Skip to main content
Opian Health
v1.0 · In testing · Q2 2026

Link

A clinical decision system for Ethiopian hospitals and health centres. Link consolidates paper records and patchwork software into one integrated workbench — from registration through consultation, pharmacy, laboratory, and reporting.

link · encounter
A. Bekele, 34
Consultation · Room 3
CBHI active
Chief complaintFever, productive cough x 5 days
VitalsT 38.7°C · HR 104 · SpO₂ 96%
CDSS flagPneumonia pathway — consider CXR + CBC
Decision support — Based on symptoms, vitals, and regional epidemiology, Link suggests the national community-acquired pneumonia protocol. Clinician retains full control.

The clinical problem

Ethiopian health facilities operate with fragmented information. Patient records live in paper files and handwritten notes. Lab results arrive on slips. Pharmacy orders are written by hand. Some facilities have basic software for registration or billing, but it doesn't talk to the paper-based clinic, the lab, or anything else. Clinicians spend time on documentation and chasing information instead of on decisions.

CDSS (clinical decision support) technology exists in other contexts — algorithms for diagnosis, workup protocols, treatment pathways. But most systems assume reliable internet, a standardised workflow, and training on systems developed for other health systems. Ethiopian facilities use CBHI (Community-Based Health Insurance), offline capacity matters, and clinicians work with Ministry-aligned primary care guidelines, not generic protocols.

The gap: clinicians know what good practice looks like. They lack integrated information at the moment of decision and tools that respect the context they actually work in.

What Link is

15 modules — registration, triage, consultation, lab, pharmacy, imaging, inpatient, billing, inventory, admin, patient portal, HEW portal, dashboards, DHIS2 integration, and a CDSS engine. One patient record. Built around Ethiopian clinical pathways and 200+ primary care algorithms aligned to Ministry guidelines.

One system

15 modules, one patient record

Link is an integrated clinical platform. Each module works on its own; together they form a complete facility information system.

Consultation & CDSS

200+ primary care algorithms aligned to Ministry clinical pathways.

Triage

Structured triage with prioritisation at reception.

Laboratory

Test orders, results, quality flags — fully traceable.

Pharmacy

Dispensing, stock, drug interactions, CBHI-aware billing.

Imaging

Radiology orders and result routing, DICOM-ready.

Billing & CBHI

Chapa payment gateway and Community-Based Health Insurance integration.

Inpatient

Ward management, rounds, handovers, discharge summaries.

Registration

National ID-aware patient registration with household linkage.

Patient & HEW portals

Mobile access for patients and health extension workers.

Plus Inventory, Admin, Dashboards, DHIS2 integration, CDSS Engine, and Link Hub on-premise runtime.

Built for the context

Where the clinic actually is

Ethiopian facilities don't always have reliable internet. They serve patients in English, Amharic, and Afaan Oromo. They report into DHIS2 every month. Link was designed around these realities, not despite them.

Offline-capable

Link Hub runs on-premise via Docker. Facilities keep working during outages and sync when the link returns.

Multilingual

Full UI in English, Amharic, and Afaan Oromo. Patient-facing materials localised.

Row-level security

Multi-tenant with strict facility-level data isolation. Each hospital sees only its own data.

DHIS2 ready

Automated report generation into national DHIS2 pipelines. No month-end spreadsheets.

The clinical decision layer

Link includes a CDSS engine with 200+ algorithms for primary care conditions. These algorithms are evidence-based and aligned to Ministry of Health clinical pathways. When a clinician logs a patient's chief complaint, vitals, and examination findings, Link surfaces suggested diagnostic workups and treatment protocols. The clinician retains full authority over the decision — algorithms suggest, humans decide.

How it works
  • 1.Clinician enters chief complaint and examination findings
  • 2.CDSS engine matches against primary care algorithms
  • 3.Suggested workups and protocols appear (aligned to regional epidemiology)
  • 4.Clinician reviews, modifies, or overrides — the system follows their decision

The intent: a junior clinician sees a patient with respiratory symptoms, and Link surfaces the pneumonia, TB, and asthma pathways worth considering based on what's been entered. It isn't telling them what to do — it's asking "have you thought about this?" That matters most where the next specialist is 200km away.

Link is in pilot testing. Clinician feedback will be published here, with consent, as pilots conclude.

How the clinical content is governed

A suggestion engine in a consultation room carries real responsibility. Link's clinical content is sourced, reviewed, and validated under an explicit governance process — not shipped like ordinary software.

Sourced

Algorithms are built from Ethiopian primary health care protocols and Standard Treatment Guidelines, with drug formulary integration. Every rule keeps a traceable link to its guideline source and version.

Reviewed

Every change that touches a clinical surface passes a dedicated clinical review in addition to engineering review — both must approve before anything ships. Safety-critical alerts and hard stops get formal hazard analysis.

Bounded

The system suggests; the clinician decides. Every CDSS action requires explicit clinician confirmation, every override is respected and recorded, and AI advisory features operate inside governed boundaries with patient data protected under Ethiopia's data protection law (Proclamation 1321/2024).

Validated

Pilot facilities test whether algorithms match clinician expectations before general availability, and a formal research validation programme — with de-identified data and staged evaluation — measures whether decision support actually improves care.

The Learn phase

What the system learns — and how.

In clinical software, learning can't mean an algorithm silently rewriting itself. Link's loop runs through people.

Recorded

Every CDSS suggestion — accepted, modified, or overridden — is captured in audit and evaluation logs. The clinician's actual decision is the data, not just the system's suggestion.

Re-examined

A protocol that clinicians keep modifying is a signal. Flagged patterns route back into the clinical governance process, where guideline experts revise the rule — reviewed and versioned, never auto-updated.

Studied

The research validation programme evaluates de-identified decision-support data in staged waves, measuring whether suggestions actually improve care. Findings feed algorithm refinement as pilots conclude.

On the roadmap: expanding the CDSS from 200+ to 500+ conditions, with refinements driven by pilot outcomes and validation evidence rather than assumption.

Technology

Link is built on a modern, proven stack that scales from single facilities to regional networks:

Frontend
React + TypeScript
Backend
Express.js + TypeScript
Database
PostgreSQL + Supabase
AI services
Python/FastAPI + Ollama
On-premise
Docker-based Link Hub
Mobile
React Native with Expo
Payments
Chapa payment gateway
Reporting
DHIS2 integration

Status and rollout

Feature-complete (March 2026): All 15 modules are built and tested internally. The CDSS engine is functional with 200+ primary care algorithms.

In testing now: Early hospitals and health centres are beginning pilots. We're validating workflows, confirming that algorithms match clinician expectations, and testing offline sync.

Initial outreach: Awareness meetings are underway with potential customers (hospital networks, regional health bureaus, MoH programs).

General availability target: Q3 2026. We're not setting a hard date; we'll move into general availability when testing feedback is incorporated and we're confident in the product.

Evidence

Link is in active testing. Current metrics:

15/15
Modules delivered
3
Languages supported
English · Amharic · Afaan Oromo
200+
CDSS algorithms
Primary care conditions

In-production metrics (completeness rates, clinical outcomes, user adoption) will be published as testing concludes and customer deployments begin.

Last reviewed: April 2026