PMID- 18494750 OWN - NLM STAT- MEDLINE DCOM- 20090729 LR - 20090715 IS - 1524-4733 (Electronic) IS - 1098-3015 (Linking) VI - 11 IP - 7 DP - 2008 Dec TI - A predictive model of health state utilities for HIV patients in the modern era of highly active antiretroviral therapy. PG - 1144-53 LID - 10.1111/j.1524-4733.2008.00326.x [doi] AB - OBJECTIVE: Existing estimates of human immunodeficiency virus (HIV)-related health state utilities are inadequate for comparing alternative treatments on the basis of regimen-specific attributes such as dosing requirements or tolerability. The objective of this study was to examine the marginal impact of dosing, adverse events (AEs), and other factors on patients' health state utilities. METHODS: Treatment naive and experienced HIV patients participating in five open-label trials of highly active antiretroviral therapy (HAART) completed the 36-Item Short Form Health Survey (SF-36) instrument at various time points. SF-36 responses were converted to utilities using a previously reported algorithm. Expected utilities were estimated as a function of patient demographics, regimen attributes, disease status, and AEs using a mixed-effects maximum likelihood model. Mean utilities for five HIV health states were derived from predicted patient utilities. RESULTS: Negative predictors of utility included greater age (-0.001), prior acquired immune deficiency syndrome-defining events (-0.036), female gender (-0.038), and injection drug use (-0.056; P < 0.01 for all). Utility also depended on CD4+ cell count (P < 0.01), but not the presence of undetectable viral load. Regimen attributes were marginally associated with changes in utility. Depression was associated with the largest decrease in utility (-0.054, P < 0.001) among the AEs examined. Using the model to generate predicted utilities from the sample provided mean estimates ranging from 0.742 (SD 0.058) to 0.798 (0.052) for CD4+ counts between 0 and 99 and > or =500 cells/mm(3), respectively. CONCLUSIONS: HIV patients' health-related quality of life may be substantially affected by clinically relevant patient-, disease-, and treatment-related factors, such as injection drug use, disease status, food/drink restrictions, and AEs. FAU - Kauf, Teresa L AU - Kauf TL AD - Department of Pharmaceutical Outcomes & Policy, College of Pharmacy, University of Florida, Gainesville, FL 32610-0496, USA. tkauf@ufl.edu FAU - Roskell, Neil AU - Roskell N FAU - Shearer, Arran AU - Shearer A FAU - Gazzard, Brian AU - Gazzard B FAU - Mauskopf, Josephine AU - Mauskopf J FAU - Davis, E Anne AU - Davis EA FAU - Nimsch, Christopher AU - Nimsch C LA - eng PT - Journal Article PT - Randomized Controlled Trial PT - Research Support, Non-U.S. Gov't DEP - 20080520 PL - United States TA - Value Health JT - Value in health : the journal of the International Society for Pharmacoeconomics and Outcomes Research JID - 100883818 SB - IM MH - Adolescent MH - Adult MH - *Antiretroviral Therapy, Highly Active MH - Female MH - HIV Infections/*drug therapy/*psychology MH - *Health Surveys MH - Humans MH - Male MH - Middle Aged MH - Predictive Value of Tests MH - *Quality of Life MH - Severity of Illness Index MH - Substance-Related Disorders MH - Viral Load MH - Young Adult EDAT- 2008/05/23 09:00 MHDA- 2009/07/30 09:00 CRDT- 2008/05/23 09:00 PHST- 2008/05/23 09:00 [pubmed] PHST- 2009/07/30 09:00 [medline] PHST- 2008/05/23 09:00 [entrez] AID - S1098-3015(10)60596-7 [pii] AID - 10.1111/j.1524-4733.2008.00326.x [doi] PST - ppublish SO - Value Health. 2008 Dec;11(7):1144-53. doi: 10.1111/j.1524-4733.2008.00326.x. Epub 2008 May 20.