Disclaimer: I’m not a doctor and this is not medical advice. My approach helps you explore and understand your data, so you can go talk with your primary care physician about treatment.

A few months ago, overwhelmed by my parents’ health situations, I started building tools to help me manage and understand their data. I wanted to understand mine too, given our ancestry risks (read my article on Health) — so I built my own health advisor. One that reads labs, tracks trends, pulls in South Asian medical studies, and syncs with glucose monitors and smartwatch health metrics, among other things.

I’m a bit of a privacy nut, but I’ve always understood that when the value trade-off is there, people are willing to exchange their data for it. We get free Google and Meta services; in exchange, our data gets used for ads — not controversial. What is controversial is health data. Most of us want to protect that.

I think the juice was worth the squeeze. My parents’ health trajectory has changed dramatically for the better and we haven’t sold out our data. Using these tools, I was able to quickly get up to speed on complex issues with my parents and ask questions on their behalf through MyChart. Since I’m far from my parents, their doctors were happy to have me join by phone as we discussed things — they assumed I was medically trained, based on the questions I was asking. I’ve become their health advocate.

How it works: three ingredients

  1. A workspace — on your encrypted and secure personal machine (an article for another day but check out EFF’s guide), create a folder per person and drop in raw records from MyChart, labs, etc.
  2. AI of choice — run an AI locally, or use Claude Cowork/Code to analyze your structured health profile and answer questions. Important: if you’re using cloud models, find the privacy settings for model training — “Help improve Claude” in Claude’s case and toggle off to opt out of data sharing (or find similar settings in ChatGPT/Gemini).
  3. Open Masala MCP — I’ve organized open-source South Asian medical studies and made them machine-accessible. When you point your AI to it, your AI calibrates to South Asian baselines instead of the general guideline defaults. To enable: claude mcp add open-masala -- uvx open-masala-mcp Full guide here.

What’s in Open Masala and what you can ask

A curated set of landmark guidelines and cohorts that actually set South Asian thresholds — WHO, IDF, ADA, MASALA, CARRS — distilled from a 113-citation South Asian cardiometabolic evidence base into 18 cited, machine-readable reference values. You’ll be able to ask questions like:

  • “My ApoB is 95. What does that mean for a South Asian body, and what’s the target?”
  • “Prep me for my cardiology appointment — what should I ask, and what tests should I request?”
  • “My last three ferritin values — is that a trend I should worry about?”
  • “I’m vegetarian and tired lately. What deficiencies are worth checking, and how do I raise them through food first?”
  • “New lab just came in — add it to my tracker and tell me what changed.”

How to build actionable materials

  1. Make one-page PDFs for your parents to reference — nutrition best practices and lifestyle tweaks at a glance. Mine were built in collaboration with a dietitian we met with.
  2. Generate questions ahead of primary or specialist visits, e.g., “My parent’s LDL is high, and lifetime vegetarian — I’m concerned about different risk factors. How might we think about a treatment approach?”
  3. Build custom fitness coaching based on your profile and your likes/dislikes when it comes to fitness (find what works for you!).

Life is busy, but if you set this up, I believe you’ll find yourself turning to it whenever health questions come up — whether you’re managing your own health, a spouse’s, your kids’, or your parents’.

Ok — that’s it, go get healthy.

South Asian woman standing with arms open with stylized landscape.

Are you a developer? Here are more ways to leverage the data.