An EMR demonstration takes place on good office broadband. The clinic where it will run has an unreliable mains supply, a generator with a fuel budget, and a connection that degrades in heavy rain.
The gap between those two environments is where most of the operational risk in a Nigerian EMR project sits, and it is rarely on the comparison sheet.
The questions to ask, precisely
Vendors will say the system "works offline". That phrase covers several very different things, so ask specifically:
- Can a clinician open an existing patient record with no connection, or only create new notes?
- Can the front desk register a new patient and take payment offline?
- What happens to a queue of offline changes if the device is closed or restarted before syncing?
- If two people edited the same record while offline, how is the conflict resolved, and who is told?
- How long can it run offline before it stops accepting work?
- Is offline available on every device, or only the mobile app?
The answers vary enormously between products that all claim the same capability.
Power is a separate problem from connectivity
They are often confused. A system that syncs beautifully is useless if the workstation is off. Practical measures:
- An uninterruptible power supply on the server or main workstation, sized to allow a graceful shutdown rather than a hard cut, which is what corrupts data.
- Laptops rather than desktops at the desk, because a laptop is its own battery backup.
- Tablets or phones for clinical staff, which survive an outage entirely.
- A router on backup power, since losing the router loses the network even when the computers are alive.
Have a paper fallback and rehearse it
Every clinic needs a defined downtime procedure: which forms are used, where they are kept, who enters them afterwards, and by when. Written down, printed, and stored somewhere that does not require electricity to reach.
Then rehearse it. A downtime procedure nobody has practised is a document, not a plan. Run it deliberately once, on a quiet afternoon, and find out what is missing before the day it matters.
Know where your data physically is
For a cloud system, ask where the servers are, because distance affects responsiveness on a constrained connection. For a locally hosted system, ask what happens if the machine in your building is stolen, flooded or fails, which is a different risk profile and needs a different backup answer.
Neither answer is automatically better. What matters is knowing which risks you have accepted, deliberately, rather than discovering them during an outage.