The 42% Query Response Rate and the 22-Year Veteran
You are the CDEO for a 24-provider multispecialty group. Q3 dashboard: 412 queries sent, 173 answered (42%); median response 11 days; internal benchmark is 80% within 5 days. The most common query topic is 'type 2 diabetes - with or without complications' (38% of queries), then 'CHF - acute, chronic, or history of.' Three providers account for 61% of unanswered queries. One of them, Dr. R, is a 22-year veteran with the highest wRVU output in the group; he tells you, 'I sign off on queries without reading them. They slow me down, and it's just coding stuff.' Group risk scores dropped 4% year over year, and two payers have flagged the group for review.
1. How would you prioritize and structure the education program given these data? Be specific about sequencing and format.
Read the worked answer
Sequence, don't broadcast. A 24-provider all-hands lecture will be ignored and will not move a 42% response rate.
Tier 1 - the topic (affects everyone). Diabetes with/without complications is 38% of query volume. That is a template problem, not a knowledge problem: the note doesn't capture complication status at the point of care. Deliverable: a one-page, two-column tip sheet using de-identified notes from this group, plus an EHR assessment prompt so the provider documents it the first time. If this works, query volume drops by roughly a third.
Tier 2 - the three outliers (61% of unanswered queries). Individual, 10-minute, data-in-hand sessions. Bring their own charts. Show the specific claim that was held and for how long, not a generic lecture.
Tier 3 - CHF acute/chronic. Second-highest volume; same approach, but sequence it after diabetes so the group isn't hit with two initiatives at once.
Format rules that matter: short, case-based, tied to the provider's own specialty and own notes; delivered at the point of care or in existing meetings, not new ones; and never framed as 'increasing risk scores.' Frame it as documentation accuracy, audit readiness, and reducing their rework.
2. Script your conversation with Dr. R. What do you show him, and what do you ask for?
Read the worked answer
Open by conceding his premise. 'You're the highest-volume provider here and queries are friction. I'm not here to add work - I'd like to take work off your plate.'
Show, don't tell - three specifics:
- His own turnaround. 'Your median response is 14 days. Every day a chart pends, it sits in my queue, your staff gets a follow-up call, and the claim doesn't go out. That's your staff's time, not just mine.'
- One chart, de-identified. Show a diabetes note where the assessment said 'DM without complication,' the medication list included gabapentin and the exam documented a monofilament test - and the query he signed without reading added a complication that was already supported by his own documentation. Point: he already did the work; the note just didn't say it.
- The audit exposure. Two payers have flagged the group. Show him what a records request looks like for one of his charts and how long it takes to defend when the note is silent.
The ask - small and concrete:
- Ten minutes to look at his three most common note templates so we can build the diabetes and CHF prompts into them.
- A commitment to answer queries within 5 days; offer to batch them so he sees one list weekly rather than pop-ups.
- Agreement on a 90-day check of his response rate and query-agreement rate.
Do not: threaten, invoke his seniority, or promise revenue. If the pattern persists after the 1:1, escalate through the medical director per policy - but escalation is the second move, not the first.
3. What metrics would you use to judge whether the program worked, and what pitfalls should you avoid in choosing them?
Read the worked answer
Process metrics (lead indicators):
- Query response rate and median turnaround - the direct targets (42% → 80%, 11 days → 5).
- Query agreement rate - if it is very high, queries are leading and need rewording; if it is very low, the underlying documentation or the query design is the problem.
- Query volume per 100 encounters - a falling volume with a stable capture rate is the success signal, because it means documentation is right the first time.
- Pended-chart days and staff follow-up touches - the operational cost that motivates providers.
Outcome metrics (lag indicators):
- Chart audit accuracy rate on a random sample, scored against the documentation.
- Payer audit findings and denial categories.
- Risk-score trajectory - reported alongside accuracy, never alone.
Pitfalls:
- Managing to risk score alone. It rewards capture and can quietly reward over-documentation. Pair every risk metric with an audit-accuracy metric.
- Response rate without quality. A provider who signs all queries in 24 hours without reading them hits the metric and creates the compliance problem.
- Educating only the outliers while leaving the broken template in place - the other 21 providers keep generating the same queries.
- Leading queries. If the redesigned diabetes query offers only 'with' and 'without,' the program is manufacturing the result. Always include 'unable to determine.'
- No feedback loop. Show providers their own numbers monthly, or the improvement decays within a quarter.