| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From New York State Psychiatric Institute (J.G.R., M.M., W.V.G., S.H.L.), Department of Psychiatry, College of Physicians and Surgeons, Columbia University (J.G.R., W.V.G.), New York, NY, and Weill College of Medicine, New York, NY (J.G.R., S.J.F.).
Address correspondence and reprint requests to Judith G. Rabkin, PhD, New York State Psychiatric Institute, Unit 51, 1051 Riverside Drive, New York, NY 10032. E-mail: jgr1{at}Columbia.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
METHODS: A total of 141 participants had semiannual neuropsychiatric, psychosocial, and medical assessments over a period of 30 months. These six occasions provided the basis for identifying patterns of employment (part-time, full-time, or unemployed). Those who completed neuropsychological testing, introduced at visit 4, constitute the sample used to identify predictors of number of hours employed, using multiple regression analysis with mixed procedure.
RESULTS: Over 30 months, 20% were continuously employed full-time, another 9% were continuously employed part-time, and 40% were continuously unemployed. Employment status changed for 31%: 4% who worked at baseline stopped, 13% started or increased their hours, 8% decreased their hours, and 6% showed a fluctuating pattern. The major parameters consistently associated with unemployment or partial employment, in order of influence, were financial (disability benefits), psychiatric (past/current diagnosis of major depression and/or dysthymia), medical (physical limitations), cognitive (executive function), and education. In contrast, age, ethnicity, laboratory markers of HIV illness status, vocational rank, and past or current substance dependence did not predict work status.
CONCLUSIONS: Overall, those who worked continued to work. However, despite improved health, most men who were unemployed at study baseline did not return to work. Structure of disability benefits, lifetime depressive disorder, physical limitations, and impairment in some areas of cognitive function each appear to represent significant barriers to work. Returning to work is evidently difficult, and clinicians may keep this in mind when recommending leaving work unless medically necessary. Specific interventions and policy changes regarding disability benefits may be needed to promote return to work for people with HIV/AIDS whose health is restored and who contemplate re-employment.
Key Words: employment, • HIV/AIDS, • depression, • neuropsychological tests, • disability benefits.
Abbreviations: AIDS = acquired immunodeficiency syndrome;; BDI = Beck Depression Inventory;; HIV = human immunodeficiency virus;; NP = neuropsychological;; SSD = Social Security Disability Insurance;; SSI = Supplemental Security Income.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
The issue of employment, which often appeared to be peripherally relevant in earlier days of the HIV epidemic, has now assumed a central place in the thinking of many HIV+ people whose health has been restored or stabilized by active antiretroviral therapy (1), and among HIV clinicians and administrators of long-term disability benefit plans. However, little is known about characteristics associated with maintenance or resumption of employment in this population. The proportion of HIV-infected people who continue to work without interruption or alteration in hours is unclear. A mail survey of nearly 2000 residents with HIV/acquired immunodeficiency syndrome (AIDS) (a 37% response rate) who had case managers at AIDS organizations in Los Angeles County found that 20% were working full-time, 15% were working part-time, and 65% were unemployed (2); whether these findings reflect continuous employment or return to work was not noted. In an ongoing study of 98 men and women who had taken some active step to return to work, only 24% had returned to work, defined as part-time or full-time paid competitive employment, 2 years later (van Gorp, Personal communication, March 2003). Although a number of studies and media articles have found high levels of interest in returning to work in the HIV community, it appears that relatively few who have stopped working actually return to work.
This apparent lack of return to the workplace is of interest for several reasons. First, work is often a source of satisfaction, self-esteem, companionship, and income. Work also serves a normalizing function, replacing the patient identity that many wish to shed. In addition, social policies and private health insurance coverage may change, especially in times of fiscal constraints. Benefits may be reduced or terminated for people receiving expensive medications who no longer are disabled by HIV illness, as noted in a 2003 article in The New York Times entitled, "Insurance Fights Grow on ‘HIV Retirement’" (3). It thus would be useful to identify factors that constitute impediments to employment among people with symptomatic HIV illness.
In the general medical literature, predictors of unemployment after illness have focused primarily on four populations: patients with traumatic brain injury, chronic pain, cardiovascular disease, and musculoskeletal disorders. Overall, using multivariate analyses, the best predictors of unemployment were current psychiatric symptoms and distress (4,5) and cognitive impairment (6–8). This literature is limited in its applicability to HIV populations both in sample characteristics and study design, although it usefully guides the selection of domains for assessment. In addition, issues specific to HIV/AIDS include broad access to health insurance and other benefits and entitlements secondary to an AIDS diagnosis, and possible problems of job re-entry associated with the stigma of the disease, issues of disclosure, the risk of future disability, and consequent financial burden to the employer.
