HIV/AIDS has been widely regarded as a young person's disease since its recognition as a major public health problem in the 1980s. First, young gay men and then intravenous drug users and their sexual partners were identified as selectively vulnerable to infection. Both sexual activity and drug use were generally associated with youth or the middle ages of adulthood. Only recently has attention been turned to the prevalence of HIV/AIDS among older adults, defined as aged 50 years or above. Recent data from the Centers for Disease Control and Prevention indicated that approximately 11% of Americans with HIV infection are at least 50 years old, including both those infected years ago who are still alive and others who were infected recently.
There is some basis for concern that older HIV-positive adults may have elevated rates of distress and disorder. As a rule, they have fewer sources of social or institutional support, fewer surviving peers, a lack parents to care for them, and their non-HIV-positive peers may consider HIV infection more stigmatizing than do younger people [1,2]. Co-morbid medical conditions are more common , and cognitive disorders associated with aging may also be present. Older people may be diagnosed later in the course of HIV illness, making successful treatment more difficult, so that the interval between diagnosis and death is shorter .
We are not aware of any formal studies of the prevalence of neuropsychiatric disorders among older people with HIV/AIDS. The purpose of this report is to review methodological issues in designing an informative study, and to summarize available data on general population and HIV-specific and age-specific rates of psychiatric disorders. Because the most commonly observed and most commonly studied disorders are mood and substance use disorders, this review is restricted to these two categories. Citations are representative rather than exhaustive.
Measurement: what measures and what timeframes to use?
Inevitably there is a trade-off between diagnostic precision and the time and personnel training required to assess disorders. In the past decade of HIV research, the most widely used method for the diagnosis of axis I disorders is the Structured Diagnostic Interview for the Diagnostic and Statistical Manual of Mental Disorders (DSM) IV (SCID) , which has a module for each diagnostic entity and is administered by a mental health professional usually with at least a master's level training. The interview takes up to 45 min or longer, depending on the number of diagnostic modules included.
The SCID and other diagnostic interviews such as the Schedule for Affective Disorders and Schizophrenia represent the gold standard for psychiatric diagnosis. However, the demands of time and staffing make this unfeasible in many studies, especially those with large enough samples to generate reliable rates of disorder. For large-scale epidemiological surveys, the Diagnostic Interview Schedule (DIS)  and its successor, the Composite International Diagnostic Interview (CIDI)  are designed to be administered by lay interviewers with one week of training, are fully structured, and use computer-based algorithms rather than clinician judgement to generate diagnoses. Like the SCID, these interviews generate psychiatric diagnoses and are time consuming.
More often, studies assessing depression and substance use disorders use either short cuts developed specifically with brevity in mind such as the PRIME-MD  or the Patient Health Questionnaire (PHQ) , both developed for use with primary care medical patients. Screening scales such as the self-report CES-D  have been widely used in HIV research, but at best they describe the presence of depressive symptoms, not disorders, and they lack diagnostic specificity. Other self-report scales to assess depression include the 21-item Beck Depression Inventory (BDI)  and a seven-item short form developed for medical patients, or the brief subscale included in the widely used SF-36 , a global measure of health-related quality of life. These self-report scales usually take approximately 5–10 min for literate respondents, and thus lend themselves to inclusion in studies with multiple assessment goals.
The choice of timeframe to be queried is quite variable between measures and studies. In recent national epidemiological studies, one study enquired about lifetime and one-month prevalence rates, whereas another used lifetime and 12-month rates. A third uniquely enquired about disorders occurring before the past 12 months. Symptom scales use the past week or two. In the HIV literature, the timeframes most often used are lifetime (including current) disorders, and one-month prevalence rates.
Sampling: who to include?
The rates of disorder vary widely depending on the population subgroup sampled and the site of enquiry. For example, rates of mood disorders are much higher in substance-using samples independent of HIV status, and in samples recruited from medical clinics, whereas sex and HIV symptom severity may also influence the rates of depression. These considerations complicate meta-analyses, and must be taken into account in describing observations regarding the prevalence of disorders.
