Build an Article 31 internal chart review sample that tests documentation continuity, location differences, staff handoffs, and corrective actions.
A useful internal chart review sample tests actual clinic workflows and preserves enough selection evidence to explain what the results do and do not represent. Pulling the easiest charts to open may make a review faster while missing the handoffs the clinic most needs to understand.
This guide describes an internal quality review approach for New York Article 31 MHOTRS programs. It is not a formal audit protocol, a legal opinion, or a statistical sampling method. Follow any applicable external review instructions separately and involve qualified specialists where the purpose requires them.
Article 31 chart audit sample: start with the question
Define what the review is trying to learn. A documentation completeness check, a treatment-planning review, a satellite workflow check, and a billing reconciliation ask different questions. Combining them without a clear scope can produce a long list of findings that nobody knows how to prioritize.
The OMH MHOTRS programmatic guidance separates case records, assessment, treatment planning, staffing, and other program topics. Use the relevant sections with the clinic's policies to build a review instrument appropriate to the question.
Write the review purpose at the top of the worksheet. State the date range, population, exclusions, and intended use of the results. This prevents a limited operational exercise from later being described as proof that every chart meets every requirement.
Define the population before selecting records
Identify the records that could have been selected. For an active-treatment review, the population might be active episodes during a specified period. For a discharge workflow review, it might be episodes closed during that period. Those populations will not necessarily produce the same findings.
Keep a dated record of the selection method. If the clinic deliberately chooses exception cases, label the review as targeted. If it uses a broader selection method, document that method accurately. Do not describe a convenience sample as random or statistically representative.
Preserve selection evidence in an authorized location. The clinic should be able to explain why each case was included without exposing patient identifiers in a general project document. Limit access to the people carrying out the review and approved follow-up.
Include workflow variation deliberately
Consider the operational differences that might affect documentation. Examples include primary and satellite sites, new and established episodes, clinician reassignment, different service types, and amended records. Use these dimensions to make blind spots visible.
A fictional sample design might include a new admission at the primary site, an established patient at a satellite, a transferred clinician assignment, a recently discharged episode, and a record with a late amendment. This is a targeted learning exercise, not a recommendation for a universal number of charts.
Keep subgroup conclusions modest. Finding a problem in one satellite chart does not establish the rate across that location. It does identify a workflow worth investigating and may justify a focused follow-up under the clinic's quality process.
Review continuity instead of isolated forms
Follow an identified need from assessment to plan to documented intervention and response. Ask whether the sequence makes sense and whether the relevant version of the plan can be identified for the service date. A set of completed forms can still tell an unclear clinical story.
Inspect authorship, dates, and amendments without treating a timestamp alone as a measure of clinical quality. If the system displays the current plan beside an older encounter, determine whether the reviewer can find the historical version that applied at the time.
Use a structured finding format: observed fact, applicable reference or policy, impact on the review question, and recommended next investigation. Separate what the reviewer saw from their interpretation. That makes disagreements easier to resolve and reduces vague findings such as poor documentation.
Keep billing and clinical findings distinct
The OMH MHOTRS billing and fiscal guidance provides a separate billing reference. If the review includes charges, have billing staff define and assess the relevant checks rather than asking a clinical reviewer to infer payment rules.
A record can have an unresolved clinical documentation issue, a billing issue, both, or neither. Classify findings accordingly. A signed note does not by itself establish coverage, and a paid claim does not prove that the clinical documentation meets every applicable expectation.
When a finding crosses teams, name one coordinating owner while preserving the separate decisions. For example, a wrong location may need operational correction, clinical amendment, and billing review. Treating it as a single generic error can leave one of those actions unfinished.
Calibrate reviewers before comparing results
Have two reviewers assess the same small set of authorized or fictional records. Compare findings and discuss disagreements about definitions, evidence, or scope. A review tool that produces inconsistent interpretations is not ready to support individual performance comparisons.
| Review question | Calibration issue to resolve |
|---|---|
| Is a current plan identifiable? | What counts as current for the service date? |
| Is a note complete? | Which policy and service-specific elements apply? |
| Is an amendment understandable? | What evidence distinguishes correction from a new decision? |
| Is a handoff resolved? | Which role accepts responsibility and where is that recorded? |
Revise the worksheet when calibration exposes ambiguity. Keep a version number and effective date so later reviews can identify which definitions were used. Otherwise an apparent trend may reflect a changed instrument rather than changed performance.
Convert findings into a tested improvement
Assign each finding a responsible owner and a next action. Distinguish a record-specific response from a process change such as training, configuration, or supervision. Follow the clinic's authorized correction policies and preserve the original record history.
After the change, retest the relevant workflow with a new sample appropriate to the question. Look for the intended result and possible unintended effects. A new mandatory field, for example, may improve completeness while encouraging copied text that does not explain the individual encounter.
Ease Health publishes this guide and sells behavioral health software. Ask vendors to demonstrate how an authorized reviewer finds historical plans, amendments, and the underlying encounter evidence. Related material includes the EHR overview, treatment plan review workflow, and appointment reconciliation guide.
The deliverable is a review whose scope, selection, findings, and follow-up are understandable. That is a stronger basis for improvement than a reassuring percentage with an unexplained denominator.
Related reading
Frequently Asked Questions
Is there one sample size that makes an Article 31 chart audit valid?
No universal sample size is established by this guide. Define the purpose, population, selection method, and applicable external instructions. Seek appropriate specialist input when statistical inference or a formal audit methodology is needed.
Should an internal review include only completed charts?
That would exclude some important workflow problems. Depending on the review purpose, include active episodes and relevant exceptions such as clinician transfers, missing documentation, amendments, and satellite encounters.
Can an internal review establish that the clinic will pass an external audit?
No. Internal review supports improvement and readiness. It does not guarantee an external finding, and external reviewers may use different scopes, methods, and standards.
What should happen after a chart review finding?
Assign a responsible owner, distinguish record-specific action from a systemic issue, document the authorized response, and retest whether the process change worked. Preserve the original finding and its resolution history.
