducted to test Hypothesis 3 (i.e., There is no change in variance of compassion fatigue when workplace variables are added to a model including individual variables) because correlations among compassi on fatigue and individual variables were not significant. Hypothesis 4 (i.e., There is no change in variance of compassion satisfaction when
72 workplace variables are added to a model includi ng individual variables) involved a stepwise regression with results indicating th at therapists with more of a se nse of control over their work environment, higher reported years of clinical work experience, and mo re personal trauma in their history had higher ratings of compassion satisfact ion. The final equation accounted for approximately 26% of the variance in compassion satisfaction.
73 Table 4-1. Descriptive Statistics fo r Categorical Individual Variables Variable Frequency (f) Percentage (%) Gender Male 25 25.5 Female 73 74.5 Ethnicity African American 2 2.0 Caucasian 90 91.8 Latino 3 3.1 Other 3 3.1 Specialized Trauma Training Yes 97 99.0 No 1 1.0
74 Table 4-2. Descriptive Statistics for C ontinuous Individual and Workplace Variables Independent Variable N Mean SD Individual Variables Years of Clinical Experience 98 18.53 9.09 Personal Trauma History 98 55.09 33.19 Workplace Variables Secondary Exposure 95 16.98 11.27 Control 98 3.23 2.58 Overinvolvement 98 6.98 3.47 Workplace Support 98 6.63 3.57 Negative Clientele 98 18.80 6.24
75 Table 4-3. Measurement Reliabili ties for ProQOL, PBI, and SLES Scale N Minimum Maximum Mean SD ProQOL Comp Fatigue 98 1 36 11.03 6.13 Comp Satisfaction 98 21 50 40.87 5.59 PBI Neg Clientele 98 3 33 18.80 6.24 Support 98 0 17 6.63 3.57 Overinvolvement 98 0 15 6.98 3.47 Control 98 0 11 3.23 2.58 SLES SLES Total 98 0 180 55.09 33.19 Note. ProQOL = Professional Quality of Life Scale; Comp Fatigue = Compassion Fatigue; Comp Satisfaction = Compassion satisfaction; PBI = Psychologists Burnout Inventory; Neg Clientele = Negative Clientele; SLES = Stressful Life Experiences Scale.
76 Table 4-4. Pearson Product Moment Correla tions between Independent and Dependent Variables Variables 1 2 3 4 5 6 7 8 9 1. CompSat __ 2. CompFat -.169 __ 3. SecondExp .086 .371* __ 4. Years .223* -.107 -.022 __ 5. TraumaHx .243* .152 .149 .213* __ 6. NegClient -.082 .118 .320** -.165 .015 __ 7. Support -.162 -.058 -.166 .027 -.021 -.328** __ 8. OverInvol -.250* .258* -.134 -.175 .019 .383** -.179 __ 9. Control -.467** .338** .073 -.108 -.037 .100 -.006 .271** __ Note. CompSat = Compassion Satisfaction; CompFat = Compassion Fatigue; SecondExp = Amount of Secondary Exposure; Years = Years of Clinical Experience; TraumaHx = Personal Trauma History; NegClient = Negative Clientele; Support = Workplace Supp ort; OverInvol = Overinvolvement. *p < .05 **p < .01
77 CHAPTER 5 DISCUSSION Overview of the Study The purpose of this stud y was to examine the relationships among workplace variables (e.g., control, overinvolvement, negative client ele, workplace support, amount of secondary exposure), individual variables (e.g., gender, ethnicity, personal trau ma history, years of clinical experience, trauma training) and levels of co mpassion fatigue and compassion satisfaction for Professional Counselors, Psychologist s, and Social Workers working with traumatized clients. The study specifically addre ssed (a) the relationships among levels of control, overinvolvement, negative clientele, workpl ace support, amount of secondary exposure and compassion fatigue; (b) the relationships among le vels of control, overinvolvement, negative clientele, workplace support, amount of seconda ry exposure and compassion satisfaction, and (c) changes in variance for compassion fatigue and compassion satisfac tion when workplace variables (e.g., control, overi nvolvement, negative clientele, workplace support, amount of secondary exposure) were added to a mode l including individual variables (e.g., gender, ethnicity, personal trauma history, years of clinical experience, trauma training). Discussion of Des criptive Statis tics The target population for this study consisted of Psychologists, Professional Counselors, and Social Workers who work with trauma survi vors in a clinical environment. Obtaining a sample of therapists from among the members of the International Soci ety of Traumatic Stress Studies (ISTSS) and Association for Traumatic Stress Specialists (ATSS) was identified by the primary investigator as an a ppropriate method for gaining a representative sample from the desired population. Members we re invited to participate through e-mail invitation which described the study and included a link to complete the survey. The surv ey included (a) the
78 Professional Quality of Life Scale (Stamm, 200 5) (b) the Psychologists Burnout Inventory (Ackerley et. al., 1988), (c) the Stressful Life Experiences Screening Short Form (Stamm, 1997), and (d) a demographics questionnaire, includi ng gender, age, ethnicit y, years of clinical experience, amount of secondary exposure (i.e., as measured by weekly hours of client contact with trauma survivors), and trau ma training, as well as several que stions to assess whether those who completed the survey accurately estim ated the population. A non-experimental, correlational design was utilized to test the research hypotheses, which were stated in the null form and tested at the .05 level of significance. The participant sample consisted of 98 ther apists, who were primarily female (74.5%) and Caucasian (91.8%). This high percentage of fema le Caucasians in a sample of therapists is similar to samples in other studies (e.g., Ka ssam-Adams, 1995; Pearlman & MacIan, 1995; Schauben & Frazier, 1995) indicating that this ma y be a somewhat accurate representation of the population with respect to gender and ethnicity. However, the sk ewed percentage of Caucasian females in the sample may indicate that member s of this gender and ethnic group may be more likely to respond to surveys and/or to be member s of professional organizations such as ISTSS and ATSS. It is also possibl e that the overrepresentation of females is a function of gender norms for expression of emotions and emotional vulnerability. Traditi onal gender norms dictate that it is more socially acceptable for wome n to express emotions and admit emotional vulnerability than it is for men. Therefore, fe males may have been more likely to acknowledge that they may be negatively impacted by trauma work and more open to disclosing this in the survey for the present study. The lack of participation of males in this study may have impacted results and the ability to accurate ly detect gender effects. Since there was a lack of diversity
79 within the sample, ethnicity and gender were not able to be us ed as individual independent variables in this study. Several items were included in the demographi cs questionnaire to en sure a representative sample of the desired population. First, part icipants were asked how they identify their profession (e.g., Psychologist, Prof essional Counselor, Social Wo rker, or Other). The final sample consisted primarily of Psychologists ( 50.0%), followed by Social Workers (26.5%), and Professional Counselors (23.5%). Approximately 19% of those who initially completed the survey indicated Other as best describing their profession. Responde nts indicating Other were deleted in order to accurately represent the ta rget population. Second, participants were asked to repor t their primary professional activity (e.g., Counseling/Therapy, Teaching/Research, Supervis ory/Administrative, Case Management, and Other). The aim was to survey clinicians who we re actively engaged in clinical activities. The majority of participants in the sample identifi ed their primary professi onal activity as counseling or therapy (77.6%). This percentage was judged by the primary investigator to be sufficient in estimating the target population for the following reasons: (1) the vast majority of the sample was primarily engaged in clinical activities (77.6%) and (2) the other 22.4% were assumed to be either previously engaged in clinical work or engaged in part-time clinical work. The assumption that most of the sample was at least en gaged in some type of clinical work was based on examination of the number of re ported hours of clinical contact wi th clients, in which all but 2 participants indicating some weekly clinical contact with traumatized clients. Finally, participants primarily in dicated that they worked with othe r professionals (66.3%) rather than alone (33.7%).
