Healthcare interface migration reconciliation is the documented process of proving that data moved between source and destination systems without losing clinical context, patient identity, or operational control. The work differs sharply by interface type: an EHR-to-EHR transition, a payer API exchange, and an imaging archive migration each define completeness, ownership, and acceptance differently. Reconciliation therefore requires named inputs, outputs, matching rules, exception owners, and cutover triggers before transfer begins. This article examines those distinctions using documented standards and vendor examples, then proposes a hypothetical validation scenario that a migration team could adapt without implying that any product has been independently tested.
Record Ownership and the Reconciliation Boundary
Reconciliation begins with a precise statement of which system owns the record at each stage. In imaging migration, Dicom Systems describes cutover as a clinical-readiness decision rather than the moment the final batch finishes, and it states that the health system retains control of acceptance criteria, rollback triggers, and decommissioning decisions [4]. That framing matters because reconciliation thresholds are operational commitments, not vendor defaults. A source PACS or VNA may remain authoritative for retrieval during coexistence, while the destination becomes authoritative only after approved proceed criteria are met.
For clinical document exchange, ownership is expressed through document templates and patient identifiers. The ONC certification companion guide for clinical information reconciliation requires a user to reconcile active medication, allergy, and problem list data from at least two sources in a single view, with source and last modification date shown as minimum attributes [2]. The MiHIN discharge medication reconciliation guide adds a stricter boundary: a C-CDA containing only a social security number root OID will be rejected, and participating organizations must send a properly formatted patient identifier from the facility or managing organization [7]. Ownership is therefore enforced before reconciliation can occur.
Inputs, Outputs, and Version Boundaries
The inputs to reconciliation are not interchangeable. ONC references C-CDA templates for Continuity of Care Document, Referral Note, and, in the inpatient setting only, Discharge Summary, with a time boundary of December 31, 2025 for certain adopted standards [2]. The same guide lists SVAP-approved versions of the HL7 CDA R2 Implementation Guide for C-CDA Templates for Clinical Notes, including Edition 4.0 and the STU Companion Guide Release 4.1 [2]. A migration team must record which template and version each source system emitted, because a later version may alter required sections or identifier handling.
Payer APIs introduce a different version boundary. CMS documentation lists HL7 FHIR Release 4.0.1, US Core IG STU 3.1.1 and 6.1.0, and SMART Application Launch Framework Release 1.0.0 and 2.0.0 for the Patient Access and Provider Access APIs [3]. Several adopted standards and derived implementation guides expired on January 1, 2026, while other guides are recommended under the proposed CMS-0062-P rule [3]. Reconciliation across a payer interface must therefore distinguish adopted from recommended versions and record which expiration date applies to each transaction set.
A Hypothetical Validation Scenario for Interface Reconciliation
A migration team could validate reconciliation by staging a synthetic patient cohort with known identifiers, medications, allergies, problems, and linked imaging studies. The team would emit a C-CDA from the source EHR, route DICOM objects from the source archive, and send a payer API test transaction, then compare each destination record against the staged manifest. For the C-CDA path, the team would confirm that the patient identifier root OID is present and that medication, allergy, and problem lists display source and last modification date in a single view, as ONC describes [2]. For imaging, the team would check count variance by object type and route any identity or metadata exception to an authorized reviewer [4].
The scenario should include a deliberate mismatch: one patient with a different medical record number in the source and destination, and one study with a broken report link. The team would document whether the mismatch is routed to manual review rather than auto-merged, and whether the broken link triggers a hold condition before cutover [4]. This is a suggested validation exercise, not a report of independent testing, and it does not establish that any vendor product would pass it.
Patient Matching as a Reconciliation Control
Patient matching is the identification and linking of one patient's data within and across health systems, accomplished at a minimum by linking demographic fields such as name, birth date, phone number, and address [5]. In migration, matching is also a reconciliation control because false matches can merge unrelated clinical records. Dicom Systems advises that when systems use different medical record numbers or assigning authorities, the migration should preserve source context, apply approved matching rules, and route uncertain records to authorized review before adding them to the destination clinical record [4].
The MiHIN guide shows how matching rules become concrete field mappings. It compares ADT PID-5.2 against the CCD patient given name path and sets a medication reconciliation indicator when the match succeeds [7]. This is a narrow, documented example of deterministic matching for one exchange. It does not establish a universal patient matching standard, and organizations with probabilistic matching or multiple assigning authorities will need separate rules and manual review queues that the MiHIN example does not address.
Exception Remediation and Cutover Evidence
Reconciliation produces exceptions, and exceptions require owners. Dicom Systems describes a remediation workflow that detects an exception through automated validation, reconciliation, or user-reported retrieval testing; classifies it as identity, metadata, DICOM conformance, missing content, duplicate content, report linkage, routing, or target-display failure; assigns an owner and severity; and corrects the rule or record through an authorized process before reprocessing [4]. This classification step is the difference between a reconciliation report and a remediation program.
Cutover evidence should be rehearsed against agreed triggers. Dicom Systems lists proceed conditions such as users retrieving current and prior studies from intended workflows with source fallback available, and hold or rollback triggers such as critical retrieval failures, unexplained losses, duplicate identifiers, broken report links, or missing related objects [4]. The same source recommends monitoring count variance by modality, department, object type, and date range, plus open exception aging and remediation throughput [4]. These metrics are evidence of control, not proof of clinical safety.
Interface Boundaries Across EHR, Payer, and Imaging Domains
The three domains reconcile different objects. EHR-to-EHR transition literature identifies data migration accuracy, methods, electronic versus manual entry, and technical difficulties as a distinct challenge domain [1]. ONC certification focuses on reconciling medication, allergy, and problem lists from transition of care documents [2]. Payer API documentation focuses on standards and implementation guides for Patient Access, Provider Access, Payer-to-Payer, and Prior Authorization APIs [3]. Imaging migration focuses on DICOM object validation, metadata exceptions, and report linkage [4].
A hospital survey data brief adds a measurement boundary: it distinguishes standards-based APIs from other methods such as proprietary EHR APIs or HL7 interfaces when hospitals receive or integrate data for clinical and administrative purposes [6]. That distinction matters for reconciliation because a proprietary interface may not expose the same source and modification attributes that a standards-based API provides. The survey data comes from the 2024 AHA IT Supplement, and the brief itself does not report reconciliation outcomes or error rates [6].
Sources & references
- Transitions from One Electronic Health Record to Another: Challenges, Pitfalls, and Recommendations - PMC · www.ncbi.nlm.nih.gov
- 170.315(b)(2) Clinical information reconciliation and · healthit.gov
- Application Programming Interfaces (APIs) and Relevant Standards and Implementation Guides (IGs) | CMS · www.cms.gov
- Healthcare Imaging Data Migration: A Seven-Phase PACS and VNA Guide — Dicom Systems · dcmsys.com
- Patient Identity and Patient Record Matching - ONC - Office of the National Coordinator for Health Information Technology · healthit.gov
- Hospital Use of APIs to Enable Data Sharing between EHRs and Third-Party Technology - ONC Health IT Research & Analysis · healthit.gov
- Exchange C-CDA: Discharge Medication Reconciliation - Inbound Implementation Guide 3.17.2026 · mihin.org