19.3 Evaluating CDI Professional Performance

Evaluating CDI Professional Performance

A useful performance review asks three connected questions: Did the specialist complete the assigned work? Was that work accurate and compliant? What conditions helped or hindered the result? Review volume alone cannot answer all three. A specialist conducting complex clinical validation, teaching providers, and reconciling codes has a different workload from a colleague assigned only initial screening.

AHIMA's CDI KPI practice brief recommends definitions and benchmarks suited to the department's objectives and responsibilities. ACDIS also describes substantial variation in productivity expectations. Published surveys and individual programs' policies can inform a comparison, but they do not establish a national ICU, medical-surgical, pediatric, or obstetric daily review quota.

Define the work before comparing staff

Write a metric dictionary that identifies the activity, counting unit, period, exclusions, data source, and responsible owner. An initial review is the first substantive review of an encounter; a follow-up review evaluates subsequent information. A post-discharge review, query follow-up, coding reconciliation, education session, and peer audit are distinct activities. Count them according to the approved definitions rather than allowing every chart opening to inflate a review total.

Use an observed workload study to establish realistic expectations. Sample different units, experience levels, record formats, and review purposes. Include time spent developing queries if that activity is part of the review assignment. Reassess expectations when staffing, technology, responsibilities, or patient complexity changes. AHIMA recommends periodically revisiting the benchmark; a number borrowed from another organization does not become a professional mandate.

Worked comparison: counts and weighted units

In a hypothetical day, a specialist completes 8 initial reviews and 12 follow-ups during 6 available chart-review hours. Raw activity is 20 reviews, or about 3.33 reviews per review hour. Keep the component counts visible: another specialist's 20 initial reviews are a different workload.

A department might pilot a locally weighted index:

$\text{Weighted units per review hour} = \frac{\text{Initial reviews} + w(\text{Follow-up reviews})}{\text{Available review hours}}.$

If a local time study supports $w=0.5$, the example produces $(8+0.5\times12)/6=2.33$ weighted units per review hour. This is not 2.33 initial reviews or 3.33 weighted units. The one-half weight is an explicit local assumption, not an AHIMA/ACDIS conversion factor. Follow-ups involving new organ dysfunction or conflicting documentation may take longer than an uncomplicated initial review. Validate the weight before using it in evaluations.

Pair productivity with quality evidence

Quality review should examine actual records and queries against applicable coding rules, the current query practice brief, and approved policies. A small scorecard can organize the evidence without turning locally selected thresholds into national requirements.

For example, 34 records with confirmed missed opportunities among 100 audited records give a 34% sample missed-opportunity rate. Before interpreting it, determine whether records were selected randomly, targeted because of suspected errors, or drawn from a particular service. A targeted result is useful for correction but cannot be presented as an unbiased estimate for all reviewed encounters.

Neither a 90%–95% compliance score nor an 80%–90% DRG match rate is a universal safe harbor. Local goals require defined methods and a response to actual defects. A high match rate can coexist with two reviewers making the same coding error. A low match rate can reflect an outdated working DRG, a procedure not included in the specialist's assignment, or later information. Reconcile discrepancies using the complete record and authoritative rules, accepting whichever result is supported, including a lower-paying DRG.

The query audit must apply the current format requirements. Multiple-choice options should be clinically supported and include an open alternative for the provider's judgment. Do not require an invented universal trio of “Other,” “Clinically Undetermined,” and “Ruled Out” in every query. Timing, re-querying, and escalation also require patient-specific review rather than judgment from a count alone.

Evaluate causes fairly and act on findings

Suppose a specialist has high review counts and a concerning audit result. Confirm the underlying queries, data extraction, sample, and reviewer agreement. Address confirmed defects promptly. Then investigate possible causes: training needs, unrealistic assignments, template design, software counting, inadequate review time, or noncompliant conduct. The rates alone do not prove deliberate gaming or superficial scanning.

A practical improvement plan identifies the specific error, the applicable rule, who will correct it, what support is needed, and how follow-up review will determine whether the change worked. Balance individual accountability with system responsibility. Compare staff performing similar assignments and account for onboarding, leave, and delegated education or audit duties.

Incentives must protect independent judgment

AHIMA's CDI ethical standards require documentation integrity regardless of financial impact and prohibit practices intended to inappropriately increase payment or distort data. ACDIS similarly prohibits directing providers solely for financial or quality-reporting impact and supports using reliable performance metrics. These principles do not categorically prohibit every financial performance measure, bonus, CMI measure, or query-volume measure.

A proposed bonus based on dollars from CC/MCC additions creates a serious risk that reviewers will pursue a desired answer, neglect downward clarification, or treat diagnoses as quotas. Bring the proposed design to compliance and appropriate leadership for review and redesign before implementation. Use accuracy, compliant review practices, appropriate coverage, education, and sustained correction as performance evidence. Financial reporting can describe program effects; it cannot justify unsupported diagnoses, leading queries, or pressure to change a provider's judgment.

An incentive arrangement by itself does not establish False Claims Act liability. The applicable statutory conduct, knowledge, and other requirements must be evaluated on the facts. The operational response is to prevent and correct improper practices, preserve evidence, and escalate concerns through the compliance process—not to pronounce automatic liability from a scorecard or compensation formula.

For CDIP decisions, choose the management action that combines defined workload, direct quality review, fair interpretation, and corrective follow-up. Raising a count or a payment measure cannot substitute for accurate documentation.