80 The mean score for the workplace variable of perceived control was 3.23 (SD = 2.58). This score was comparable to previous studies (e.g., M = 2.59; Ackerley et. al., 1988). The workplace variable of overinvolvement was a mean score of 6.98 (SD = 3.47), which is somewhat lower than scores found in prior st udies (e.g., M = 9.52; Ackerley et. al., 1988). Participants average score for workplace support was 6.63 (SD = 3.57), which is also comparable to prior studie s (e.g., M = 5.02 and M = 7.28; Ackerley et. al., 1988). Finally, participants reported a mean score of 18.80 (SD = 6.24) for the workplace variable of negative clientele, which is slightly lower than mean scores found in previous research (e.g., M = 23.76; Ackerley et. al., 1988). Thus, overall mean scor es for the workplace variables as measured by the Psychologists Burnout Inventory in the cu rrent study are comparable to scores found by Ackerley and colleagues (1988). Amount of secondary exposure, the average amount of time that participants worked with trauma clients per week, was 16.98 hours (standard deviation = 11.27). The array of responses for this item was from 0 to 50 hours, which repres ents a wide variety in the amount of secondary exposure among participants. This variety of res ponses indicates that part icipants in this study ranged from no clinical interactions with trauma c lients (N = 2) to full-time clinical contact, with the average respondent being more of a part-time clinician. These more part-time clinicians may have been also involved in other workplace ac tivities in addition to client contact (e.g., supervisory, paperwork). Since all but 2 particip ants were actively engaged in direct clinical work with traumatized clients, this researcher determined that th e sample accurately reflected the desired population with respect to clinical involvement. Ther apists reported an average of approximately 18 years of experience (standard deviation = 9.09), with a range from 1 to 36
81 years. This wide range of clinicians expe rience among the participan ts indicates that the sample included novice to senior clinicians. The average total score for personal trauma history, as measured by the SLES (Stamm, 1997), was 55.09 (SD = 33.19). The total score possi ble for the SLES-Short Form is 200, which would indicate that the participant experienced a ll 20 trauma experiences (i.e., a score of 10 for each item meaning exactly like my experience). Thus, the therapists who participated in this survey had some trauma experiences, but not subs tantially high amounts of trauma experiences. The individual variable of trauma training was no t able to be used in further analyses because 99% of participants indicated they had some type of trauma training. This finding further demonstrates that the sample for the present stud y consists primarily of specially trained trauma therapists, whereas the entire population of trauma therapists may include more therapists who are not specially trained. For the dependent measure of compassion fatigue therapists in the current study reported an average score of 11.03 (SD = 6.13). The mean score found for compassion fatigue in this study is comparable to normative data provided by Stamm (2005), in which the mean score for compassion fatigue was 13 (SD = 6.3). According to categories of risk levels devised by Conrad and Kellar-Guenther (2006), this co mpassion fatigue score was consid ered to be very low risk. In Conrad and Kellar-Guenthers (2006) study, mo re than 50% of child protection workers who were studied fell into the high risk category (i.e., total score of 51 to 75). Thus, the population of child protection workers was found to be more hi gh risk for negative outcomes than participants in the current study. Other researchers have f ound more varying levels of compassion fatigue (e.g., M = 28.78, SD = 13.15; Stamm, 2002). In the current study, trauma therapists were studied as identified thro ugh their participation in a trauma-spe cific professional organization. It
82 is possible that members of these organizations fare slightly better due to their high motivation and vested interest in the field as evidenced by their participation in the organization. In future studies, efforts should be made to access a different sample to estimate the population or to control for these potentially rela ted factors (e.g., motivation level, interest in the field). For the dependent variable of compassion sati sfaction, participants reported an average score of 40.87 (SD = 5.59). According to Stamms (2005) normative data for compassion satisfaction, the average score in this study is similar to that found in the normative data (M = 37, SD = 7.3). According to Conrad and Kellar-Guenthers (2006) rating system for compassion satisfaction, this level was considered to be low potential (i.e., a range from 0 to 63). Approximately 50% of child protection workers in their study were placed in the category of good potential for compassion satisfaction. This higher average level of compassion satisfaction is similar to another study that utiliz ed data from South Afri can professionals trained in crisis de-briefing, mental he alth caregivers, and Canadian ra pe crisis workers (i.e., M = 92.10, SD = 16.04; as analyzed by Stamm, 2002). Thus respondents in the current study seemed to have slightly lower levels of co mpassion satisfaction as well. Discussion of Correlational Re lationships among Variables Several significant correlationa l relationships were found a mo ng study variables. First, higher levels of the dependent va riable of compassion fatigue we re found to be related in the following directions to the following independent variables: le ss perceived control over the workplace, more overinvolvement with clients, and more secondary exposure to traumatized clients. All correlations found for compassion fatigue were co nsistent with the expected directions of the relationship s based on prior research. Second, higher levels of the dependent variab le of compassion satisfaction were found to be associated in the following directions to the following independent variables: more perceived
83 control over the work environmen t, less overinvolvement with clients, more years of clinical experience, and greater amount of personal trauma experiences. Findings that more compassion satisfaction was related to more perceived cont rol, less overinvolvement and more years of clinical experience were consistent with the expected directions of the relationships. However, the association between compassion satisfaction a nd personal trauma history was counter to the anticipated direction of the relationship. For both compassion fatigue and compassion satisfaction, perceived control was found to be an important factor. Therapists who percei ved that they had the ab ility to use their own initiative at work, were able to structure thei r work day, and meet their own work expectations had both lower levels of compassion fatigue an d higher levels of compassion satisfaction. Compassion fatigue was also associated with more secondary exposure, op erationally defined as the amount of weekly client cont act hours. Specifical ly, therapists who saw more clients had higher levels of compassion fati gue. Compassion fatigue was also related to overinvolvement, with therapists who reported that they take their work home, take on excessive responsibility for their clients, and feel they work harder for change than their clients having higher levels of compassion fatigue. Compassion satisfaction was also related to ov erinvolvement, but in the opposite direction as compassion fatigue. In addition to perceived control, compassion satisfaction was related to the individual factor of years of clinical experience. Therapists with more years of experience were associated with higher levels of compassion satisfaction. The exp ected direction of the relationship between the dependent variables and years of clinical experience was as follows: more experience would be associated with lower levels of compassion fatigue and higher levels of compassion satisfaction. Prior research meas uring years of clinical experience and negative
84 therapist outcomes (e.g., burnout, vicarious trau matization, compassion fa tigue) indicated mixed findings, with some studies dete cting a significant positive relationship (e.g., Ackerely et. al., 1988; Pearlman & MacIan, 1995), and some st udies detecting no relationship (e.g., KassamAdams, 1999). No prior studies have focused on years of clinical experience and compassion satisfaction or other positiv e therapist outcomes. It was expected that therapists with more personal trauma history would have lower levels of compassion satisfaction. Research findings ha ve indicated a consistent relationship between higher levels of personal trauma and higher le vels of compassion fatigue (e.g., Pearlman & MacIan, 1995; Kassam-Adams, 1995). The pos itive relationship found between compassion fatigue and personal trauma histor y suggests that, for this group of trauma therapists, those who have experienced significant traumas in their life may find trauma work w ith clients particularly satisfying and rewarding. This finding also indicates an indirect relationship between co mpassion fatigue and compassion satisfaction, rather than a direct inverse re lationship as was originally proposed by Stamm (2002). In previous studies higher levels of personal trauma history were shown to be associated with higher levels of compassi on fatigue, whereas the present study found higher levels of personal trauma history to also be associated with higher levels of compassion satisfaction. If the relationshi p between compassion fatigue and compassion satisfaction were direct and inverse, it would be expected that the relationship between personal trauma history and compassion satisfaction would be negative (i.e ., in the opposite direction as was found). An indirect relationship between co mpassion fatigue and compassion satisfaction has been alluded to in recent studies (Conrad & Kellar-Guenther, 2006), wherea s the original construct of compassion satisfaction was developed to be a positive parallel to the construct of compassion
85 fatigue (Stamm, 2002). In the current study, personal trauma hi story was uniquely related to compassion satisfaction, but was not a relevant factor for compassion fatigue. It is possible that depending upon how therapists uniquely respon d to and process their personal trauma experience, they may either have compassion satis faction or compassion fatigue. Therapists in the current study may have more effectively cope d with and/or processed their personal trauma and thus had more positive outcomes (i.e., compa ssion satisfaction) than negative outcomes (i.e., compassion fatigue). Previous researchers (e.g., Conrad & Kellar-Guenther, 2006) have found that compassion satisfaction may have a moderati ng effect, with compassion satisfa ction being associated with lower levels of compassion fati gue. Although the current researcher did not find a correlational relationship between compassion fa tigue and compassion satisfacti on, findings indicate that the workplace variables of perceived control and ove rinvolvement are conversely related to each of the dependent variables. Because the current study is correlational, it is un able to be determined whether certain levels of overinvolvement and control cause compassion fatigue or compassion satisfaction or whether the two dependent variables cause therapists to have less or more perceived control and more or le ss overinvolvement. Findings in the current study indicate that the relationship between compassion fatigue a nd compassion satisfaction may be mediated by the workplace variables of c ontrol and overinvolvement. Discussion of Hypotheses Relationships among Control, Overinvolveme nt, Negative Clientel e, Workplace Support, amount of Secondary Exposure, and Compas sion Fatigue The first hypothesis was a null statement about the relationships between the workplace variables of control, overinvolvement, and amount of secondary exposure and compassion fatigue. Examination of the result s of the data analysis conducted revealed that therapists with
86 less perceived control over their work activities, more overinvolvement with clients, and a higher amount of secondary exposure reported higher le vels of compassion fatigue. Thus, the null hypothesis was rejected. Workplac e variables related to control, overinvolvement and amount of secondary exposure accounted fo r a substantial amount of vari ance in compassion fatigue. Therefore, there is a strong association among the workplace factors of perceived control, overinvolvement with clients and amount of exposure to clients trauma material with therapists experience of compassion fatigue. The relationship between amount of secondary exposure and compassion fatigue found in this study is consistent with Lees (1995) findings for the relationship between secondary exposure and vicarious traumatization. Several other studies have repo rted an association between therapists amount of secondary exposure (i.e., operationalized as higher trauma clients on their caseloads) and their levels of compa ssion fatigue (Brady et. al, 1999; Chrestman, 1999; Schauban & Frazier, 1995). The workplace vari ables of control and overinvolvement are consistent with studies about the relations hip between workplace variables and burnout (Ackerley et. al, 1988). Ackerley et. al (1988) found that lack of control in the workplace and overcommitment to clients was associated with more burnout. In a more recent study, Rupert and Morgan (2005) found that solo and group priv ate practitioners had less burnout related to more control in the workplace. These prev ious research findings involving control, overinvolvement, and secondary exposure are rela ted to the findings for compassion fatigue in the present study due to a focus on negative ps ychological outcomes for therapists. Although these constructs differ in some respects (e.g., speed of onset, length of recovery, specific symptomology), all are negative psychologica l outcomes for therapists.