For older HIV-positive samples, appropriate comparison groups might include patients without HIV but with other serious or life-threatening illnesses common in older age. Because of the non-specificity of depression, co-morbid conditions aside from HIV (e.g. coronary heart disease, cancer) may play a causal or contributing role in prevalence rates observed in HIV elderly. Analysis of the rates of depression in such comparison groups would allow us to disaggregate the specific contribution of HIV/AIDS.
What disorders to assess?
The available literature indicates that the most common psychiatric disorders observed in HIV samples are mood and substance use disorders. Within each category, however, there are multiple choices about what to include. For example, with respect to depressive disorders, one can assess major depression (always done), dysthymia (commonly included), subthreshold major depression (rarely included), minor depression and adjustment disorder with depressed mood. The last is not usually included in American studies, but when it is, the overall rates of depression zoom. This also applies to substance use disorders: especially in HIV-positive samples, the inclusion of cannabis abuse substantially elevates the overall rates as marijuana is used medicinally as well as recreationally, to enhance appetite and to relieve neuropathic pain. Overall, the more disorders assessed, the higher the overall prevalence rate (and the longer it takes to conduct the assessment).
Differential diagnosis in HIV
The diagnosis of depression may be complicated by somatic symptoms common to depression, HIV infection, co-morbid illness such as hepatitis C virus, or to concurrent medications. Fatigue, poor appetite, weight loss, low libido or insomnia may be symptoms of depression, systemic HIV illness or the multiple medications most HIV-positive patients are prescribed. However, affective/cognitive symptoms of depression are not usually affected (with exceptions such as alpha interferon). Clinicians familiar with the criteria for depressive disorders, with HIV manifestations and HIV medication side-effects can reliably diagnose depression even for patients with advanced HIV illness, based on structured evaluations with expert clinician consensus. Depression can also be accurately diagnosed for patients with other serious illnesses (e.g. amyotrophic lateral sclerosis) and even among hospice patients as death approaches [12,13].
Diagnosing depression in elderly adults may also pose challenges (although the large majority of ‘older’ HIV-positive adults are at the younger end of the ‘old age’ distribution, most being in their 50s). As noted by Mulsant and Ganguli , some authors have proposed that depression presents differently, with older patients emphasizing somatic rather than psychological symptoms. In addition to HIV infection, older adults are more likely to have co-morbid physical ailments and illnesses such as diabetes, which may cause somatic symptoms. These must be sorted out before attributing symptoms to a depressive diathesis. This can be done by experienced clinicians, but represent a source of error in self-rating scales such as the CES-D, in which such distinctions are not made.
The foregoing issues warrant consideration in the design and interpretation of prevalence studies. No assessment method, timeframe or sample composition is necessarily right or wrong, but needs to match the research goals of the study. Furthermore, cross-study comparisons and analyses should be made only when the study designs, measures and samples are similar. Although expensive, the pay-off may be worthwhile for the establishment of some sort of consortium of multi-site studies with a standard protocol and uniform rigorous, comprehensive assessment of psychiatric disorders.
Total prevalence rates
General population surveys
Three large American epidemiological studies of the prevalence of mental disorders have been conducted in the past 20 years: the Epidemiologic Catchment Area (ECA) study , the National Comorbidity Survey (NCS) , and the National Longitudinal Alcohol Epidemiologic Survey (NLAES) . The ECA study used the DIS, which as noted is a fully structured interview schedule, and the diagnostic criteria of DSM-III. Over 20 000 community residents in five urban areas were assessed. Diagnoses were generated by a computer algorithm. Timeframes assessed were lifetime, the past 12 months and the past month. In contrast, the NCS used DSM-III-R, interviewed a nationally representative study of 8098 persons aged 15–54 years, and assessed prevalence for lifetime and the past 12 months. The assessment interview was the CIDI, a ‘second generation’ version of the DIS used in the ECA study. The NLAES focused specifically on the assessment of alcohol use disorders, major depression and their co-morbidity, using DSM-IV, and a fully structured interview administered by trained lay interviewers to a nationally representative sample of 42 862 adults.
