ABA Scheduling Data Quality Audit

Define fields and authoritative sources

Create a data dictionary for each field, allowed value, owner, source, effective-date rule, validation, and downstream consumer. A clinical plan, payer record, workforce record, room calendar, and client communication preference may each control different fields.

Document each field contract

A field contract should include name, definition, type, allowed values, null meaning, source, transformation, effective dates, update owner, validation, consumers, retention, and access rule. Use one stable identifier across systems where possible and record approved mappings when identifiers differ.

Avoid generic approved, active, or ready flags. Store which authority approved what, for which person, service, provider, location, modality, date range, and source. This makes stale or overbroad values easier to detect.

Lock a representative sample

Select records before reviewing errors. Cover sites, services, payers, modalities, time bands, staff roles, recurring and one-time visits, changed records, and known risk areas. Keep excluded records and reasons visible. Use a full cohort for high-risk releases when sampling would miss material harm.

Choose audit tests by risk

Test identity, date and time, service, provider, supervisor, setting, modality, payer state, access support, client notice, release state, version, and downstream reconciliation. Add targeted tests after system changes, imports, incidents, or repeated errors.

Record the population, sample method, sample size, period, tests, source versions, auditor, and limitations before opening records. A convenience sample of easy visits should not be described as representative.

Test source-to-schedule accuracy

Compare exact values and dates. The BACB Ethics Code addresses documentation and accountability for covered professionals. HealthCare.gov cautions that preauthorization does not promise cost coverage, so schedule, authorization, and payment states need separate fields.

Classify errors without changing the source

Use missing, invalid, stale, conflicting, wrong mapping, wrong scope, unauthorized change, failed synchronization, or downstream mismatch. Preserve the observed value and evidence. An auditor can identify the discrepancy, while the authorized owner determines and records the correction.

Rate severity by immediate safety, access, service, pay, payer, privacy, and operational consequence. Route urgent issues immediately while the broader audit continues. Do not delay a required action until the sample report is finished.

Route corrections by authority

Operations can correct operational fields under approved policy. Qualified clinicians correct clinical records and decisions within scope. Payer, privacy, payroll, access, and workforce owners resolve their evidence. Preserve original value, corrected value, reason, author, date, version, and downstream impact.

Use correction episodes and retests

Create one episode per source error and affected set. Link all visits sharing that exact cause. Record containment, approved correction, application, downstream systems, people notified, and retest. Keep individual exceptions visible when the common fix does not apply.

A retest should compare the corrected value with the authoritative source and confirm that unrelated fields stayed unchanged. Reopen the episode when a downstream consumer still holds the old value.

Govern the sample and evidence trail

Write the audit population, period, unit, inclusion and exclusion rules, risk strata, selection method, and expected source before drawing records. Include random items for an ordinary error estimate and targeted items for known high-risk conditions, then report them separately. Preserve selected identifiers and unavailable records. Replacing a missing item with a convenient one changes the sample and can hide the very control failure the audit is meant to find.

For each test, retain the schedule value, authoritative source value, source timestamp, comparison rule, outcome, reviewer, and correction link. A second reviewer should reproduce material calculations and sample a subset of conclusions. Protect client and workforce information in the audit file through role-limited access and minimum necessary detail for the operating purpose. The report should allow a later reviewer to understand the evidence without editing the source records.

Fix the creation and propagation control, not only the field

Trace each material error backward to where the value was collected, approved, entered, transformed, copied, synchronized, displayed, and used. A wrong time zone may originate in intake, a default, an interface, or a report conversion. A stale authorization date may reflect a missing update, failed queue, overwritten source, or an unclear ownership rule. Correct the affected records through their authorized routes while keeping the audit finding intact.

Choose prevention that matches the cause: structured input, clearer ownership, validation, reconciliation, versioning, access changes, training, interface monitoring, or a removal of duplicate entry. Define the affected cohort and test the change on both ordinary and failure paths. Then resample after a meaningful period. Closure requires evidence that the control works and the known cohort is reconciled, not simply that one displayed value now looks right.

A fictional audit sample

Brookside ABA locks 60 visits. Fifty-two match every tested authoritative field. Four have outdated location values, two have provider mismatches, one lacks an access-support field, and one has a stale payer state. First-pass accuracy is 52 of 60, or 86.7%. All eight errors remain in corrective follow-up.

Report record and field errors separately

Eight visits fail at least one tested field, so record-level first-pass accuracy is 52 of 60. If a visit had two wrong fields, failed-field count would exceed eight while failed-record count would stay eight. Report both when useful.

For each field, use records tested for that field as the denominator. Access-support completeness may apply only to a defined cohort, but the missing field still requires correction and a review of whether the sampling rule captured everyone who needed support.

Retest and reconcile

Confirm corrected fields across schedule, notifications, staff time, documentation, authorization usage, charges, and claims. Report accuracy by field, error age, correction completion, repeat error, affected visits, service loss, and downstream impact. A corrected dashboard value is incomplete until every affected system and person is addressed.

Turn audit findings into prevention

Identify whether errors came from unclear ownership, stale source data, interface mapping, manual entry, permissions, weak validation, bulk changes, or delayed updates. Assign a preventive control and test it on future records. Avoid responding to a system cause only with staff reminders.

Version the data dictionary and audit logic. Compare repeat errors under the same rule and maintain an exception route for legitimate cases. Close the audit only after urgent consequences and corrective episodes have accountable owners.

Owner audit questions

• Is the population locked with clear units, period, source systems, exclusions, and risk strata?
• Are random and targeted samples reported separately and reproducibly?
• Does every field have a source, owner, definition, valid range, freshness rule, and correction route?
• Can a second reviewer reproduce material tests without changing the underlying record?
• Are record error rates, field error rates, unavailable evidence, and corrected items kept in distinct denominators?
• Did remediation address creation and propagation controls before the follow-up sample?

The audit closes only after the known affected cohort reconciles and a later test shows the preventive control working. A cleaner display alone does not establish better data.

Related resources

• ABA Schedule Demand Heatmap
• ABA Weather Disruption Scheduling Plan
• ABA Staffing Capacity Buffer Policy
• ABA Travel Delay Response Workflow

Sources

• Council of Autism Service Providers, Organizational Guidelines public overview
• Behavior Analyst Certification Board, Ethics Code for Behavior Analysts
• HealthCare.gov, Preauthorization glossary