Why Every Health Plan Fails DOL Review and What to Do About It
The Mental Health Parity and Addiction Equity Act requires health plans to perform and maintain comparative analyses demonstrating that non-quantitative treatment limitations applied to mental health and substance use disorder benefits are no more restrictive than those applied to medical and surgical benefits. The Department of Labor has reviewed these analyses across four consecutive Reports to Congress and found that not a single one was sufficient on initial submission.1 This paper examines the regulatory framework that drives the requirement, the six-element comparative analysis that must be performed, the eight specific patterns of insufficiency that DOL reviewers identify, the structural limitations of the consulting model that produces these analyses, and the case for automation. The objective is not to argue for a specific product. It is to establish that the current approach to NQTL comparative analysis is structurally incapable of producing compliant documentation at scale, and that the nature of the work is suited to systematic, repeatable, tool-driven analysis.
MHPAEA (2008) established the foundational requirement. The 2013 Final Rule added the "as written and in operation" standard. CAA 2021 converted the analytical obligation into a documentation mandate with a 10-business-day production requirement. The September 2024 Final Rule expanded data requirements — and while its enforcement was paused in May 2025, the underlying obligations remain fully in effect.
The Paul Wellstone and Pete Domenici Mental Health Parity and Addiction Equity Act of 2008 established the foundational principle: group health plans and health insurance issuers that provide mental health and substance use disorder (MH/SUD) benefits may not impose financial requirements or treatment limitations on those benefits that are more restrictive than the predominant financial requirements and treatment limitations applied to substantially all medical and surgical (M/S) benefits.2
The statute applies to both quantitative treatment limitations (QTLs) such as visit limits and copayment amounts, and non-quantitative treatment limitations (NQTLs) such as prior authorization requirements, step therapy protocols, network composition standards, and medical necessity criteria. The QTL analysis is mechanical. The NQTL analysis is not.
The tri-Department Final Rule of November 2013 implemented the NQTL requirements in detail. It established the "as written and in operation" standard: parity must be demonstrated not only in the plan's written policies but in how those policies are actually applied in practice.3
The 2013 rule defined the analytical framework. For each NQTL, the plan must identify the processes, strategies, evidentiary standards, and other factors used in applying the limitation. It must then demonstrate that those factors are applied comparably to MH/SUD and M/S benefits and that the limitation is not applied more stringently to MH/SUD than to M/S.3
This framework created an obligation that is conceptually straightforward but operationally massive. A plan that applies prior authorization to both MH/SUD and M/S benefits must compare the authorization criteria, turnaround times, denial rates, appeal processes, and reviewer qualifications across both benefit types. It must do this for every NQTL category, in every benefit classification where the limitation applies.
Section 203 of the CAA 2021 converted the analytical obligation into a documentation mandate. Plans must now perform and document a specific comparative analysis for each NQTL and produce that documentation to regulators upon request. The statute specifies six elements that every comparative analysis must contain.4
The CAA also required the DOL, HHS, and Treasury to issue annual Reports to Congress documenting plan compliance with the comparative analysis requirement. These reports have become the primary evidence base for understanding the state of NQTL compliance across the industry.
The September 2024 Final Rule codified additional requirements, most notably the "no more restrictive" standard for NQTLs applied in operation. Plans must collect and evaluate relevant data (denial rates, out-of-network utilization rates, network adequacy metrics) as part of the comparative analysis. The rule also codified the "meaningful benefits" standard and expanded the definition of NQTLs to explicitly include network composition and provider reimbursement rates.5
In May 2025, the tri-Departments paused enforcement of the 2024 Final Rule pending ongoing litigation challenging several of its provisions. This created significant confusion in the market. Some plans interpreted the pause as a suspension of parity obligations generally.
The enforcement pause applies to the 2024 Final Rule's new provisions. The underlying MHPAEA statute, the 2013 Final Rule, and the CAA 2021 Section 203 comparative analysis requirement remain fully in effect and fully enforceable. Plans must still maintain NQTL comparative analyses. The DOL can still request them. State enforcement continues independently.6
The pause is not a pardon. It is a temporary hold on one layer of regulation while the foundational obligations continue to apply.
CAA 2021 specifies six elements every NQTL comparative analysis must contain: identify the NQTL, document the factors used, document the evidentiary standards, perform the comparison (as written and in operation), state findings and conclusions, and include a remediation plan if non-compliant. A plan with 9 NQTL categories across 6 benefit classifications faces up to 54 comparison points, each requiring all six elements.