The findings of the three studies differed substantially. In the ECA study, for major depression, lifetime prevalence was 4.9%, compared with 17.1% in the NCS, and 9.9% in the NLAES. The 12-month prevalence was 2.7% in the ECA study, 10.3% in the NCS, and 3.3% in NLAES. Other diagnostic categories varied similarly. In an effort to reconcile these disparate findings, Narrow et al.  applied the criterion of ‘clinically significant’ symptoms, and arrived at an overall estimate for the 12-month prevalence of major depression of 4.5%.
The rates and patterns of major depression were assessed in an international study including 38 000 respondents in 10 countries . As in the ECA study, the DIS was used to generate DSM-III diagnoses. The lifetime rate of major depression in the United States was found to be 5.2%, with an annual (12-month) rate of 3.0%.
Substance use disorders
Substance use disorders were assessed in the ECA and NCS surveys, and all three surveys assessed alcohol dependence. The rates of alcohol dependence are consistent across the three studies, ranging from 13.6 to 14.1% for lifetime prevalence. Any drug dependence was similarly consistent between the ECA and NCS studies (6.2 and 7.5%), respectively.
Formal diagnostic studies of the prevalence of psychiatric disorders have been conducted in several US states over the past 15 years, including community volunteers of both HIV-positive and HIV-negative status. The first was published in 1988 , using the DIS, DSM-III criteria and a 6-month timeframe in addition to lifetime. The subsequent cohort studies of non-intravenous-drug-using gay men [21–25] used the SCID, DSM-III-R criteria, and a one-month timeframe in addition to lifetime rates. No statistically significant differences were found between HIV-positive and HIV-negative samples recruited from the same community, although current (4–9%) and especially lifetime rates (29–61%) of major depression were elevated in both HIV-positive and HIV-negative samples.
In a national survey of 2864 HIV-positive adults (the HIV Cost and Services Utilization Study) , a psychiatric screen rather than formal syndromal assessment was used. The CIDI ‘brief screener’ was used to ask about the past year. No HIV-negative comparison group was included. Thirty-six per cent of respondents screened ‘positive’ for major depression.
Fewer investigators have assessed rates of axis I disorders in women. Two small studies [27,28] of women in recovery or women in the military, respectively, found low rates of current major depression (2 and 5%). In contrast, Morrison et al.  studied 93 HIV-positive and 62 HIV-negative women in rural Florida using a modified SCID with DSM-IV criteria and found rates of major depression, but not dysthymia, significantly elevated for the HIV-positive women (19.4 versus 4.8% for HIV-negative women). Both groups reported high rates of lifetime major depressive disorder (47 versus 39%). Fifty-one per cent of HIV-positive women and 31% of HIV-negative women had a history of substance use disorders. In contrast, in a New York City sample of 39 HIV-positive and 37 HIV-negative intravenous drug using women, Lipsitz et al.  found high rates of depression regardless of the HIV serostatus.
Several of the foregoing studies included participants whose health status ranged from asymptomatic to late-stage AIDS [20,23,25]. Cumulatively, they showed no differences in rates of depressive or anxiety disorders regardless of the degree of immunosuppression (CD4 cell count) or the presence of an AIDS-defining condition.
Ciesla and Roberts  conducted a meta-analysis of 10 studies including HIV-positive and HIV-negative samples, none of which individually found a difference in the rate of depression associated with serostatus. By combining these studies with disparate sites, the choice of disorders that were assessed, sex and risk groups, they found that the rate of major depression for HIV-positive respondents was nearly double that of HIV-negative respondents, although dysthymia did not significantly differ by HIV status. Although this observation may be valid, it is perhaps problematical to combine such disparate populations because interactions cannot be noted.
Many studies have used self-report scales of depressive symptoms . As a rule, the rates of depressive symptoms are high (in the 30–60% range), independent of HIV status, in clinic samples and in the community where inner city, drug-using or impoverished participants were targeted for evaluation [33,34]. As noted above, the symptom ratings are designed to be highly sensitive at the expense of specificity, so that these estimates cannot be translated into diagnoses of clinical syndromes; they are best considered as indices of general distress.