87 The remaining workplace variables related to negative clientele and workplace support were not included in the regres sion analysis because no significan t correlational relationships were found between these variables and comp assion fatigue. Studies examining workplace support, social support, and other negative outcom es for therapists work ing with traumatized clients (e.g., vicarious traumatization and burnout) have often found an inverse relationship between these negative outcomes and support (Etzion, 1984; Pearlman & MacIan, 1995). However, a study by Landry (2001) did not find a relationship be tween compassion fatigue and social support. These research ers utilized different scales to measure support and perhaps different operational definitions of social support. Further inve stigation into how social support mitigates compassion fatigue is needed, using clearly defined constructs and psychometrically strong scales to measure workplace and other forms of social support. In the current study, the researcher did not find any association be tween negative client behaviors and compassion fatigue, which indicates that the variable of negative clientele (e.g., client behaviors/conditions such as high lethal ity, major psychopathology, re sistance in therapy, and non-compliance with aspects of therapeutic treatment) may not be related to the experience of compassion fatigue. However, Rupert a nd Morgan (2005) found that working with less negative or disturbed clie ntele was associated with lower leve ls of burnout for psychologists who worked independently and in gr oup private practice. The lack of findings for a relationship between negative clientele and compassion fatigu e may be due to the different experience and onset of burnout versus compassion fatigue. Burnout tends to have a more gradual and insidious onset than compassion fatigue, which is a more acute trauma response triggered by clients traumatic material (Figley, 1995). Thus, negati ve client behaviors, as measured by the
88 Psychologists Burnout Inventory (e .g., making suicidal statements, being late for appointments), may not have any association with the experience of compassion fatigue as compared to burnout. In discussing prior research associated with other negative therapist outcomes (e.g., burnout), it should be noted that th ese studies were not focused on trauma therapists, whereas the present study was specific to therapists working w ith a significant amount of traumatized clients. This difference in specific populations may account for some disparity in findings between some prior research and the findi ngs in the current study. Relationships among Control, Overinvolveme nt, Negative Clientel e, Work place Support, amount of Secondary Exposure, and Compassion Satisfaction The second hypothesis was a null statement a bout the relationships between workplace variables (e.g., control and overinvolvement) and compassion satisfaction. Therapists with greater perceived control over their work activities and less overinvolvement with clients were found to have higher levels of compassion satisfa ction. Thus, for the second hypothesis, the null hypothesis was rejected. However, regression anal yses indicated that ov erinvolvement was not significantly related to compassion satisfaction, whereas control was significantly related. The relationship between control and compassion satisfac tion found in this study is consistent with previous research and theore tical propositions about the correlates of positive psychological outcomes for therapists. For exam ple, one qualitative analysis by Schauben and Frazier (1995) revealed that a majority of therap ists felt their work is important. Therefore, it follows that therapists who perceive high levels of control over their work activities and responsibilities may feel better a bout the importance of that wor k. Stamm (2002) proposed that positive collegial support should combat a decreased sense of competency and control. Thus, the finding of a positive relationship between control and compassion satisfaction, which is related to a sense of competency, is consistent with Stamms (2002) assertion. Although the finding of a
89 relationship between control and compassion satis faction was consistent with Stamm (2002), the finding that support was not rela ted to compassion satisfaction was inconsistent with Stamm (2002). In the current study, significant relationshi ps between negative clientele, workplace support, and amount of secondary exposure we re not found. Findings of this study are inconsistent with assertions of other researchers that social support and positive collegial support are associated with compassion satisfaction (Sta mm, 2002). With respect to negative clientele and amount of secondary exposure, there does no t appear to be a rela tionship between these factors and compassion satisfaction. However, negative clientel e and secondary exposure have been shown to be associated with negative r eactions for therapists (e.g., compassion fatigue, burnout), to which compassion satisfaction was c onstructed to be a positive parallel. The difference between factors that have been a ssociated with compassion fatigue and those associated with compassion satisfaction further indicates that the relationship among these two factors (i.e., compassion fatigue and compassion satisfaction) is not a direct and/or inverse one and may be more complex with, as yet undetermin ed, mediating and or moderating factors. Also, as mentioned previously, many prior studies related to workplace variables and negative therapist outcomes were focused on general therapists rather than trauma therapists. This difference in specific populations may account for some inconsistencies between significant findings of relationships between some of the workplace variables and compassion fatigue and compassion satisfaction. Since the current study was focused on compassion fatigue and compassion satisfaction, which are trauma specific, the population of therapists who work with trauma survivors was targeted.
90 Relationships among Individual and Workplace Variables and Compassion Fatigue The third hypothesis was a null statem ent about changes in variance for compassion fatigue after workplace variables were added to a model including individual variables. A stepwise regression was to be used to test this hypothesis; however init ial correlational analyses revealed no significant relationshi ps between compassion fatigue and any individual variables (e.g., gender, ethnicity, trauma training, personal trauma history, y ears of clinical experience). The results from the initial correlational an alyses of the relationships between compassion fatigue and individual variables ar e inconsistent with some research concerning related factors of negative psychological outcomes for trauma therapis ts. First, personal tr auma history has been frequently found to be positively associated with negative therapist outcomes. Vicarious traumatization and compassion fatigue have been shown to be associated with more personal trauma history (Pearlman & MacIan, 1995; Kass am-Adams, 1995). However, one study did not find a relationship between compassion fatigue and personal trauma history (Schauben & Frazier, 1995). Schauben and Frazi er (1995) concluded that future studies could be improved by including a scale to measure personal trauma hi story rather than a dich otomous variable (i.e., whether they have experienced personal trauma or not). The use of a scale to measure the extent to which a therapist had encountered trauma in th eir own life was used in the present study based on the suggestion of Schauben and Frazier. The lack of variance in responses (i.e., the majority of respondents indicated low levels or no trauma in their history) found in this study related to personal trauma experiences may have contributed to the lack of dete ction of a significant relationship. These results also may indicate that when using a valid scale to measure personal trauma history as a continuous variable, no rela tionship exists between personal trauma history and compassion fatigue. Thus, previous detection of a relationship between personal trauma history and compassion fatigue may have been due to measuring personal trauma as a
91 dichotomous variable (e.g., related to confusi on about the definition of trauma, inaccurate measurement of personal trauma). It is unknown whether therapists in this study have worked through or processed their traumatic experiences as well as the subjective se verity of their trauma experiences, which may have accounted for the lack of findings of a relationship between personal trauma history and compassion fatigue. Specifically, if therapists effectively processed thei r personal trauma, they may not be as high risk for developing comp assion fatigue. Also, it is unknown whether participants experienced the even ts indicated on the Stressful Li fe Events Scale Short Form (Stamm, 1997) as traumatic, which can be subjec tive. Future research ers should attempt to account for these potentially relate d factors (e.g., processing of trau ma, subjective severity of the trauma). In the current study, a predominately female and Caucasian sample was obtained. The relationship between gender and ethnicity and co mpassion fatigue was not examined because of the lack of diversity with resp ect to gender and ethnicity in th e sample. Previous research examining gender and compassion fatigue as well as other negative outcomes (e.g., burnout and vicarious traumatization) has re vealed mixed findings. Specifi cally, some researchers found a significant relationship (e.g., Etzion, 1984; Kassam-Adams, 1995; Meyers & Cornille, 2002), whereas others did not (e.g., Wee & Myers, 2002). No studies to date have indicated ethnicity as a related factor of compassion fatigue, most likely due to lack of ethnic diversity in samples of previous research. Future studies should use specific methods (e.g., quota sampling) to obtain a more diverse sample to examine gender and ethnic ity effects. In the current study, the primary investigator attempted to obtai n a diverse sample by surveying members of an international organization. However, due to this researcher not using a specific sampling method such as
92 quota sampling, the organization members who co mpleted the surveys were predominantly female and Caucasian. A significant relationship between compassion fati gue and years of clinical experience also was not found. Prior studies that have examined vicarious traumatization and burnout have found an association between less clinical expe rience and higher levels of these negative psychological outcomes (e.g., Pearlman & MacI an, 1995; Etzion, 1984). Only one known study revealed a significant relationship between compa ssion fatigue and years of clinical experience (Landry, 2001). These findings indi cated that there may be differenc es between the constructs of vicarious traumatization, burnout, and compassion fatigue that is related to years of clinical experience. The relationship be tween negative psychological outco mes for therapists and years of clinical experience may be rela ted to a potential mediating variab le that could be examined in future studies. As reviewed in detail in the discussion of Hypothesis One results, there were no significant relationships between the workplace variables of negative clientele or workplace support and compassion fatigue. Relationships among Individual and Workplace Variables and Compassion Satisfaction The fourth and final hypothesis was a null statement about changes in variance for com passion satisfaction after workplace variables (e.g., control, overinvolvement) were added to a model including individual variables (e.g., pers onal trauma history a nd years of clinical experience). Thus, the fourth null hypothesis wa s rejected. The final regression model included personal trauma history and contro l as the factors that best acc ounted for variance in compassion satisfaction. Though years of c linical experience was correlated with compassion satisfaction and was retained in the regression equation, the relationship was not shown to be statistically significant by the regression analysis. Because therapists with more personal trauma experience
