Blog
Migrating Legacy Hospital Software: Data Transfer and Cutover Plan
How to plan a hospital software migration: what data to move, how to choose a cutover model, what goes wrong, and when to bring in outside help.
- bottom
- hospital
If your hospital is replacing old software, the short answer is this: the new system is rarely the hard part. The hard part is deciding which data to move, proving it arrived correctly, and choosing how to switch over without stopping patient services. Plan for a rehearsed, measured migration with a clear rollback option, not a single heroic weekend.
What you are actually moving
“Migrating data” sounds like one task. In a hospital it is several, and each has a different owner:
- Patient master data: names, identifiers, contact details, and the duplicates that years of registration have created.
- Clinical and visit history: past visits, orders, results and notes. Decide how much needs to be live in the new system and how much can stay in a read-only archive.
- Financial data: open invoices, deposits, insurance and corporate account balances, and price lists.
- Master lists: doctors, departments, tariffs, drug and item catalogues, and user accounts.
- Stock and asset data: inventory balances, batches and expiry dates.
Not everything has to move. A common and sensible trade-off is to migrate current patients and open balances fully, and keep older history searchable in an archive. That keeps the project smaller and the cutover shorter.
Choose a cutover model deliberately
Industry guides describe three common approaches, and each has a real cost (overview of cutover strategies):
| Model | Advantage | Downside |
|---|---|---|
| Big bang (everything switches at once) | Shorter timeline, no temporary links between old and new | Highest risk: a serious problem affects the whole hospital |
| Phased (by module or department) | Risk is contained, lessons carry into later phases | Longer project, temporary links between systems, staff working in two systems |
| Parallel running (both systems live) | Strong safety net, outputs can be compared | Double data entry, staff fatigue, highest cost |
For most private hospitals, a phased approach works better than it first sounds: for example, start with finance and inventory, then registration and outpatient, then inpatient and the more complex departments. Parallel running is realistic for narrow areas such as billing, and rarely for the whole hospital. A big bang can still make sense for a small hospital with a simple setup, but only with a rehearsed plan.
Rehearse before you commit
The most useful practice in published migration guidance is the dry run. Vendor and case-study material consistently describes test migrations as the way to find mapping errors and performance problems while there is still time to fix them, and to measure how long the real transfer will take so you can size the cutover window (migration guidance). A published hospital example describes two fully played-out dress rehearsals before the switch (T-Systems case flyer). Treat these as vendor-published material, but the logic holds: if you have not rehearsed it, you do not know how long it takes.
A practical sequence:
- Extract data from the old system into a staging area.
- Clean it: merge duplicate patients, standardize spellings, retire inactive items and map old codes to the new system’s lists.
- Load into a test copy of the new system.
- Validate with counts, totals and samples. Have finance reconcile balances and have clinical staff check high-risk records themselves.
- Repeat until the results are stable, then schedule the real cutover.
What goes wrong
- Duplicate and dirty patient records. Moving them as they are means the new system starts with the old system’s problems.
- Forgotten connections. Labs, imaging, pharmacy, payment and accounting links each need to be listed early. Missing one tends to show up on day one.
- Undocumented workarounds. Years of informal practices, custom fields and spreadsheets sit beside the old system. Staff interviews usually find more than the documentation does.
- Finance does not reconcile. If opening balances differ between old and new, trust in the new system erodes quickly.
- Training too late. Staff who first see the system at go-live will fall back to paper or to the old system.
- No rollback plan. Decide in advance what conditions mean “we go back”, and who decides.
Plan for the first days after go-live
Go-live is not the finish line. Budget for extra support in the first weeks: a visible help desk, named super users in each department, a daily issue list and a clear route for urgent fixes. Keep the old system available in read-only mode for a defined period so staff can look things up.
Data protection and records
Patient data is sensitive, and a migration copies it into staging areas, test environments and sometimes vendor hands. Indonesia’s personal data protection law (Law No. 27 of 2022) treats categories such as medical information as “specific personal data”, which carries higher risk, and commentary on the law notes that large-scale processing of such data calls for a data protection officer and an impact assessment, though “large scale” is not yet defined (Norton Rose Fulbright summary). In practice:
- Limit who can access migration copies, and delete test copies when finished.
- Put data-handling terms in writing with any vendor that touches the data.
- Check how long medical records must be retained before deciding what to archive or discard. Retention rules are set by health regulations and have changed over the years, so confirm against the current official text with your legal or medical records team.
This is general information, not legal advice.
Cost and effort
Costs vary too widely by hospital size, data quality and number of connected systems for any honest single figure. The main cost drivers are: how messy the data is, how many departments go live, how many connections to other systems must be rebuilt, how much training is needed, and whether you run parallel for a period. Dirty data and unclear ownership cost more than the software licence in many projects. Ask any vendor to price the migration separately from the software, and to state what happens if a rehearsal fails.
When to get help
Get outside help if you do not have a project owner with authority across departments, if nobody can say what the old system actually contains, if the old vendor is uncooperative about data export, or if you are combining a new hospital system with Odoo for finance, inventory or HR and need the two to agree on opening balances and master data. A short assessment of the old data before you sign a contract is cheap compared with discovering the problems mid-project.
Talk to us
Teknologia Solutions is an Indonesia-based software and Odoo consultancy. We offer our own clinic and hospital management systems, and our hospital system can integrate with Odoo, and we also build custom software. If you are weighing a migration and want a second opinion on scope, cutover model or data readiness, we are happy to have a no-charge conversation. There is no obligation.
Sources: Implementation and cutover strategies; EHR data migration, validation and cutover; T-Systems hospital conversion flyer; Norton Rose Fulbright, Indonesia PDP law. Several sources are vendor material or secondary commentary. Verify legal and record-retention points against the primary regulations and with qualified advisers.