Hospital systems are evaluated on module coverage and bought on price, then discovered to be inadequate on the handful of things that are structural rather than functional. These are the ones worth being uncompromising about.
Patient identity that holds over time
A returning patient must be findable — by name variants, by phone, by old record number. Duplicate registrations are the most common data problem in hospitals and they are clinically dangerous, because a clinician looking at one record does not know another exists.
Insist on: duplicate detection at registration, a merge process that preserves both histories, and searchable legacy identifiers. Retrofitting identity management onto a system with years of duplicates is a project in itself.
An audit trail on clinical records
Who recorded what, when, and what it said before it was changed. Records should be amendable with the original preserved, not overwritten.
This matters for clinical safety, for any dispute, and for accreditation. A system allowing silent edits to clinical notes is not acceptable, and this is not a setting that can be added later — it depends on how the data model was built.
Interfaces, named and demonstrated
Ask which specific lab analysers, radiology systems and pharmacy dispensing units are already integrated — by manufacturer and model, not "we support HL7".
Standards support means it is possible. An existing, working integration with your equipment means it will happen without a project. The difference is months.
Where an interface does not exist, ask what building one costs and who has done it before.
Billing that handles how you are actually paid
Few hospitals are paid a single way. A system needs to handle cash, insurance, corporate panels and government scheme patients — frequently within one admission.
- Package rates versus itemised, and the point where one converts to the other
- Scheme-specific rate cards and pre-authorisation workflows
- Split billing where a scheme covers part and the patient pays the balance
- Claim documentation assembled from the record rather than retyped
Scheme billing is where hospitals lose money through claim rejection, and the cause is usually documentation assembled by hand under time pressure.
Statutory registers, produced not exported
Birth and death registers, notifiable disease reporting, blood bank records, controlled substance registers. Each has a prescribed format and a legal deadline.
Ask to see each one generated. "Exportable to Excel and formatted manually" is a recurring staff cost and a compliance risk when someone is on leave.
Role-based clinical access
A pharmacy assistant should not see full clinical notes. A visiting consultant should see their own patients. Access needs to be granular by role and by relationship to the patient, and it needs to be logged.
Blanket access with a shared login is common in smaller hospitals and is both a privacy exposure and an accountability gap.
Availability expectations, written down
A hospital runs at 3am. Ask what support exists overnight, what the response commitment is, and what happens during an outage — is there a documented downtime procedure with paper forms that reconcile into the system afterwards?
Every hospital system will be unavailable at some point. The ones that cope have agreed in advance what happens.
About the author. Kislay Chandan has worked with schools since 2014. Rather than gathering requirements from the outside, he worked inside a school to see what actually goes wrong day to day, and built what he found into the software. He still works with Shemrock Glitters in Patna — most changes to the product start with something that came up there.
Gallery