93 were found to have had higher le vels of compassion fatigue in prior research findings, it was expected that therapists with more personal tr auma experiences in this study would have lower levels of compassion satisfaction. However, it wa s found that therapists with more experiences of personal trauma had higher levels of compassion satisfacti on. Thus, the findings of the current study do not support or contradict prior research on the topic of compassion satisfaction. Findings associated with perceived control and compassion satisfac tion were discussed previously in the discussion of Hypothesis Two. Theoretical Implications Findings of the present study lend substantial s upport to several of the theories underlying the constructs of com passion fatigue and compas sion satisfaction. The research findings that higher levels of compassion fati gue were strongly related to less perceived control in the workplace, more overinvolvement with clients, and more secondary exposure provide support for aspects of Figleys (1995) causal model for compassion fa tigue and Maslach and Leiters (1997) work-climate perspective. An additiona l discovery that higher levels of compassion satisfaction were related to more perceived control and greater e xperience with personal trauma can be explained by aspects of resilience theo ry, posttraumatic growth, and cognitive selfdevelopment theory. Figleys (1995) Compassion Fatigue model outlines a proc ess for the development of compassion fatigue, beginning with exposure to a trauma survivor, then empathic connection, and then the development of compassion stress. Compassion stress may become the more severe form of compassion fatigue when protectiv e factors (e.g., sense of achievement and disengagement) are not present and/or exacerbat ing factors (e.g., prolonged exposure, traumatic recollections, other life disruptions ) are present. In the present study, strong associations were found between compassion fatigue and amount of s econdary exposure to clie nts trauma. This
94 finding lends support for Figleys assertion that the amount of s econdary exposure to clients trauma (e.g., prolonged exposure) is a risk factor for the development of compassion fatigue in therapists. Additionally associations found in the current study between compassion fatigue and overinvolvement, defined as the opposite of engagement, support Figleys claim that disengagement is a protective factor that can ameliorate compassion fatigue. Associations found between compassion fatigue and perceived control in the workplace are consistent with Maslach and Le iters (1997) work-climate view of burnout and other negative consequences for therapists. Maslach and Leiters theory purports that la ck of control in the workplace as well as work overload are risk factors for burnout. In the current study, compassion fatigue was found to be related to pe rceived workplace control. The relationship found in the current study between compassion fatigue and perceived control indicates a conceptual link between compassion fatigue and burnout, with both constructs being related to the workplace factor of perceived control. Compassion satisfaction was shown in the pres ent study to be strongly associated with perceived control in the workplace and therapis ts personal trauma history. In building an integrated model for compassion satisfaction, perceived workplace control is an important contributing factor. It is as yet undetermin ed whether higher percei ved control is caused by having high compassion satisfaction, or vice versa. An unexpected finding in this study was the association between therapists personal trauma history and comp assion satisfaction, with more personal trauma experiences associ ated with higher levels of compassion satisfaction. According to Cognitive Self-Development Theory, changes in core beliefs about self, others, and the world can lead to a sense of helplessness, lack of cont rol, and disturbance in identity through secondary exposure to client trauma. As these cognitive changes occur, therapists w ho have already been
95 through a personal trauma may have previously adopted more resilient core beliefs that allow for integrating the possibility of trauma. Whereas changes in core beliefs post-trauma may increase resilience and make the development of negative therapist outcomes (e .g., compassion fatigue) less likely, this idea does not fully explain the association between therapists amount of personal trauma experiences and compassion satisfaction. It is possible that therapists who have been through traumatic experiences and coped effectively are better able to work with traumatized clients and may feel more of a sense of achievement in helping someone else heal from trauma. This ability to turn personal suffering into a sense of meaning and pu rpose is a core concep t in the theory of Posttraumatic Growth (Tedeschi & Calhoun, 1996) Thus, there may be a conceptual link between the constructs of posttraumatic growth and compassion satisfaction as both concepts relate to the ability to turn suffering into a positive outcome. This preliminary conclusion needs further clarification throu gh future research. Another possible explanation for the finding that more personal trauma experiences related to higher levels of compassion sa tisfaction is that trauma therapists, who were targeted in the current study, may have been more drawn to ther apeutic work with trauma survivors because of their own experiences with trauma Thus, some trauma therapis ts may have higher levels of compassion satisfaction because they find their work especially meaningful in light of their own experience overcoming trauma. Future researchers attempting to replicate the finding that more personal trauma experiences is related to higher levels of co mpassion satisfaction may obtain different results if using a sample of ge neral, not trauma-specific, therapists. Practical Implications The present study and associated findings c ontribute to knowledge and understanding of the im portance of workplace factors in influenc ing both positive and negative psychological
96 outcomes for therapists. Specifi cally, the findings of this stud y underscore the importance of therapists, particularly those working with trauma survivors, having more perceived control over their work environment (e.g., using their own initiative at work, making decisions about the work they do). Perceived control in th e workplace was found in this study to be associated with higher levels of compassion satisfaction and lower levels of compassion fatigue. Associations between workplace control and burnout have been demons trated in previous research (e.g., Rupert & Morgan, 2005). Thus, allowing ther apists to have more control over their work activities may be preventative for both burnout and compassion fatigue Increasing therapis ts workplace control may also enhance positive psychological outcom es such as compassion satisfaction. Based on the current study findings, specific recommenda tions for workplace settings to enhance therapists sense of control include the following: (1) allowing therapists to be democratically part of decision making; (2) allowing therapists to provide input about their caseloads (e.g., both amount of cases and types of cases). The practical implications for the finding conc erning the importance of perceived control are extensive given the directi on that the mental health fiel d is moving with managed care becoming the norm. Specifically, managed care often involves therapists having less of a sense of control over their therapeutic work as they ha ve to answer to thirdparty payers to justify continued coverage. The findings in this study of the importance of control in both enhancing positive outcomes (e.g., compassion satisfaction) and decreasing negative outcomes (e.g., compassion fatigue) highlight aspe cts of managed care that may be detrimental to the well-being of therapists and in-turn to client s and the organizati on as a whole. The recommendation of allowing therapists to provide input about their caseloads relates to the current research findings concerning therapists secondar y exposure to clients trauma
97 material and compassion fatigue. In todays comm unity mental health environments, therapists in a variety of settings are enco uraged to have higher and highe r caseloads to meet financial demands for the organization an d/or practitioner. The pres ent study results highlight the potentially negative outcomes if th erapists, especially those worki ng with trauma survivors, have higher amounts of weekly client contact hours. Therapists with higher client caseloads have been shown in previous studies to show increased levels of burnout for therapists (Ackerley et al., 1988; Rupert & Morgan, 2005). Thus, having higher caselo ads of trauma clients puts therapists at greater risk for compassion fati gue and burnout. An increased risk for both compassion fatigue and burnout is of concern since the combination of these two conditions has been shown to have the most negative psychol ogical outcomes for therapists (Stamm, 2005). Although the present study did not find workplace support to be associated with compassion fatigue or compassion satisfaction, this researcher still encourages workplace support to be made a priority. The encouragement of workplace support is consistent with existing recommendations in the literature (Collins & Long, 2003; Skovholt, 2001; Stamm & Pearce, 1995). Future researchers should examine the role of workplace support using alternative measures as there we re some weaknesses in the PBI w ith respect to reliability. Overinvolvement with clients was associated with increased compassion fatigue and, to a lesser degree, with decreased compassion satis faction. Therefore, it is recommended that employers provide adequate supervision a nd consultation for therapists. Another recommendation based on findings in this study is to provide therapists with opportunities for personal counseling to discourage psychological overi nvolvement with clients and assist them in developing healthy ways to disengage. In a ddition to collegial support through supervision and consultation, professional development oppor tunities and workshops on self-care are
98 recommended to enhance therapists professi onal preparation and ability to take better psychological care of themselves. These opportun ities are intended to prevent therapists from becoming overinvolved with clients and developing compassion fatigue. Additional opportunities for paid time-off a nd vacation days may also enhance therapists ability to effectively disengage, enhance self-care, and to mitigate negative outcomes. Overinvolvement with clients should also be discouraged as part of the organization or agencys philosophy and practices. Some workpl ace settings require therapists to take on too much with respect to involvement with and advoc acy for clients, at the potential expense of the therapist, clients, and even the agency itself. Ag encies should set the precedent for therapists to be able to disengage and empower clients to be advocates for themselves whenever possible. Finally, an unexpected finding in this study was that therapists with more personal trauma were associated with higher levels of compa ssion satisfaction. This finding indicates that therapists who have experienced their own traumatic experiences may find greater purpose and sense of achievement in their work with trauma su rvivors. As surmised in the previous section, alterations and expansions in therapists core beliefs that often occu r post-trauma may allow them to better defend against negative outcomes and enhance positive outcomes. Although the degree to which therapists effectively processed their personal trauma was not examined in the current study, the development of more resilien t core beliefs is one possible explanation. Employers are encouraged to offer personal counse ling as well as supervision to help therapists effectively process personal trauma and receive support, which may encour age higher levels of compassion satisfaction. As mentioned previously, it is possible that therapis ts in this study were specifically drawn to trauma work due to their personal trauma experiences, and they thus may experience their work as more rewa rding and satisfying.
