interference; social support and social isolation; and abuse among women with physical .... referred clinic patients to onsite study staff who met with each.
Archives of Physical Medicine and Rehabilitation journal homepage: www.archives-pmr.org Archives of Physical Medicine and Rehabilitation 2013;94:2410-6
ORIGINAL ARTICLE
Depressive Symptoms in Women With Physical Disabilities: Identifying Correlates to Inform Practice Susan Robinson-Whelen, PhD,a,b Heather B. Taylor, PhD,a,c Rosemary B. Hughes, PhD,d Margaret A. Nosek, PhDb From the aSpinal Cord Injury and Disability Research Center, The Institute of Rehabilitation and ResearcheMemorial Herman, Houston, TX; b Department of Physical Medicine and Rehabilitation, Center for Research on Women with Disabilities, Baylor College of Medicine, Houston, TX; cDepartment of Pediatrics, University of Texas Health Science Center at Houston, Houston, TX; and the dRural Institute on Disabilities, University of Montana, Missoula, MT.
Abstract Objective: To examine correlates of depressive symptomatology in a sample of women with diverse physical disabilities to inform practice of modifiable risk factors that warrant attention and intervention. Design: Interview survey. Setting: Outpatient chronic care clinics. Participants: Racially and ethnically diverse women (NZ415) aged 18 to 64 years living with physical disabilities. Interventions: Not applicable. Main Outcome Measure: Center for Epidemiologic Studies Depression Scale. Results: Depressive symptoms were high with more than half the women exceeding an established cutoff for clinically significant depressive symptomatology. In hierarchical multiple regression analyses, demographic, disability, and health variables explained significant variance in depressive symptoms; however, modifiable variables (pain interference, social support, abuse) contributed significantly to depression scores over and above demographic, disability, and health variables. Analyses examining predictors of depression classification revealed similar findings. Conclusions: Depression is a significant problem for many women with physical disabilities. Modifiable contributors to depressive symptoms may provide intervention opportunities for researchers and clinicians. Clinicians need to attend closely to pain, particularly perceptions of pain interference; social support and social isolation; and abuse among women with physical disabilities. It may be valuable to include pain selfmanagement, social networking and social skill development, and safety and abuse prevention training when designing depression intervention programs for this population. Archives of Physical Medicine and Rehabilitation 2013;94:2410-6 ª 2013 by the American Congress of Rehabilitation Medicine
Women with disabilities represent 19.6% of the population of U.S. women.1 According to Healthy People 2010,2 women with disabilities are at an elevated risk for depression compared with men with disabilities, women without disabilities, and people in general. Although research has documented elevated rates of Presented in part to the American Congress of Rehabilitation Medicine, October 12, 2012, Vancouver, BC, Canada. All authors were affiliated with Baylor College of Medicine’s Center for Research on Women with Disabilities (CROWD) at the time data were collected. Supported by a grant awarded to Baylor College of Medicine from the Centers for Disease Control and Prevention (grant no. RO4/CCR614142). No commercial party having a direct financial interest in the results of the research supporting this article has conferred or will confer a benefit on the authors or on any organization with which the authors are associated.
depression in women with disabilities,3-5 there are challenges to diagnosing and treating depression in this population. The co-occurrence of depression and physical disability presents diagnostic challenges because somatic symptoms of depression may also be symptoms of the underlying disabling condition. In a study5 on secondary conditions among women with physical disabilities (WWPD), 94% reported fatigue and 80% reported sleep problems during at least 1 bimonthly interview over the course of a year. Such commonly experienced secondary conditions can mask an underlying mood disorder, leaving depression undiagnosed. WWPD face tremendous barriers to accessing health care services, including treatment for depression. Barriers include a lack of providers informed about depression in the context of
0003-9993/13/$36 - see front matter ª 2013 by the American Congress of Rehabilitation Medicine http://dx.doi.org/10.1016/j.apmr.2013.07.013
Depression in women with disabilities disability, inadequate insurance coverage resulting from disparities in employment, facility and programmatic inaccessibility, and transportation. One study6 found that only 44% of the WWPD classified as depressed had received recent treatment for depression. Failing to recognize and treat depression has a devastating effect on quality of life.7,8 It is imperative that we improve the identification and treatment of depression in WWPD. Although a wide range of variables have been found to predict depression in people with disabilities,9-13 few studies have examined correlates of depression among women with diverse physical disabilities.4,6 Research has demonstrated that people with diverse physical disabilities share many of the same secondary health conditions, life experiences, and life challenges.5,14-17 Turner and Noh18 noted that physical disability, regardless of the cause, is itself a chronic strain with adverse mental health implications. Coyle et al14 argued that despite some differences between disability subgroups, the common secondary conditions experienced by WWPD can be addressed in nondisability-specific health promotion programs. With many health care and other providers offering services to and providing care for people with diverse and multiple chronic disabling conditions, there is value in understanding correlates of depression that are important across disability types. While understanding such correlates can help health care providers better identify those at risk for depression, research must also aim to identify modifiable risk factors. As Bombardier et al19 point out in their study examining depression after spinal cord injury, a better understanding of modifiable risk factors can aid in the development of more effective interventions. The primary goal of this study was to examine the extent to which potentially modifiable variables were associated with depressive symptoms above and beyond more static variables. The constructs under consideration (pain, social network, abuse) were considered modifiable and, thus, important for informing intervention development and clinical practice. A link exists between depression and each of these constructs in the literature on the general population20,21 and people with physical disability22-24; thus, we hypothesized that greater depressive symptomatology would be related to more pain/pain interference, a weaker social network, and recent abuse. Based on the literature,25-27 we anticipated that pain interference would be more strongly related to depression than pain severity. Finally, we hypothesized that these potentially modifiable variables would be significantly associated with symptom severity above and beyond the contribution of demographic, disability, and health variables.