In the HIV field, only a few investigators have examined reasons for leaving work or characteristics of those who continue to work despite HIV infection. Heaton et al. (9) found that, among 289 nondemented HIV+ subjects, those who were impaired on neuropsychological tests were twice as likely to be unemployed as nonimpaired subjects, even after excluding those with potentially medically disabling symptoms. Further, those with neuropsychological impairment who continued to work were five times more likely to complain of difficulties performing their jobs. In another study, Albert et al. (10) found that neuropsychological impairment in medically asymptomatic HIV+ adults was associated with unemployment.
The most likely explanation for leaving work would appear to be the direct effect of HIV illness, including opportunistic infections and chronic symptoms like fatigue, wasting, or uncontrolled diarrhea. However, in a study of 100 HIV+ patients, of whom 45 were unemployed, the main causes of work cessation were psychological distress (not further defined), not physical symptoms or an AIDS-defining condition (11). Blalock et al. (12) conducted a cross-sectional assessment of 200 men and women with HIV/AIDS who attended a medical clinic serving primarily indigent African American patients who at some time had a CD4 cell count less than 200 cells/mm3. Sixty percent were unemployed, 25% were employed part-time, and 15% were employed full-time. Of those employed, most had worked continuously throughout the course of HIV infection. Ethnicity, education, psychiatric or substance use histories, or current distress levels were not associated with differences in employment status, although those who were unemployed had on average a lower CD4 cell count and a higher viral load. Those who worked reported significantly better quality of life, although whether this was the cause or consequence of employment was not determined.
The current study was designed to evaluate the degree of association, over time, between five domains and hours worked at study endpoint. The domains are sociodemographic characteristics, laboratory and symptom indices of HIV illness, past or current mood and substance use disorders, neuropsychological functioning, and financial benefits. We examined these relationships first with the entire sample of 141 gay men, and then included only those employed at the study endpoint. A secondary interest was to describe longitudinal patterns of employment in terms of who stopped working and who returned to work.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Measures
Medical Measures
CD4 cell count and HIV RNA viral load determined by polymerase chain reaction were measured after every study visit at a local commercial laboratory (Quest). For the HIV RNA assay, the lower limit of detection was 400 copies per milliliter of serum. Undetectable test results were conservatively scored as 399 (2.6 log) copies. Antiretroviral medications and incident opportunistic infections and cancers were recorded at each visit. HIV-related symptoms were assessed by self-report with a 14-item scale of signs and symptoms commonly associated with HIV illness such as fatigue, night sweats, and oral candidiasis. The score is the sum of symptoms. The scale has been used in other HIV+ cohort studies (14). In addition, because restricted mobility often characterizes symptomatic HIV illness, we administered the Physical Limitations Scale from the RAND Medical Outcomes Study (15), which lists 10 activities in descending order of difficulty. Items are summed to obtain a total score, with higher scores indicating more physical limitations.
Psychiatric Measures
The full Structured Clinical Interview for DSM-IV (16) was administered by trained postdoctoral fellows at study baseline, and the mood and substance use modules were repeated annually (visits 3 and 5). Severity of depressive symptoms was assessed with the self-rated Beck Depression Inventory (BDI) (17), on which scores can range from 0 to 63.
Neuropsychological Assessment
The neuropsychological (NP) test battery assessed the domains of attention, concentration, verbal memory, psychomotor speed, and executive function (set shifting-cognitive flexibility). Tests administered at visits 4 through 6 were the Trail Making Test Parts A and B (18), California Verbal Learning Test (19), WAIS Digit Symbol Test (20), Stroop Color and Word Test (21), and Grooved Pegboard (22). Test–retest reliabilities reported for these tests over a period of 6 to 12 months are as follows: Trail Making Test Parts A and B, 0.64 and 0.72; California Verbal Learning Test, 0.59; Wechsler Adult Intelligence Scale (WAIS) Digit Symbol Test, 0.82; Stroop Color and Word Test, 0.73; and Grooved Pegboard, 0.68 to 0.78. For descriptive purposes, subjects were classified as NP-impaired if they scored 1+ SD in the impaired direction relative to age, education, and gender-matched standardized norms (where available) on two or more nonredundant NP tests.
