Spravato prior authorization denials by payer: a 2026 playbook for specialty clinics
Where Spravato denials cluster, what treatment-resistant depression means at UnitedHealthcare, Cigna, Aetna, and Medicare as of mid-2026, the medical-versus-pharmacy fork, and the REMS documentation that fails an audit weeks later.
Spravato (esketamine) sits at one of the hardest billing intersections in behavioral health. It is REMS-only, so patients cannot take it home. The visit runs at least two hours. Coverage splits between the medical benefit and the pharmacy benefit depending on the plan. And every major payer keeps its own definition of treatment-resistant depression. First-pass denials are common, and most land on a clinic that already bought the drug, scheduled the visit, and held the room.
This is a payer-by-payer playbook: where Spravato denials cluster, what treatment-resistant depression means at the big commercial plans as of mid-2026, the medical-versus-pharmacy fork that drives the billing path, and the REMS documentation that fails an audit weeks later.
Where Spravato denials cluster
- 01Prior auth never landedThe request was never submitted, was still pending at the visit, or approved the wrong dose tier or provider NPI.
- 02Treatment history misses the payer's TRD definitionThe documented antidepressant trials do not match the number, classes, dose, or duration the payer requires.
- 03REMS documentation gapsThe two-hour observation window has a start time but no discharge time, the field auditors look for.
- 04Drug-and-administration code mismatchBuy-and-bill and specialty-pharmacy clinics use different code sets, and crossing them produces unbundled or duplicate denials.
- 05Frequency and reauthorization limitsInduction caps and maintenance reauthorization windows differ by payer, and a missed reauth creates a coverage gap.
What treatment-resistant depression means by payer
The most expensive Spravato denials come back saying medical necessity not established. Almost always that means the documented treatment history does not line up with the payer's specific definition of treatment-resistant depression. As of mid-2026 the major commercial plans converge on at least two failed antidepressant trials at adequate dose and duration in the current episode, with variations worth knowing. Payer policies rotate, so confirm the current policy before you rely on any specific number.
| Payer | Trial requirement | Worth knowing |
|---|---|---|
| UnitedHealthcare / Optum | At least 2 antidepressants, adequate dose and duration; some plans specify more | Prior authorization required; verify the exact count on the member's plan |
| Cigna / Evernorth | At least 2 antidepressants from different classes, documented nonresponse | Often expects an oral antidepressant alongside Spravato |
| Aetna | At least 2 failed antidepressant trials | A separate pathway covers depression with acute suicidal symptoms; cite the right criteria set |
| Medicare | At least 2 adequate antidepressant trials | National coverage governs billing; do not unbundle the G-codes |
The operator move is to count discrete generic names and classes when you pull intake. Two SSRIs is one class. An SSRI then an SNRI then bupropion is three trials across three classes. A patient who discontinues an oral antidepressant mid-course can trip a reauthorization denial at payers that expect concomitant therapy, so flag that before the maintenance request goes out. For Aetna's acute-suicidality indication, reference the suicidality criteria, not the treatment-resistance checklist, or the reviewer reads the wrong list.
The medical-versus-pharmacy fork
Spravato's defining complexity is the benefit fork. The same patient's plan may cover it under the medical benefit, where the provider buys and bills the drug, or under the pharmacy benefit, where a specialty pharmacy dispenses it and the clinic bills only the services. The two rails differ in prior authorization, billing path, and dispensing logistics, so both have to be checked. The determination is parseable from the payer's, the plan's, and the state's policy, and it drives how the auth and the claim should route.
Two coverage rails, two different claims.
The clinic buys the drug and bills it alongside administration. Route the prior authorization through the payer or clearinghouse.
A specialty pharmacy dispenses and bills the drug. The clinic bills only the administration and evaluation, never the drug again.