99 Research Implications and Future Directions A m ajor finding in the current study is the association of th erapists perceived workplace control with both compassion fatigue and co mpassion satisfaction. Although perceived workplace control has been shown to be associated with burnout in previous studies (Rupert & Morgan, 2005), the present study is the firs t one known to demonstr ate the link between therapists perceived workplace control a nd compassion fatigue as well as compassion satisfaction for therapists who work primarily with trauma. Future researchers should attempt to replicate the present findings using different methods of measuring workplace control. The Psychologists Burnout Inventory (PBI, Ackerley et. al., 1988) is an infrequently used and therefore underdeveloped instrument with somewhat lower reliability. Thus, it is imperative to determine if therapists perceived workplace cont rol is associated with compassion fatigue and compassion satisfaction when measured in different ways. Workplace support was not found to be associat ed with compassion fatigue or compassion satisfaction in the present study. However, th e Psychologists Burnout Inventory had lower reliability for the subscale of Workplace Support, which may have affected the ability of the measure to detect a relationship between workplace support and compassion fatigue and compassion satisfaction. An additional valid and reliable measure needs to be developed to address workplace support. Existing measures we re found that measure various forms of social support, however no measures were found that addressed perceived workplace support for therapists except for the Psychologi sts Burnout Inventory. Future researchers should examine indivi dual and workplace f actors and compassion fatigue and compassion satisfacti on with different samples that are representative of a broader population of trauma therapists. As will be discussed in the Limitations section, the ISTSS and the ATSS members may have been un ique with respect to their le vel of training, investment in
100 the field, and education level. These differe nces in the study sample may have differed systematically (i.e., an undetermined confounding variable) as compared to the population of all therapists working with traumatized populations Future studies should attempt to obtain samples with varying levels of training, which may impact levels of compassion fatigue and compassion satisfaction as well as associated factors. The primary investigator in the present study a ttempted to measure specialized training as an individual factor. Because th is variable was measured dichot omously, it was not able to be used in analyses due to dispr oportional nature of the data. Future researchers should include specialized training as a variable and measure it continuously (e.g., aski ng for the number of hours of specialized training) or perhaps through qualitative inquiry to determine the nature of training. Additional research needs to be done to furt her explore the interr elationships between compassion fatigue, compassion satisfaction, and burnout (i.e., subscales on the ProQOL). Previous research has indicated that levels of compassion satisfaction may moderate levels for both compassion fatigue and bur nout (Conrad & Kellar-Guenther 2006). In the present study, the interrelationships among compassion satisf action, compassion fatigue, and burnout were not examined. However, there was no significant correlation found between compassion fatigue and compassion satisfaction in the present study. Future research is needed to clarify the nature of these relationships. Future researchers should attempt to repli cate the finding that more personal trauma experiences for therapists are associated with higher levels of compassion satisfaction. Additionally, therapists core beliefs should be examined to test the hypothesis that more resilient core beliefs (i.e., post-trauma core beliefs) enhance compassion satisfaction and protect
101 against compassion fatigue. Post-traumatic grow th should be explored as a mediating factor between personal trauma experiences and compassi on satisfaction to clarify the conceptual link among these constructs. Finally, future researchers s hould investigate therapists pe rsonal trauma history as it relates to compassion fatigue usi ng different samples of trauma therapists. It is possible that other samples of trauma therapists may have more or less trauma experiences than those in the sample for this study. An association between personal trauma history and compassion fatigue was not found in this study. This relationship ha s been found in many previous research studies (Kassam-Adams, 1995; Pearlman & MacIan, 199 5), although not in a ll previous studies (Schauben & Frazier, 1995). The present study was among the first to util ize Figleys (1995) model of compassion fatigue as a conceptual framework. The resear cher in the current st udy found aspects of this model to be significantly rela ted to compassion fatigue, incl uding prolonged exposure to secondary trauma (i.e., amount of secondary expo sure and years of clin ical experience) and disengagement by measuring overinvolvement, defi ned as its opposite. Future researchers also should examine other aspects of Figleys (1995) model, including empathy, sense of achievement, traumatic recollections, and amount of life stress. A struct ural equation modeling method could be used to specify a model of comp assion fatigue and to dete rmine the accuracy of the proposed directions of the constr ucts as articulate by Figley. Limitations Several lim itations will be discussed in th is section that may have compromised the validity and generalizability of the findings of this study. First, re liabilities of the subscales of the Psychologists Burnout Inventory (PBI) were somewhat low. According to Nunnally (1978), an alpha level of .70 is the accepted standard for instruments in basic rese arch. Reliabilities for
102 the PBI in the present study are as follows: the Control subscale had an internal co rrelation of .68, Workplace Support was .55, and Overinvolveme nt was .64. The Types of Negative Clientele subscale had a sufficiently high intern al consistency. Thus, findings related to overinvolvement, workplace support, and control va riables may have been compromised because the measures were not sufficiently reliable. These measurement problems may explain why relationships were not found between workplace s upport and overinvolvement and either of the dependent variables: compassion fatigue nor compassion satis faction, even though previous researchers detected significant relationships among these variable s. A significant relationship was found between both compassion fatigue and co mpassion satisfaction and perceived control in the workplace. However, the control subscale had marginally low internal consistency (i.e., closer to the accepted standard of .70), compared to lower internal consiste ncies of the other two subscales. Therefore, this more marginal level of internal consistency lends greater support to the validity of findings related to therapists perceived workplace control, as compared to findings related to overinvolve ment and workplace support. A second limitation of this study relates to the generalizability of the sample to the target population. The population to whic h findings were intended to generalize consisted of all therapists working primarily with traumatized clie nts in a clinical setti ng. The sample obtained for this study consisted of ther apists (e.g., primarily psychologists, professional counselors, and social workers) who were members of profe ssional organizations aimed at keeping members abreast of research, training, and professional credentialing in the field of trauma. It is possible that the sample obtained for this study differs from the target population in terms of their professional training, interest in the field, and specific professi onal activities which could have accounted for different responses than a different sa mple of trauma therapists. Participants in
103 this sample may have had lower levels of comp assion fatigue as compared to other studies (e.g., Conrad & Kellar-Guenther, 2006) due to their high motivation and interest in the field, and therefore may be different from the entire populati on of trauma therapists. Thus, generalizability of the present findings to all trauma therapists may be somewhat limited. Third, limitations exist whenever using self-repo rt measures. It is impossible to know the level of accuracy with which par ticipants responded to items on th e survey. Social desirability biases may have occurred in the participants res ponses. Also, due to the use of internet and email, it is not possible to know for sure if the person who responded to the survey was the person intended from the e-mail list. A fourth limitation involved th e use of a survey methodology and correlational design. Correlational methods of inquiry allow conclusi ons to be drawn about associations between variables but do not permit conclusions to be drawn about whether changes in one variable cause changes in another. Though the use of multiple re gression analyses allows a researcher to draw a stronger conclusion implying causality, causation s till cannot be determined because variables are naturally occurri ng and not experimenta lly manipulated. A fifth and final limitation of the present study was the response rate, which implies a potential response bias. A respons e bias refers to the inability to determine whether those who participated in the survey were systematically different in unknown vari ables or study variables from those who chose not to participate. In the present study, a re sponse rate of 17% was calculated from the first sample and 11.62% from th e second sample. It is possible that those who chose not to respond to the survey had highe r levels of compassion fatigue, which may have impacted the findings in this study. Those with high levels of compa ssion fatigue may have chosen to avoid sharing about their difficulties or were perhaps fearful of being psychologically
104 triggered by the survey items. Also, thirteen members who partially completed the survey were not included in the data analyses It is impossible to know if those who did not respond to the survey, or those who partially co mpleted the survey, differed systematically from those who did in any way. Summary In summ ary, this chapter provided a discus sion of the results, recommendations, and limitations of the present study. The current st udy examined the following relationships: (a) workplace variables and compassion fatigue, (b) workplace variables and compassion satisfaction, (c) individual and workplace variables and compassion fatigue, (d) individual and workplace variables and compassion satisfaction. F actors found to be significantly related to compassion fatigue and compassion satisfaction were discussed as well as possible explanations for study variables that were not related. Based on conclusions drawn from this studys findings, future directions for research as well as theore tical and practical impli cations were discussed. Finally, limitations of the presen t study were detailed in the conclusion of this chapter.
