A study arrives at PACS with the wrong patient name. Or the right name but the wrong date of birth. Or the right patient but the accession number from a completely different exam. For a radiologist, a wrong demographic is more than an annoyance — it is a potential patient safety event. For the IT team, it means manual correction, potential report re-issue, and a lot of explaining to clinical governance.
Demographics mismatches are almost never random. They follow predictable patterns, and understanding those patterns is the first step to eliminating them.
The Four Ways Demographics Get Wrong
1. Worklist Not Used — Manual Entry
The technologist types patient details directly into the modality instead of selecting from the MWL. Typos in names, wrong dates, transposed digit in the MRN. This is the most common source of mismatches in sites with poor MWL adoption.
The fix is policy and enforcement, not just technical. Configure modalities to require MWL selection where possible. Report weekly on "studies without accession number" as a proxy metric for manual entry.
2. Wrong Worklist Item Selected
The technologist selects the wrong patient from the worklist — scanning SMITH, JOHN but selecting SMITH, JANE because the list is sorted in a way that puts similar names together.
The fix is UI configuration. Configure worklists to show enough disambiguation fields (DOB, Patient ID, Procedure) to prevent selection errors. Some modalities support an ADT query at selection time to confirm identity.
3. ADT Feed Not Populating the RIS Correctly
Demographics in PACS are only as good as demographics in the RIS, which are only as good as the ADT feed from the EMR. If the HL7 A08 (Update Patient Information) messages aren't flowing, a name correction in the EMR never reaches the RIS or the modality worklist.
Check the Patient's Name tag and Patient ID tag in arriving images against the RIS record. Systematic differences (e.g., name always in wrong case) indicate a transformation issue in the HL7 pipeline.
4. Patient Identity Not Resolved Before Imaging
Emergency patients sometimes arrive before registration is complete. The modality receives an order for "Unknown, Unknown" with a temporary MRN. The exam is performed under the temporary ID. When the patient is subsequently identified and their record is updated in the EMR, an A40 (Merge) message should flow to PACS to reassign the study — but often doesn't.
The Correction Workflow
When a mismatch is identified, the correction process depends on the severity:
Category A — Minor typographic error (single character, no clinical ambiguity)
- PACS admin corrects the demographic in the PACS patient index
- Documents the correction in the change log
- Notifies radiologist if the report has already been issued
Category B — Wrong name or DOB (patient still correctly identified by other fields)
- Check the original MWL item in the RIS — was it wrong there too, or did the modality introduce the error?
- If RIS was wrong: trace back to the HL7 ADT feed and identify the root cause
- Correct PACS patient index and link the study to the correct patient
- If a report has been issued: radiology department must decide if amendment/re-issue is required
- Document in incident management system
Category C — Study potentially attributed to wrong patient
This is a clinical incident. Escalate immediately to the radiologist, the clinical governance lead, and follow your organisation's patient safety reporting procedure. Do not attempt to correct silently.
The 3 Tags That Define Patient Identity in DICOM
Every PACS patient identity issue traces back to these three tags:
(0010,0010) — Patient's Name
Format is Family^Given^Middle^Suffix^Prefix. Casing inconsistencies between systems (SMITH vs Smith) cause duplicate patient records in PACS if the matching algorithm is case-sensitive. Define a canonical casing policy and enforce it at the HL7-to-MWL transformation layer.
(0010,0020) — Patient ID
The primary matching key. Leading zeros, spaces, and system prefixes are the most common inconsistency. 00123456, 123456, and MRN-123456 will create three separate patient records in a PACS that uses exact-match Patient ID logic.
(0010,0030) — Patient's Birth Date
Used as the tiebreaker when Patient IDs collide. A wrong DOB (e.g., 19000101 for unknown) effectively disables birth date as a matching key.
Check our DICOM tag reference for the full attribute specification for each of these tags.
Building a Prevention Strategy
1. MWL compliance monitoring. Track the percentage of incoming studies with a populated Accession Number (0008,0050). Target > 98% compliance. Any study without an Accession Number was either manually entered or picked up incorrectly.
2. Demographic consistency checks at ingestion. Configure your PACS or ingestion gateway to compare Patient ID and Patient Name in arriving DICOM files against the corresponding order in the RIS. Alert on mismatches instead of silently storing wrong data.
3. Pre-admission demographic verification. For scheduled outpatients, have the scheduling team verify demographic accuracy in the RIS against the EMR record the day before the appointment. Catch issues before the patient arrives.
4. A40 merge processing verification. Test and periodically re-test that A40 merge messages are correctly processed by PACS. Run a monthly audit: pick 10 recently merged patients in the EMR and verify their imaging history is unified in PACS.
5. Post-go-live demographics audit. In the first 60 days after any new modality or interface go-live, run a daily report of studies with missing or non-standard Accession Numbers and Patient IDs. Catch systematic issues early.
Demographics accuracy is a clinical quality issue, not just an IT hygiene issue. Build it into your go-live acceptance criteria, your change management process, and your ongoing operational metrics.
Related: PACS Go-Live Checklist · Modality Worklist Troubleshooting · Patient ID tag reference