OBJECTIVES: To determine if the SF-12 survey (using the PCS and MCS) can discriminate between predefined HIV severity states (as defined by CD4 and viral load) in a Kenyan population.  The discriminative ability of the SF-12 quality of life instrument has been tested in patients with HIV/AIDS.  In a US study, patients were stratified by CD4 count and viral load as a threshold for severity of disease.  Associations between severity and mental component score (MCS)/physical component score (PCS) were investigated.  The study found an association between CD4 threshold and PCS score.  Using similar methods, we tested the same hypotheses using a Swahili translation of the SF-12 in a Kenyan population of HIV positive patients.  METHODS: This cross-sectional study was based on data from a randomized controlled trial conducted in Nairobi, Kenya involving HIV patients.  CD4 count, viral load, and SF-12 survey results were collected on patients initiating anti-retroviral therapy at 3 clinics.  Clinical variables were stratified as follows: CD4≥200 cells/mm3compared to CD4<200 and viral load>55000 copies/mL compared to viral load ≤55000.  The PCS and MCS scores means were compared across these strata using a t test. RESULTS: CD4 counts were measured in 280 patients and viral load in 497 patients.  All patients completed the SF-12 instrument.  PCS scores showed significant discriminative ability in the CD4 comparison (p<0.01) with a mean(SD) of 41.3(11.42) for CD4<200 cells/mm3compared to 45.1(11.2) for CD4≥200.  Comparison in viral load strata and MCS scores were not significant.  Mean and standard deviation for both MCS and PCS in each severity state were similar in value to those reported in the earlier American trial. CONCLUSIONS: These findings suggest that a Swahili translation of the SF-12 is a useful instrument in discriminating across limited clinical variables of HIV disease progression and severity.