Financial Benefits
Financial benefits include Supplemental Security Income (SSI), a needs-based benefit, and Social Security Disability Insurance (SSD), which is for workers who have paid Social Security tax for at least 10 quarters of employment, thus reflecting a stronger work history than recipients of SSI. Those with SSI who return to work stop receiving payments when their income exceeds a modest monthly sum, whereas those returning to work with SSD continue to receive payments for 6 months, and their case remains open for as long as 45 months. In all instances, these benefits were initiated when the study participants were more severely ill, predating their study participation.
Employment Status
Hollingshead-Redlich scoring (23) was used to assess level of vocational attainment, with higher numbers indicating higher levels. Categorical data regarding work status (unemployed; part-time, at least 4 hours/week; or full-time, 35 or more hours/week) was collected at every study visit. Actual number of hours of paid competitive employment worked regularly (on or off the books) was systematically collected only at time 6 and constitutes the dependent variable in analyses of predictors of work status. For the 19 men not seen at time 6, we calculated number of hours worked at last visit as follows: if unemployed at last visit, they were assigned a value of zero hours. If working full-time at last study visit, they were assigned 40 hours, and if working part-time at last study visit, they were assigned 20 hours per week.
Procedures
Between July 1995 and January 1998, subjects were seen at six semiannual visits, each lasting 3 to 4 hours. Medical information was elicited by postdoctoral fellows under the supervision of a physician who worked in the hospital HIV/AIDS clinic (S.J.F.), and neuropsychological assessments were supervised by a board-certified neuropsychologist (W.V.G.). The study was approved by the Institutional Review Board of Weill Medical College, Cornell University. After complete description of the study to the subjects, written informed consent was obtained. Subjects were paid $30 to $40 for each study visit.
Statistical Analyses
Log10 transformations were used for HIV RNA as is the convention. We first included the entire sample (N = 141). Analyses were then repeated to include only those working 4 or more hours per week at the study endpoint (N = 70). The dependent variable was number of hours worked per week, including "none" in the first analysis and 4 to 70 hours per week in the second analysis. The study endpoint was time 6, except for the 19 men who left the study after times 4 or 5.
The independent variables, except for lifetime mood and substance use disorder diagnoses (assessed at study baseline and again at time 3), were taken from times 4 to 6. Independent variables individually correlated with the outcome variable, and those considered likely to have a mediating or moderating effect were included. The 21 independent variables in this preliminary model included five domains with multiple measures: sociodemographic variables (age, education, Hollingshead-Redlich vocational rank, race-ethnicity [white vs. other]), medical measures (CD4 cell count, HIV RNA viral load, and physical limitations), psychiatric measures (BDI scores and Axis I diagnosis of lifetime and current depressive disorders [major depression, dysthymia] and drug or alcohol dependence), financial measures (SSI, SSD), and neuropsychological measures (scores on the seven neuropsychological tests). The final (reduced) model shows the independent variables that were statistically significant predictors of number of hours worked at study endpoint.
These multiple regression analyses were performed using SAS Proc MIXED (24) with maximum likelihood estimate and a type 3 F test that partialed out (adjusted for) the other covariates in the model and retained the unique contribution of each covariate. The final, reduced model retained the covariates with substantial contribution (p < .10) to the model.
To maximize the information in this data set, we conducted five separate regression estimates using various time points for the independent variables, as follows: (1) time 4 only (1 year before endpoint); (2) mean values across times 4 to 6; (3) difference (time 6 minus time 4) values; (4) time 5 only (6 months before study endpoint), and (5) concurrent (time 6).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Medical Characteristics
At study baseline, 115 (82%) had an AIDS diagnosis. Mean time since notification of HIV status was approximately 5 years. By visit 4, 65% were prescribed three or more antiretroviral medications, mean CD4 cell counts had risen from a mean of 242 cells/mm3 (SD = 223) at baseline to 304 cells/mm3 (SD = 237), median log viral load dropped from 4.9 copies (SD = 0.85) to 3.7 copies (SD = 1.0), and the percent with undetectable viral load increased from 3% at baseline to 31% at visit 4. Incident or recurrent opportunistic infections and cancers declined by 50% (Rabkin et al. (25) provide more detailed information about health changes in this sample).
Psychiatric Characteristics
Across the six visits, mean BDI scores were in the not depressed range and did not change significantly over time. DSM-IV diagnoses for lifetime including current disorders were assessed at study baseline and visits 3 and 5 for current disorders. Fifty-seven men (40%) had a lifetime diagnosis of major depressive disorder (with or without dysthymia), and another six (4%) had a diagnosis of dysthymia. Seventy-four men (52%) had lifetime diagnoses of substance dependence, primarily alcohol, cocaine, and polysubstance use, but also including marijuana, sedatives or hypnotics, and noninjected opiates.