The CAA 2021 and the implementing regulations specify that every NQTL comparative analysis must contain six distinct elements. Understanding what each element requires is essential to understanding why the analyses fail.
The analysis must identify the specific NQTL at issue and the MH/SUD and M/S benefits to which it applies, in each benefit classification. The six benefit classifications are: inpatient in-network, inpatient out-of-network, outpatient in-network, outpatient out-of-network, emergency care, and prescription drugs.7
This element seems simple. In practice, it requires the plan to map every limitation to every benefit classification where it applies. A plan that requires prior authorization for inpatient MH/SUD admissions must determine whether comparable prior authorization exists for inpatient M/S admissions in the same classification. The mapping itself is a substantive analytical exercise.
The analysis must identify the factors used to determine the application of the NQTL to MH/SUD and M/S benefits. Factors include the sources and types of clinical criteria used, the qualifications of the individuals applying the criteria, the process for determining medical necessity, the frequency and timing of reviews, and similar operational parameters.8
The factors differ by NQTL type. Prior authorization factors include the scope of services requiring authorization, turnaround time standards, clinical criteria sources, and reviewer qualifications. Network composition factors include provider-to-member ratios, geographic access standards, reimbursement rate methodologies, and credentialing requirements. A generic factors list does not satisfy this element.
The analysis must identify the evidentiary standards used to define the factors. If a plan requires prior authorization based on clinical criteria, the evidentiary standard is the specific criteria set (InterQual, MCG, proprietary guidelines) and the evidence base supporting those criteria. If the plan sets network adequacy standards based on provider ratios, the evidentiary standard is the data source and methodology used to establish those ratios.8
This element requires more than naming the criteria source. It requires documenting why that source was selected, how it is applied, and whether the same evidentiary rigor is applied to both MH/SUD and M/S benefit determinations.
This is the core element. The analysis must demonstrate that the factors and evidentiary standards are applied no more stringently to MH/SUD benefits than to M/S benefits, both as written and in operation.9
"As written" requires comparing the formal policies side by side. "In operation" requires comparing actual outcomes. If prior authorization denial rates for MH/SUD are materially higher than for M/S, the limitation may be applied more stringently in operation regardless of what the written policy says.
The "in operation" analysis is where most comparative analyses break down. It requires plans to collect and present data that many do not systematically track by MH/SUD vs. M/S segmentation.
The analysis must state its findings and conclusions. Is the NQTL applied comparably? If not, where are the disparities? The findings must be specific, supported by the comparative data, and stated clearly enough for a DOL reviewer to evaluate.9
If the analysis identifies non-compliance, the plan must include a corrective action plan with specific steps and timelines for achieving compliance. A conclusion of non-compliance without a remediation plan is itself an insufficiency.10
The DOL's insufficiency determinations reveal a consistent pattern across plans of all sizes. The eight most common findings: conclusory statements without analysis, missing "in operation" data, failure to identify factors and evidentiary standards, generic analysis not specific to each NQTL, missing benefit classification mapping, no remediation plan, reliance on delegated entity documentation, and stale documentation. These are structural failures, not random ones.
The DOL's Reports to Congress and insufficiency determination letters reveal a consistent pattern. The same deficiencies appear in analysis after analysis, across plans of all sizes and across every NQTL category. These are not random failures. They are structural.
The following eight findings represent the most common insufficiency determinations documented across the DOL's published enforcement correspondence and Reports to Congress.111
These eight findings recur because the analysis itself is genuinely difficult. A mid-size plan with 9 NQTL categories across 6 benefit classifications faces up to 54 comparison points. Each comparison point requires identifying category-specific factors, evaluating evidentiary standards, comparing written policies, comparing operational outcomes, and documenting findings. A comprehensive analysis with full "in operation" data across all NQTL/classification combinations is a project involving thousands of data points.
The DOL is not finding plans negligent. It is finding that the scale of the required analysis exceeds what manual processes can reliably produce.
The predominant market response has been consulting engagements at $150,000 to $500,000 each. Five structural limitations make this model inadequate: cost (prohibitive for most of the market by plan count), time (3-6 months, longer than a DOL response window), staleness (point-in-time documents that go stale immediately), non-scalability, and opacity.
The predominant market response to the NQTL comparative analysis requirement has been consulting. Large actuarial and legal advisory firms have developed parity compliance practices. Plans engage these firms at substantial cost to produce the required documentation. The model has five structural limitations.
