Why Patient Matching Matters
Accurate patient identification underlies every downstream health information function, from billing to clinical decision support. A single duplicate or overlaid record can cause medication errors, delayed diagnoses, and reimbursement denials. The RHIA exam expects candidates to understand how the master patient index (MPI) uses matching algorithms to keep one patient tied to one unique identifier across every encounter and facility.
Deterministic Matching
Deterministic algorithms compare specific identifying fields, such as last name, date of birth, and Social Security number, and require an exact or near-exact match on a defined set of attributes. This method is straightforward and fast but is vulnerable to typographical errors, nicknames, and missing data, which can create false negatives and unnecessary duplicate records.
Probabilistic Matching
Probabilistic algorithms assign weighted scores to multiple data elements based on the statistical likelihood that a match is correct. Fields that are highly unique, like a full Social Security number, carry more weight than common fields like a last name. Records that score above a defined threshold are considered matches, records below a lower threshold are considered non-matches, and records that fall between the thresholds are flagged for manual review by an HIM professional.
Duplicate and Overlay Errors
A duplicate record occurs when a single patient is assigned more than one medical record number. An overlay error occurs when two different patients are merged under a single identifier, which is far more dangerous because it can mix clinical data between two individuals. HIM departments must run regular duplicate detection reports and follow a documented merge and unmerge procedure to correct these errors without losing data.
Best Practices for Data Quality
- Standardize data entry fields at registration to reduce variability
- Use address and phone validation tools at the point of intake
- Conduct routine MPI integrity audits with a defined duplicate rate threshold
- Establish a data quality management team responsible for MPI governance
- Train registration staff on the downstream impact of identification errors
Exam Tip
Expect scenario-based questions asking you to distinguish between a duplicate and an overlay, and to identify which type of algorithm would best resolve ambiguous matches. Remember that probabilistic matching is generally preferred in large health information exchanges because it tolerates data variability better than strict deterministic rules.