105 APPENDIX A PROFESSIONAL QUALITY OF LIFE SCALE Compassion Satisfaction and Fatigue SubscalesRevision IV [Helping ] people puts you in direct c ontact with their liv es. As you probably have experienced, your compassion for those you [help] has both positive and negative aspects. We would like to ask you questions about your experiences, both positive and negative, as a [helper] Consider each of the following questions about you and your current situation. Select the number that honestly reflects how frequently you expe rienced these charac teristics in the last 30 days 0=Never 1=Rarely 2=A Few Times 3=Somewhat Often 4=Often 5=Very Often 1. I am happy. 2. I am preoccupied with more than one person I [help] 3. I get satisfaction from being able to [help] people. 4. I feel connected to others. 5. I jump or am startled by unexpected sounds. 6. I feel invigorated af ter working with those I [help] 7. I find it difficult to separate my personal life from my life as a [helper] 8. I am losing sleep over traumatic experiences of a person I [help] 9. I think that I might ha ve been infected by the tr aumatic stress of those I [help] 10. I feel trapped by my work as a [helper] 11. Because of my [helping] I have felt on edge about various things. 12. I like my work as a [helper] 13. I feel depressed as a result of my work as a [helper] 14. I feel as though I am experienci ng the trauma of someone I have [helped] 15. I have beliefs that sustain me. 16. I am pleased with how I am able to keep up with [helping] techniques and protocols.
106 17. I am the person I always wanted to be. 18. My work makes me feel satisfied. 19. Because of my work as a [helper] I feel exhausted. 20. I have happy thoughts and feelings about those I [help] and how I could help them. 21. I feel overwhelmed by the amount of work or the size of my case [work] load I have to deal with. 22. I believe I can make a difference through my work. 23. I avoid certain activities or situations because they remind me of frightening experiences of the people I [help] 24. I am proud of what I can do to [help] 25. As a result of my [helping] I have intrusive, fr ightening thoughts. 26. I feel bogged down by the system. 27. I have thoughts that I am a success as a [helper] 28. I can't recall important parts of my work with trauma victims. 29. I am a very sensitive person. 30. I am happy that I chose to do this work. Copyright Information B. Hudnall Stamm, 1997-2005. Professional Quality of Life: Compassion Satisfaction and Fatigue Subscales, R-IV (ProQOL) http://www.isu.edu/~bhstamm. This test m ay be freely copied as long as (a) author is credited, (b) no changes are made other than those authorized below, and (c) it is not sold. You may s ubstitute the appropria te target group for [helper] if that is not the best term. For example, if you are worki ng with teachers, replace [helper] with teacher. Word changes may be made to any word in italic ized square brackets to make the measure read more smoothly for a particular target group.
107 APPENDIX B PSYCHOLOGISTS BURNOUT INVENTORY 0 1 2 3 4 5 6 Never Always 1. I work with clients who make su icidal statements or gestures. 2. I receive constructive feedback fr om coworkers or supervisors. 3. I take work home. 4. I find myself feeling responsible for my clients well-being. 5. I work with clients who frequently are late or miss appointments. 6. I have the opportunity to use my own initiative at work. 7. I work with clients who are self-critical and/or unsure of their identities. 8. I have control over what I do and when I do it during the work day. 9. I work with clients who defe nsively withdraw and withhold. 10. I can confer with someone about a problem with a case. 11. I am meeting my own work expectations. 12. I work with clients who have compulsive behaviors. 13. I share work responsibilities with my coworkers. 14. I work with clients who make psychopathic statements. 15. I feel that at times Im working ha rder for change than the client. Ackerley, G.D., Burnell, J., Holder, D.C., Kurdek, L.A. (1988). Burnout among licensed psychologists. Professional Psychology: Research and Practice, 19 (6), 624-631.
108 APPENDIX C STRESSFUL LIFE EXPERIENCESSHORT FORM We are interested in learning about your experiences. Below is a list of experiences that some people have found stressful. Please fill in the num ber that best represents how much of the following statements describe your experiences. If you are not sure of your answer, just give us your best guess. Describes your Experience (Use in Describe Experiences Column): 0 1 2 3 4 5 6 7 8 9 10 ______________________________________________________________________ I did not A little like my somewhat like exactly like my experience experiences my experiences experiences Describe Experience Life Experience I have witnessed or experienced a natural disaster; like a hurricane or earthquake. I have witnessed or experienced a hu man made disaster like a plane crash or industrial disaster. I have witnessed or experience d a serious accident or injury. I have witnessed or experienced chemical or radiation exposure happening to me, a close friend or a family member. I have witnessed or experienced a life threatening illness happening to me, a close friend or a family member. I have witnessed or experienced the death of my spouse or child. I have witnessed or experienced the death of a close friend or family member (other than my spouse or child). I or a close friend or family member has been kidnapped or taken hostage. I or a close friend or family member has been the victim of a terrorist attack or torture. I have been involved in combat or a war or lived in a war affected area. I have seen or handled dead bodi es other than at a funeral. I have felt responsible for the serious injury or death of another person. I have witnessed or been attacked with a weapon other than in combat or family setting. As a child/teen I was hit, spanked, choked or pushed hard enough to cause injury. As an adult, I was hit, choked or pushed hard enough to cause injury. As an adult or child, I have witne ssed someone else being choked, hit, spanked, or pushed hard enough to cause injury. As a child/teen I was forced to have unwanted sexual contact. As an adult I was forced to have unwanted sexual contact. As a child or adult I have witnessed someone else being forced to have unwanted sexual contact
109 I have witnessed or experienced an extremely stressful event not already mentioned. Please Explain: B. Hudnall Stamm Traumatic Stress Research Group, 1996, 1997, http://www.isu.edu/~bhstamm/index.htm. This form may be freely copi ed as long as (a) authors are credited, (b) no changes are made, & (c) it is not sold.
