What’s a normal denial rate, days in AR, and clean-claim rate for a mental health practice?

What’s a normal claim denial rate for a mental health practice?

There is no benchmark specific to outpatient mental health, but the strongest published anchor puts claim denials at roughly one in five — and behavioral health plausibly runs at the worse end, for reasons set out below. Treat about 1-in-5 as the ACA marketplace baseline, which is the population that figure actually covers: not all payers, and not a behavioral-health number.

The strongest anchor is federal marketplace data. In KFF’s analysis of 2024 HealthCare.gov data, insurers denied about 19% of in-network claims and 37% of out-of-network claims, for a combined 20% — with 3% of insurers denying 30% or more in network. That is HealthCare.gov qualified health plans only — no Medicare, no Medicaid, no employer coverage — and every specialty pooled together. It is a reference point for “normal,” not a behavioral-health figure.

Prior-authorization denials run on their own clock and their own rate. The HHS Office of Inspector General found that Medicaid managed care plans denied one of every eight prior-authorization requests (about 12.5%) in 2019, with 12 of 115 plans denying more than 25% — and Medicaid is the dominant payer for much of behavioral health and substance-use treatment. Note that this figure sits below the marketplace baseline above — a reminder that these two numbers measure different things and should not be stacked: one counts authorization requests, the other adjudicated claims, on different populations in different years. The honest bottom line: no denial rate specific to behavioral health is published yet, so treat the marketplace figure as a reference point rather than a target you are failing to hit.

What’s a normal days in AR, and what counts as “aged”?

Days in AR is roughly how long a dollar waits before you collect it; AR is “aged” once it sits past the point it should have resolved, sorted into buckets (0–30, 31–60, 61–90, 91–120, 120-plus) from the date you billed. There is no clean behavioral-health-specific public benchmark — the practical signal is how much of your AR sits past 90 days and whether the trend is worsening.

The credible source here defines the metric without setting a target. HFMA’s MAP Keys — the industry-standard revenue-cycle definitions — specify Net Days in A/R and Aged A/R as a percentage of total billed A/R as consistent, comparable metrics, but HFMA publishes the definitions, not a threshold. The numbers you will see quoted as “good” — days in AR in the 30-to-50 range, single-digit-to-low-teens percent of AR past 90 days — are general-RCM rules of thumb, not behavioral-health benchmarks, and a residential per-diem program will not look like an outpatient practice.

So judge your own trend and the shape of your buckets rather than a borrowed number, and treat the share of AR sitting past 90 days as the health signal — that is where claims quietly cross filing and appeal deadlines and become write-offs. What to do about the balances already sitting there is its own process: how to triage and recover aged A/R.

What’s a normal clean-claim (first-pass) rate?

Clean-claim or first-pass rate is the share of claims that pass every payer edit and adjudicate without manual rework on the first submission. It is a defined standard metric, but there is no official or behavioral-health-specific target — RCM teams commonly treat the mid-90s percent as the “clean” bar, as a rule of thumb rather than a published benchmark.

HFMA’s MAP Keys define the Clean Claim Rate as claims passing edits with no manual intervention divided by total claims — an all-industry definition, again with no threshold attached. The widely-repeated “95%” is a general operating target, not an HFMA number and not a behavioral-health one. It matters here because outpatient mental health works against exactly the things that drag first-pass rates down: behavioral-health carve-outs with separate payer IDs and filing rules, medical-necessity documentation, time-based psychotherapy codes such as 90837 where the note has to establish the time, and telehealth billing conventions that vary by payer. The gap between your first-pass rate and that general bar is itself the signal — it points straight at where claims are failing before they ever reach adjudication.

Why would behavioral health numbers run worse than the cross-industry benchmarks?

Because most denials are administrative and procedural rather than clinical — and behavioral health plausibly carries more of exactly those triggers. An outpatient practice runs into authorization renewals on visit-limited plans, prior authorization for services such as psychological testing and TMS, carve-out payers with their own rules, and telehealth conventions that change. Each of those is a place a claim can be denied or delayed for a reason that has nothing to do with the care.

The reason mix is consistent with the first half of that. In the KFF 2024 marketplace data, only 5% of in-network denials were for lack of medical necessity, while 25% were administrative and another 36% were filed under “other.” The denial process is mostly procedural — and behavioral health arguably has more procedure: authorizations that have to be renewed rather than granted once, carve-out payers with their own clocks, and stricter documentation bars. A meaningful share of those denials are simply wrong: the OIG found that 13% of denied Medicare Advantage prior-authorization requests (2019 data) met Medicare coverage rules and would likely have been approved under original Medicare.

