BACKGROUND: Evidence on patient-reported outcomes of newly diagnosed HIV patients is scarce, and largely cross-sectional. This prospective cohort study describes the prevalence of, and changes in, patient-reported outcomes in the three months after HIV diagnosis, in 11 HIV outpatient centres in Kenya and Uganda. METHODS: Adults were recruited within 14 days of result, completing self-report measures four times at monthly intervals. Multilevel mixed-effects linear regression (quality of life continuous outcomes) and ordinal logistic regression (symptoms and concerns categorical outcomes) modelled change over time, with repeated observations grouped within individuals adjusted for demographic/clinical characteristics, and multiple imputation for missing data. RESULTS: 438 adults were enrolled and 234 (53·4%) initiated ART. Improvement was found for MOS-HIV physical health (from 46·3 [95% CI 45·1-47·3], to 53·7 [95% CI 52.8-54·6], p < 0.001), and mental health (from 46·4 [95% CI 45·5-47·3] to 54·5 [95% CI 53·7-55·4], p < 0.001). POS subscale 'interpersonal problems' improved but remained burdensome (OR = 0·91, 95% CI = 0·87-0·94, p < 0.001; 22·7% reported severe problems at final time point). The scores for the existential POS subscale (OR = 0·95, 95% CI = 0.90-1·00, p = 0.056) and physical/psychological problems POS subscale (OR = 0·97, 95% CI = 0.92-1·02, p = 0.259) did not improve. Participants who initiated ART had worsening physical/psychological (OR = 0·64, 95% CI = 0·41-0·99, p = 0·045) and interpersonal problems (OR = 0·64, 95% CI = 0·42-0·96, p = 0·033). CONCLUSION: Although some self-reported outcomes improve over time, burden of interpersonal problems remains substantial and existential concerns do not improve.