Methods Study design and procedures This study involved secondary analyses on data from a crosssectional survey study. Participants were recruited through 3 public and 2 private chronic care clinics, which were selected to provide diversity in socioeconomic status, race/ethnicity, and
List of abbreviations: CES-D Center for Epidemiologic Studies Depression Scale SF-36 Medical Outcomes Study 36-Item Short-Form Health Survey WWPD women with physical disabilities
www.archives-pmr.org
2411 disability type. Clinic specialties included spinal cord injury, multiple sclerosis, rheumatology, and musculoskeletal disorders. Participants completed a 20- to 30-minute interview about their health and abuse experience. Although the focus of the study was abuse and health-related concerns of WWPD, a history of abuse was not an eligibility requirement. To be eligible, women had to be aged 18 to 64 years, have a physical disability for at least 2 years that substantially limited their daily self-care and mobility, be able to communicate in English or Spanish, and have no known cognitive, communication, or mental health impairments that would significantly limit their ability to participate. Once institutional review board approval was received, medical staff referred clinic patients to onsite study staff who met with each woman in a private room, confirmed eligibility, obtained informed consent, and conducted the interview in the woman’s preferred language (English or Spanish). Although an assessment of depressive symptoms was not in the original study design, a measure of depressive symptoms was added approximately 3 months after the initiation of data collection. The sample for this study includes 415 women (of the original 511) for whom data on depressive symptoms were available.
Sample characteristics The sample was diverse in terms of race, ethnicity, education, and income, but most completed the interview in English (table 1). The most frequently reported disabilities were joint or connective tissue disease, multiple sclerosis, spinal cord injury or other spinal impairment, and neuromuscular disorder including postpolio syndrome and cerebral palsy. Most participants had lived with their disability for many years, required some personal assistance, and used at least 1 assistive device, such as a cane (38%), manual wheelchair (34%), walker (28%), or power wheelchair (16%).
Primary outcome measure Depressive symptoms were measured using the Center for Epidemiologic Studies Depression Scale (CES-D),28 a 20-item self-report measure of depressive symptoms experienced during the past week. A number of studies29-34 have shown the CES-D to be a reliable and valid measure of depressive symptomatology among people with physical disabilities, chronic health conditions, and other illness conditions, with several studies specifically noting that somatic items were not a threat to validity29 or that removal of somatic items is not advised.31 Although not a diagnostic tool, high scores on the CES-D indicate greater depressive symptomatology. The most common cut point used to indicate clinically significant depressive symptoms is 16 or greater.28
Correlate measures Demographic variables Demographics including age, race, ethnicity, marital status, education, employment, and income were collected through a selfreport questionnaire. Disability and health variables Primary disability was determined by providing participants a list of disabling conditions and asking them to indicate which conditions they had. Those reporting more than 1 condition were asked to indicate which condition was most limiting; this
2412 Table 1 (NZ415)
S. Robinson-Whelen et al Sample demographic and disability characteristics
Variable
Mean SD or %
Age (y) Years of education Less than high school High school or GED Some postehigh school College or graduate degree Race/ethnicity White, non-Hispanic Hispanic African American Other Married* Personal income ($) Household income ($) Employed Interview conducted in Spanish Disability type Joint and connective tissue disease Multiple sclerosis Spinal cord injury/spinal impairment Neuromuscular including postpolio and cerebral palsy Stroke Amputation Traumatic brain injury Other Missingz Disability duration (y) Mean age at onset (y) Need for assistance No assistance needed Assistance with IADLx only Assistance with ADLjj and IADL Use assistive device(s)
46.2211.92 12.743.86 25 26 23 25 47 21 24 8 44 7,000y 15,000y 21 14 32 25 17 16 4 2 1 1 1 14.3913.49 31.7515.66 14 39 47 82
Abbreviation: ADL, activities of daily living; GED, General Educational Development; IADL, instrumental activities of daily living. * Reflects the percentage of participants who are married or in a marriage-like relationship. y Median rather than mean income is presented because of the skewed distribution of this variable. z Six women had multiple disabilities and were unable to identify one as primary. x Instrumental activities of daily living, such as shopping, house cleaning, or meal preparation. jj Activities of daily living, such as bathing, dressing, or eating.
