Background Early childhood diagnosis of HIV is a challenge in many developing countries, including Mozambique. Approximately 50% of exposed children and HIV-positive are lost to follow-up, i.e. during Postpartum Consultation (CPP), at-risk child consultation (CCR) or ART consultation in the country. The objective was to carry out an intervention to reduce the loss to follow-up of children exposed to and positive for HIV in Manica and Sofala provinces.

Methods Intervention study in HIV-positive women and their children in CPP and CCR in six health facilities in 2016. Stepped-wedge design with 3 cohorts for 3 months of two health facilities randomly selected. Interventions included activist allocation, telephone calls to contact the mothers, guide the mothers with exposed child from CPP to CCR, active outreach to missed mothers, and initiation of ART in the CCR for 3 months. Data were collected from the health facilities and study books. Analysis was binomial logistic regression model with mixed effects.

Results Of the aggregated data, PCR +was 7.7%, and proportion of HIV-positive women in CPP 17.4%. In the control group only 24% of the mothers had more than 2 visits with us, compared with 60% in the intervention group [OR=2.05 (95% CI: 1.60, 2.62)]. In the intervention group, more children were transferred to CCR 52% vs 32% in the control [OR=1.7 (CI: 1.3–2.41)], 65% of the mothers in intervention group reached at CCR vs 57% of the mothers in the control group [OR=1.69 CI: 1.27–2.41)] and returned to receive the PCR result of their child, 6.7% in the control vs 8.2% in intervention [OR=2.3 (CI: 1.36, 3.87)].

Conclusion The intervention had a greater impact on the number of visits to CPP, the transfer of mothers from CPP to CCR, and the reception of PCR results in CCR by the companion.