To prevent duplicate patient records, every patient should receive a unique patient identifier at their first visit, and registration staff should always search for an existing record before creating a new one. The identifier alone is not enough: you also need a rule for entering names, fixed search criteria and a procedure for merging the duplicates already in your system. Here is the method, adapted to the realities of facilities in the region.
Key takeaways
- An identifier assigned once, never reused and meaningless in itself (no date of birth, no initials).
- A systematic search before any new record is created, using at least two criteria.
- A written data-entry rule for surnames, other names and approximate dates of birth.
- A patient card or booklet showing the number, issued at the first visit.
- A regular duplicate review by a named person.
Why duplicate records are a problem
A duplicate is one patient registered under two or more records. Their history sits in one file and their latest lab result in the other. A doctor who opens the wrong record decides without knowing about an allergy or current treatment, when the whole point of an electronic patient record is to bring that information together.
The consequences go beyond care. Billing is scattered, claims to insurers become hard to justify and statistics are distorted: a returning patient is counted as a new one. Example: a 40-bed clinic in Lagos believes it registers many new patients each month, when some of them are known patients registered again because nobody searched first.
Common causes of duplicates in the region
Duplicates do not come from careless staff. They come from everyday situations the procedure did not anticipate.
| Cause | Typical situation | Remedy |
|---|---|---|
| Variable spellings | The same name spelled two ways by different staff | Check against an ID document, also search by phone and date of birth |
| Surname and first name swapped | The first name typed in the surname field | Written rule: surname first, in capitals |
| Everyday name | The patient gives a name that differs from their ID | Record the name on the ID, add the everyday name |
| Unknown date of birth | The patient only knows an approximate year | One written convention, applied by everyone |
| Shared phone | Several family members give the same number | Never identify a patient by phone alone |
| Unnamed newborn | Registered as “baby of” the mother, never updated | Mandatory update as soon as the name is known |
Building a reliable unique patient identifier
Rules for the identifier
A good unique patient identifier is assigned once, never reused, even after a death, and carries no information about the patient. A number that includes the date of birth or initials becomes wrong as soon as that data is corrected.
National ID numbers as a complement
When the patient has a national identification number, such as the NIN issued by the National Identity Management Commission in Nigeria or the Ghana Card number issued by the National Identification Authority, it can be recorded as an extra search criterion. The NHIS membership number in Ghana or NHIA enrolment details in Nigeria can play the same role. None of these replaces your facility’s identifier: not every patient carries them or knows them.
The patient card or booklet
Issue a card or booklet with the patient’s number at the first visit. Example: a health centre in Ho writes the number on the cover of the patient’s health booklet and asks them to bring it to every visit. Finding the record then takes seconds.
A 6-step registration routine to prevent duplicates
- Ask whether the patient has been to the facility before, even long ago, even in another department.
- Ask for the patient card or booklet. If the number is there, open the record and check the name and date of birth.
- Without a number, search using at least two criteria (name and date of birth, then phone), trying similar spellings.
- Check identity against an ID document when the patient has one.
- Create a new record only if every search has failed, following the written data-entry rule.
- Issue or update the patient card with the assigned number.
This routine takes a little time at the desk, but saves much more later. It fits into the organisation described in our article on reducing waiting times at reception.
Detecting and merging existing duplicates
Even with a good routine, duplicates will appear, especially while migrating old registers as described in our method for moving from paper registers to electronic records. You therefore need a regular review, handled by one or two trained people.
- Flag suspect records: same name and date of birth, or same phone and a similar name.
- Compare both records and call the patient if in doubt.
- Choose the main record, usually the oldest.
- Bring the information together in the main record and deactivate the other, deleting nothing.
- Log the merge: records involved, reason, author and date.
Example: in a district hospital, the records officer spends one hour every Friday on this review. He also reports recurring errors to registration, which fixes the cause rather than the symptom.
Patient identification and MediFlow
MediFlow does not yet offer patient records: this module is planned. Today, MediFlow covers on-call management (rosters, attendance, lateness, handovers, operational checks) and the stock room (catalogue, stock movements, inventories, batches, expiry dates, reorder alerts, stock-out forecasts).
The identification procedure in this article applies to your registers right now: duplicates fall because of the data-entry rule, more than the tool.
Frequently asked questions
Can we use the phone number as the patient identifier?
No. One number can be shared by several family members, and patients change numbers. The phone is a good search criterion, combined with name and date of birth, but not an identifier.
What if the patient does not know their exact date of birth?
Adopt one convention, written into the registration procedure, for example recording the estimated year and flagging the date as approximate. What matters is that every staff member applies the same rule, otherwise the same patient will be registered with different dates.
Who should be allowed to merge two records?
One or two named and trained people, often the head of registration or the medical records officer, with a doctor’s opinion when clinical data is in doubt. Every merge must be documented: records involved, reason, author and date.
Want to see how MediFlow fits your facility? Request a free demo.