Abstract
Fibromyalgia (FM) is a lifelong central nervous system disorder that is precipitated by a range of biological, psycho-cognitive, and social factors. The aims of this exploratory study were to (a) identify biopsychosocial and cognitive factors that may affect an individual’s response to FM, (b) determine whether individuals with FM can be grouped into homogeneous subgroups based on biopsychosocial factors associated with response to FM, and (c) compare subgroup differences in health outcomes and life satisfaction. This study included 302 participants with FM. Principal components analysis yielded three sets of biopsychosocial factors that may affect response to FM (i.e., protective, cognitive-affective, and physical factors). Based on these three factors, a cluster analysis was performed, which produced three homogeneous subgroups: (a) the moderate amount of problems group, (b) the least amount of problems group, and (c) the many problems group. Analysis of variance (ANOVA) results indicated that these three subgroups differed significantly in terms of health outcomes and life satisfaction. The findings of this study broaden the existing literature related to understanding FM from a multidimensional symptom response perspective and contribute to the development and validation of biopsychosocial interventions for people with FM.
Keywords
Fibromyalgia (FM) is a lifelong central nervous system (CNS) disorder of undetermined cause, characterized by chronic musculoskeletal pain (Smedema et al., 2016). Clinical symptoms associated with FM include widespread pain, stress, fatigue, sleep disturbances, depression, and cognitive impairments (Smedema et al., 2016). It is also associated with several secondary symptoms and health conditions including dizziness, neurological disturbances, headaches, cognitive impairments, irritable bowel syndrome, Raynaud’s syndrome, restless leg syndrome, and temporomandibular joint disorder (Bennett et al., 2007; Smedema et al., 2016). These symptoms significantly affect physical, mental, and social functioning and interfere with community participation (e.g., employment, social activity, and leisure time activity), resulting in reductions in quality of life and life satisfaction (Henriksson & Liedberg, 2000; Liedberg & Henriksson, 2002). A Humana research study of its members enrolled in the commercial and Medicare Advantage plans revealed that patients with FM have greater health care utilization rates and higher health care costs than those without FM (Palacio et al., 2010). Empirical literature also indicates that persons with FM have higher rates of absenteeism, work injury, and mental health issues in the workplace than do individuals without FM (Penrod et al., 2004; Robinson et al., 2003).
The American College of Rheumatology (ACR) has developed classification criteria for physicians to diagnose FM (Wolfe et al., 2010). The criteria require tenderness on pressure (tender points) in at least 11 of 18 specified sites and the presence of widespread pain for diagnosis. Widespread pain is defined as axial pain, left- and right-sided pain, and upper and lower segment pain. Although the physical symptoms of FM can be profound, these symptoms represent only one component of the FM experience. The disability experience associated with FM is the result of complex interactions among a number of biopsychosocial factors, including health condition factors, personal–psychological factors, and socio-environmental factors (Abeles et al., 2007; Loevinger et al., 2012; Smedema et al., 2016). Medical diagnosis alone is not useful for determining the impact of FM on functioning and participation levels of persons with FM living in the community, and there are ongoing debates regarding FM classification, causation, pathophysiology, management, and recovery outcomes (Garg & Deodhar, 2012; Koch & Rumrill, 2017; Masi et al., 2002). Not surprisingly, there is currently no well-validated FM classification system that can be utilized by rehabilitation and health professionals to assess person–environment contextual factors that interact with physical factors to affect response to FM. Developing an integrative model to classify clients’ responses to FM from a biopsychosocial perspective will be valuable for case conceptualization, assessment, rehabilitation planning, and selection of evidence-based psychosocial and vocational interventions for clients with FM living in the community. However, there is still a gap in research on the physical, psychological, and environmental factors aimed at improving diagnosis, developing treatment strategies, and hence helping individuals with FM attain better health and rehabilitation outcomes. Research should focus on investigating physical factors, cognitive/psychological features, and environmental contexts and how they interact to influence FM (Koch & Rumrill, 2017).
The Biopsychosocial Model
Under the biopsychosocial model, a disease is viewed as an objective event of a biological nature, and illness is viewed as a subjective acknowledgment that a disease is present (Gatchel et al., 2007). Within this framework, functioning and disability have a complex interaction among the individual’s biological health condition (e.g., pain), the psychological/cognitive aspects of the illness (e.g., anxiety), and the contextual factors of the environment (e.g., social stigma). A comprehensive taxonomy of FM symptoms and understanding of how individuals affected by FM respond to this syndrome are fundamental to implementing effective treatment interventions. FM is best conceptualized as a multilevel syndrome that encompasses the interaction of physical, individual, and environmental features (Suls & Rothman, 2004; Turk & Flor, 1999). This approach can help organize the vast diversity and complexity of multilevel mechanisms affecting individuals and heterogeneous subgroups of people with FM. Validating a classification system based on factors that may affect an individual’s response to FM from this proposed framework assumes that the functionality of the individual is affected by some form of physiopathology at the biological or physical level (e.g., affecting muscles and nerves), as well as at the cognitive/psychological level (e.g., stress and cognition). Individual characteristics (e.g., coping and resilience) influence responses to FM either by ameliorating or exacerbating symptomology. In addition, environmental factors in part affect responses in individuals with FM (e.g., stigma and social support). The proposed levels of responses are the following: (a) functioning and impairment, (b) individual characteristics, and (c) environment characteristics.
