Blog
Migrating Patient Records to a New System Without Losing Data
How Indonesian clinics can move patient records to a new system safely: what to clean, test and verify, what it costs, what goes wrong, and when to get help.
- bottom
- clinic
Short answer: you can move patient records to a new system without losing data, but only if you treat it as a project with a test run, a checklist and a rollback plan, not a file copy on a weekend. Most losses come from skipped cleaning, unchecked field mapping and switching off the old system too early.
This guide is for clinic owners, medical directors and operations or IT managers who are about to leave a spreadsheet, a paper-plus-Excel setup, or an older clinic application.
What “migration” actually involves
Migration means extracting data from the old source, reshaping it to fit the new system’s structure, loading it, and proving it arrived intact. The data usually includes:
- Patient demographics and contact details
- Visit history and clinical notes
- Prescriptions, allergies and diagnoses
- Invoices, payments and outstanding balances
- Scanned documents, lab results and images
- Appointment history and future bookings
Each of these has a different shape and a different risk. Demographics are easy to move. Free-text clinical notes and attachments are where problems hide.
Step 1: Inventory what you have
Before talking to vendors, list every place patient data lives: the old application, Excel files on staff laptops, WhatsApp photos of lab results, paper folders, a separate billing tool. Many clinics discover that “the system” is really four systems.
Decide for each source whether it will be migrated, archived read-only, or left behind. Not everything needs to go into the new system. Ten-year-old paper charts may be better scanned and attached than retyped.
Step 2: Clean before you move
Industry guidance on health-record migration is consistent on one point: moving unclean data just relocates the mess. Typical problems in Indonesian clinics:
- The same patient registered two or three times, with spelling variations of the name
- Inconsistent date and phone formats
- Missing ID or contact details
- Free-text fields where the new system expects a structured choice
- Different code or naming conventions for the same service or drug
Budget real staff time for duplicate review. A tool can flag likely duplicates, but a person who knows the patients should decide on merges, because wrongly merging two patients’ records is worse than leaving a duplicate.
Step 3: Map the fields explicitly
Mapping is the written agreement of “this column in the old system goes to this field in the new one”. Gaps appear here: the old system has a field the new one lacks, or the reverse. Decide in advance what happens to data with no home. Options are an extra custom field, a notes field, or an archived export kept alongside.
Guides also stress that terminology and codes must be mapped, not assumed to match between systems.
Step 4: Run a full test migration
Ask for a complete trial load into a test copy of the new system before any go-live date is fixed. Guides treat mapping, validation and testing before go-live as core phases, often repeated, so fix mapping gaps before the trial load.
Then verify in two ways:
- Counts and totals. Number of patients, visits and invoices before and after. Outstanding balance totals should match to the rupiah.
- Spot checks by people. Have a doctor, a nurse and a front-desk person each open 20 to 30 real patients and compare against the old system: allergies, current medications, last visit, balance. Guidance notes that migration QA should involve real clinicians, not just scripts comparing record counts.
Expect to repeat the test at least once. A first trial that is perfect is unusual.
Step 5: Choose your cutover style
| Approach | How it works | Trade-off |
|---|---|---|
| Big bang | Everyone switches on one day | Fastest, but needs a tested fallback |
| Phased | One branch, department or function at a time | Safer, takes longer, two systems for a while |
| Parallel run | Both systems used for a period | Lowest risk, highest staff effort and double entry |
For a single small clinic, a big-bang switch over a quiet weekend is common. For multi-branch groups, piloting one branch first is usually worth the extra weeks.
Step 6: Keep the old system, read-only
Do not delete or cancel the old system on go-live day. One vendor guide suggests keeping it in read-only mode for at least 90 days after cutover before decommissioning. Problems often surface only when a returning patient with a complicated history comes in.
Also take a full, verified backup of the old data and store it securely before you start.
Retention and privacy
Two points to check with your own advisers, since we are not a law firm.
Retention. Secondary sources report that Permenkes 24/2022 on medical records sets a minimum of 25 years from the patient’s last visit for electronic records. The practical lesson: do not discard old data casually during a migration. Confirm the current rule against the regulation text.
Personal data. Indonesia’s Personal Data Protection Law (Law No. 27 of 2022) treats health information as “specific” personal data, and commentators note it may attract additional obligations. For a migration this means: limit who can touch the export files, avoid sending patient spreadsheets over personal email or chat, use encrypted transfer, delete temporary copies afterwards, and make sure any vendor handling the data is bound by a written agreement. Verify details against the primary text and, if needed, with a lawyer. A law-firm summary is a reasonable starting point but not a substitute.
Costs and timelines
We will not quote numbers that vary this much. Cost and duration depend on data volume, how messy the source is, how many sources there are, whether scanned documents move too, and how much staff time you can free up.
What is reliably true is where the money goes:
- Staff time for cleaning and testing (often the largest hidden cost)
- Vendor or consultant time for mapping, scripting and repeat trial loads
- Training and a temporary productivity dip for the first week or two
- Possible parallel-running costs for two systems
Ask any vendor for a written scope that states how many trial migrations are included and what happens if the data turns out to be messier than expected.
What typically goes wrong
- Duplicates carried over, so staff cannot tell which record is current
- Truncated or misformatted fields, such as long notes cut off or dates swapped between day and month
- Attachments left behind because only the database was moved
- Open balances and future appointments missed, which patients notice immediately
- Old system switched off too soon
- No named owner, so cleaning and sign-off fall between departments
- Staff not trained on where migrated data now appears
When to get help
Handle it in-house if you have one source, a few thousand patients, and someone with time and care. Bring in help if you have multiple branches, several data sources, years of unstructured notes or images, a legacy system with no export function, or no one who can own the project. If the target system is custom or heavily configured, mapping and trial loads are the part to put in writing.
Talk to us
If you are planning a move and want a second opinion on scope, risks or vendor quotes, Teknologia Solutions is happy to have a no-charge conversation. We build custom software and also offer our own clinic management system, so we can discuss whether a ready-made, custom or Odoo-based route fits your clinic, with no obligation.
Sources: Mindbowser, EHR data migration; Thinkitive, EHR data migration guide; Tilleke & Gibbins, personal data protection and healthcare services in Indonesia; CMS, compliance with Indonesia’s PDP Law; Legalita journal article on medical record retention. Several sources are vendor or secondary commentary. Verify retention periods and legal obligations against the primary regulation texts before acting.