Mental health parity rules, set out in the Mental Health Parity and Addiction Equity Act of 2008 (MHPAEA), are usually explained with a simple question: “Does the plan cover therapy the same way it covers a doctor visit?” That’s a fair starting point. But it’s not where most plans get burned anymore.
Today, the most expensive parity mistakes come not from an obvious exclusion in the plan document but from the way the benefit is operated: the approvals, routing, network rules, vendor handoffs, and administrative hoops that quietly make mental health and substance use disorder (MH/SUD) care harder to access than medical/surgical (M/S) care.
If you want the most practical way to understand parity in 2026, frame it as an operating requirement rather than a benefit design rule: your benefits system shouldn’t create extra friction for MH/SUD, especially when you can’t explain, measure, and defend that friction.
The parity shift: what’s written vs. what happens in practice
A few years ago, parity conversations centered on “quantitative” limits: deductibles, copays, out-of-pocket maximums, and visit limits. Most employers have made meaningful progress there.
Where parity risk lives now is in Non-Quantitative Treatment Limitations (NQTLs). These are the policies and operational practices that control access to care without showing up as a clean dollar figure or visit count.
Common NQTLs that create parity exposure
- Prior authorization rules (what requires approval, and when)
- Medical necessity criteria (which guidelines are used, and how exceptions work)
- Concurrent review frequency (how often continued care must be re-justified)
- Network admission standards (credentialing, reimbursement rates, panel closures)
- Step therapy or “fail-first” protocols (especially in SUD and pharmacy)
- Claims editing, coding rules, and reimbursement methodologies
- Provider documentation requirements and administrative burden
- Site-of-care and level-of-care rules (often a major factor in SUD)
The key idea is simple: a plan can look neutral on paper and still fail parity if MH/SUD care is managed more stringently than comparable M/S care in real life.
The under-discussed root cause: parity breaks in the seams
Most employers don’t have “one plan” in practice. They have an ecosystem. And ecosystems create seams: hand-offs between vendors where policies, data, and workflows don’t line up cleanly.
A typical setup might include a medical administrator/TPA, a behavioral health partner (carve-out or subcontractor), an employee assistance program (EAP), a tele-therapy vendor, an SUD program, and a PBM. Each one can bring its own playbook, criteria, and operating rhythm.
This is where parity becomes a systems architecture problem. Even if every vendor is “reasonable” in isolation, the combined employee experience easily turns into: more steps, more transfers, more documentation, and more delays for MH/SUD than for M/S.
And from a plan governance standpoint, “the vendor did it” doesn’t end the conversation. Under ERISA principles, the employer’s job is to run a prudent process, which means visibility, oversight, and documentation that the ecosystem is behaving equitably.
Enforcement reality: parity now expects proof, not good intentions
Parity scrutiny has moved into what many teams underestimate: comparative analysis. Since the Consolidated Appropriations Act, 2021 (CAA 2021) amended MHPAEA, plans that impose NQTLs must perform and document a comparative analysis of how those limits are designed and applied for MH/SUD versus M/S benefits, and produce it to federal or state regulators on request. Regulators aren’t only asking what your policy says. They’re asking how it performs, and whether it performs differently for MH/SUD than for M/S.
Parity is starting to look like an A/B test you didn’t realize you were running: MH/SUD on one side, M/S on the other, with your plan’s rules and vendor workflows as the “experiment.”
The operational metrics that tell the real story
- Prior authorization volume and approval rates
- Denial rates (initial and after appeal)
- Time-to-decision for authorizations
- Out-of-network utilization (often an access red flag)
- Allowed amount patterns and reimbursement methodologies
- Appointment availability proxies and provider panel capacity
- Provider directory accuracy and update cycles
- Member “drop-off” after provider searches or call-center interactions
If you can’t see these measures in a unified way, especially across carved-out behavioral health, you may be left with a compliance story but no compliance proof. WellthCare, the first Health-to-Wealth Benefit System, operates alongside the existing plan and maintains compliance-grade records of the care it delivers, so employers have a documented trail to support their parity process.
Where enforcement stands in 2026
The rulebook is in motion, and that matters when you are sizing up your parity exposure. On September 9, 2024, the Departments of Labor, Health and Human Services, and the Treasury issued final rules under MHPAEA that spelled out content requirements for NQTL comparative analyses and added data-collection and meaningful-benefits provisions, with staggered applicability for plan years starting January 1, 2025 and January 1, 2026.
