In simple terms, “mental health parity” is a federal requirement that employer health plans and insurers cover mental health and substance use disorder care on the same terms as they cover physical health care. That means comparable co-pays and deductibles, but it also means comparable rules behind the scenes, like how strict a prior-authorization process is or how a plan decides which providers count as “in network.” A plan cannot make it noticeably harder to get therapy or addiction treatment covered than it is to get a knee surgery covered. “Compliance” is the ongoing work employers, insurers, and their administrators do to prove, on paper and in practice, that those rules really do match up.

The U.S. Department of Labor’s Employee Benefits Security Administration published two new documents on September 8, 2026, that together reshape how the agency says it will police mental health parity compliance: Field Assistance Bulletin No. 2026-03, which sets out “guiding principles” for enforcing the nonquantitative treatment limitation (NQTL) comparative-analysis requirements of the Mental Health Parity and Addiction Equity Act (MHPAEA), and an updated Self-Compliance Tool plan sponsors and issuers can use to test their own coverage against the law.

The bulletin is an internal EBSA policy memorandum, from Assistant Secretary Daniel Aronowitz to the agency’s enforcement staff, and by its own terms creates no enforceable rights for plans, issuers, or participants. But it is the clearest public signal yet of how EBSA intends to direct its MHPAEA investigations following an 18-month stretch of regulatory limbo, and it follows through on an enforcement priority the agency flagged on January 15, 2026, when it overhauled its national enforcement projects for fiscal year 2026 to include barriers to mental health and substance use disorder (MH/SUD) benefits alongside cybersecurity, surprise billing, and benefit distributions.

Some background explains why EBSA felt the need to clarify things. In September 2024, the Departments of Labor, Health and Human Services, and the Treasury issued a final rule under MHPAEA, as amended by the Consolidated Appropriations Act, 2021 (CAA), adding new requirements for the NQTL comparative analyses plans and issuers must prepare (the 2024 Final Rule, 89 Fed. Reg. 77,586). Four months later, the ERISA Industry Committee sued in the U.S. District Court for the District of Columbia, arguing the rule was arbitrary and capricious and exceeded the agencies’ statutory authority. Then came Executive Order 14219, directing agencies to identify and de-prioritize enforcement of regulations seen as imposing outsized compliance burdens (90 Fed. Reg. 10,583). In May 2025, the three Departments responded to both developments with a formal nonenforcement statement, agreeing not to enforce the new provisions of the 2024 Final Rule until the litigation concludes, plus an additional 18 months, while stressing that MHPAEA’s underlying statutory obligations remain fully in effect.

That litigation has since moved further away from the 2024 Final Rule rather than toward defending it. According to a joint status report the Departments and the ERISA Industry Committee filed with the court in late March 2026, the Departments have now decided that, rather than defend the rule as written, they will issue a new proposed rule with anticipated substantial revisions to the challenged provisions, with a notice of proposed rulemaking targeted for no later than December 31, 2026 (see the March 30, 2026 joint status report in ERISA Indus. Comm. v. Dep’t of Health & Hum. Servs., No. 1:25-cv-00136 (D.D.C.)). In the meantime, plans and issuers are left navigating NQTL compliance under the pre-2024 regulatory framework, without a finalized replacement rule.

It is against that backdrop that the new bulletin narrows EBSA’s enforcement focus to three categories the agency says carry the highest potential for participant harm. The first is separate treatment limitations, including blanket exclusions of MH/SUD treatments where comparable medical or surgical treatments are covered; EBSA says it will prioritize wholesale exclusions but may still pursue narrower ones, especially in response to complaints. The second is medical necessity standards and review, with particular attention to prior authorization, concurrent review, and retrospective review; plans may rely on proprietary clinical guidelines, the bulletin notes, but must make them available on request during investigations and to participants. The third is network adequacy, with emphasis on provider admission standards and reimbursement methodologies, on the theory that a thin MH/SUD network pushes participants toward costlier out-of-network care. EBSA says it may still investigate other categories of NQTLs as complaints arise, but these three will get the bulk of its attention.

The companion Self-Compliance Tool, a roughly 40-page document plan sponsors, plan administrators, issuers, and state regulators can use as a self-audit checklist, is required to be updated every two years under Section 13001(a) of the 21st Century Cures Act. This edition walks through MHPAEA’s six benefit classifications, the “substantially all” and “predominant” tests used to evaluate financial requirements and quantitative treatment limits, and a four-step method for analyzing NQTLs: identifying the limitation, the factors behind it, the evidentiary sources for those factors, and whether the whole process is applied comparably to MH/SUD and medical/surgical benefits, both on paper and in practice. It also folds in DOL’s existing guidance on medication-assisted treatment for opioid use disorder and eating-disorder benefits, both of which the tool says remain subject to MHPAEA’s parity requirements notwithstanding the rulemaking uncertainty.

For insurers and self-funded plans specifically, the tool’s most operationally significant addition may be Appendix II, a framework for benchmarking provider reimbursement rates against Medicare as a way of self-testing network-related NQTLs before DOL auditors do it for them. The tool flags, as possible warning signs warranting further review, MH/SUD reimbursement rates set at or near Medicare levels while medical/surgical rates run well above Medicare, and psychiatrists reimbursed less than other physicians for identical evaluation-and-management billing codes. The tool is explicit that outcomes and denial-rate disparities are not, by themselves, proof of a violation, but it treats them as red flags that can trigger closer scrutiny of the underlying process. The Self-Compliance Tool also complements benchmarking resources the National Association of Insurance Commissioners has developed for state regulators and issuers.

Employee-benefits practitioners who reviewed the bulletin in the days after its release were, by and large, measured about how much it actually changes. One law firm client alert characterized the bulletin as reaffirming existing enforcement priorities rather than establishing new compliance standards, and separately noted that neither the bulletin nor the tool fills the gap left by the 2025 nonenforcement policy on what a compliant NQTL comparative analysis must contain, since a new final rule remains pending. For employers and issuers, the practical takeaway is that while the 2024 Final Rule’s new provisions remain unenforced pending litigation and rulemaking, MHPAEA’s statutory NQTL comparative-analysis obligations under the CAA are still live and enforceable, and EBSA has now told the regulated community, in some detail, where it plans to look first.

This summary is for general informational purposes only. Readers should consult the full Field Assistance Bulletin and Self-Compliance Tool, and the underlying rules and court filings linked above, for complete details and methodology and consult with legal counsel before making compliance decisions.