SF-20 score and item distributions in a human immunodeficiency virus-seropositive sample

Abstract

The question of whether the full range of possible health states is measured by the Medical Outcomes Study (MOS) 20-Item Short-Form Health Survey (SF-20) in human immunodeficiency virus (HIV)-seropositive individuals is examined in this article. Ninety-five HIV-seropositive men (37 with asymptomatic infection, 58 with symptomatic infection) from two primary care practices were enrolled. Patients completed the SF-20 evaluating six dimensions of health status. Asymptomatic patients reveal substantial skew in score distributions for the dimensions of physical (-1.60), role (-1.19), and social (-1.13) functioning; no substantial skew is exhibited by symptomatic patients. Both subgroups reveal ceiling effects for physical, role, and social functioning, and pain dimensions; asymptomatic patients' ceiling effects are higher (physical functioning: 65% versus 31%; role functioning: 73% versus 41%; social functioning: 54% versus 43%; and pain: 41% versus 24%). Patients from both subgroups reveal floor effects for the role functioning dimension (asymptomatic patients, 22%; symptomatic patients, 34%). When looking at items rather than scales, asymptomatic patients' item distributions for the physical, role, and social functioning, and pain dimensions reveal clustering toward positive health states in most items; distributions of symptomatic patients are similar. Because this HIV-seropositive sample exhibits substantial ceiling effects in four of six SF-20 dimensions, effects that particularly are notable for asymptomatic patients, these dimensions should be revised for use in HIV-seropositive individuals or a disease-specific quality of life instrument should be constructed.