Health data at district scale — leading 21 officers on a CDC-funded program

The problem

Public-health programs live or die on their data. On a CDC-funded HIV program across an entire district, dozens of facilities report on TB/HIV, cervical-cancer prevention (CECAP), gender-based violence (GBV), and Option B+ (PMTCT); and if that data is late, inconsistent, or wrong, decisions and funding suffer. My job was to make sure the district's data was accurate, timely, and actually used.

What I did

I served as district data officer on the Afya Hatua program (THPS) and led a team of 21 data officers across the district. Concretely, I:

• Engineered and automated the reporting pipelines into DHIS2 and DATIM; weekly, monthly, and quarterly, aiming at improving submission accuracy and timeliness against donor and government requirements.
• Ran the analysis (SQL + visualization) on key health performance indicators to surface actionable insight for program decisions.
• Built the quality layer; regular Data Quality Assessments (DQAs) and validation protocols to keep data consistent across systems.
• Led the team of 21 on collection, cleaning, and entry into the central systems.

Results

• Designed an automated S.E.A.L.S reporting template that streamlined weekly analysis and made trend/anomaly detection early instead of after the fact.
• Engineered a custom viral-load (VL) eligibility reporting system that improved facility-level follow-up tracking and contributed to 95%+ district VL coverage, month over month.

What I learned

Leading 21 people taught me that data quality is a human system before it's a technical one; the best pipeline in the world fails if the people upstream don't trust it or understand it. It's the conviction I carry into everything I build now: the model is the easy part; earning reliable data is the real work.