During the period of observation, 37 men (26%) had a current depressive disorder on at least one occasion. Of these, only seven had new-onset disorders, and the rest were recurrent episodes. Of the 74 men with a lifetime diagnosis of substance dependence, 21 had a current diagnosis on at least one occasion during the study. Two had new-onset substance dependence diagnosed during this period.
Neuropsychological Characteristics
When neuropsychological assessment was introduced at time 4, 74 men (52%) were classified as NP-impaired. This definition is widely used in the HIV/neuropsychiatric literature and is one criterion for diagnosis of HIV dementia or minor cognitive motor disorder. By itself, it represents a low threshold and does not necessarily reflect clinically significant impairment, but it does suggest a probable decline in overall cognitive function from pre-HIV status. This proportion declined over time: 44% were outliers at time 5, and 32% were outliers at time 6. Factors associated with this decline include practice effects and increased use of antiretroviral therapy by time 6.
Patterns of Employment
Over the 3-year time frame, 28 men (20%) were continuously employed full-time, another 12 (9%) were continuously employed part-time, and 57 men (40%) were continuously unemployed. Six men (4%) stopped working, 19 (13%) returned to work or increased their hours from part-time to full-time, 11 men (8%) decreased from full-time to part-time, and the remaining 8 (6% of the total sample) were unemployed at both their first and last study visits but worked at some point between.
Return to Work
Nineteen men returned to work, defined as moving from unemployment to full-time employment (N = 2), unemployment to part-time employment (N = 12), or part-time to full-time employment (N = 5). Six men stopped working, and 11 decreased their hours. These small numbers preclude statistical analysis, but by inspection, there appears to be no common thread.
Of those who returned to work, two did so because of improved health. The others resumed working because opportunities presented themselves or because of financial considerations. Overall, change in health status accounted for only a minority of changes in employment status in this cohort. Only one relinquished benefits, which were needs-based, when he resumed work.
Predictors of Work: Entire Sample
Table 1 shows the independent variables significantly associated with number of hours worked at endpoint, both for the entire sample (N = 141) and for those working 4 or more hours per week at the study endpoint (N = 70). Across all five strategies for assessing the independent variables, the same eight variables (and no others) emerged, although not all were significant predictors in every regression analysis. The eight independent variables represent all five domains studied: sociodemographic (education), medical (physical limitations), psychiatric (BDI scores and lifetime mood disorder diagnoses), financial (SSI or SSD), and neuropsychological (Trail Making or Stroop Color and Word Test).
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
It may be argued that subjects receiving disability benefits were not working because of their disability, and that the observed association between these benefits and employment is actually circular. However, by time 4, study participants were no longer acutely ill. The correlations between medical markers of illness progression (CD4 cell count and HIV RNA viral load) with SSI and SSD show only weak relationships (correlations between 0.10 and 0.19). The observed link between benefits and unemployment is congruent with a survey our group conducted to determine perceived advantages and disadvantages to working. Among 75 men and 53 women with HIV/AIDS whose health had stabilized, loss of benefits was the primary barrier identified (26). The finding is also consistent with an analysis of Social Security data which found that "nationally, the proportion of people who become disabled by an illness and who then return to work is very small" (27), whatever the illness. Not everyone receiving disability benefits is well enough to return to work. However, a considerable number of subjects in this study (and in the HIV population) have substantially improved health. Within the HIV community, in fact, it is acknowledged that many of those whose health has been restored have not returned to work, and in some circles, the question of "riding the disability gravy train" (28) has evoked considerable controversy. On the other hand, those receiving either SSI or SSD were found to be disabled by the government agencies responsible for approving such claims, and in fact may have been more seriously ill than those who never applied for or received such benefits. However, the number of current medical symptoms was not associated with hours worked in this sample in any analysis, including cross-sectional analyses (data not shown).
The problem of returning to work is complex and entails more than a simple reluctance to work. HIV+ people recognize many benefits of employment, including finding structure and purpose in their lives, rejoining society as a functional member, and finding companionship and meaning at work. Reluctance to relinquish benefits often reflects uncertainty about maintenance of restored health; lack of confidence that benefits, once canceled, will be reinstated; and anxiety about adequate health insurance for the extraordinarily expensive medications and medical care required to maintain good health. It would appear essential to resolve these issues at a policy level to promote the return to paid competitive employment of significant numbers of HIV+ recipients of disability benefits.
