Objective: Early infant diagnosis (EID) is the first step in HIV care, yet 75% of HIV-exposed infants born at 2 hospitals in Mozambique failed to access EID.
Design: Before/after study.
Setting: Two district hospitals in rural Mozambique.
Participants: HIV-infected mother/HIV-exposed infant pairs (n = 791).
Intervention: We planned 2 phases of improvement using quality improvement methods. In phase 1, we enhanced referral by offering direct accompaniment of new mothers to the EID suite, increasing privacy, and opening a medical record for infants before postpartum discharge. In phase 2, we added enhanced referral activity as an item on the maternity register to standardize the process of referral.
Main outcome measure: The proportion of HIV-infected mothers who accessed EID for their infant <90 days of life.
Results: We tracked mother/infant pairs from June 2009 to March 2011 (phase 0: n = 144; phase 1: n = 479; phase 2: n = 168), compared study measures for mother/infant pairs across intervention phases with χ², estimated time-to-EID by Kaplan-Meier, and determined the likelihood of EID by Cox regression after adjusting for likely barriers to follow-up. At baseline (phase 0), 25.7% of infants accessed EID <90 days. EID improved to 32.2% after phase 1, but only 17.3% had received enhanced referral. After phase 2, 61.9% received enhanced referral and 39.9% accessed EID, a significant 3-phase improvement (P = 0.007). In adjusted analysis, the likelihood of EID at any time was higher in the phase 2 group versus phase 0 (adjusted hazard ratio: 1.68, 95% confidence interval: 1.19 to 2.37, P = 0.003).
Conclusions: Retention improved by 55% with a simple referral enhancement. Quality improvement efforts could help improve care in Mozambique and other low-resource countries [added].