https://youtu.be/jlKMU-PveYA

1. who we heard from

“i’m a clinical officer… i work as a clinician, a general clinician and also a public health practitioner… mostly in urban and peri-urban private clinics and under-resourced settings.” dr. martin muthare

17 years on the ground, martin has run hiv programs, built spreadsheets to track cohorts, and still fights the same three dragons every shift: missing diagnostics, missing records, and missing time.


2. raw pain points (verbatim)

problem signal direct quote
no timely diagnostics “you’ll be required to just treat empirically… if you require a test it will take you… six hours or a day… some tests even two weeks… a month because of queues.”
lost patient history “patients are highly mobile… you wouldn’t know what this patient received in a different facility.”
follow-up collapse “since you didn’t have contact information… high mobility… that poses a very significant challenge of follow-up.”
long queues & stock-outs “long queues and drug stock-outs… impact on the care quality.”
data-sharing fear “anything that tries to make the facilities become interconnected… is normally met with a lot of opposition.”

external context backs him up:


3. opportunity briefs (problems worth building for)

# problem statement why now user signal
1. 20-minute labs anywhere peri-urban clinics lack point-of-care analyzers that hit <30 min TAT for hiv viral load, malaria, full blood count portable molecular readers now <$500; ai-assisted image recognition cuts reagent cost “lack of timely diagnostic support… you send the sample to a different facility.”
2. pocket patient history clinicians need medication & diagnosis history that travels with the patient, not the facility sim-based personal data vaults (m-health wallets) run offline; QR sync prevents massive infra “they never recovered… but you don’t know what was given out there.”
3. queue & stock radar lack of visibility into neighbour-clinic load causes overflow and stock-outs simple USSD or WhatsApp bot can broadcast live queue length & drug inventory snapshots “long queues and drug stock-outs… impact on care quality.”
4. outbreak early-warning feed clinicians hear about outbreaks on the evening news instead of in-clinic national dhis2 already houses aggregate alerts; push a geo-filtered rss / webhook to phones “if you could be having a tool that also gives you information… you’d plan drug stocks faster.”
5. gentle recall engine chronic patients miss 3-month check-ins; default labels hide real compliance low-literacy voice + sms reminders tied to outcome-based financing “a system that tracks and reminds both the clinician and the patient… would reduce readmissions.”

4. design guardrails

  1. low-friction data capture – typing killed the spreadsheet; favour auto-logging (camera scan, bluetooth analyzer dumps).
  2. interoperability first – publish FHIR-lite json; plug straight into the national HMIS roadmap (see Musabi 2024) SCIRP.
  3. trust & consent loops – build opt-in visibility (patient can mask HIV status by default); borrow participatory design lessons from iris-scan pilots.