Care that adapts to the individual.
Our personalized-medicine agents use structured patient data, diagnostics and genetics to guide patient-specific decisions. We started where personalization genuinely changes the treatment — sickle cell disease.
- Genotype-aware
- Explainable scores
- Doctor-in-the-loop
- Starting with sickle cell
Sickle cell is where personalization changes the treatment.
Personalized medicine only matters when the individual's data genuinely alters the decision. Sickle cell disease is one of the clearest cases — and in Nigeria, it is also the most urgent.
Nigeria is the global epicenter
More people live with sickle cell disease in Nigeria than anywhere else on earth. The need is concentrated exactly where we build.
Genotype testing is cheap & normalized
SS, SC and AS genotyping is routine and affordable — so genotype-aware care is realistic at scale, not aspirational.
Hydroxyurea titration is truly n-of-1
The right dose is found per patient through dose-response and CBC monitoring — the textbook definition of personalized medicine.
The data already exists
Genotype already lives on the patient health profile, so the same structured record powers both the app and the hospital EHR.
The principle
We do not personalize for its own sake. We personalize only where the individual's genetics, diagnostics and history should move the clinical decision — and we prove the case before we expand to the next condition.
A complete sickle cell care system.
Built and running for the telemedicine side today — doctors use it, the admin panel tracks outcomes — and being brought into the hospital EHR next.
Genotype-aware protocols
Care thresholds adapt to SS, SC and AS genotypes — personalization that changes the clinical decision.
Explainable risk engine
Every risk score breaks down into named, weighted factors — including live local weather — so clinicians can see the why.
Crisis diary
Patients log pain, sites, triggers and hospitalizations; clinicians get the trend, not just the moment.
Hydroxyurea titration
Dose-response tracking with CBC safety flags — hold-dose guidance to the doctor, never self-dosing for the patient.
Daily care agent
A monitoring agent nudges patients and escalates to the care team when risk rises.
Outcomes & research data
Admin outcomes dashboards plus anonymized, reference-only data for research and pilots.
One structured patient profile drives all of it — genotype, diagnostics, crisis history and daily readings feed every agent, score and brief.
Every score breaks into named factors.
There is no black box. A patient's risk level is the sum of named, weighted factors — including live local weather such as the harmattan season — so the clinician can always see exactly why the number moved.
Safety rails are never personalized. Personalization tunes thresholds and nudges — but the hard safety boundaries are fixed rules that the model cannot move.
Weighted, not opaque
Each factor carries a visible weight.
Clinician-facing
The why travels with the score.
Current level
Moderate
Score
77
logged 4 days ago · long bones
below personal baseline
live local weather · dry & dusty
94% · trending down
3 days without a check-in
Week 7 — CBC flag: ANC low. Hold-dose guidance sent to the treating doctor for review.
n-of-1 medicine, in practice.
Finding the right hydroxyurea dose is the clearest example of personalized medicine in sickle cell care. Doctium tracks the response and the safety signals — but the doctor always decides, and the patient is told, plainly, never to self-dose.
Track the dose-response
Each patient's response to hydroxyurea is followed over time — the dose that works for one is rarely the dose for another.
Watch the CBC safety flags
Counts are monitored against safe ranges, and the agent raises a flag the moment a value crosses a safety threshold.
Guidance goes to the doctor
When a hold-dose is indicated, the recommendation is surfaced to the clinician — who makes the call. Patients never self-dose.
Proven on the app. Coming to the EHR.
The same structured patient data drives personalized care in the telemedicine app today and in the hospital EHR next — no separate model, no separate record.
On the telemedicine app
Built and running for the patient-facing side. Doctors use the full sickle cell toolkit during consultations, and the admin panel tracks outcomes across the program.
- Doctors see genotype-aware risk in the consult
- Patients log crises and daily readings
- Admin outcomes dashboards track the cohort
Inside the hospital EHR
We are bringing the same personalized-medicine engine into the Hospital OS — so inpatient and outpatient teams work from one structured record, not a separate tool.
- Same genotype-aware thresholds and scores
- Pre-visit brief ready in the clinical station
- Shared longitudinal chart, no double entry
One profile, two front doors. Cleaner structured data from every encounter is what makes personalized medicine possible across both.
See the Hospital OSPersonalized — but never unsafe.
Personalization changes how care is tuned to the individual. It never changes the safety boundaries, and it never removes the clinician from the decision.
Rules outrank the model
Hard clinical safety rules take precedence over any LLM output. When a rule and the model disagree, the rule wins — every time.
Safety rails are never personalized
Personalization tunes thresholds and nudges. It never relaxes a safety boundary — those limits are fixed and the same for everyone.
Doctor-in-the-loop, always
Every recommendation, hold-dose flag and risk score is surfaced to a licensed clinician. AI assists; the doctor decides.
Research data is reference-only
Anonymized outcomes power dashboards and research — stripped of names and contacts, used for reference and pilots, never to identify a patient.
The same discipline runs through the whole platform — AI assists, recommends and monitors, while licensed humans stay responsible for every final clinical decision.
Personalized medicine, built on structured data.
See the explainable risk engine, hydroxyurea titration and the daily care agent in a live walkthrough — built for your hospital and your patients.