In summary, the rates of depressive and substance use disorders among HIV-positive respondents may or may not be elevated compared with HIV-negative respondents from the same community, but they are clearly greater than the rates for the general population. Depressive symptoms (including somatic symptoms) are common, and self-rating scales that include somatic symptoms (and nearly all of them do so) may show misleadingly high rates of ‘depressive’ symptoms. The available evidence, although limited, does not show increasing rates of psychopathology with advancing HIV illness. Finally, it is noteworthy that most respondents with lifetime or current depressive disorders report an onset that preceded the knowledge of HIV status and most likely preceded infection.
Prevalence by age
General population surveys
In contrast to the disparate findings of depression rates in the major epidemiological studies of the general population cited above, the rates of depression have consistently been observed to decrease substantially by age and birth cohort. For example, in the ECA study, the lifetime rates of major depression declined from a high of 7.5% for the 30–44 year age range, to 1.4% for respondents aged over 65 years. Similar patterns were found for dysthymia and for one-year prevalence rates of major depression . Overall, the decline with age is dramatic, not subtle, in these community samples. It should be noted that medically ill older adults may have higher rates of depression than those who are healthy, which is true in all age categories.
Substance use disorders
The rates declined consistently and substantially with increasing age in all three national surveys. The rates are dramatically lower for the oldest age group in each study. For alcohol dependence, rates declined in the NLAES from 13.3% in the 18–29 year age range to 3.4% for respondents aged over 65 years . In the ECA study, the alcoholism rates for men declined progressively from 26.6% for men aged 18–29 years, to 13.5% for men aged over 65 years, and for women the rates declined from 4.8% in the youngest group to 1.5% in the oldest .
ECA rates of lifetime drug dependence declined from 9% in the 18–29 year age group of men to 0.12% for men over 65 years, whereas the rates for women declined from 5.5 in the youngest age group to 0.05% in the oldest . Similarly, in the NCS data set, drug use and dependence were found to be more common in cohorts born after World War II than those born before 1945, and odds ratios show a significant decline in rates for any substance use disorder with increasing age.
HIV-positive samples with HIV-negative comparison groups: under and over 50 years
We were unable to find any studies that assessed the rates of psychiatric disorders among older HIV-positive individuals. One study of a national sample of 640 HIV-positive adults under 50 years and 73 aged over 50 years examined medical, psychological, environmental and a host of other variables . The mental health variables included the mental health subscale of the SF-36 (a health survey questionnaire widely used to measure various components of health-related quality of life), and the CES-D, a screening scale for depressive symptoms. The investigators did not detect any differences between younger and older participants on any of the multiple measures examined, with the exception of physical functioning (poorer in older respondents) and medical co-morbidities (more common).
In the absence of published studies, we have re-analysed two of our own data sets that included HIV-positive and HIV-negative respondents under 50 years and over 50 years in order to generate preliminary data regarding prevalence rates for depressive and substance use disorders. Both samples were community cohorts of non-intravenous drug-using gay men in New York City, in which the same methods were used for recruitment and assessment (SCID). The first author was a co-investigator of the Columbia study  and adopted the same methods for psychiatric evaluation in the Cornell study , of which she was principal investigator. Data were combined to enlarge the sample of men aged over 50 years. The combined data set includes 308 HIV-positive men of whom 42 (14%) were aged over 50 years, and 134 HIV-negative men, of whom 20 (18%) were aged over 50 years. In the ‘over 50’ HIV-positive sample, the mean age was 54 years (SD 4.5), range 50–67. In the HIV-negative ‘over 50’ group, the mean age was 55 years (SD 5.3), range 50–71.
The lifetime major depression rates were, respectively, 32, 36 and 39% for HIV-positive men under and over the age of 50 years and HIV-negative men under the age of 50 years. The rate for HIV-negative men over 50 years was 20%.
Current (past month) major depression rates were 6, 5, 6% and zero, for HIV-positive men under and over 50 years, and HIV-negative men under and over 50 years. In both of these comparisons, HIV-negative men had lower rates, as expected according to general population surveys, but HIV-positive men over 50 years did not.
Substance use disorders
Lifetime substance use disorders (dependence) in this data set were analysed for alcohol dependence, drug dependence (including prescription drugs), and any substance dependence. As shown in Table 1, HIV-negative men aged over 50 years had lower lifetime rates compared with the other three groups, particularly for drug dependence.