Current depression has been associated with unemployment in every study in which it was assessed (10,29). However, we did not expect that a lifetime history of depressive disorders would have a nearly equivalent impact on employment, especially because most men with a history of depression were not currently depressed when assessed at any time during the study. The burden of lifetime depression has, however, been documented as the leading cause of worldwide disability (30) and is increasingly recognized as a major source of missed days or unemployment in the public health literature (31).
We and others previously found that those who were unemployed after major illness or injury had more neuropsychological deficits than those who worked (32,33). In our cross-sectional study of 130 men in this same cohort, we found that unemployed men performed more poorly than employed men on tasks of memory, executive function, and psychomotor speed when controlling for age, CD4 cell count, and physical limitations. In this longitudinal analysis, we replicated the role of executive function but not psychomotor speed or memory to predict the number of hours worked, although their impact was not as strong as the effects of benefits and depression history. These data, along with those in the present study, suggest that highly specific and selective deficits in neurocognitive function may serve as barriers to employment, and could be lost if one looks only at global neuropsychological functioning or aggregate measures. The data in the present study suggest that specific impairments in speed of information processing and frontal lobe functions are key domains that could potentially serve as barriers to return to work or full-time work in a person who is HIV+. It thus appears that specific difficulties in higher-order (executive) tasks of problem-solving and shifting of mental set are important in successful achievement of obtained and sustained employment.
The absence of effect on work status of laboratory markers of HIV illness progression is consistent with earlier findings (14,34) that these markers are not associated with depressed mood, although behavioral indices of illness (such as fatigue) have been found associated with receiving disability benefits (35). Although in general more advanced HIV illness, as signified by laboratory markers, is often accompanied by more physical symptoms, there is not a direct correspondence. Furthermore, the significance of these numbers depends on context. For the man whose CD4 count has declined consistently and has now, for example, 200 cells/mm3, the significance is more grave than the same value for a man who once had CD4 cell counts in the single digits and whose numbers have risen as a consequence of antiretroviral therapy.
Study limitations include the restricted sample; no women were included, and men with intravenous drug use histories were excluded. In addition, the study was conducted in a state (New York) and at a time when disability criteria were more flexible than may be the case today and in other locales. Further, we did not study those HIV patients who may never have had gainful employment and for whom skills acquisition may need to precede job searches.
Nevertheless, our findings have clinical and policy implications. Care providers have been quick to recommend leaving work on the theory that stress and distress will be thereby reduced; perhaps this recommendation should be more cautiously evaluated considering that returning to work is more difficult than was initially apparent. In terms of policy, termination of needs-based benefits including health insurance when the formerly disabled return to work appears counterproductive, and even the more flexible SSD plan may appear daunting. In addition, it is likely that programs actively assisting return to work, including both support and specific preparation with financial counseling, training in handling interviews, and assistance in finding jobs, may be needed to promote successful re-employment.
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Received for publication April 9, 2003.
Revision received August 25, 2003.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  R Dray-Spira, A Gueguen, F Lert, and the VESPA Study Group* Disease severity, self-reported experience of workplace discrimination and employment loss during the course of chronic HIV disease: differences according to gender and education Occup. Environ. Med., February 1, 2008; 65(2): 112 - 119. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Dray-Spira, A. Gueguen, J.-F. Ravaud, and F. Lert Socioeconomic Differences in the Impact of HIV Infection on Workforce Participation in France in the Era of Highly Active Antiretroviral Therapy Am J Public Health, March 1, 2007; 97(3): 552 - 558. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. J. Ferrando and C. Nims HIV-Associated Mania Treated With Electroconvulsive Therapy and Highly-Active Antiretroviral Therapy. Psychosomatics, March 1, 2006; 47(2): 170 - 174. [Full Text] [PDF]
|
|
|
|
 |
|
 R. Dray-Spira, A. Persoz, F. Boufassa, A. Gueguen, F. Lert, T. Allegre, C. Goujard, L. Meyer, and The Primo Cohort Study Group Employment loss following HIV infection in the era of highly active antiretroviral therapies Eur J Public Health, February 1, 2006; 16(1): 89 - 95. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. G. Rabkin, M. C. McElhiney, R. Rabkin, P. J. McGrath, and S. J. Ferrando Placebo-Controlled Trial of Dehydroepiandrosterone (DHEA) for Treatment of Nonmajor Depression in Patients With HIV/AIDS Am J Psychiatry, January 1, 2006; 163(1): 59 - 66. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