Parity law is the clearest evidence that this disparity is real and recognized: the Mental Health Parity and Addiction Equity Act (MHPAEA) exists precisely because payers applied tougher rules to behavioral health than to medical or surgical care, and the 2024 federal rule added a requirement that plans analyze their non-quantitative treatment limitations — prior-authorization standards and medical-necessity criteria among them. That said, the new provisions of that rule are under a federal non-enforcement policy announced by the Departments of Labor, Health and Human Services, and the Treasury in 2025, so parity is not the quick fix it sounds like.

Be precise about what that supports, though. Parity law governs coverage rules — benefit design, treatment limitations, authorization standards — not claim adjudication, days in AR, or clean-claim rates. It does not predict your denial rate. What it does is establish that the friction is real and concentrated in the procedural layer, which is where to look for the gap rather than in the clinical work.

Why isn’t there a clean behavioral-health-specific benchmark yet?

Because the public data that exists is not broken out for behavioral health, and the bodies that standardize these metrics publish definitions, not behavioral-health targets. Three facts explain the whole gap.

First, the federal transparency data is not split out yet. KFF notes that as part of the plan-year 2027 certification process insurers “will report to CMS additional data elements: whether claims received and denied were for behavioral health or non-behavioral health services” — so a clean federal behavioral-health denial rate does not exist today. Second, the industry standard is a set of definitions, not numbers: HFMA’s MAP Keys standardize how to calculate clean-claim rate, days in AR, and denial rate so organizations can compare consistently, but they set no target threshold — let alone a behavioral-health one.

Third, routine denial data is thin enough that denials have to be studied specially. The OIG’s 2023 Medicaid managed care review had to assemble denial rates plan by plan and flagged that oversight of plan denials is weaker in Medicaid than in Medicare Advantage. Until behavioral-health-specific public data matures, the honest answer to “what is normal for behavioral health” is: measure your own and compare it to the cross-industry standard, treating that figure as a reference point rather than a target you are failing.

How should a practice actually use these benchmarks?

Compute your own four numbers, then read the gaps — against the cross-industry standard and against your own trend, not a borrowed target. The four that matter: first-pass (clean-claim) rate, net days in AR, the share of AR past 90 days, and your denial rate. Segment each by payer and by service (therapy, medication management, testing, telehealth), because one payer or one service usually drives the problem.

Each gap points somewhere specific. A low first-pass rate points upstream — eligibility and benefits verification (VOB), authorization, coding. AR stacking past 90 days points at unworked denials and no-response claims nobody had time to chase. A denial rate concentrated in a few payers points at carve-out rules and lapsed authorizations. And the gap is measured in dollars: the AHIMA Journal reports the industry average that as many as 60% of returned claims are never resubmitted, and puts the average cost of reworking a denied one at $25 for a practice (the $181 figure beside it is the hospital number). Those are all-industry figures, not behavioral-health ones, and they are averages the journal repeats from 2017–2019 trade sources rather than measurements anyone took — but they are why an unworked denial is not a neutral event: it usually ends as work you delivered and were never paid for.

Read this way — by payer and by dollar, rather than against a generic figure — the four numbers stop being a report card and start telling you which upstream step to fix first. In a small practice that is usually one payer and one recurring documentation gap, not a systemic failure.

Key takeaways

• No clean behavioral-health-specific public benchmark exists for denial rate, days in AR, or clean-claim rate — the credible numbers come from other populations. Behavioral health can be expected to run worse — it carries more procedural triggers, and parity law exists because disparate treatment is a recognized problem — but that expectation is reasoning, not a published measurement.
• ACA marketplace anchor: KFF found insurers denied about 19% of in-network marketplace claims in 2024 (37% out-of-network) — HealthCare.gov plans only, all specialties pooled, not behavioral-health-specific. A federal behavioral-health breakout is due to arrive with plan-year 2027 certification.
• HFMA’s MAP Keys define the metrics (clean-claim rate, net days in AR, denial rate) but publish no targets — the “95% clean” and “sub-X days” figures are general rules of thumb, not benchmarks.
• Most denials are not clinical calls — in KFF’s in-network mix, 5% were for medical necessity, 25% administrative, and 36% went unclassified — and behavioral health plausibly carries more of the procedural triggers (authorization renewals, carve-outs, documentation bars).
• A real share of denials are simply wrong: the OIG found 13% of denied Medicare Advantage prior-authorization requests met coverage rules in 2019.
• Measure your own four numbers — first-pass rate, net days in AR, AR past 90 days, denial rate — segment by payer and service, and fix the upstream step, since the industry average is that as many as 60% of returned claims are never resubmitted (all-industry, not a behavioral-health figure).