E/M Level Shift After an EHR Template Update
Northside Multispecialty has 45 providers and uses a common EHR. After an EHR upgrade six months ago, level-4 established-patient visits rose from 32% to 58% of encounters, and level-5 visits rose from 6% to 14%. The new template auto-populates review of systems and exam findings from prior notes; clinicians must actively delete text. Billing staff add modifier 25 to about 40% of same-day office visits with a procedure. A payer issues an additional documentation request for 20 claims. The compliance officer audits 40 randomly selected claims: 18 lack documentation supporting the billed level, 7 have a supportable level but wrong modifier, and 15 are supported. Two high-volume providers account for 12 of the 18 unsupported claims. No refunds have been made. The COO asks whether the group should repay the payer immediately or wait for the payer’s audit results.
1. Analyze the audit results. What can and cannot be concluded, and how should the compliance officer quantify the potential overpayment?
Read the worked answer
The audit cannot be extrapolated to all claims without a statistically valid sample and a defined population. Forty claims is a small convenience sample; 18 unsupported is a red flag, but the confidence interval is wide. Separate error types: level unsupported (18), modifier error (7), and supported (15). The modifier errors may be overpayments or underpayments depending on the procedure and payer rules. Quantify the difference between billed and supported level for each claim, add modifier-related differences, and annualize only as an estimate, not a liability. Identify the population - all claims since the upgrade, by provider and payer - stratify by provider and level, and use a random sample with a documented methodology. Review the two high-volume providers separately but do not assume intent. Check whether the template auto-population caused cloned documentation; compare note timestamps and copy-forward patterns. The compliance officer should also determine whether the payer’s 20-claim ADR overlaps the sample. Potential overpayment is not the same as confirmed overpayment; confirm with a clinician coder and calculate per claim. Document the methodology, findings, and limitations.
2. Develop a root-cause and corrective action plan covering templates, provider education, modifier use, and monitoring.
Read the worked answer
Root causes include template auto-population and copy-forward, unclear modifier rules, production pressure, insufficient prebill review, and no provider feedback loop. Corrective actions: turn off auto-population of review of systems and exam or require active attestation, and add an I-reviewed-and-edited prompt; revise templates to require medical decision-making support; re-educate all providers on E/M level selection and modifier 25, which requires a significant, separately identifiable E/M service, not just a same-day visit; require diagnosis-procedure linkage review; add prebill edits for level-5 and modifier 25 claims, with 100% review for 90 days for the two high-volume providers; audit 30 claims per provider per quarter, then reduce based on results; monitor level distribution monthly against a baseline and peer benchmarks; and document training attendance and audit results. Assign owners and dates. The plan should distinguish correction of past claims from prevention of future claims. Include a feedback loop to the compliance committee and a re-audit within 90 days to test effectiveness.
3. Advise on repayment and payer communication, including whether to wait for the payer audit. Explain the reasoning and the risks of each option.
Read the worked answer
Do not wait passively for the payer audit, but do not repay blindly either. The organization has an obligation to return identified overpayments within 60 days of identification for Medicare and many payers; state and commercial rules vary. The internal audit identified 18 unsupported claims and 7 modifier errors, but the amount is not yet calculated. First, calculate the overpayment per claim and validate with a second coder. Second, decide whether the 40-claim sample is representative enough to estimate a broader refund; a voluntary refund of the confirmed claims is defensible, while extrapolated repayment should be discussed with counsel and the payer. Third, communicate with the payer: disclose the internal audit, the corrective action, and the confirmed overpayment; ask whether the payer wants a broader look-back. Waiting for the payer audit risks larger extrapolation, penalties, and loss of credibility; immediate blanket repayment risks overpaying for claims that are supportable. A middle path is to repay confirmed overpayments promptly, offer a plan for the look-back population, and document good faith. Consider self-disclosure only if the conduct suggests systematic fraud; ordinary coding errors are usually handled through repayment and corrective action.