Moving a small hospital off paper: a sequence that works

Small hospitals moving off paper rarely fail because the software was wrong. They fail because the rollout ignored how a hospital actually runs — continuously, with tired people, at night, when nobody from the vendor is reachable.

The sequence below is deliberately conservative, and it is conservative for reasons that have been earned.

Start where the transaction is simplest

Registration and OPD billing.

High volume, short transactions, immediate feedback, and a mistake is easily corrected. Staff learn quickly because they repeat the same action a hundred times a day, and the value is visible within a week.

Starting with IPD instead — the intuitive choice, since it is where the money is — means learning a complex, long-running, multi-department process while still learning the software. It is the commonest sequencing error.

Run parallel, but not for long

Parallel running is a safety net and it is also a trap. Two systems means double work, and under pressure people abandon the new one — which is the wrong one to abandon.

Two weeks is usually enough for OPD. Set the end date before starting, and stop on it unless something is genuinely broken. An open-ended parallel period never ends.

Plan for the night shift specifically

The gap nobody budgets for.

Training happens during the day. Support is available during the day. Then a night-duty nurse hits something unfamiliar at 2am with no one to ask, and does it on paper — correctly, because the patient comes first.

The problem is what happens next: that paper record either gets entered later or does not, and either way there is now a gap.

Three things help. Train the night staff separately rather than expecting a handover. Give them a documented fallback that captures what happened for later entry. And make someone reachable at night during the first fortnight, even informally.

Migrate less than you think

The instinct is to bring across all historical records. It is expensive, slow, and produces messy data that undermines confidence in the new system.

A more practical position: migrate active patients and current stock, keep historical records accessible in their existing form, and let history accumulate naturally going forward.

Hospitals that insist on full historical migration usually spend months on it and then find the migrated data is not trusted anyway.

Expect the second week to be the hardest

Week one runs on novelty and vendor presence. Week two is when the vendor is less visible, the novelty is gone, and staff hit the cases training did not cover.

Planning support presence for week two rather than week one is a small change that meaningfully affects whether adoption holds.

The sequence

1. Registration and OPD billing — two weeks parallel, then stop
2. Pharmacy — starting with OPD dispensing, then IPD issue
3. Investigations — order to bill
4. IPD — admission, charges, continuous billing
5. Everything else — once the above is habitual

Each stage should be genuinely settled before the next begins. Settled means staff no longer ask how, not that it technically works.

The one thing that predicts success

Whether someone inside the hospital owns it.

Not the vendor, not the doctor who is interested in technology. Someone with authority who decides how a process should work and makes that decision stick when a department pushes back.

Hospitals with that person succeed with mediocre software. Hospitals without it struggle with excellent software.

If you want a rollout scoped this way, see MediNexus Lite or talk to us about your sequence.