Component errors hidden by a global fidelity score occur when correct completion of many routine steps offsets failure on one clinically important part of the protocol. Break implementation into operational components, report each numerator and denominator, and classify error type and consequence. A 90% global score can coexist with zero performance on a critical step, so qualified review should never rely on the average alone.

Define meaningful components

Break the protocol into observable steps tied to clinical logic, access, safety, or the proposed mechanism.

Score each component

Record correct steps divided by eligible opportunities for that component rather than one session-wide checkbox.

Classify error type

Separate omission, incorrect action, sequence, timing, dosage, and added nonprotocol behavior.

Identify critical steps prospectively

Name components whose failure requires hold, coaching, safety action, or clinical review before seeing the score.

Report global and component views

Keep the overall summary for orientation while displaying every critical or weak component beside it.

Build Sol's component-level integrity audit

For the component errors hidden by global fidelity score question, create a versioned component-level integrity audit. Preserve the target, client priority, operational definition, observation state, service and entry times, author, numerator, denominator, missingness, graph, integrity protocol, component data, context, access, decision rule, qualified owner, snapshot, correction, implementation, and follow-up. Another reviewer should be able to reconstruct Sol's evidence and decision without guessing which values were available.

Work through Sol's data example

Sol's ten-step teaching protocol is scored in one session. Nine steps occur correctly, producing 9 of 10, or 90% global fidelity. The missed step is the partner's required response to Sol's accessible stop message. That component is 0 of 1 and triggers immediate coaching and plan review despite the high total. Show every count, denominator, state, and date before summaries. This fictional clinic teaching-protocol review example illustrates one workflow and does not establish a universal maturity threshold, fidelity target, review frequency, plan change, or treatment recommendation.

Audit Sol's evidence trail

Sol's audit stores all ten steps, opportunity definitions, one missed component, error type, client response, observer, timing, severity, feedback, and follow-up. It does not average the stop-message failure into an approval state. The audit also checks definition and protocol versions, source-record access, correction history, graph axes, session spacing, invalid states, observer evidence, calculation precision, review permissions, and downstream dependencies. Unresolved discrepancies remain visible and hold the exact decision they affect.

Address Sol's main data risk

Component lists can become overly granular and burdensome. Sol's clinician defines steps that matter to treatment logic, communication access, safety, and interpretation and tests whether observers can score them reliably. A metric or alert can surface a concern. Qualified reviewers interpret measurement, outcome, integrity, client experience, context, and risk together. One score cannot establish treatment fit, clinical importance, causation, authorization, or completion.

Choose Sol's next action

The team restores the required partner response, provides focused training, and samples the component again. It also reviews whether the protocol or environment made the step difficult to perform. Record the action, rationale, owner, due date, support, and review trigger. Keep preliminary evidence, finalized evidence, treatment integrity, outcome, client input, and clinical decisions as separate states so one cannot silently substitute for another.

Protect Sol's access and participation

Keep Sol's AAC, interpreters, mobility, food, water, bathroom use, prescribed care, health support, rest, relationships, and emergency help available. Use accessible consent and assent processes when applicable and respond to withdrawal, dissent, or distress. Data collection, fidelity observation, or review timing never authorizes staff to delay urgent care or remove ordinary supports.

Apply current sources to Sol's review

Sol's page uses direct research showing that global treatment-integrity measures can mask important component errors. The BACB ethics hub and CASP public summary provide professional context, while the BCBA Test Content Outline identifies examination content on measurement, integrity, and data-based decisions. The WWC handbook supplies research-design context. Research on graphing fidelity with rate, integrity reporting, and integrity effects shows why implementation evidence matters. A single-case design review and evidence-based ABA framework describe analysis and decision context. ASHA supports continuous AAC access.

Rehearse Sol's workflow

Test the component-level integrity audit with fictional preliminary records, a late correction, a missing denominator, no integrity opportunity, measured zero fidelity, high fidelity with low use, a critical component miss, an overdue review, and a no-change decision. Confirm that states, due cohorts, calculations, snapshots, permissions, alerts, and qualified routes behave as intended. Store expected results, software version, reviewer notes, and corrections before live use.

Close Sol's data review

Review the component-level integrity audit with Sol, the responsible clinician, and specialists required by the question. Preserve source data, versions, graphs, maturity status, integrity coverage, components, client input, access and safety evidence, decision, implementation, corrections, and later outcomes. Keep the page draft and noindex until every manifest-named review is complete.

Related resources

Sources