110 APPENDIX D DEMOGRAPHIC QUESTIONNAIRE What is your gender? a) Male b) Fe male What is your ethnicity? a) Black/African American b) White/Caucasian c) Latino/Hispanic d) Asian/Pacific Islander e) Other ____________ Which of the following professions best describes your profession? a) Professional Counselor b) Psychologist c) Social Worker d) Other _____________ Which of the following most accurately describes your PRIMARY professional activity ? a) Counseling/therapy and/or other clinical activities b) Case Management c) Supervisory/Administrative d) Teaching and/or Research activities e) Other ____________
111 Which of the following most accurately desc ribes your clinical work environment? a) I am the only professional in my work environment b) I work with other professionals Do you consider your c linical work to be PRIMARILY with traumatized clients? a) Yes b) No How many hours per week are spent working direct ly with clients who have been traumatized? ______ How many years have you been doing clinical work including internship and practicum experiences? ______ Have you had any specialized training in wo rking with traumatized populations? a) Yes b) No
112 APPENDIX E INFORMED CONSENT Protocol Title: T rauma Therapists Quality of Life: Individual and Workplace Factors associated with Compassion Fatig ue and Compassion Satisfaction Purpose of the research study: The purpose of this study is to examine individual and workplace factors that may be associated with compassion fatigue and compassion satisfac tion in mental health professionals working with trauma survivors. What you will be asked to do in this study: You will be asked to complete a demographics ques tionnaire and complete three short surveys. Time requirement: The entire survey should take about 15 to 20 minutes. Risks and Benefits: A potential risk of participating in this study in cludes emotional reactiv ity due to the sensitive nature of some of the questions. Because the topic of the study involve s reactions to trauma work, you may find yourself experien cing emotions associated with trauma, particularly if you have a history of traumatic experiences. If you feel that you are experiencing any distress, you may choose to stop completing the survey. If you feel concerned about em otional reactivity due to recent traumatic events you have experienced you may choose not to co mplete the survey at this time. Potential benefits include increasi ng self-awareness about your feelings and thoughts about your work. Compensation: There is no compensation for participating in this study. Confidentiality: Your identity will be kept confidential to the ex tent provided by law. Your information will be assigned a code number. Your name will not be used in any report. Voluntary participation: Your participation in this study is completely voluntary. There is no penalty for not participating.
113 Right to withdraw from the study: You have the right to withdraw from th is study at anytime without consequence. Whom to contact if you have questions about the study: Lindsay G. Leonard, Ed.S., Doctoral Student Department of Counselor Education, (352) 392-0731 Sondra Smith, PhD, Department of Couns elor Education, (352) 392-0731 ext. 239 Whom to contact about your rights as a research participant in this study: UFIRB Office, Box 112250, University of Florida, Gainesville, Fl 32611-2250; (352) 392-0433 Agreement: I have read the above and by clicking belo w I give my consent to participate in the study.
114 LIST OF REFERENCES Abraido-Lanza, A.F., Guier, C., & Colon, R.M. (1998). Psychologi cal thriving am ong Latinas with chronic illness. Journal of Social Issues, 54 (2), 405-424. Ackerley, G.D., Burnell, J., Holder, D.C., Kurdek, L.A. (1988). Burnout among licensed psychologists. Professional Psychology: Research and Practice, 19 (6), 624-631. Arnold, D., Calhoun, L.G., Tedeschi, R., & Ca nn, A. (2005). Vicarious posttraumatic growth in psychotherapy. Journal of Humanistic Psychology, 45 (2), 239-263. Baird, S., & Jenkins, S. (2003). Vicarious trau matization, secondary traumatic stress, and burnout in sexual assault and domestic violence agency staff. Violence and Victims, 18 (1), 71-86. Beaton, R.D., & Murphy, S.A. (1995). Work ing with people in crisis: Research implications. In C.R. Figley (Ed.), Compassion fatigue: Coping with secondary traumatic stress disorder in those who treat the traumatized, pp.51-81. New York: Brunnel/Mazel. Bell, H. (1998). The impact of counseling battered wo men on the mental health of counselors. Unpublished raw data. Re trieved on May 12, 2006 from http://www.arte-sana.com/artic les/social_worker_burnout.htm Bell, H., Kulkarni, S., & Dalton, L. (2003). Organizational prevention of vicarious traum a. Families in Society, 84 (4), 463-470. Bonanno, G.A. (2004). Loss, trauma, and human resilience: Have we underestimated the human capacity to thrive afte r extremely aversive events? American Psychologist, 59 (1), 20-28. Brady, J.L., Guy, J.D., Poelstra, P.L., & Broka w, B.F. (1999). Vicarious traumatization, spirituality, and the treatment of sexual abuse survivors: A national survey of women psychotherapists. Professional Psychology: Research and Practice, 30 (4), 386393. Campbell, D.T. & Fiske, D.W. (1959). C onvergent and discriminant validation by the multi-traitmultimethod matrix. Psychological Bulletin, 56 81-105. Cerney, M.S. (1995). Treating the heroic treaters. In C.R. Figley (Ed.), Compassion fatigue: Coping with secondary traumatic st ress disorder in those who treat the traumatized, pp.131-149. New York: Brunnel/Mazel. Chrestman, K.R. (1999). Secondary exposure to trauma and self-reported distress among therapists. In B.H. Stamm (Ed.), Secondary traumatic stress: Self-care issues for clinicians, research ers and educations. Luterville, MD: Sidram Press.
115 Clark, M.I., & Gioro, S. ( 1998). Nurses, indirect trauma and prevention. Image: Journal of Nursing Scholarship, 30, 85-87. Cohen, J., & Cohen, P. (1983). Applied MRC analysis fo r the behavioral sciences (2nd ed.) Hillsdale, NJ: Erlbaum. Collins, S., & Long, A. (2003). Working with the psychological effects of trauma: Consequences for mental health-care workers-A literature review. Journal of Psychiatric and Mental Health Nursing, 10 417-424. Conrad, D., & Kellar-Guenther, Y. (2006). Compassion fatigue, burnout, and compassion satisfaction among Colo rado child protection workers. Child Abuse and Neglect, 30 (10), 1071-1080. The Commonwealth Fund (1999). Health Concerns Across a Womans Lifespan: 1998 Survey of Womens Health Corey, M.S., & Corey, G. (1989). Becoming a helper. Pacific Grove, CA: Brooks/Cole. Creamer, T.L., & Liddle, B.J. (2005). Seconda ry traumatic stress among disaster mental health workers responding to the September 11 attacks. Journal of Traumatic Stress, 18 (1), 89-96. Cunningham, M. (2003). Impact of trauma wo rk on social work clinicians: Empirical findings. Social Work, 48 (4), 451-459. Cushway, D., & Tyler, P. (1996). Stress in clinical psychologists. International Journal of Social Psychiatry, 42, 141-149. Duan, C., & Hill, C.E. (1996). The cu rrent state of empathy research. Journal of Counseling Psychology, 43 (3), 261-274. Dutton, M., & Rubinstein, F.L. (1995). Work ing with people with PTSD: Research implications. In C.R. Figley (Ed.), Compassion fatigue: Coping with secondary traumatic stress disorder in those who treat the traumatized, pp.82-100. New York: Brunnel/Mazel. Edwards, D., Burnard, P., Coyle, D., Fothergi ll, A., & Hannigan, B. (2000). Stress and burnout in community mental health nur sing: A review of the literature. Journal of Psychiatric and Mental Health Nursing, 7 7-14. Etzion, D. (1984). Moderating effect of social support on the stress-bur nout relationship. Journal of Applied Psychology, 69 (4), 615-622. Figley, C.R. (1989). Helping traumatized families San Francisco: Jossey-Bass.
116 Figley, C.R. (1995). Compassion fatigue: Coping with secondary traumatic stress disorder in those who treat the traumatized. New York: Brunnel/Mazel. Freudenberger, H.J. (1974). Staff burnout Journal of Social Issues, 30 (1), 159-165. Gladstein, G.A. (1983). Understanding empathy: Integrati ng counseling, developm ental, and social psychology perspectives. Journal of Counseling Psychology, 30 (4), 467-482. Hellman, I.D., & Morrison, T.L. (1987). Practice setting and type of caseload as factors in psychotherapist stress. Psychotherapy: Theory, Research, Practice, Training, 24 (3), 427-433. Herman, J.L. (1992). Trauma and recovery New York: Basic Books. Jenkins, S.R., & Baird, S. (2002). Secondary traumatic stress and vicarious trauma: A validational study. Journal of Traumatic Stress, 15 (5), 423-432. Kahill, S. (1988). Interventions for burnout in the helping professional: A review of the empirical evidence. Canadian Journal of Counselling Review, 22, 310-342. Kassam-Adams, N. (1995). The risk of treatin g trauma: Stress and secondary trauma in psychotherapists. In B. Stamm (Ed.) Secondary traumatic stress: Self-care issues for clinicians, researchers, and educators, pp. 37-48. Lutherville, MD: Sidran Press. King, M., King, D., Fairbanks, J., & Adams, G. (1998). Resiliencerecovery factors in post-traumatic stress disorder among female and male veterans: hardiness, post war social support and additional stressful life events. Journal of Personality and Social Psychology, 74, 420-434. Kobasa, S.C., Maddi, S.R., & Kahn, S. (1982). Hardiness and health: A prospective study. Journal of Personality and Social Psychology, 42 168-177. Landry, L.P. (2001). Secondary traumatic stre ss disorder in the th erapists from the Oklahoma City bombing. Dissertation Abstracts International, 61 3849. Lee, C.S. (1995). Secondary traumatic stress in therapi sts who are exposed to client traumatic material Unpublished dissertation, Florid a State University, Tallahassee. Maslach, C. (2001). The truth about burnout. San Francisco: Jossey-Bass. Maslach, C., & Jackson, S.E. (1986). Maslach Burnout Inventory manual (2nd ed.). Palo Alto, CA: Consulting Psychologists Press.