Functioning and Impairment Characteristics
Functioning is intrinsically related to the body’s systems and structures and to the physical functioning of these structures, along with activities and participation in society (Chan et al., 2019). In FM, the core features of pain and tenderness are almost always accompanied by additional symptomology such as sleep problems, fatigue, and morning stiffness (Abeles et al., 2007). In addition to physiological aspects, anxiety and depression are the most frequent psychiatric complaints in FM and are associated with feelings of fatigue, concentration problems, restlessness, and muscle tension (Koch & Rumrill, 2017). These symptoms are stressors that can negatively affect the individual’s mental and physical well-being. Cognitive difficulties (e.g., memory and attention) are among the chief complaints reported by individuals with FM (Bennett et al., 2007). All of these symptoms encompass the biological component.
Individual Characteristics
Individual factors are internal traits unique to the individual, such as character strengths in people with FM (Muller et al., 2017). These positive human traits influence the perception of health and illness, impact symptom responses, and have shown significant results regarding their relationships to FM. Factors such as resilience, hope, optimism, coping styles, and self-efficacy can influence symptom responses as well as treatment engagement and management (Muller et al., 2017).
Environment Characteristics
Environmental factors relate to features in the physical, social, or attitudinal world, ranging from the immediate to a more general environment, that either facilitate or hinder functioning. Because illness is conceptualized as the individual’s experience within the context of the environment, intrinsic factors in the immediate or general physical, social, or attitudinal milieu must be considered. For example, social support is a multidimensional construct that plays an essential role in an individual’s treatment adherence and engagement, leading to health and well-being (Chronister et al., 2006). The biopsychosocial model posits the influence of the interaction between the individual and the environment on life experiences, functioning, and outcome (Dunn & Elliott, 2008).
Outcome Variables
Turk et al. (1998) acknowledged the need for a better classification system for FM. They recommended the use of FM-related biopsychosocial variables to identify homogeneous subgroups with different levels of health, well-being, and employment outcomes. The same concept has been applied to other disorders in an attempt to understand the relationships between clinical features and outcome variables (Ryan et al., 2007; Thong et al., 2008). The comparison of homogeneous subgroups aids in determining the effectiveness of responses to FM on health and rehabilitation outcomes and informs evidence-based practice (EBP). Therefore, it is important to explore the relationship between homogeneous subgroups and outcome variables such as functional disability, community participation, health-related quality of life (HRQOL), and life satisfaction. FM can affect both work disability and functional disability, including performing activities of daily living and participating in social events (Liedberg & Henriksson, 2002). Individuals with FM also often experience a variety of affective conditions such as depression (Arnold et al., 2007; Belt et al., 2009). Therefore, life satisfaction and HRQOL can be compromised for people with FM because multiple symptoms can affect so many areas of the individual’s life (Boonstra et al., 2012).
Purpose of the Present Study
The primary purpose of this study was to validate a classification of individuals’ responses to FM through clinical features within a biopsychosocial approach and to determine how those clinical features relate to health outcomes. The aim of the study was threefold: (a) to identify the number of dimensions underlying the construct response to FM, (b) to determine whether individuals with FM can be classified into a manageable number of homogeneous subgroups based on biopsychosocial factors that may affect their response to FM, and (c) to compare subgroup differences in terms of psychological and health outcomes. This study was guided by the following research questions.
Method
Procedures and Participants
The Institutional Review Board (IRB) at a Midwestern university approved this study, and the National Fibromyalgia and Chronic Pain Association (NFMCPA) agreed to assist with recruitment of volunteer research participants. To be eligible for this study, participants had to meet the following criteria: (a) diagnosed with FM for at least 1 year prior to survey completion, (b) between 18 and 65 years of age, and (c) have a self-reported sixth-grade reading level or above. Participants were excluded from this study if they were diagnosed with traumatic brain injuries or substantial intellectual disabilities. Research information was disseminated in a monthly electronic newsletter by the NFMCPA, and potential participants were provided with an online survey platform. In addition, all instruments and procedures were pilot tested by five volunteers to determine necessary instrument revisions or procedural problems before data collection. The final sample included 302 individuals with self-reported FM from the NFMCPA’s client registry. Detailed information for these participants is presented in Table 1.
Participant Demographic and Fibromyalgia Characteristic (N = 302).
Note. SSI = supplemental security income; SSDI = social security disability insurance.
Measures
Eleven measures were used to assess biopsychosocial factors that may affect an individual’s response to FM, including self-efficacy, resilience, social support, coping, social stigma, pain intensity, fatigue, sleep problems, perceived stress, anxiety, and cognitive impairment. Four outcome variables (functional disability, participation, life satisfaction, and HRQOL) were included in this study.
Self-efficacy
Self-efficacy was measured by the 10-item General Self-Efficacy Scale (GSE; Schwarzer & Jerusalem, 1995; for example, “Thanks to my resourcefulness, I can handle unforeseen situations”). Each item is rated on a 4-point Likert-type scale from 1 (hardly true) to 4 (exactly true). The internal consistency reliability coefficients (Cronbach’s alpha) for the GSE range from .86 to .94 (Luszczynska et al., 2005). For this study, the Cronbach’s alpha was computed to be .92.
Resilience
Resilience was measured by the six-item Brief Resilience Scale (BRS; Smith et al., 2008; for example, “I tend to bounce back quickly after hard times”). Each item is measured on a 5-point Likert-type scale rated from 1 (strongly disagree) to 5 (strongly agree). The Cronbach’s alpha coefficients range from .80 to .91 (Smith et al., 2008). For this study, the Cronbach’s alpha was computed to be .88.
