Behavioral health CPT codes introduce an audit risk that most other specialties don’t face. Providers must document and code time with precise accuracy—and even attempts at conservatism can backfire, potentially leading to audits.
In this article, you’ll learn about new levels of scrutiny and how to limit compliance risks at your organization.
{{CTA-BLOCK}}
Time-based CPT coding: accuracy vs. Audit Risk
While time-based psychotherapy codes like 90832, 90834, and 90837 are standard, the risk extends to other services such as diagnostic evaluations (e.g., 90791) and E/M services (e.g., CPT 99214).
When billing for behavioral health, CPT codes are primarily time-based, reflecting the total duration of a session. For example, the standard time bands used for individual psychotherapy are:
- CPT 90832: 16–37 minutes
- CPT 90834: 38–52 minutes
- CPT 90837: 53 minutes or more
These time bands offer a straightforward link to a correct code, but it’s critical to document the actual start and stop times of each session so that documentation and coding align.
Time-stamped precision billing also makes it easier for data analytics algorithms to identify routine discrepancies at scale: Payers use automated and, increasingly, AI-driven systems to flag discrepancies that will put your organization at greater risk of triggering an audit.
Audit triggers often stem from mismatches between service types and the place of service (POS). Telehealth modifiers, such as Modifier 95, must accurately align with the documented POS code to prevent automated denials.
Furthermore, diagnostic coding must reflect the highest ICD-10 specificity and map directly to DSM-5 criteria. Payers are increasingly auditing whether the frequency and duration of services are clinically justified based on these documented diagnoses.
Accordingly, providers should also avoid rounded estimates and EHR timestamps and exclude administrative tasks and waiting periods from clinical session times. Additionally, avoid cloning time entries across sessions, as identical durations can trigger automated reviews for non-compliance.
AI enables evolving scrutiny of behavioral health providers
An OIG case study shows how scrutiny has evolved: In 2022, a New York City-based psychiatrist came under fire for consistently failing to document start and end times. Routine data analysis surfaced the provider’s noncompliance before the era of AI review, and the provider was subject to mandatory corrective actions, including repayment of more than $1 million.
In the age of AI, a singular study isn’t as remarkable. Charta identifies a 2026 behavioral health trend where payers, such as UnitedHealthcare and Humana, deploy tools to drive sharp increases in denials. Time-based CPT codes create an audit vector that payers can easily review at scale.
Payers only accept precise claims compliance
Payers care about paying exactly the right amount for exactly the care delivered. This means undercoding won’t protect you from scrutiny, either. While some behavioral health organizations may choose to undercode as a risk-mitigation strategy, payer AI has changed the landscape.
Ultimately, undercoding creates a "compliance gap" and just results in capturing less revenue without reducing audit risk. Even if time is documented correctly, documentation that fails to reflect actual patient acuity can trigger denials during medical necessity audits. You’re less likely to avoid the cost sink of an audit when AI can comprehensively scan and interpret your documentation and coding.
How payer AI increases behavioral health audit risk
Payer AI can audit billing patterns across an entire provider’s claim history, impacting revenue integrity while also creating ongoing exposure risk through continuous reviews that may extend for years.
And, because discrepancies accumulate invisibly and at scale, RCM leaders can’t remediate the issue through traditional manual spot-checks. If your providers aren’t actually documenting and coding time consistently, payer AI can catch it.
Protect your organization with your own AI capabilities
A critical component is maintaining treatment plan integrity. Modern payer AI reviews now link progress notes back to the treatment plan. Documentation must include measurable, time-bound goals tied to specific diagnoses to withstand this scrutiny.
Your main risk-mitigation tactic is to match payer capabilities with your own autonomous, pre-billing AI chart review to identify and correct variances before submission. These tools validate the Golden Thread—ensuring that progress notes reflect movement toward goals and automatically verifying documentation elements like POS and diagnostic specificity. Integrating AI tools into your existing EHR workflows enables comprehensive review at the point of care, allowing providers to fix errors and gaps before sending information downstream.
This lets your organization discover where specialty-specific pitfalls, like time documentation and coding issues, could otherwise undermine your revenue integrity. Providers can also receive personalized feedback at scale to support performance improvements that help keep charts compliant.
By aligning charts with clinical reality, organizations resolve both revenue and compliance vulnerabilities simultaneously, ensuring every claim is fully defensible before submission.
Learn how implement AI for RCM operations
Vanessa Miller was one of the earliest adopters of AI for the revenue cycle, enabling her team at Family Care Centers to scale operations by 5x without scaling headcount.
Learn how Vanessa evaluated and implemented AI solutions for her team to gain efficiencies across the revenue cycle.
