Reliability and validity of the SF-36 in HIV-infected homeless and marginally housed individuals E.D. Riley 1 , D.R. Bangsberg 1 , S. Perry 2 , R.A. Clark 2 , A.R. Moss 2 & A.W. Wu 3 1 Epidemiology and Prevention Interventions Center (E-mail: eriley@epi-center.ucsf.edu); 2 Department of Epidemiology and Biostatistics, University of California, San Francisco General Hospital, San Francisco, CA, USA; 3 Department of Health Policy and Management, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA Accepted in revised form 25 October 2002 Abstract Objective: To assess the reliability and validity of the Short-Form 36 (SF-36) health survey as a health status indicator among HIV-infected homeless and marginally housed (HMH) individuals. Methods: Between July 1996 and May 2000, a sample of HMH individuals completed interviews that included the SF-36. Responses to the SF-36 were analyzed for missing data, range, internal consistency, and construct validity. Results: Among 330 individuals interviewed, 83% were male, 43% were African-American, and the median age was 39 years. All internal consistency reliability coefficients exceeded 0.70, all item–scale correlations exceeded 0.40, all items were more strongly correlated with their hypothesized scale than any other scale, and all reliability coefficients exceeded inter-scale correlations for the same scale. Three of four physical health scales were significantly associated with CD4 cell count and HIV viral load. All scales were sig- nificantly associated with depression. Discussion: We found that scales were internally consistent, items correlated to an acceptable degree with their hypothesized scales, items were distinct from other scales, physical scales were associated with CD4 cell count and viral load, and all scales were associated with depression. These analyses provide evidence for the reliability and validity of the SF-36 as a measure of health status in HIV-positive HMH individuals. Key words: Health status, HIV, Homeless, Reliability, Short-Form 36 (SF-36), Validity Introduction Homelessness is a problem faced by people in many US metropolitan cities including San Fran- cisco where between 8500 and 15,000 people are homeless on any given night [1]. Homelessness is associated with poor access to health care [2] and a high prevalence of medical problems [3], including HIV infection [4], and an increased risk of mor- tality [5–7]. Therapeutic advances such as highly active antiretretroviral therapy (HAART) have significantly reduced opportunistic complications and improved survival of HIV-infected individuals [8–10]. However, improvements have been smaller in marginalized populations such as the urban poor [11, 12]. The discrepancies in health status and in the benefits made possible by available health care warrant the development of both health delivery systems and health evaluations targeted toward the homeless. Poor literacy, mental illness, drug intoxication, and potential cultural inappropriateness make the validity and reliability of traditional evaluation measurements questionable in indigent populations. Measures of self-reported health status such as physical functioning (PF), mental health (MH), and social-role function are important indicators of overall health [13]. Such self-rated health measu- rements offer information not captured by clinical assessment [14], and give a more comprehensive Quality of Life Research 12: 1051–1058, 2003. Ó 2003 Kluwer Academic Publishers. Printed in the Netherlands. 1051