117 Maslach, C., & Leiter, M.P. (1997). The truth about burnout San Francisco: JosseyBass. McCann, L., & Pearlman, L.A. (1990). Vicar ious traumatization: A framework for understanding the psychological eff ects of working with victims. Journal of Traumatic Stress, 3 (1), 131-149. McMillen J.C., & Fisher, R.H. (1998). The Perceived Benefit Scales: Measuring perceived positive life changes after negative events. Social Work Research, 22 (3), 173-186. Meldrum, L., King, R., & Spooner, D. (2002). Compassion fatigue in community mental health case managers. In C.R. Figley (Ed.) Treating compassion fatigue, (pp.85-106). New York: Brunner/Routledge. Meyers, T.W., & Cornille, T.A. (2002). The trauma of working with traumatized children. In C.R. Figley (Ed.), Treating compassion fatigue, pp.39-55. New York: Brunner-Routledge. Munroe, J.F., Shay, J., Fisher, L., Makary, C., Rapperport, K., & Zimering, R. (1995). Preventing compassion fatigue: A team treat ment model. In C.R. Figley (Ed.), Compassion fatigue: Coping with secondary trauma tic stress disorder in those who treat the traumatized, pp.209-231. New York: Brunnel/Mazel. Neuman, D.A., & Gamble, S.J. (1995). Issu es in the professiona l development of psychotherapists: countertr ansference and vicarious tr aumatization in the new trauma therapist. Psychotherapy, 32 341-347. Nunnally, J.C. (1978). Psychometric theory (2nd ed.). New York: McGraw-Hill. Park, C.L., Cohen, L.H., & Murch, R.L. (1996) Assessment and prediction of stressrelated growth. Journal of Personality, 64 (1), 71-105. Pearlman, L.A., & Maclan, P.S. (1995). Vicario us traumatization: An empirical study of the effects of trauma work on the trauma therapist. Professional Psychology: Research and Practice, 26 (6), 558-565. Pearlman, L.A., & Saakvitne, K.W. (1995). Trauma and the therapist: Countertransference and vicarious traumatization in psychothe rapy with incest survivors. New York: Norton. Pines, A.M. (1993). Burnout. In I. Goldberger & S. Brezawitz (Eds.), Handbook of stress: Theoretical and clinical aspects (2nd ed.). New York: Free Press. Pines, A.M., & Aronson, E. (1988). Career burnout: Causes and cures. New York: Free Press.
118 Raquepaw, J.M., & Miller, R.S. (1989). Psychotherapist burnout: A componential analysis Professional psychology: Research and practice, 20 (1), 32-36. Rogers, C.R. (1959). A theory of therapy, pers onality and interpersona l relationships as developed in the client-centered framework. In S. Koch (Ed.), Psychology: A study of a science. Study 1. Conceptual and systematic: Vol. 3. Formulations of the person and the social context (pp.184-256). New York: McGraw-Hill. Rothschild, B. (2000). The body remembers: The psychophysiology of trauma and trauma treatment. New York: Norton. Rupert, P.A., & Morgan, D.J. (2005). Wo rk setting and burnout among professional psychologists. Professional Psychology: Research and Practice, 36 (5), 544-550. Saakvitne, K.W. (1995). Therapists responses to dissociative clients: Countertransference and vicarious traumatizatio n. In L.M. Cohen, J.N. Berzoff, M.R. Elin, & M.D Lanham, Dissociative identity disord er: Theoretical and treatment controversies, pp.467-492. Lanham, MD: Jason Aronson. Saakvitne, K.W., Tennen, H., & Affleck, G. (1998). Exploring thriving in the context of clinical trauma theory: Constructivist self development theory. Journal of Social Issues, 54 (2), 279-299. Salston, M., & Figley, C.R. (2003). Seconda ry traumatic stress of working with survivors of criminal victimization Journal of Traumatic Stress, 16 (2), 167-174. Schauben, L.J., & Frazier, P.A. (1995). The effects on female counselors of working with sexual violence survivors. Psychology of Women Quarterly, 19 49-54. Skovholt, T.M. (2001). The resilient practitioner: Burnout prevention and self-care strategies for counselors, therapist s, teachers and health professionals. Needham Heights, MA: Allyn & Bacon. Sexton, L. (1999). Vicarious traumatisation of counselors and effects on their workplaces. British Journal of Guidance and Counselling, 27 (3), 393-403. Soderfeldt, M., Soderfeldt, B., & Warg, L. E. (1995). Burnout in social work. Social Work, 40, 638-646. Stamm, B. (2002). Measuring compassion satisfaction as well as fatigue: Developmental history of compassion fatigue and sati sfaction test. In C.R. Figley (Ed.), Treating Compassion Fatigue pp.1-20. London: Taylor & Francis. Stamm, B.H. (2005). The Professional Quality of Life Scale (ProQOL) Manual. Idaho State University: Sidran Press.
119 Stamm, B.H. (1997). Stressful Life Experiences Screening Short Form Idaho State University: Sidran Press. Stamm, B., & Pearce, F.W. (1995). Crea ting virtual community: Telemedicine Applications for self-care. In B. Stamm (Ed.) Secondary traumatic stress: Self-care issues for clinicians, researchers, and educators, pp.230-246. Lutherville, MD: Sidran Press. Stamm, B.H., Rudolph, J.M., Dewane, S., Gain es, N., Gorton, K., Paul, G., McNeil, F., Bowen, G., & Ercolano, M. (1996). Ps ychometric Review of Stressful Life Experiences Screening. In B.H. Stamm (Ed.) Measurement of stress, trauma, and adaptation. Lutherville, MD: Sidran Press. Tedeschi, R.G., & Calhoun, L.G. (1996). Th e Posttraumatic Growth Inventory: Measuring the positive legacy of trauma. Journal of Traumatic Stress, 9 (3), 455472. Trippany, R.L., White Kress, V.E., & Wilc oxon, S.A. (2004). Preventing vicarious trauma: What counselors should know when working with trauma survivors. Journal of Counseling and Development, 82 (1), 31-37. Vredenburgh, L.D., Carlozzi, A.F., & Stei n, L.B. (1999). Burnout in counseling psychologists: Type of practice se tting and pertinent demographics. Counseling Psychology Quarterly, 12 (3), 293-302. Wee, D.F., & Myers, D. (2002). Stress respon ses of mental health workers following disaster: The Oklahoma City bombing. In C.R. Figley (Ed.), Treating compassion fatigu e, pp.57-83. New York: Brunner-Routledge. Wilson, J.P. & Lindy, J.D. (1994). Countertran sference in the treatment of PTSD. New York: Guilford Press. Young, M. (2001). Community crisis response team training manual (3rd ed). Washington, D.C.: National Organiza tion of Victim Assistance.
120 BIOGRAPHICAL SKETCH Lindsay Gwyn Leonard was born in 1981 in Ma dison, W isconsin. Lindsay is the middle of three children, growing up mostly in Jacksonv ille and Orlando, Florida. She graduated with honors from Dr. Phillips High School in 1999. Li ndsay earned a B.S. in Psychology from the University of Florida in 2003. She then went on to earn an M.Ed. and Ed.S. in Marriage and Family Therapy, also from the University of Florida, Department of Counselor Education. During her graduate training in Counselor E ducation, Lindsay worked in a variety of counseling settings. She first worked at the Alachua County Crisis Ce nter serving the local community. Lindsay also worked as an intern at the University of Flor ida Counseling Center in 2004-2005 and with PACE Center for Girls in 2 006-2007. During this clinical training, Lindsay pursued a doctoral degree in Couns elor Education at the University of Florida. She decided to specialize in Crisis Intervention and Post-Trauma Counseling, which reflected the majority of her clinical work with the Crisis Center throughout her professional studies. Beginning in September of 2007, Lindsay joined the clinical team at the University of North Floridas Counseling Center in Jacksonville, Florida. Curre ntly, Lindsay is working as a therapist for college students a nd as a coordinator for group ther apy. After completing the Ph.D. program, Lindsay plans to pursue a teaching position at the Univer sity in addition to her ongoing clinical work. Lindsay curren tly lives with her partner, Sh ane McKim and their dog Charley. She enjoys painting, kickboxing, running, reading, writing, and spending tim e with her family.