Social support
Social support was measured by the 12-item Multidimensional Scale of Perceived Social Support (MSPSS; “My family really tries to help me” [family subscale], “My friends really try to help me” [friends subscale], and “There is a special person who is around when I am in need” [significant other subscale]). Each item is rated on a 7-point Likert-type scale from 1 (very strongly disagree) to 7 (very strongly agree). The Cronbach’s alpha coefficients range from .84 to .92 (Zimet et al., 1988). For this study, the Cronbach’s alpha was computed to be .95.
Coping
Coping was measured with the Brief COPE (Carver, 1997) with a total of 28 items that measure 14 features of coping. For this study, five subscales representing adaptive coping styles were used and endorsed on a 4-point Likert-type scale ranging from 0 (I haven’t been doing this at all) to 3 (I’ve been doing this a lot) with higher scores representing higher use of adaptive coping strategies. Sample items of the five subscales are as follows: “I take action to try to make the situation better” (active coping); “I’ve been learning to live with it” (acceptance); “I concentrate my efforts on doing something about the situation I’m in” (planning); “ I’ve been getting help and advice from other people” (instrumental support); “ I’ve been trying to see it in a different light, to make it seem more positive” (positive reframing). The Cronbach’s alpha coefficients were the following: active coping .68, acceptance .57, planning .73, positive reframing .64, and using instrumental support .64 (Lin, 2009). For this study, the Cronbach’s alpha for the adaptive coping styles scale was computed to be .88.
Social stigma
Stigma was measured by Green’s (2001) eight-item adapted version of the Devaluation-Discrimination Scale (Link et al., 2001; for example, “Most people in my town feel nervous and/or awkward when they meet someone with a disability”). Each item was rated on a 5-point Likert-type scale from 1 (strongly disagree) to 5 (strongly agree). The Cronbach’s alpha coefficients range from .73 to .78 (Green, 2001). In this study, the items were reverse scored so that high scores reflect lower levels of perceived stigma. For this study, the Cronbach’s alpha was computed to be .79.
Pain
Pain intensity was measured by the Numeric Pain Scale (NPS; Farrar et al., 2001). This scale asks individuals to rate their pain experience on a scale ranging from 0 (no pain) to 10 (severe pain). The instrument asks participants to rate the amount of pain they had experienced during the two previous weeks. Test–retest reliability of the NPS was .95 (Alghadir et al., 2018; Williamson & Hoggart, 2005). We did not collect data to compute the test–retest reliability for this study.
Fatigue
Fatigue was measured by the nine-item Brief Fatigue Inventory (BFI The first three items asks individuals to rate the severity of their fatigue at its “worst,” “usual,” and “now” during normal waking hours, ranging from 0 (no fatigue) to 10 (fatigue as bad as you can imagine). The six interference items assess the degree to which fatigue has interfered with diverse aspects of the individual’s life during the past 24 hr, ranging from 0 (no fatigue) to 10 (fatigue as bad as you can imagine). The six interference items are also rated on an 11-point Likert-type scale ranging from 0 (does not interfere) to 10 (completely interferes). The Cronbach’s alpha coefficient was .96 (Mendoza et al., 1999). For this study, the Cronbach’s alpha was computed to be .91.
Sleep problems
Sleep difficulty was measured by the 12-item Medical Outcomes Study–Sleep instrument (Viala-Danten et al., 2008). The two first items ask for information about how long the individual took to fall sleep and how many hours the individual slept per night. The next 10 items ask questions pertinent to the dimensions of sleep and are rated from 1 (all of the time) to 6 (none of the time). The Cronbach’s alpha coefficients range from .71 to .81 (Viala-Danten et al., 2008). For this study, the Cronbach’s alpha was computed to be .68.
Perceived stress
Stress was measured by the four-item Perceived Stress Scale (PSS-4; Cohen et al., 1983; for example, “In the last month, how often have you felt confident about your ability to handle your personal problems?”). Each item is rated on a 5-point Likert-type scale ranging from 0 (never) to 4 (very often). The Cronbach’s alpha coefficient was .72 (Cohen et al., 1983) and .72 for this study.
Anxiety
Anxiety was measured by the seven-item General Anxiety Disorder (GAD-7; Löwe et al., 2008; for example, “Feeling nervous, anxious or on edge”). Each item is rated using a 4-point Likert-type scale ranging from 0 (not at all) to 3 (nearly every day). The Cronbach’s alpha coefficient was .90 (Löwe et al., 2008). For this study, the Cronbach’s alpha was computed to be .89.
Cognitive impairment
Cognitive impairment was measured by the 10-item Cognitive Failures Questionnaire (CFQ; Wallace et al., 2002; for example, “Do you find you forget whether you’ve turned off a light or a fire or locked the door?”). Each item is rated on a 5-point rating scale ranging from 0 (never) to 4 (very often). The Cronbach’s alpha coefficient was .91 (Wallace et al., 2002). For this study, the Cronbach’s alpha was computed to be .90.
Functional disability
Disability was measured by the 12-item version of the World Health Organization Disability Assessment Schedule II (WHODAS-II; for example, “Taking care of your household responsibilities?”), and each item is rated on a 5-point scale ranging from 1 (no difficulty) to 5 (extreme difficulty/cannot do). The Cronbach’s alpha coefficient was .89 (Luciano et al., 2010). For this study, the Cronbach’s alpha was computed to be .87.
