Equity-by-design clinical decision support

Know which patient
to follow up first.

StrokeRisk AI ranks the hypertension patients in your clinic by stroke risk, using the fourteen clinical fields you already record. It returns a risk band, the reasons behind it, and guideline-aligned next steps — in your own language, in under two seconds, on any phone or tablet.

85.3%
Detection rate
equal in women and men
10,092
Nigerian patients
in development cohort
0.776
Out-of-fold
AUC-ROC
11
Languages
available
Built and trained in Nigeria Clinician reviews every output No patient data leaves your facility Works on any Android device
Registered on this device

Welcome back.

Register to continue

Before you begin

We record who is using the tool, in what role and at which facility. This gives every assessment a clear provenance and helps us improve the tool for the people actually using it. It takes about thirty seconds and you only do it once on this device.

Privacy notice — what we collect and why
Controller. Quantium Insights Nig. Limited, a CAC-registered Nigerian company, is the data controller for this registration.
  • What we collect: your email, optionally your name, your cadre, facility, state, optional LGA and preferred language.
  • Why: to attribute each assessment to a registered user, to understand which cadres and settings use the tool, and to improve it. It also supports the independent evaluation of the service.
  • Lawful basis: your consent, given below. You may withdraw it at any time by writing to the address below.
  • What we never collect here: no patient names, no patient records, no clinical data and no location beyond the facility, LGA and state you type in. We do not read your device location.
  • Retention: 36 months from your last use, then deletion.
  • Your rights under the Nigeria Data Protection Act 2023 include access, correction, erasure, objection and portability. Write to dramobionovo@quantiuminsights.com.
  • Sharing: aggregate counts only — never your email — are shared with the independent evaluator. Your details are not sold or used for marketing.
This registration is not a security check or a clinical credential. It does not verify your professional registration and does not authorise clinical use.
What the tool does

Built for the clinic you actually work in

StrokeRisk AI was developed on individual-level records from 10,092 adults managed for hypertension in Enugu State, and it asks only for information a Nigerian hypertension visit already produces. No new test, no laboratory, no specialist equipment.

🎯

Prioritisation, not prediction

The tool ranks your patients so that scarce follow-up capacity reaches the people most likely to benefit. If you can intensively follow one patient in five, that group contains around 60% of the strokes that will arise in your hypertensive population.

⚖️

Equal detection for women and men

A single cut-off would have detected 95% of cases in men but only 71% in women. Sex-specific thresholds bring detection to 85.0% and 85.7%. Equitable detection was treated as a design requirement, not a trade-off.

🗣️

Guidance in your own language

Clinical summaries, guideline-referenced recommendations and a patient-facing explanation, composed in the language you choose — including Nigerian Pidgin, Hausa, Yoruba and Igbo — rather than translated from English afterwards.

Clinical safety comes before the model

Three rules are evaluated before any score is calculated. Reported FAST symptoms trigger emergency escalation immediately. A documented prior stroke or TIA triggers the secondary-prevention pathway. Systolic pressure at or above 180 mmHg, or diastolic at or above 110 mmHg, triggers same-day clinical escalation.

The model may raise a patient's risk band. It can never lower it below the standard of care indicated by clinical assessment alone. Every result records whether the band came from the model or from a clinical rule.

This tool does not diagnose stroke and does not replace clinical judgement. It is a screening and prioritisation aid for clinicians. Developed and internally validated on Nigerian clinical data; prospective external validation is underway.