Whats the highest number of documents you have seen for any offline user without hampering the offline functionality

Implementation
1

I’m trying to understand real-world upper limits and experiences with performance at scale, especially in contexts where health workers manage a large catchment area.

Things like:

  1. Docs per user.
  2. Lessons learned/ Best practices you can share would be super helpful.
  3. Device recommendations / device specs in use
  4. Server recommendations / serves in use

The largest I have seen in projects in Nepal was a user with 40k documents. Multiple users trying to sync together caused performance issues at times, and at various times, we have had app slowness that might be attributed to that. Nevertheless, there are other CHWs with more than the 10k docs, the amount CHT recommends an offline user have in their CHT App, but the app is working as expected.

2

What is the real world use case that users with such large numbers of documents replicated are trying to solve?

And what data (in real world terms, not documents) is actually necessary for those use cases?

3

It’s unavoidable in scenarios where longitudinal care is provided to people and/or in cases where a healthcare worker is assigned to an area with hundreds of households (this is unavoidable).

For example, let’s take a CHW responsible for ~500 households. Over the course of a year:

  • Pregnancy & Delivery Care

    • 1 pregnancy per household results in ~8 ANC tasks, 1 Delivery task, and 4 PNC follow-ups.

    • With 500 households, that’s already ~6,500 tasks linked just to pregnancies.

  • Newborn & Child Health

    • Each newborn is registered, tracked for growth and nutrition, and followed up for immunizations up to 5 years (e.g., ~10+ reports/tasks per child).

    • If 500 households yield ~300–400 births annually, that’s an additional 3,000–4,000+ child-related documents per year.

  • NCD & Chronic Care

    • Adults with hypertension, diabetes, asthma, or TB require repeated follow-up tasks, reports, and longitudinal tracking.

    • In some catchments, NCD follow-ups can account for thousands of documents annually.

We could explore ideas of creating a patient history and deleting older docs, but that doesn’t seem to be possible in CHT either, as far as I know.