Community participation
Community participation was measured by the World Health Organization Disability Assessment Schedule II (WHODAS-II; Garin et al., 2010). The WHODAS-2 contains 36 items on functioning and disability with a recall period of 30 days, and it is composed of seven domains. For this study, only eight items (e.g., “How much of a problem did you have in joining in community activities [e.g., festivities, religious or other activities] in the same way as anyone else can?”) from the participation in society subscale were used, and each item is rated on a 5-point Likert-type scale from 1 (no difficulty) to 5 (extreme difficulty or cannot do). The Cronbach’s alpha coefficient was .82 (Garin et al., 2010). For this study, the Cronbach’s alpha was computed to be .84.
HRQOL
HRQOL was measured by the 12-item version of the Medical Outcomes Survey (MOS) Short Form Health Survey (SF-12v2). Item 1 (general health) is rated on a 5-point Likert-type-type scale ranging from 1 (excellent) to 5 (poor). Items 2 and 3 (physical functioning) are rated on a 3-point Likert type scale ranging from 1 (yes, limited a lot) to 3 (no, not limited at all). Items 4 to 12 (e.g., “Limitations due to physical health problems”) are rated on a 5-point Likert-type scale from 1 (all of the time) to 5 (none of the time). All 12 items are used to calculate the Physical Component Summary (PCS-12) and the Mental Component Summary (MCS-12), and a total score, with low scores indicating poor health and high scores reflecting well-being. The Cronbach’s alpha coefficients range from .77 to .80 (Luo et al., 2003). For this study, the Cronbach’s alpha was computed to be .74.
Life satisfaction
Life satisfaction was measured by the five-item Satisfaction With Life Scale (SWLS; Diener et al., 1985; for example, “In most ways my life is close to my ideal”). Each item was rated on a 7-point Likert-type scale ranging from 1 (strongly disagree) to 7 (strongly agree). The Cronbach’s alpha coefficient was .87 (Pavot & Diener, 1993). For this study, the Cronbach’s alpha was computed to be .88.
Data Analysis
The Statistical Package for the Social Sciences (SPSS version 22.0) for Windows was used to perform all data analyses. Principal components analysis (PCA), a data reduction technique, was used to identify the optimal number of dimensions underlying the construct biopsychosocial responses to FM (Floyd & Widaman, 1995; Hair et al., 2006). For sample size estimate, Gorsuch (1983) stated there should be at least five participants per variable and that a sample size of at least 200 is preferred. Cluster analysis was used to partition the data into homogeneous subgroups to produce an empirical classification of biopsychosocial response to FM. There is no consensus on how to compute sample sizes for cluster analysis. However, cluster analysis is a data mining technique frequently utilized by business researchers for market segmentation analysis. Dolnicar et al. (2014) conducted a simulation study using artificial data of known structure and many simulated data circumstances. They concluded that a sample size of 70 times the number of variables is deemed adequate for cluster analysis in business research. In our cluster analysis, we used three variables (protective response, physical response, and cognitive–affective response) to classify participants into homogeneous subgroups. Using Dolnicar et al.’s recommendation, the minimum sample size needed for the cluster analysis in this study was 70 × 3 = 210. A series of one-way analyses of variance (ANOVAs) was also computed to compare the FM subgroups in terms of health and biopsychosocial outcomes (Nunnally, 1978). The minimum number of participants required was determined by an a priori power analysis, medium effect size (f = .25), α = .05, power = .80, and number of groups = 3, and the sample size for ANOVA was computed to be 159. With a sample size of 302 participants, the study had sufficient sample size for all of the analyses.
Results
PCA
The Kaiser–Meyer–Olkin measure of sampling adequacy was computed to be .74 (>.50), suggesting that a factor analysis is useful with the data in this study. The Bartlett’s test of sphericity, which tests the null hypothesis that the correlation matrix is an identity matrix, was statistically significant χ2 (55, N = 302) = 609.48, p < .01, indicating that variables in this study are related and therefore suitable for structure detection (Tabachnick & Fidell, 2007). The Kaiser–Guttman criterion of eigenvalue >1.00 indicated a three-factor structure. The three-factor solution accounted for 53% of the total variance and appeared to provide a measurement structure that is parsimonious, psychologically meaningful, and interpretable. Consequently, the three-factor solution was retained for this study. With a cutoff of factor loadings of .40 for inclusion of a variable in interpretation of a factor, every variable was appropriately loaded on a specific factor, and there was no cross-loading variables. Following extraction, the retained factors were rotated to a simple structure using oblique rotation to make them more interpretable.
Information on factor loadings and percentage of variance are shown in Table 2. Variables are ordered by size of their loadings to facilitate interpretation. The first factor appears to measure “protective response,” which is characterized by a combination of positive human traits (e.g., generalized self-efficacy, resilience, and positive coping) and environmental factors (e.g., social support and low levels of perceived social stigma). Five out of the 11 variables loaded onto the “protective response” factor, and this factor accounted for 27.2% of the total variance. The second factor appears to measure “physical response” such as pain intensity, fatigue, and sleeping difficulty; it accounted for 15.3% of the total variance. The third factor appears to assess “cognitive–affective response” (e.g., perceived psychological stress, anxiety, and cognitive impairments) and accounted for 10.2% of the total variance. Average scores were computed for each FM response factor to retain the scale metric, which allows for easier interpretation (DiStefano et al., 2009).
Factor Loadings, Communalities (h2), and Percent of Variance for Principal Components Extraction and Oblique Rotation on Response to Fibromyalgia Variables, Three-Factor Solution.
Note. F1= protective response; F2 = physical response; F3 = cognitive–affective response.