condition constituted the participant’s primary disability. Participants also were asked the age at onset of their primary disability. Disability severity was assessed by asking participants if they used any assistive devices and if they needed assistance with activities of daily living (eg, bathing, eating) and instrumental activities of daily living (eg, shopping, house cleaning). In addition, the 10-item physical functioning subscale of the Medical Outcomes Study 36-Item Short-Form Health Survey (SF-36)35 and the 9-item mobility subscale of the Craig Handicap Assessment and Reporting Technique36 were used to assess mobility and functional limitations. Global health was measured using the 5-item
general health subscale of the SF-36,35 which asked participants to rate their overall health and rate the extent to which they (1) get sick easier than others; (2) are as healthy as others; (3) expect their health to get worse; and (4) consider their health to be excellent. Modifiable risk factors, psychosocial resources, and vulnerabilities The variables considered to be modifiable resources or vulnerabilities were pain, social network, and past year abuse. Pain was assessed using the 2-item bodily pain subscale of the SF-36,35 which captures pain severity and the extent to which pain interferes with normal activities. Social resources were assessed using 2 measures. First, social isolation was measured using 3 items from the Human Population Laboratory Study of Alameda County37 that ask respondents how many close friends and relatives they have and how many of them they see at least once a month. Second, social support was assessed using an abbreviated version of the Medical Outcomes Study Social Support Scale,38 which consists of 2 items from each of the 4 subscales plus an additional item assessing overall satisfaction with social relationships. Past year abuse was assessed with the Abuse Assessment ScreeneDisability,39 which asks about physical, sexual, and disability-related abuse (eg, having devices broken/ withheld). In the current study, a dichotomous variable, reflecting whether the participant reported 1 or more forms of abuse in the past year, was used in analyses.
Analysis plan First, correlations were calculated to examine the relation between CES-D score and the continuous variables of interest. For categorical variables, analyses of variance were conducted to examine group differences in CES-D scores. Variables significantly related to CES-D were included in subsequent regression analyses, with variables entered in blocks: step 1, demographics; step 2, disability and health; and step 3, modifiable resources and vulnerabilities. As each block was entered, the extent to which the block contributed to depression scores was examined. If the block accounted for significant variance, then the individual variables in the block were examined. Variables that did not contribute statistically to explaining CES-D score were systematically eliminated until only the most statistically relevant variables in the block remained. This process was repeated with each step, with variables from previous blocks retained in later analyses regardless of whether they remained significant. Hierarchical multiple regression analyses were run separately for each group of modifiable variables (ie, pain, social network, abuse), and then 1 final analysis was conducted with all modifiable variables entered together as step 3. Finally, logistic regression analyses were run, using the same variables and the same sequence of procedures and steps, to determine the extent to which the variables predicted depression classification (CES-D 16).
Results The mean CES-D score SD for the sample was 20.8112.26 (median, 18.95), a value that exceeds the commonly used cut point of 16 representing clinically significant depressive symptomatology. Pearson correlations revealed that depressive symptoms decreased as age, education, and household income increased (table 2). Depressive symptoms were higher among women who www.archives-pmr.org
Depression in women with disabilities
2413
were not married or in a marriage-like relationship (22.9012.36 vs 18.1411.66) and were of minority race (22.5312.36 vs 18.8811.88). CES-D scores were also higher among women reporting more than 1 disability (24.7612.96 vs 19.7511.86), and women who required assistance with activities of daily living (23.3513.12 vs 18.5811.08). CES-D scores were inversely related to disability duration and were higher among those with greater functional/mobility limitations and poorer health perceptions. Depressive symptoms were not significantly related to any of the remaining demographic or disability variables, including primary disability type. Depressive symptoms were significantly related to all of the modifiable variables. CES-D scores were significantly correlated with pain and both social network
Table 2 Summary of Pearson correlations and analysis of variance used to examine the relation between risk factors/predictors and CES-D score (NZ415) Variable Demographic variables Age Years of education Minority statusy Marital statusy Household incomex Personal incomex Employment statusy Disability and health variables Disability type{ Multiple disabilitiesy Disability duration (in years) Functional limitations Mobility limitations Assistance needed with ADLy,# General health Modifiable variables Pain severity Pain interference Social support Social isolation Experience of abuse
Pearson r
F (df)
.16* .16* 9.34* (1,413) 15.93z (1,412) .20z .09
3.45jj (1,409) 1.67 (4,410) 11.88z (1,413)
.13* .17z .30z
16.17z (1,413)
.35z .34z .40z .31z .27z
5.72z (1,413)
Abbreviations: ADL, activities of daily living; df, degrees of freedom. * P