Those new provisions are now contested. The ERISA Industry Committee sued in January 2025, and in May 2025 the Departments announced they would not enforce the 2024 final rule’s new provisions while the litigation is pending, plus 18 months after a final decision. The Departments are revising the rule, with proposed regulations targeted for the end of 2026.
None of that pauses the older, statutory duty. The CAA 2021 comparative analysis requirement remains in force, and the Employee Benefits Security Administration keeps requesting and reviewing comparative analyses; the 2025 report to Congress describes letters requesting analyses covering 77 NQTLs. Plans still have to perform and document the analysis, and produce it on request. The safe posture is to treat the rulemaking as a reason to build the discipline now, not a reason to wait.
The “invisible” parity failure: network economics
One of the most common real-world parity breakdowns starts not with a rule like “we don’t cover that” but with economics and operational drag.
If reimbursement levels, credentialing timelines, claims payment rules, and administrative requirements make participation unattractive for therapists and SUD providers, the result is predictable: a fragile network, narrower panels, longer wait times, and more out-of-network utilization.
Employees don’t experience that as “network strategy.” They experience it as “I can’t get care.”
From a parity perspective, this matters because a network that functions poorly for MH/SUD can operate as a de facto limitation, whether or not anyone intended it that way.
Parity is also a communications problem (because confusion is friction)
Another overlooked parity driver is what employees are told, and how consistently they’re guided. Behavioral health access often fails in mundane ways: inaccurate directories, confusing instructions, inconsistent call routing, and unclear appeal language.
Those issues are bad experience on their own, but when they recur they create a pattern where MH/SUD members face more obstacles than M/S members, which is exactly what parity is meant to prevent.
Why parity is hard for employers: the governance model is usually wrong
Parity lives at the intersection of clinical policy, claims operations, vendor management, employee experience, and fiduciary oversight. Most organizations don’t manage those as one connected system.
HR may own relationships. Finance may own cost containment goals. Vendors own day-to-day decisions. Legal is pulled in late. The result is predictable: parity is treated as a “compliance project” rather than an ongoing operating discipline.
The employers who are safest treat parity more like an assurance program, closer to SOC-style oversight than a one-time memo. That is the level of oversight this complexity demands.
A practical approach: run a parity “friction audit”
If you want to reduce parity risk without turning your benefits strategy into a legal exercise, focus on one thing: identify where MH/SUD care requires more work to access than M/S care, and then fix it.
Step 1: Map your NQTLs like an engineer
Build a single inventory of every meaningful “gate” for MH/SUD and the comparable M/S gate. You’re looking for extra steps, tighter thresholds, or more frequent review on the behavioral side.
Step 2: Require parity telemetry from vendors
Don’t settle for “engagement” reports. Ask for operational parity measures that show how the system is behaving, and put your vendors’ obligations in writing. The CAA 2021 requires plans to maintain written arrangements with service providers to support the comparative analysis and furnish the data behind it.
Step 3: Treat routing rules as NQTLs
If mental health members must go through extra layers (mandatory EAP screening, forced digital-first pathways, additional documentation, more frequent reviews), treat those as parity-relevant rules. Routing can be clinically appropriate, but it needs consistent rationale, comparable standards, and outcomes you can defend.
Step 4: Close the loop with documented remediation
Parity defenses are built on process and proof. When you find a disparity, document the fix and re-test. That documentation is the backbone of a credible fiduciary process.
- Document what you found (the gap and where it shows up operationally).
- Document what changed (policy, workflow, vendor configuration, network approach).
- Document when it changed and who approved it.
- Measure whether the change reduced the disparity.
The takeaway
Mental health parity is the lived reality of how quickly someone can find a provider, get authorized (if required), receive care, have the claim paid, and understand their next step.
When MH/SUD care takes more steps than M/S care (more phone calls, more denials, more delays, more confusion), parity risk rises. The best way to lower that risk is to treat parity for what it has become: a standard for how your benefits operating system runs, measured and managed across vendors, workflows, and data.
This article is for general information only and is not legal, tax, or medical advice. Employers should consult their own advisors.