The protective factor had a moderately strong inverse relationship with the cognitive–affective factor (r = −.31, p < .001), indicating that higher levels of protective response are associated with lower levels of anxiety, stress, and cognitive impairments. The protective factor had a significant but smaller association with the physical response factor than with the cognitive–affective factor (r = −.19, p < .001). The physical response factor showed a stronger relationship with the cognitive–affective factor (r = .36, p < .001) than with the protective response factor.
Cluster Analysis
Cluster analysis was used to identify prominent clusters of participants on the basis of their mean factor scores on each of the three “response to FM” factors—protective response, physical response, and cognitive-affective response. The three response dimension scores were standardized as T-scores. Ward’s hierarchical agglomerative clustering method was used with squared Euclidean distance as the index of pairwise similarity-dissimilarity between participant profiles. To identify an optimal grouping of participants in the clustering hierarchy, the agglomeration schedule was examined to find a late stage in the hierarchy with a relatively small number of participant clusters, where the error sum of squares coefficients increased dramatically at subsequent stages in the hierarchy after relatively small increases at previous stages (Berven & Hubert, 1977). The stage producing three clusters of participants had relatively small increases of 1,558, 2,539, and 2,725 in the error sum of squares from one stage to the next at the three stages preceding the three-cluster stage, as compared with increases of 5,341, 7,188, and 10,995 at the three subsequent stages in the hierarchy. Thus, cluster homogeneity dropped substantially after this stage, after much smaller decreases at previous stages, and it seemed to provide a reasonable compromise between maximizing cluster homogeneity while also providing a limited number of participant clusters. The three-cluster solution was selected in this study as the optimal set of participant clusters (n = 204, 35, 63, respectively). The mean scores for the three-cluster solution are provided in Figure 1. The profile characteristics of the three biopsychosocial responses to FM are described as follows.
Biopsychosocial classification of response to fibromyalgia, three-cluster solution.
Cluster 1: Individuals who perceived themselves to have some biopsychosocial problems (moderate amount of problems subgroup; n = 204)
Persons with FM in this subgroup perceived themselves to have some amount of problems adjusting to the demands of living with FM. They displayed average physical, cognitive–affective, and protective responses to FM relative to other participants in this study. In addition, it is the largest subgroup, representing 68% of the total sample. The T-scores in this cluster were close to 50 on all three factors: protective response (M = 50.39), physical response (M = 50.28), and cognitive–affective response (M = 48.38). The results regarding cognitive–affective response indicate that this subgroup would be considered as having moderate levels of pain and anxiety and higher levels of perceived stress, sleep problems, cognitive impairment, as compared with a general national sample (Cohen & Williamson, 1988; Hays et al., 2005).
The participants in this subgroup, although scoring in the average range, had lower levels of social support, participation, self-efficacy, and resilience than the general population. Demographically, 98.5% of the people in this group were women, 61.3% were married, and 91.7% were white. The people in this group had an average age of 47.7 years (SD = 11.0), with a mean age at FM onset of 30.6 years (SD = 12.3) and an average time since diagnosis of 17.1 years (SD = 11.6). A total of 27.5% had an associate’s degree or vocational/technical degree, 23.5% had a bachelor’s degree, 15.2% had a master’s degree, 44.6% were currently employed, and 25.0% were receiving social security disability insurance (SSDI) benefits.
Cluster 2: Individuals who perceived themselves to have very few biopsychosocial problems (least amount of problems subgroup; n = 35)
People with FM in this subgroup exhibited low negative physical response (M = 35.95), low negative cognitive–affective response (M = 43.65), and higher protective response (M = 54.03) compared with other individuals with FM in this study. The above-average score on the protective component along with lower average scores on the cognitive–affective and physical component suggested that participants in this subgroup had the least amount of problems and the highest amount of positive assets among these three subgroups. The symptom response to FM is less affected by symptoms from the cognitive–affective (anxiety, perceived psychological stress, and cognitive impairment) and physical (pain, fatigue, and sleep disturbances) components, which may be more detrimental as clinical symptoms experienced by people with FM. Although participants in this cluster exhibited better levels of adjustment to FM than the other two groups in this study, they still would be considered to have mild anxiety and moderate pain intensity levels. When compared with the general population, the participants in this cluster exhibited significantly higher stress levels, decreased cognitive functioning, and more sleep disturbances. In the protective component, although this subgroup exhibited higher psychosocial traits, it still displayed lower levels of such characteristics when compared with average, general population samples.
The higher level of positive protective response can serve as an ameliorating component to deal with the symptoms more effectively and therefore improve overall adjustment to FM, as compared with the participants in the other two clusters. The fact that this group had the highest reported rate of employment among the three clusters demonstrates that they have achieved a standard of socio-economic success that is often considered an index of better adjustment to the challenges of living with a disability.
Within this cluster, 97.1% were women, 68.6% were married, and 91.4% were white. The average age was 51.5 (SD = 9.3), the average age of FM onset was 34.7 years (SD = 13.1), and the average time since diagnosis was 16.9 (SD = 14.1). Approximately 5.7% had an associate’s degree or vocational/technical degree, 40.0% had a bachelor’s degree, 28.6% had a master’s degree, 74.3% were employed, and 2.9% were receiving SSDI benefits.