This small data set is of interest because of the inclusion of younger HIV-positive and older HIV-negative samples, and because of the rigor of diagnostic procedures. Cumulatively, the results suggest but certainly do not prove that among older HIV-positive adults, the rates of psychopathology are not elevated, but the expected substantial declines in lifetime diagnoses of depression and substance dependence associated with older age in HIV-negative adults are not apparent when compared with their younger counterparts.
This review illustrates both the perils and promise of attempts to assess the prevalence of psychiatric disorders in population subgroups. Several conclusions can be drawn from the prevalence studies of psychopathology in HIV/AIDS conducted to date, which, it should be noted, is probably more extensive than the body of evidence regarding nearly any other disease category. First and foremost, appropriately selected comparison groups are essential to the interpretation of findings, because HIV-positive patients also have other characteristics likely to be associated with a differential risk of depression and substance use disorders. With respect to a subgroup such as the one we are interested in, two comparison groups are needed to pinpoint the specific role of age and HIV status: younger HIV-positive adults and older HIV-negative adults. In order to identify rates that are higher or lower than expected for this age group, the prevalence rates of psychiatric disorders for the general population over the age of 50 years, and for younger HIV-positive adults are needed as context. Such a study remains to be conducted.
1. Nokes K, Holzemer W, Corless I, Bakken S, Brown M, Powell-Cope G, et al
. Health-related quality of life in persons younger and older than 50 who are living with HIV/AIDS. Res Aging
2. Heckman T, Kochman A, Sikkema K, Kalichman S, Masten J, Bergholte J, Catz S. A pilot coping improvement intervention for late middle-aged and older adults living with HIV/AIDS in the USA. AIDS Care
3. Justice A, Weissman S. The survival experience of older and younger patients with AIDS. Res Aging
4. Spitzer RL, Williams JBW, Gibbon M, First M. Structured Clinical Interview for DSM-IV.
Washington DC: American Psychiatric Press; 1995.
5. Robins LN, Helzer J, Croughan J. National Institute of Mental Health Diagnostic Interview Schedule. Arch Gen Psychiatry
6. Robins LN, Wing J, Wittchen H-U. The Composite International Diagnostic Interview: an epidemiologic instrument suitable for use in conjunction with different diagnostic systems and in different cultures. Arch Gen Psychiatry
7. Spitzer RL, Williams JBW, Kroenke K. Utility of a new procedure for diagnosing mental disorders in primary care: the PRIME-MD 1000 study. JAMA
8. Spitzer RL, Kroenke K, Williams JBW. Validation and utility of a self-report version of PRIME-MD: the PHQ primary care study. JAMA
9. Radloff L, Locke B. The community mental health assessment survey and CES-D scale.
In: Weissman MM, Meyers J, Ross C, editors. Community surveys of psychiatric disorders
. New Brunswick, NJ: Rutgers University Press, 1986. pp. 177–187.
10. Beck AT, Weissman A, Mendelson M. An inventory for measuring depression. Arch Gen Psychiatry
11. Ware JE, Sherbourne CD. The MOS 36-item short-form Health Survey (SF-36): conceptual framework and item selection. Med Care
12. Goy E, Ganzini L. Delirium, anxiety and depression.
In: Morrison R, Meier D, Capello C, editors. Geriatric palliative care
. New York, Oxford University Press; 2003. Chapter 20, pp. 286–303.
13. Chochinov H, Wilson K, Enns M, Lander S. Are you depressed? Screening for depression in the terminally ill. Am J Psychiatry
14. Mulsant B, Ganguli M. Epidemiology and diagnosis of depression in late life. J Clin Psychiatry
1999, 60 (Suppl. 20)
15. Robins L, Regier DA. Psychiatric disorders in America
. New York: Free Press; 1991.
