Practices are not imagining it: industry surveys over recent years consistently put average initial denial rates around 10–12% and trending upward — a meaningful climb from the single-digit averages of a decade ago. The physician on the receiving end experiences it as friction without explanation: the same care, the same coding habits, more denials.
What Changed on the Payer Side
Denying got cheap. Payers have automated claim review at scale — edit engines that apply coding policies to every line, algorithms that flag high-level E/M patterns, authorization requirements extended across service categories that never carried them. A human reviewer costs money per claim; an automated edit costs nothing per million. When the marginal cost of denial approaches zero and a large share of denials are never contested, the economics drive exactly the trend the surveys measure.
Medicare Advantage growth amplifies it: half of Medicare now runs through plans that bring managed-care review to a population traditional Medicare barely touched.
What Doesn’t Work
Indignation, for one — payer behavior responds to appeals and data, not frustration. Neither does working harder inside a broken structure: a billing team hand-fighting each denial individually loses to automation on volume arithmetic alone.
The Response That Works
Match systems with systems. Prevention first: categorize your denials by root cause, and eliminate the preventable categories upstream — verification cadence for eligibility, scheduling gates for authorizations, edits built from your own denial history for coding conflicts. Preventable categories routinely constitute a large share of total denials; eliminating them changes the whole equation.
Then contest programmatically: appeal wrongly denied claims in volume with documentation, because the published data on appeal outcomes is striking — large shares of appealed denials succeed, while most denials are never appealed at all. Payers’ automated denials are calibrated to that non-response; practices that respond systematically get different results and, over time, different treatment.
Finally, measure per payer: when one plan’s denial behavior is the outlier, that data belongs in escalation, contracting conversations, and — where behavior crosses lines — regulator complaints.
The trend is not reversing on its own. But the practices treating denials as a managed production system, rather than an ambient misfortune, are holding their rates in the mid single digits while the averages climb. The difference is structure, and structure is buildable.