The codes, and the ones that have changed
Two procurement models mean two different code sets, and the crossover is where most recurring denials live.
| Code | What it is | Note |
|---|---|---|
| J0013 | Esketamine drug code, billed per 1 mg | Effective 1/1/2026; replaced S0013. Not payable by Medicare, which bundles the drug into the G-code |
| G2082 | Up to 56 mg: drug, administration, and 2-hour observation | Bundled; do not unbundle on Medicare |
| G2083 | More than 56 mg: same bundle | Bundled; do not unbundle on Medicare |
| S0013 | Former esketamine drug code | Discontinued at the end of 2025; clinics still submitting it get silent clearinghouse rejects |
For buy-and-bill on Medicare, use the G-code bundle and do not bill the drug separately. For commercial buy-and-bill, J0013 is the 2026 drug code, though some plans still route an unclassified code, so verify per plan. On the pharmacy-benefit rail, do not bill the drug at all: the specialty pharmacy already billed it, and a second drug line is a duplicate that often closes the door on appeal. Bill the administration and the evaluation instead. Add modifier 25 to the evaluation when a separate, identifiable assessment is performed, or payers bundle it into administration and pay nothing for the cognitive work. Prolonged E/M service codes (99417, G2212) are generally not separately reportable with the G-code bundles, since those bundles already include the administration and two-hour observation time.
REMS documentation: the denial nobody catches in time
The two-hour monitoring requirement is REMS-mandated and is the single most common avoidable Spravato denial. It rarely denies on the first day. It denies weeks later when the payer pulls a sample and audits.
Two more REMS patterns recur. A patient who feels well and leaves at minute 95 does not qualify for the full administration code, so document the early departure with a clinical reason and bill the appropriate reduced service. And REMS site certification renews on a schedule; when the admin who owns it leaves and the renewal lapses, claims keep going out and the payer denies retroactively against the dates of service. Run certification renewal as a monthly checklist owned by a named person with a backup.
How Foresight handles this
Two surfaces matter for Spravato. Pre-encounter readiness checks prior-authorization status, REMS enrollment, and reauthorization windows before the patient arrives, so a gap surfaces the day before rather than at check-in, when a two-hour visit cannot be made up later in the week. Pre-submission claim scrubbing validates against the payer's live policy as the claim leaves, which is where a missing modifier 25, a discontinued S0013, or a note-versus-code mismatch gets caught before it becomes a clearinghouse reject or a documentation audit weeks later.
The ePA auto-answer assists the form work, matching fields and handling signature initials, with the operator reviewing and a maybe answer routed to a human rather than guessed. For the broader behavioral-health context, see how Cigna's 2026 TMS prior-auth change reshaped adjacent workflows.
01How long does Spravato prior auth take?
Standard prior authorization at the major commercial payers generally runs one to three weeks for non-urgent cases. Aetna's acute-suicidality pathway can be expedited but requires explicit clinical-urgency documentation in the request.
02Can one NPI bill approved Spravato across all payers?
Prior authorization is plan-specific. The same NPI can administer, but the approval is tied to a specific patient and payer, so switching payers requires a fresh request against the new plan's policy.
03Why do Spravato claims get rejected so often on first pass?
Three drivers: a prior auth that expired before the next dose, formulary or step-therapy requirements not fully documented, and the medical-versus-pharmacy split, where the medical benefit covers administration but the pharmacy benefit denies the drug. Aligning the claim with the approval window resolves most of them.
04Is the two-hour monitoring billable separately?
On Medicare, no. G2082 and G2083 bundle the drug, administration, and observation. On commercial buy-and-bill, the observation time can support a higher evaluation level, but it is not a freestanding observation code, and prolonged-services codes are generally not separately reportable with the G-bundles.
05What if the patient leaves before 120 minutes?
Document the early departure with clinical context and do not bill the full administration code if monitoring was not completed. This is the most common takeback during retrospective REMS audits.
Bottom line
Generic Spravato guides tell you to have prior auth, document REMS, and use the right codes. That advice has been out there for years and the denial rate has not moved. The operator move is to treat each payer's medical policy as its own checklist, run a pre-encounter readiness pass to catch prior-auth and REMS gaps before the patient walks in, and scrub claims against live payer policy before submission so the bundling and modifier mistakes do not leak out. If your Spravato denial rate is high, the gap is almost never code knowledge. It is the workflow that connects the chart to the claim.