16. Kessler R, McGonagle K, Zhao S, Nelson C, Hughes M, Eshleman S, et al
. Lifetime and 12-month prevalence of DSM-III-R psychiatric disorders in the United States: results from the National Comorbidity Survey. Arch Gen Psychiatry
17. Grant B, Harford T. Comorbidity between DSM-IV alcohol use disorders and major depression: results of a national survey. Drug Alcohol Dependence
18. Narrow W, Rae D, Robins L, Regier D. Revised prevalence estimates of mental disorders in the United States. Arch Gen Psychiatry
19. Weissman M, Bland R, Canino G, Faravelli C, Greenwald S, Hwu H, et al
. Cross-national epidemiology of major depression and bipolar disorder. JAMA
20. Atkinson JH, Grant I, Kennedy C, Richman D, Spector S, McCutchan J. Prevalence of psychiatric disorders among men infected with HIV virus: a controlled study. Arch Gen Psychiatry
21. Perry SW, Jacobsberg L, Fishman B, Frances A, Bobo J. Psychiatric diagnosis before serologic testing for the HIV virus. Am J Psychiatry
22. Williams J, Rabkin J, Remien R, Gorman J, Ehrhardt A. Multidisciplinary baseline assessment of homosexual mnen with and without HIV infection: standardized clinical assessment of current and lifetime psychopathology. Arch Gen Psychiatry
23. Rosenberger P, Bornstein R, Nasrallah H, Para M, Whitaker C, Hass R, Rice R. Psychopathology in HIV infection: lifetime and current assessment. Comp Psychiatry
24. Perkins D, Stern R, Golden R, Murphy C, Naftolowitz D, Evans D. Mood disorders in HIV infection: Prevalence and risk factors in a nonepicenter of the AIDS epidemic. Am J Psychiatry
25. Rabkin JG, Ferrando S, Jacobsberg L, Fishman B. Prevalence of axis I disorders in an AIDS cohort: a cross-sectional, controlled study. Comp Psychiatry
26. Bing E, Burnam A, Longshore D, Fleishman J, Sherbourne C, London A, et al
. Psychiatric disorders and drug use among HIV-infected adults in the United States. Arch Gen Psychiatry
27. Goggin K, Engelson E, Rabkin J, Kotler D. The relationship of mood, endocrine and sexual disorders in HIV+ women: an exploratory study. Psychosom Med
28. Brown GR, Kendall S, Ledsky R. Sexual dysfunction in HIV+ women without AIDS. J Psychol Hum Sexuality
29. Morrison M, Petitto J, Ten Have T, Gettes D, Chiappini M, Weber A, et al
. Depressive and anxiety disorders in women with HIV infection. Am J Psychiatry
30. Lipsitz J, Williams JBW, Rabkin J, Remien R, Bradbury M, el Sadr W, et al
. Psychopathology in male and female intravenous drug users with and without HIV infection. Am J Psychiatry
31. Ciesla J, Roberts J. Meta-analysis of the relationship between HIV infection and risk for depressive disorders. Am J Psychiatry
32. Griffin K, Rabkin J. Psychological distress in people with HIV/AIDS: prevalence rates and methodological issues. AIDS Behav
33. Moore J, Schuman P, Schoenbaum E, Boland B, Solomon L, Smith D. Severe adverse life events and depressive symptoms among women with or at risk for HIV infection in four cities in the United States of America. AIDS
34. Knowlton A, Latkin C, Chung S, Hoover D, Ensminger M, Celentano D. HIV and depressive symptoms among low-income illicit drug users. AIDS Behav
35. Weissman M, Bruce M, Leaf P, Florio L, Holzer C. Affective disorders.
In: Robins L, Regier D, editors. Psychiatric disorders in America: The Epidemiologic Catchment Area Study
. NY: Free Press; 1991. Chapter 4, pp. 53–81.
36. Helzer J, Burnam A, McEvoy I. Alcohol abuse and dependence.
In: Robins L, Regier D, editors. Psychiatric disorders in America: The Epidemiologic Catchment Area Study
. NY: Free Press; 1991. Chapter 5, pp. 81–115.
37. Anthony J, Helzer J. Syndromes of drug abuse and dependence.
In: Robins L, Regier D, editors. Psychiatric disorders in America: The Epidemiologic Catchment Area Study.
NY: Free Press; 1991. Chapter 6, pp. 116–154.