Cluster 3: Individuals who perceived themselves to have many biopsychosocial problems (many problems subgroup; n = 63)
The symptom response profile for this subgroup indicated high negative physical response (M = 56.90), high negative cognitive–affective response (M = 58.31), and low positive protective response (M = 46.49) relative to people with FM in the total sample for this study. The high scores on the physical and cognitive–affective domains and low scores in protective responses are indicative that this subgroup has many problems that could affect daily functioning. The slightly above average score on the cognitive–affective response indicates that within this cluster, the most troublesome symptom response to FM lied within the cognitive spectrum (e.g., anxiety, perceived psychological stress, and cognitive impairment). Participants in this subgroup not only scored higher in the cognitive–affective component compared with the other two subgroups, but they would also be considered as having severe anxiety and significantly higher levels of stress and cognitive impairments than the general population, while exhibiting significantly lower levels of social support, self-efficacy, resilience, and positive coping strategies.
Participants in this subgroup also had high scores on the physical response component, which included high levels of pain, fatigue, and sleep disturbances. The fact that this subgroup showed a low positive protective response, while showing higher negative responses in the other two components, makes this subgroup more susceptible to poorer adjustment to FM. A higher positive response in the protective domain (e.g., resilience, self-efficacy, positive coping, social support, and FM as non-devaluating) could serve as a buffer for the effects of symptom response on the physical and cognitive–affective components. When compared with the subgroup with moderate problems and the subgroup with least amount of problems, this subgroup had the lowest percentage of employment and the highest percentage of participants receiving SSDI benefits. This finding further supports the idea that poor adjustment to FM in this subgroup leads to low employment and higher percentage of people with FM in this group receiving SSDI benefits.
In this subgroup, 98.4% were women, 50.8% were married, and 87.3% were White. The mean age was 48.9 (SD = 8.8), the mean age at FM onset was 30.4 (SD = 11.6), and average time since FM diagnosis was 18.5 (SD = 9.9). In this cluster, 31.7% had an associate degree or vocational/technical degree, 19.0% had a bachelor’s degree, 12.7% had a master’s degree, 25.4% were employed, and 54.0% were receiving SSDI benefits.
The results of the analysis in this study supported the assumption that the participants could be clustered into homogeneous subgroups to produce an empirical classification. Based on participants’ profile scores on their response dimensions, they could clearly be classified into one of the three subgroups: (a) the least amount of problems subgroup, (b) the moderate amount of problems subgroup, and (c) the many problems subgroup.
ANOVA
One-way ANOVAs followed by Bonferroni’s post hoc comparisons tests were performed to determine whether participants in the three response to FM groups differed in functional disability, community participation, HRQOL (physical and mental well-being), and life satisfaction.
Functional disability
One ANOVA showed that the effect of response to FM on functional disability was significant, F(2, 299) = 58.82, p < .001. Post hoc analyses using the Bonferroni post hoc criterion for significance indicated that the functional disability score was significantly lower for the least amount of problems group (M = 21.06, SD = 5.77) than for the moderate amount of problems group (M = 28.58, SD = 6.13) and the many problems group (M = 34.42, SD = 5.14). The moderate amount of problems group also had significantly lower average functional disability scores than the poorly adjusted group.
Community participation
The result of the ANOVA for response related to participation was significant F(2, 299) = 52.76, p < .001. In the post hoc analysis, the average score for the poorly adjusted group was significantly lower (M = 18.50, SD = 4.82) than for the moderate amount of problems group (M = 23.79, SD = 5.70) and the least amount of problems group (M = 30.14, SD = 4.97). The least amount of problems group had significantly higher average participation than the moderate amount of problems group.
HRQOL (physical)
In relation to HRQOL (physical component), the ANOVA result was significant F(2, 299) = 21.07, p < .001. The post hoc analysis indicated a significantly higher (M = 35.15, SD = 9.13) average score for the least amount of problems group than for the moderate amount of problems group (M = 26.45, SD = 8.47) and the many problems group (M = 24.17, SD = 7.08).
HRQOL (mental)
The ANOVA result showed that the effect of response to FM on the HRQOL (mental component) was also significant F(2, 299) = 34.93, p < .00. Post hoc analyses indicated that the HRQOL (mental component) score was significantly lower for the many problems group (M = 30.19, SD = 8.72) than for the moderate amount of problems group (M = 39.88, SD = 10.36) and for the least amount of problems group (M = 46.40, SD = 9.57). The moderate amount of problems group also had significantly lower average scores than the least amount of problems group.
Life satisfaction
The effect of response to FM was also significant on life satisfaction F(2, 299) = 19.88, p < .001. The post hoc analysis revealed that the average score for the least amount of problems group was significantly higher (M = 21.68, SD = 8.20) than for the moderate amount of problems (M = 15.16, SD = 6.98) and the many problems (M = 12.19, SD = 7.16) groups. The average score for the moderate amount of problems group was significantly higher than for the many problems group.
Discussion
FM is a multifaceted disorder that requires an individual’s active engagement in treatment and illness management. Using a biopsychosocial model provides an FM framework that allows practitioners and researchers to organize the heterogeneous presentation of FM through the evaluation and treatment of intersecting factors to understand FM and the influence and interdependence of multiple construct categories (e.g., individual, environmental, cognitive/psychological, and physical/biological factors) is imperative due to FM’s complexity and heterogeneity. The biopsychosocial approach has a potential impact on the way the illness is conceptualized, and alters the ways in which FM is currently conceptualized by a biomedical approach or the diagnostic criteria established by the ACR. In addition, this approach provides a more effective research framework from which to study FM by expanding FM research to investigate a broader range of biopsychosocial attributes that contribute to the severity and functional impact of FM beyond that rather than associated with physical pain as the sole focus of the syndrome. This model also facilitates broader and more holistic treatment interventions that address important protective factors, or a combination of positive human traits and environmental factors, and the development of therapeutic approaches. Due to FM’s heterogeneity, comorbidities, and the intercorrelations among clinical characteristics, classifying individuals into homogeneous subtypes could add valuable insight into the nature of this complex illness.
