The Overview

Since August 2024, I have been working as a software engineer with Ubunifu Solutions (contracted by PATH) on the Zanzibar Laboratory Information System (ZanLIS). This national-scale project involves maintaining and extending a LIMS based on Senaite to serve the Zanzibar Ministry of Health.

The Problem

Before ZanLIS, laboratory operations relied heavily on manual processes. Sample referrals and result recording were done on paper, leading to significant challenges:

• High risk of human error in data entry.
• Longer Turnaround Times (TAT) for test results.
• Difficulty in tracking samples and retrieving historical data.
• Lack of real-time insights for decision-making.

The Solution & Implementation

ZanLIS digitizes the entire laboratory workflow, ensuring data accuracy and efficiency. My role involved managing the project, conducting user training, and collaborating with diverse teams (UCC, UCSF, UNICEF) to implement critical integrations:

Key Integrations:

• CTC System: Integrated with the Care and Treatment Clinic system (with UCC) for seamless HIV/AIDS patient data flow.
• SHAMS: Implemented referrals for culture tests to designated testing facilities.
• ZanEMR: Integrated with the Electronic Medical Record system to supplement the laboratory module.
• Lab Instruments: Built machine interfacing API in ZanLIS to interface with machines like GeneXpert and Abbott for automatic result uploading via a machine interfacing middleware (TilleLab).

Data Engineering & Visualization:

I architected a data warehouse solution to unlock insights from laboratory data:

• ETL Pipeline: Middleware to extract data from all laboratory instances, transform it, and load it into a ClickHouse database.
• Analytics: Created aggregated views in ClickHouse and built interactive dashboards using Apache Superset for KPI tracking and trend analysis.

The Impact

The system has transformed laboratory operations by automating result recording, enabling real-time data access, and facilitating data-driven decisions. It ensures that when a sample arrives, results are processed, verified, and sent directly to integrated systems without manual intervention.