An estimated $150,000 to $500,000 per engagement is the industry range for a comprehensive NQTL comparative analysis. This cost is borne annually or whenever plan design changes materially. For a large plan sponsor with multiple plan designs, the aggregate cost can exceed $1 million per year. This cost is accessible to the largest national payers and plan sponsors. It is prohibitive for mid-market plans, regional insurers, and self-funded employers, which represent the majority of the market by plan count.
A typical consulting engagement spans 3 to 6 months from kickoff to deliverable. During this period, the plan's operations continue to evolve. By the time the analysis is complete, the operational data it relied upon may no longer reflect current practice. The 3-to-6-month timeline also means that a plan receiving a DOL request cannot produce a current analysis within the 10-business-day response window specified by the CAA.4
The output of a consulting engagement is a document. It reflects plan operations at a specific point in time. The moment the plan changes a prior authorization policy, updates its clinical criteria source, modifies its network adequacy standards, or adjusts its reimbursement methodology, the analysis becomes stale. The stale document is not wrong. It is incomplete. And an incomplete NQTL comparative analysis is an insufficient one.1
The consulting model depends on senior professionals spending significant time understanding each plan's specific benefit design, operational processes, and data environment. This expertise does not scale. There are a limited number of qualified parity consultants in the market. As enforcement increases and more plans seek analysis, the bottleneck tightens. Wait times for engagements extend. Quality may vary as firms staff lower on the experience curve.
The consulting deliverable is typically a narrative document. The plan receives findings and recommendations but limited visibility into the analytical methodology, the specific comparisons performed, or the thresholds applied. When DOL requests revisions, the plan must return to the consultant for explanation and re-work, generating additional fees and delays. The plan does not own the analytical engine. It owns a document.
The NQTL comparative analysis is a structured comparison exercise. The framework is defined. The factors are defined. The comparison methodology is prescribed. The documentation format is specified. What can be automated: data structuring, factor identification, the comparison itself, gap identification, and documentation generation. What requires human judgment: data accuracy, clinical criteria evaluation, legal interpretation, and remediation design.
The NQTL comparative analysis is, at its core, a structured comparison exercise. For each NQTL and benefit classification, the analyst collects a defined set of data points, applies a defined comparison framework, evaluates stringency on defined dimensions, and documents the findings in a defined format. This is the signature of a process that can be systematized.
Data structuring. The six-element framework defines what data must be collected for each NQTL type. The collection can be structured into category-specific input forms that ensure completeness. Prior authorization inputs differ from network composition inputs. The tool can enforce this distinction.
Factor identification. For each NQTL category, the relevant factors are defined by regulation and DOL guidance. An automated system can present the required factors, prompt for the plan's data on each, and flag gaps in input completeness.
Comparative analysis. The comparison of MH/SUD and M/S limitations on each factor is a defined operation. Given two sets of structured inputs, the system can evaluate whether the MH/SUD limitation is more restrictive, less restrictive, or comparable. This evaluation can incorporate both "as written" (policy comparison) and "in operation" (outcome comparison) dimensions.
Gap identification. When the comparison identifies that MH/SUD is more restrictive on a given factor, the system can generate a specific gap narrative describing the disparity, citing the regulatory provision, and recommending the remediation action.
Documentation generation. The six-element structure is a documentation template. An automated system can populate this template from the analysis results, producing documentation that is formatted for DOL review standards.
Automation does not eliminate the need for professional judgment. Several aspects of the process remain inherently human.
Data accuracy. The quality of the analysis depends on the quality of the inputs. The plan's compliance team must ensure that the data entered accurately reflects current plan operations. No tool can verify that the denial rate entered matches the actual denial rate in the claims system.
Clinical criteria evaluation. Determining whether one clinical criteria set is "more restrictive" than another requires clinical and regulatory judgment. An automated system can flag disparities in criteria source, update frequency, and application process. Whether a specific clinical guideline is substantively more restrictive than a comparator guideline requires expert evaluation.
Legal interpretation. The NQTL comparative analysis exists within a legal and regulatory framework. Whether a specific plan design decision constitutes a parity violation is ultimately a legal determination. An automated tool provides the analytical foundation. ERISA counsel renders the legal conclusion.
