Patient matching is the quiet failure underneath most interoperability programmes. Two records for one person are a clinical risk; one record for two people is worse.
Matching is probabilistic by nature in the absence of a nationwide identifier, and most organisations run it with defaults nobody has reviewed since go-live. The result is a match weight set by a vendor, a threshold nobody can justify, and no record of what was merged or why.
The work is to make the rules explicit and auditable. Which demographic fields participate, how each is normalised before comparison, what the deterministic rules are, where the probabilistic band sits, and what happens to a merge: is it reversible, who approved it, what survives the unmerge.
I also design for the two failures that matter. A false negative sends a clinician to an incomplete record and they make a decision on half the history. A false positive merges two people and the correction is expensive and sometimes impossible. The design has to treat those asymmetrically rather than optimise one accuracy number.
fixed-scope work with a written output. Assessment first, then a sequence of changes, with the criteria for each step agreed in advance so the outcome is measurable rather than asserted.
The common failure is a matching threshold tuned once and never revisited. Source systems change, demographics change with a new registration workflow, and the match rate drifts for a year before anyone measures it.
Engagements run statewide from Las Vegas: Reno, Sparks, Carson City and the rural counties, where a critical-access hospital often has one IT director covering infrastructure, security and the EHR. Remote delivery is the default; travel is done when the work genuinely needs a room with the people in it. Statewide work usually involves the state Medicaid programme, which changes the interface requirements materially compared with commercial payers.
Scope is a sequence of work with a measurable acceptance criterion per step. Each step is signed off before the next begins.
Charles Trinh works from Las Vegas, Nevada and delivers remotely across the United States. Engagements are scoped in writing, fixed in fee where possible, and measured against the interface or control they were meant to produce. Where the work touches patient data, the data-handling terms are agreed before access is granted.
Should we use a commercial enterprise master patient index? If the volume and number of sources justify it. Below a certain size the rules matter more than the product, and a well-specified ruleset over an existing store outperforms a badly configured EMPI.
What match rate is good enough? A single number is the wrong target. What matters is the split between false positives and false negatives and whether the organisation has agreed which it can tolerate, because the remediation cost of the two is not the same.
How do we handle a patient who has records in two systems? Deterministically, with a documented merge that can be reversed, and with the patient informed. Silent merging is how an organisation ends up unable to reconstruct what its own record said last month.
What does this engagement produce? A written design or assessment tied to a specific interface, control or revenue workflow, with the evidence behind each finding.
How is it delivered? Remotely, with a scope call first and a fixed fee agreed before work begins.
Working with organisations across Nevada and the surrounding area — on site, remote, or both.