Part 1 of an ongoing series on OSS interoperability layers, workflow architectures, and automation pipelines I've been building.
Over the past several months, I've been studying how real-world healthcare systems actually operate beneath the dashboards, reports, and interfaces.
One thing became very clear:
Most healthcare systems aren't failing because the software is missing. They're failing because the software ecosystems don't work together.
The invisible operational chaos
From the outside, many national and regional healthcare systems look digitally mature. You'll commonly find:
- DHIS2 for national reporting and analytics
- OpenMRS/ OpenClinic for clinical records
- RapidPro for messaging workflows
- SMS gateways and mobile reporting systems
- Laboratory information systems
- Surveillance dashboards
- Community reporting channels
On paper, this looks modern.
But inside operations, a different reality appears
- Data gets entered multiple times across disconnected systems
- Reporting chains still depend on Excel handoffs
- Surveillance workflows run on delayed batch syncs, not real-time streams
- Manual validation loops eat operational time
- Alerts exist but aren't wired to a specific response action
- Lab systems don't automatically update surveillance platforms
- Field teams re-enter data manually into centralized dashboards
The systems exist. The orchestration layer doesn't. [1]
Research on integrated care systems consistently shows this isn't a technology gap; it's a coordination failure between tools built in isolation. [2]
What Rwanda's eIDSR quietly revealed
One of the most instructive examples I found was Rwanda's electronic Integrated Disease Surveillance and Response system(eIDSR) built on DHIS2.
The system is genuinely impressive. It captures disease surveillance data in near real-time, generates outbreak alerts, and supports epidemic response coordination.
But what stood out wasn't the technology.
It was what the implementation documentation openly acknowledged. [3]
Even after successful deployment, the system still faced:
- Interoperability friction with laboratory systems
- Fragmented reporting workflows at the facility level
- Follow-up on data inconsistencies
- Excessive manual validation steps
- Disconnected response tracking; alerts without linked actions
- Operational burden on frontline workers
In other words, even a strong digital health system struggles when workflows remain fragmented across disconnected operational layers.
That's not a software problem. That's a systems integration problem.
The real problem isn't data collection
Healthcare has no shortage of tools. The real bottleneck is what happens between them.
Most operational inefficiencies happen at the seams:
- Between labs and surveillance systems
- Between clinics and reporting layers
- Between field workers and centralized dashboards
- Between alerts and actionable response workflows
When systems don't talk, humans fill the gap. Clinicians, data clerks, and coordinators become what I'd call the "human API layer" manually moving information between disconnected platforms.
At scale, that creates invisible operational load: [1]
- Transcription errors
- Duplicated effort
- Delayed surveillance loops
- Incomplete follow-ups
- Reporting fatigue
- Inconsistent decision signals
What I started building
This is where my work began, not replacing existing systems, but building the orchestration layer that makes them work as one.
The stack I'm working with:
- OpenMRS · DHIS2 · RapidPro · Africa's Talking
- Custom n8n nodes as the workflow bridge
- Event-driven routing replacing batch sync
- FHIR-normalized data before it hits the orchestration engine
- Reusable interoperability pipelines for fragmented environments
The shift in practice looks like this:
A real workflow walkthrough is coming in the next post.
Not because people worked harder. Because the systems stopped fighting each other.
Beyond healthcare
Healthcare is the first domain I'm exploring publicly, but the same fragmentation patterns appear across:
Fintech for fragmented payment settlement workflows
Agriculture for siloed field data and supply chain reporting
Logistics for disconnected supply chain visibility layers
Public sector operations ....
Different industries. Same orchestration problem.
Why OSS matters here
These ecosystems already run on open-source infrastructure. That's the advantage because it means interoperability doesn't have to be locked behind expensive proprietary middleware.
With open tools, we can build:
- Reusable connectors between platforms
- Shared workflow architectures
- Event-driven automation layers
- Interoperable public infrastructure
- Africa-first operational systems designed for real constraints
That's the part that genuinely excites me.
Final thought
Systems thinking isn't about building more dashboards. It's about reducing the invisible friction between the platforms we already depend on.
Sometimes the biggest innovation isn't a new platform. It's finally making existing platforms communicate properly.
Top comments (0)