The quick answer: Single-system EHR tools are good at deduplicating patients within one system, but they can't see across two different EHRs used by affiliated clinics in the same referral network — that's an interoperability gap, not a data-entry problem. A tool that processes patient panels entirely client-side (never uploading PHI to a server) is the safest way to reconcile that gap and de-identify the shared result, without a full interoperability platform project.
Why This Is a Different Problem Than Single-System Dedup
Most EHR systems handle duplicate-patient detection reasonably well within their own database — same system, same patient ID scheme, straightforward matching. The harder, less-discussed problem shows up when two affiliated practices in the same referral network each use a different EHR: each system assigns its own patient ID, so the same real person can appear as two completely unrelated records when the two systems are compared.
That's not a data-entry mistake at either clinic. It's a structural interoperability gap — and it means a genuinely useful comparison has to happen across the two exports, matched on something more forgiving than patient ID (typically name plus date of birth, tolerant of minor formatting differences), rather than inside either system alone.
The Options
1. Your EHR's Built-In Dedup
What it does: Flags likely duplicate patients within a single system's own database.
Sohovi automatically finds every duplicate in your dataset — including near-matches — and shows you exactly which rows are affected.
Where it breaks down: Has no visibility into a second, independent EHR used by an affiliated clinic. If the overlap you're trying to find is between two systems, not within one, this tool literally cannot see the other side.
2. A Full Health Information Exchange (HIE) / Interoperability Platform
What it does: Standardizes and continuously synchronizes patient records across many participating organizations at scale, typically via FHIR or similar standards.
How it differs: Built for permanent, ongoing exchange across a large network of participants — a genuinely valuable long-term investment for a large health system.
Dealbreaker: Implementation timeline and cost are built for a large, ongoing multi-organization initiative, not a smaller practice group that periodically needs to know "how much overlap actually exists between these two clinics' patient panels."
Sohovi gives you the data quality picture you need to make the case for fixing it — and to track improvement over time.
3. Sohovi's Reconciliation + De-Identify Tools — Best for Periodic, Point-in-Time Reconciliation
What it does: Compare two patient-panel exports — one from each clinic's EHR — matched flexibly on name and date of birth rather than each system's own ID, surfacing exactly which patients overlap and where the name or date formatting differs between the two records. Once reconciled, the free de-identify tool applies tunable k-anonymity (rather than blunt Safe Harbor field-stripping) before the merged list is shared further.
Best for: Practices that need a periodic, defensible answer to "how much patient overlap exists between these locations" without standing up permanent interoperability infrastructure — and want the comparison to happen without any PHI leaving the browser to do it.
Dealbreaker: This is a point-in-time comparison, not continuous, automated synchronization — for that, an HIE-grade platform is genuinely the right long-term tool.
Comparison
| | EHR built-in dedup | HIE / interoperability platform | Sohovi reconciliation + de-identify | |---|---|---|---| | Sees duplicates within one system | Yes | Yes | Yes (via export) | | Sees overlap across two different EHRs | No | Yes | Yes | | Continuous, automatic synchronization | No | Yes | No (point-in-time) | | PHI ever leaves the local environment | Depends on system | Depends on architecture | No — browser-only | | De-identifies with tunable k-anonymity (vs. blunt Safe Harbor) | No | Varies | Yes | | Setup time | Already in place | Months | Same day |
How to Choose
- Is the overlap you're worried about within one EHR, or between two different systems used by affiliated clinics? If it's between two systems, your EHR's own dedup tool has no way to see it — you need a cross-export comparison.
- Do you need this continuously synchronized, or periodically checked? Continuous, large-scale synchronization across many organizations is what an HIE platform is for. A periodic, defensible check between a handful of affiliated locations doesn't require that scale of investment.
- Do you need to share the reconciled list further — for a joint quality initiative, a referral analysis, or research? De-identify it with tunable k-anonymity first, which preserves more analytical usefulness than Safe Harbor's blanket date/geography stripping while still passing a k-anonymity check.
Frequently Asked Questions
Q: Is uploading patient data to a browser-based tool actually HIPAA-safe? The relevant distinction is whether the raw data is transmitted to a server at all. Sohovi processes files entirely client-side — the raw file and rows never leave the browser, only the score and results are saved — which is a meaningfully different exposure profile than uploading PHI to a cloud-hosted service. Confirm this fits your practice's own HIPAA compliance policy and, if required, your Business Associate Agreement posture before using any tool with real patient data.
Q: Why not just use Safe Harbor de-identification like we already do? Safe Harbor's approach — stripping dates to year-only, generalizing geography — is simple but blunt, and can still leave a meaningful share of records re-identifiable in practice while destroying the ability to do seasonal or regional analysis. Tunable k-anonymity is a more defensible middle ground for many use cases, but check which standard your specific research or sharing context actually requires.
Q: Does this replace our EHR vendor's interoperability features? No — if your EHR vendor already offers a direct interoperability connection to the other clinic's system, that's likely the better long-term answer. This is for the common case where no such connection exists yet, and the two exports are what you actually have to work with today.
If you need to know how much patient overlap actually exists between two affiliated clinics' EHRs — without either clinic's PHI leaving a browser tab to find out — both the reconciliation and de-identify tools are free at sohovi.com/tools.