ABA schedule export reconciliation proves that a defined schedule cohort left the source, reached the intended destination, and retained the correct identity, version, status, time, assignment, and permitted details. The process locks the expected rows, compares each transformation, tracks secure delivery, confirms downstream use, and keeps missing, duplicate, altered, and rejected rows open until an owner resolves them.

Define the export's purpose

An operational roster, payroll file, clinician calendar, payer submission aid, family calendar, and analytics feed serve different purposes. Name the recipient, approved use, minimum fields, covered period, source version, export time, format, delivery route, retention rule, and owner before generating the file. ABA schedule export reconciliation starts with that contract. Reusing one broad export for every destination increases exposure and invites downstream teams to treat fields as authoritative beyond their intended scope.

Lock the expected cohort

Capture the exact visit identifiers and count expected at the extraction cutoff. Record which statuses qualify, how same-day edits behave, and whether canceled, held, tentative, or deleted visits appear. A moving query cannot support a fair comparison because the schedule may change between extraction and review. If the business needs a live feed, reconcile by sequence or event window and preserve a watermark. Keep records that fail extraction visible as source exceptions rather than removing them from the denominator.

Specify every field transformation

Build a map from source field to destination field, including data type, format, time zone, allowed values, default, null handling, truncation, and sensitivity. Stable IDs should survive whenever the destination supports them. Translate statuses through an approved crosswalk and identify lossy transformations. For example, combining held and canceled into inactive may work for one report while failing an operations handoff. Give each derived value a documented rule, version, test case, and owner so reviewers can reproduce it.

Control access and delivery

Classify the data and recipient before sending it. For HIPAA covered entities and business associates, the HHS Security Rule overview frames safeguards for electronic protected health information. Use approved encrypted routes, named recipients or service accounts, least-role access, authentication, delivery logging, and retention controls. Avoid shared links with indefinite access. When a vendor or downstream service receives protected information as a business associate, confirm the applicable agreement and security responsibilities through qualified review.

Compare at three levels

At the file level, compare checksum, size, row count, generation time, and schema. At the row level, match stable identifiers and detect missing, duplicate, or unexpected records. At the field level, compare critical values and approved transformations. A row-count match can hide one missing row and one duplicate. A successful delivery receipt can confirm transport while saying nothing about content. Keep these tests separate so a green transport result does not close a data discrepancy.

Handle corrections and reruns

Define whether a corrected export replaces the entire file, supplies a delta, or creates a new version under the destination's accepted process. Give each run a unique identifier and mark superseded files so they cannot be mistaken for current work. When the destination already processed a faulty row, open a separate downstream correction task and preserve its result. Avoid editing a delivered file in place because its checksum and recipient evidence would no longer match. The reconciliation record should link the original run, discovered defect, corrected source or rule, replacement run, recipient acknowledgment, and final business disposition. This lineage matters when payroll, billing, or client communication has already acted on the first file.

Design for failed delivery

A secure transfer can fail because of authentication, file size, unavailable service, expired certificate, invalid schema, or recipient-side processing. Define retries, backoff, duplicate prevention, escalation, and a final delivery deadline. Verify whether a retry creates a new artifact ID or reuses the first one, then reconcile the destination accordingly. Keep the locked row cohort unchanged across technical retries. If the approved fallback is a manual transfer, apply the same recipient, encryption, logging, retention, and row-comparison controls. Record the actual route used so later reviewers do not attribute the result to an automated path that never completed.

Review data minimization after each use

After the recipient confirms the intended use, ask whether every exported field remains necessary. Remove unused sensitive columns through a new approved mapping version and retest the destination. Review saved copies, temporary folders, email attachments, vendor storage, and analyst workbooks under the applicable retention and deletion rules. A broad extract often survives long after its original purpose, increasing both confusion and exposure. Data minimization also improves reconciliation because reviewers can focus on the fields that actually drive the destination's decision. Record why any free-text or clinical field is included and prefer stable coded values when they meet the purpose.

Protect clinical and schedule meaning

The BACB Ethics Code supports accurate documentation, confidentiality, and qualified professional decisions for covered people. An export may carry a clinician's recorded decision, yet the export process does not create or reinterpret it. Preserve authorship, date, effective period, and source. When a destination cannot represent a supervision, access, safety, or clinical status, hold the relevant use or provide a governed companion field rather than compressing the distinction away.

A fictional reconciliation

Riverstone ABA exports 180 visits for a two-week payroll review. The file contains 180 rows, but stable-ID matching finds 178 expected visits, one duplicate, and one unexpected canceled visit. Exact row yield is 178 of 180, or 98.9%. Field checks then find two duration conversions that round incorrectly. The practice corrects the export rule and regenerates the full locked cohort. The second file matches all 180 expected visits with no extras, and the first file is marked superseded.

Confirm downstream receipt and use

Ask the destination to confirm the file identity, accepted row count, rejected row count, processing time, and resulting version. Reconcile any proprietary response to the original IDs. When the destination changes a value, require a reason and preserve both states. Confirm that a superseded file cannot be processed later. If the export supports payroll, billing, or client communication, reconcile that downstream result under the separate rules for that function rather than treating file acceptance as final business correctness.

Close exceptions with evidence

Give every missing, duplicate, unexpected, rejected, or altered row a category, severity, owner, next action, and due date. Correct the owning source when its fact is wrong; correct the export rule when transformation is wrong; correct the destination when its import is wrong. Preserve the original evidence and link the resolution. Closing an exception through deletion alone loses the learning needed to prevent recurrence. Sample closed rows in the next run to confirm the repair persisted.

Measure reliability by run and field

Report reconciled expected rows divided by the locked cohort, missing and duplicate rows, critical-field accuracy, delivery and processing completeness, exceptions by age, superseded-file controls, and recurrence by rule version. Segment by export type and destination because a payroll file and family calendar carry different fields and risks. Pair percentages with counts and severity. One wrong client identity or visit time deserves direct investigation even when the overall export rate rounds to 100%.

Related resources

Sources