The first aim of this study was to validate a cognitive and behavioral classification of response to FM. The factor analysis results showed three distinct factors. Factor 1 represented the protective dimension, which included positive individual characteristics (e.g., general efficacy, resilience, positive coping) and social characteristics (e.g., social support, perception of disability as non-devaluating). This factor is highly consistent with the well-documented role that core self-evaluations play in rehabilitation planning and outcomes with people with FM and other disabling conditions (Smedema et al., 2016). Factor 2 represented the physical response dimension, which included pain intensity, fatigue, and sleeping difficulty. These common physical responses among individuals with FM have been supported by other studies (Bennett et al., 2007). Koch and Rumrill (2017) noted that subjective pain is the most debilitating feature of FM from the point of view of those who are affected by the disorder. Rutledge et al. (2009) found that pain, fatigue, and morning stiffness can be grouped into one factor; however, De Souza et al. (2009) found that the impact of sleep did not load onto the physical dimension factor. Factor 3 represented the cognitive and affective dimension, which included perceived psychological stress, anxiety, and cognitive impairments. These results compliment previous research including De Souza et al. (2009), who found that differences in psychological distress (anxiety and depression) and level of morning fatigue marked the heterogeneity between two FM subgroups. In addition, Rutledge et al. (2009) found that cognitive traits (concentration) loaded on two factors: distress (depression, anxiety) and dyscognition. Wilson et al. (2009) performed a factor analysis and found that physical and psychological symptoms (anxiety, depression, anger) and cognitive problems (forgetfulness, concentration) grouped into one factor.
The second aim of this study was to explore whether individuals with FM can be classified into homogeneous subgroups based on their symptom responses. The three subgroups that emerged in the cluster analysis had distinctive biopsychosocial (protective, cognitive-affective, and physical) profiles: (a) moderate amount of problems group, (b) least amount of problems group, and (c) many problems group.
Three explanations for the differences in symptom responses among the participants in this study emerge from close analysis of the results. First, as FM progresses, symptom severity increases (Koch & Rumrill, 2017). This provides a possible explanation for the many problems group, who had the longest time elapsed since FM diagnosis, followed by the moderate amount of problems and least amount of problems groups, respectively. The level of somatic and cognitive/affective problems displayed by each subgroup also reflects this order, suggesting that physical problems increase over time (Andrew & Andrew, 2017). The least amount of problems group also exhibited the lowest level of functional disability, and receipt of SSDI benefits, whereas the mean age of participants was the highest (51.5) versus the many problems group (48.9) and the typical-adjusted group (47.7). Consequently, time elapsed since diagnosis better represents the level of symptom severity than age.
Second, it is plausible that the level of psychosocial characteristics serves as a protective factor for least amount of problems participants (Livneh et al., 2019). These characteristics (social support, perception of disability as non-devaluating, resilience, positive coping, self-efficacy) could have a buffering effect and serve to minimize the negative effects of physical and cognitive symptoms on health and life satisfaction outcomes (Smedema et al., 2016). The least amount of problems group had the highest level of social support among all groups, and the mean on this characteristic was the greatest among all the other social and psychological features within the cluster. In addition, the least amount of problems group showed a greater degree of social participation and higher percentage of participants who are married, which are important factors that can serve as a proxy for social support (Smart, 2016). Results from this study support Turk et al.’s (1998) findings that the least amount of problems group responded better to FM symptoms due to higher levels of positive traits, such as adaptive coping, self-efficacy, and social support.
Third, emotional distress, including stress and anxiety, can affect the level of physical symptomology (Andrew & Andrew, 2017; Koch & Rumrill, 2017). The moderate amount of problems group differed from the least amount of problems group in their level of anxiety and depressive symptoms. Although the difference was not significant on the level of stress, the moderate amount of problems group also scored higher on this trait. However, the moderate amount of problems participants showed a higher physical response than cognitive–affective response. In comparison, the many problems group exhibited higher levels of psychological distress and showed lower levels of physical symptoms. Therefore, it is possible that individuals with FM can be classified into two distinctive subgroups based on their level of psychological distress and physical-related symptoms. De Souza et al. (2009) identified two clusters using the FM Impact Questionnaire: (a) FM-Type I (high pain, fatigue, and stiffness, but low morning tiredness, anxiety, and depressive symptoms) and (b) FM-Type II (high levels of all physical features and emotional distress). The authors concluded that the differences between these two clusters were driven by the differences in psychological distress and morning fatigue, supporting the idea that individuals with FM may fall into two separate subgroups.
The third aim of this study was to determine if the three cluster groups identified would differ in terms of health outcomes. The findings revealed that the three clusters differed significantly from each other on the four outcome variables. The results underscored that positive responses to FM lead to positive health outcomes and life satisfaction, and negative responses to FM are detrimental to health outcomes and life satisfaction. These findings underscore the heterogeneity of individuals with FM. At the same time, the results substantiate the proposition that individuals with FM can be classified into homogeneous subgroups representing diverse physical, cognitive/affective, and protective features, and these profile groups vary in their relationship to health outcomes and life satisfaction, with positive health outcomes and improved life satisfaction being pre-eminent end goals of the rehabilitation process (Bishop, 2012; Roessler et al., 2018).
