Every day, I work with SmartCare Pro exports — big Excel files that track HIV program data. My daily routine usually involves:
Filtering the ART Register to calculate TX_CURR (patients currently on treatment)
Splitting results into age groups (<15, 15+)
Checking the QAQI report for viral load coverage and suppression
Preparing lists of patients who are due for viral load or have missed appointments
This used to take me 30+ minutes a day with lots of filtering, copying, and manual counting. It was repetitive, slow, and prone to small mistakes.
That’s why I built TxLens — a local web app that automates these steps. Instead of wasting time in Excel, I upload my reports and get:
Instant TX_CURR counts with age disaggregation
A clear viral load cascade chart
Ready-to-export patient lists (due for VL, missed appointments, IIT)
How Kiro Helped Me Build It
The most powerful part of using Kiro was how it turned me into the architect of the project, not just a coder grinding through syntax. Here’s how:
Vibe Coding: I started by describing my daily routine, and Kiro generated an initial architecture with components for parsing Excel, calculating indicators, and generating reports.
Spec-Driven Development: Once the basics were working, I structured my needs as simple specs. For example, I told Kiro: “Make the lists print-friendly with headers and spacing.” Kiro expanded that into a full requirement with a print button, A4 formatting, and CSV export. It even helped me add an Upcoming Appointments Report by just describing the idea.
Agent Hooks: I added a hook to check for Flake8 compliance, so every time Kiro generated code it was immediately checked for style and consistency. That saved me time and gave me confidence in the code quality.
Spec-to-Code: The most impressive code Kiro generated was the calculation engine. I only explained how I filter TX_CURR or identify patients due for viral load in Excel, and Kiro turned that into reusable functions that matched my manual results exactly.
In short, I guided, steered, and directed — Kiro handled the grunt work. That sped up development and let me focus on the real problem: helping clinicians save time and make better decisions.
The Impact
What used to take me half an hour of manual work every morning now takes under a minute. Clinicians get cleaner patient lists, program managers see the indicators faster, and errors are reduced.
TxLens is open source, runs locally to comply with data laws, and is designed to be practical for real healthcare settings.
Closing
Kiro changed the way I approach development. Instead of worrying about syntax and boilerplate, I spent my energy on design, direction, and accuracy. That’s the future of development — and with TxLens, I got to experience it firsthand.
Thanks @kirodotdev for being my coding partner in this journey 🚀
Top comments (0)