INTRODUCTION:
Missed opportunities for early HIV and TB testing among pregnant women hinder Nigeria’s maternal health and epidemic control goals. In high-burden Akwa Ibom state, service delivery and data quality gaps limit community testing. This study, under the KNCV Global Fund GC7 grant, evaluated Epcon AI’s ability to enhance data integrity and identify hotspots for reaching pregnant women with HIV testin

METHODS:
From January to August 2025, Epcon AI was deployed across LGAs in Akwa Ibom State. The system integrated routine program data, demographic profiles, and community-level vulnerability indicators to predict priority hotspots for testing. Teams implemented HIV testing interventions in both AI-identified hotspots and comparison sites selected through standard non-AI outreach processes. Logistic regression was used to evaluate the association between AI-guided targeting and two primary outcomes: HIV testing uptake and HIV positivity yield. Results are reported as adjusted odds ratios (aORs) with 95% confidence intervals (CIs).

RESULTS:
A total of 4,215 pregnant women were reached during the study period, with 2,540 (60.2%) tested in AI-predicted hotspots and 1,675 (39.8%) in non-AI sites. Testing uptake was significantly higher in AI-guided communities (82.5% vs. 64.1%, p < 0.001). After adjustment, women reached in AI-identified hotspots were nearly twice as likely to be tested (aOR = 1.94; 95% CI: 1.52–2.46). HIV positivity yield was also significantly higher in AI-targeted areas (6.8% vs. 4.1%, p = 0.002), with increased likelihood of testing positive (aOR = 1.73; 95% CI: 1.22–2.45).

RECOMMENDATIONS:
Epcon AI enhanced efficiency, testing uptake, and case-finding among pregnant women while improving accountability and data integrity. Scale-up should include continuous algorithm validation, staff capacity building, and strong ethical safeguards to mitigate risks related to confidentiality and bias. Responsible AI integration can strengthen equity and program effectiveness.