Limitations
There are several limitations to consider when interpreting the findings of this study. First, the generalizability of the findings is limited by the use of an online survey format and a convenience sample. Generalizability of these findings is also limited because participants were predominantly White, well-educated, female individuals who are members of the NFMCPA. Individuals who responded to this survey may also have higher levels of functioning than non-respondents, which may have impacted completion of many of the survey items. The survey was also self-report, and as a result, responses are vulnerable to bias and error (Kobau et al., 2008). It is also possible that the physical and psychological/cognitive limitations associated with FM impacted the individual’s ability to self-report on symptoms and related severity. Finally, this was a cross-sectional study; therefore, causality and directionality cannot be ascertained. Based on literature and the biopsychosocial framework, assumptions were made about some aspects of directionality when interpreting the findings.
Clinical Implications
The findings of this study contribute to the existing literature related to assessing biopsychosocial factors that may affect a person’s response to his or her FM condition. Rehabilitation professionals can first use an individual’s scores to determine to which of the three generated clusters (very few problems, moderate amount of problems, many problems) each client belongs. Then, they can use the assessment data to work conjointly with at-risk clients to identify vulnerability and protective factors affecting response to FM. This important information can be used to develop treatment goals and identify evidence-based interventions to help at-risk clients cope with the myriad challenges and stressors that inhere to living with FM (Koch & Rumrill, 2017; Smedema et al., 2016).
For example, self-management training and health promotion strategies (e.g., exercise, physical activity, nutrition, smoking cessation, relaxation training) can be used to reduce FM-related symptoms (e.g., pain, stress, sleep problems, and fatigue; Busch et al., 2011; Giesser, 2015). Social skills training can be used to increase social support and quality of life (Miller & Chan, 2008). Positive psychology interventions (e.g., self-efficacy, resilience, positive coping training) can be used to increase personal strengths and social support resources (Chou et al., 2013). Psychological and mental health counseling can be used to help clients reduce their depression and anxiety (Bennett & Nelson, 2006; Hind et al., 2014). Cognitive remediation therapy and assistive technology can be used to help clients reduce the impact of FM on cognitive functioning (Blake & Bodin, 2002; Galvez-Sánchez et al., 2018; Mitolo et al., 2015). Financial counseling can help clients manage stress related to financial difficulties (National Fibromyalgia Association, 2020; National Multiple Sclerosis Society, 2017; Yael et al., 2019). For clients with few or mild physical, social, and cognitive problems, the assessment data can be used to work conjointly with these clients to identify positive person–environment contextual factors and interventions that can help them continue to cope effectively with chronic daily stressors and challenges associated with living with FM (Chou et al., 2013).
More broadly, the incorporation of physical, cognitive, and behavioral factors that may affect response to FM into clinical interview and treatment protocols will help rehabilitation clinicians in several areas, including case conceptualization, assessment, treatment planning, and treatment interventions (Roessler et al., 2018). Positive human traits, such as self-efficacy, resilience, and positive coping, will advance the adjustment process of those with FM while increasing functioning levels (Livneh et al., 2019). Social support should also be a primary focus when implementing treatment strategies (Bishop, 2012; Smart, 2016). Thus, rehabilitation clinicians should help clients by encouraging them to develop skills to increase their participation in social activities and enlarge their social networks (Koch & Rumrill, 2017). Increasing these psychosocial features can help enhance the positive attributes of the protective factors identified in this study, thus leading to better coping mechanisms for physical symptoms; decreased effects of cognitive–affective symptom responses such as anxiety, stress, and cognitive difficulties; and enhanced health outcomes and subjective well-being.
Rehabilitation professionals can utilize several additional strategies to help individuals with FM in their rehabilitation process, including illness management, health education, and promotion of positive human traits and character strengths. For instance, as noted previously, individualized exercise regimens, water and massage therapies, and other changes in lifestyle may be beneficial. Improving social support and developing/increasing the utilization of intrinsic, personal attributes (self-efficacy, resilience, positive coping) are also important to the enhancement of illness self-management. Accordingly, rehabilitation professionals should familiarize themselves with these forms of therapies and related EBP interventions to improve illness self-management.
Finally, rehabilitation professionals must help individuals gain knowledge about FM and the ramifications of the illness (Andrew & Andrew, 2017). Health education will help individuals with FM increase their understanding of the overall impact of the disorder on their daily living activities, while increasing awareness of other issues related to FM (e.g., decrease misunderstanding and stigma of FM, increase social support, advocate for needed accommodations).
Implications for Future Research
Findings from this study support the use of a biopsychosocial approach and the understanding that individuals with FM differ in their psycho-cognitive and behavioral profiles, which are influenced by biological, personal, and social features. It also substantiates the hypothesis that people with FM can be classified into homogeneous subgroups. Future research should focus on identifying homogeneous subgroups and related treatments to improve health and employment outcomes. Furthermore, because the findings of this study suggest that the least amount of problems subgroup coped better with FM symptomology than the other two subgroups, research should continue to explore the positive characteristics of this subgroup to expand knowledge that can lead to positive symptom responses for all individuals.
Future research should center on determining causality, and emphasizing this aspect would facilitate the development of more targeted treatment strategies. In addition, research can explore different psychosocial characteristics that improve individuals’ responses to FM. Future research using a longitudinal design will improve understanding of the illness course of FM and related treatment strategies.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by the National Institute on Disability, Independent Living, and Rehabilitation Research (Grant no.: H133B13001).
