Five Revenue Cycle Workflows Behavioral Health Providers Never Test Before Signing an EHR

By Easton Hallock, Managing Partner, Saint Health Group
LinkedIn: Easton Hallock
LinkedIn: Saint Health Group

Most behavioral health EHR demos follow the same script: a clean intake, a polished note, a claim that drops with one click. The workflows that decide whether a program gets paid rarely make it on screen, because they only appear when real patients move between levels of care, payers and clinicians. I am in the middle of an EHR build for an Oregon outpatient mental health and substance use program right now, and it has confirmed what I tell every leadership team: clinical setup gates billing. If the chart cannot carry the data, the claim cannot either.

Here are five to test in a sandbox before you sign:

1. Level of care documentation a surveyor can actually read
Oregon’s outpatient behavioral health rules require that any change to an ASAM level of care placement be justified within an update to the multidimensional assessment, including the ASAM Dimensional Admission Criteria (OAR 309-019-0135), and that every transfer document the date, the reason, the level of care recommendation and the severity of risk (OAR 309-019-0140). The rule text still cites The ASAM Criteria, Third Edition, even though ASAM released the Fourth Edition in 2023.

Test it: Admit a test patient to intensive outpatient, step them up to partial hospitalization, then back down. Ask the vendor to print the record a surveyor would see. Each transition should show dimensional ratings, justification and dates without anyone assembling them by hand. Ask which ASAM edition the template follows.

2. Payer enrollment status that lives outside the chart
Federal rules require states to screen and enroll every network provider of a Medicaid managed care plan, and a plan may contract with a provider whose enrollment is pending for no more than 120 days (42 CFR 438.602(b)). Yet the credentialing tracker and the EHR’s provider table rarely talk. A new associate goes on the schedule on day one, sees Medicaid members for weeks, and every claim carries a rendering provider the plan does not recognize yet.

Test it: Build a clinician who is active with one payer and has an effective date 30 days out with another. Schedule them with members of both. Does the system warn at scheduling, hold the claim, or let it go out? Then ask whether rendering and supervising provider logic for associates can follow payer-specific rules instead of one global setting.

3. Group notes that are clinically sound and uncodable
The program I am working with runs more than 20 groups. Oregon requires a note for each service that ties it to that individual’s service plan and records the actual time rendered, the setting, and the clinician’s name, credentials and signature (OAR 309-019-0140). A group of twelve is twelve notes and twelve charge lines.

Test it: Run a 90-minute group with ten attendees, two who leave early, co-facilitated by an intern. Confirm each member gets an individualized note, units calculate from that member’s own start and end times, and each charge routes to the right payer under the right rendering provider. Then look at the claims, not the notes.

4. Authorization tracking that assumes one authorization per episode
Oregon tells Oregon Health Plan members that behavioral health services do not need prior authorization, while many commercial plans authorize residential and partial hospitalization days in short blocks with concurrent review. The system in my current build allows one level of care per client with multiple programs attached. That works until a step-down creates overlapping authorizations, and tracking quietly moves to a spreadsheet.

Test it: Load two overlapping authorizations for one patient with different payers, units and date ranges. Bill against both. Does it decrement the right one and alert someone before the days run out, not after the denial?

5. Self-pay and pending verification patients
Patients awaiting a benefits check usually enter the same queue as insured patients. If verification never resolves, the balance becomes a write-off nobody decided to take. Uninsured and self-pay patients are also generally entitled to a good faith estimate when services are scheduled at least three business days ahead, and a patient billed at least $400 above that estimate can dispute the bill.

Test it: Admit a patient with verification pending, then convert them to self-pay on day five. Does the system produce a good faith estimate, apply your sliding fee schedule, re-bill earlier sessions correctly and send a statement? Who gets the work queue alert?

The question behind all five

No EHR handles all of this out of the box. What matters is whether you find the gaps during the sale, when the vendor still has reason to fix them, or after go-live, as denials. Bring your own scenarios, payer mix and billing staff to the demo, and ask to see the claim.