Remediation design. When a gap is identified, the specific operational change required to remediate it depends on the plan's structure, vendor relationships, and benefit design constraints. An automated system can describe the nature of the required change. Implementing that change requires organizational decision-making.
TurboTax does not replace accountants for complex tax situations. It automates the structured comparison and documentation that every filer must perform, flags the issues that require professional attention, and generates the required documentation in the format the IRS expects. It does not pull data from your employer. You enter it. But the analysis, the cross-referencing, and the documentation are what the software does.
NQTL comparative analysis occupies a similar position. The regulatory framework defines what must be compared. The comparison factors are specific. The documentation format is prescribed. The human judgment required is real but bounded. This is work that benefits from automation, not work that automation threatens to do poorly.
Regardless of tools used, every plan subject to MHPAEA should follow a seven-step structured approach: inventory all NQTLs, identify M/S comparators, document factors and evidentiary standards, perform the "as written" comparison, perform the "in operation" comparison, document findings and action plans, and establish a maintenance cadence for continuous compliance.
Regardless of the tools used, every plan subject to MHPAEA should follow a structured approach to NQTL comparative analysis. The following framework reflects the regulatory requirements and the patterns of insufficiency documented in DOL enforcement actions.
The NQTL comparative analysis requirement is the most consequential compliance obligation in behavioral health policy. Four years of DOL review produce the same finding: the current approach is structurally inadequate. The regulatory framework itself provides the solution architecture — a structured process with defined factors, prescribed methodology, and specified format. The question for the industry is not whether to automate. It is when.
The NQTL comparative analysis requirement is the most consequential compliance obligation in behavioral health policy. It touches every health plan in the country. It requires sustained analytical effort across dozens of comparison points. And the evidence from four years of DOL review is unambiguous: the current approach to meeting this obligation produces insufficient results.
This is not a failure of intent. Plans are not ignoring parity. They are investing hundreds of thousands of dollars and thousands of staff hours in the effort. The failure is structural. The analysis is too large, too nuanced, and too dependent on current operational data to be produced reliably through episodic consulting engagements or internal spreadsheet exercises.
The regulatory framework itself provides the solution architecture. The six-element comparative analysis is a structured process. The factors are defined. The comparison methodology is prescribed. The documentation format is specified. Work this structured is work that automation handles well.
The question for the industry is not whether to automate. It is when. Plans that adopt systematic, tool-driven approaches to NQTL comparative analysis will be better positioned for DOL review, state enforcement, and private litigation. Plans that continue to rely on episodic consulting will continue to produce the same insufficient results documented in every Report to Congress to date.
The compliance gap is real. It is documented. And it is, for the first time, addressable.
Equipoise automates the NQTL comparative analysis required under MHPAEA and CAA 2021. The tool implements the DOL's six-element comparative analysis framework, provides category-specific analysis for 9 NQTL types across 6 benefit classifications, and generates documentation-ready output with regulatory citations and remediation recommendations.
For more information: Get in touch → · equipoise.health
References
[1] U.S. Department of Labor, Employee Benefits Security Administration, "Report to Congress: Compliance with the Mental Health Parity and Addiction Equity Act of 2008," 2022, 2023, 2024, and 2025 editions. EBSA issued over 199 request letters and over 180 insufficiency determinations. No NQTL comparative analysis was found sufficient on initial submission across any reporting period.
[2] Paul Wellstone and Pete Domenici Mental Health Parity and Addiction Equity Act of 2008, Pub. L. 110-343, Division C, Title V, §512, codified at 29 U.S.C. §1185a, 42 U.S.C. §300gg-26, and 26 U.S.C. §9812.
[3] 29 CFR §2590.712(c)(4), "Final Rules Under the Paul Wellstone and Pete Domenici Mental Health Parity and Addiction Equity Act of 2008," 78 Fed. Reg. 68,240 (Nov. 13, 2013). The "as written and in operation" standard is codified at 29 CFR §2590.712(c)(4)(i).
[4] Consolidated Appropriations Act, 2021, Pub. L. 116-260, Division BB, Title II, §203, "Strengthening Parity in Mental Health and Substance Use Disorder Benefits," codified at 29 U.S.C. §1185a(a)(8). The 10-business-day production requirement is at §203(a)(8)(B).
[5] "Requirements Related to the Mental Health Parity and Addiction Equity Act," 89 Fed. Reg. 77,586 (Sept. 23, 2024). The final rule added requirements for data collection and evaluation as part of the "in operation" analysis and codified the "no more restrictive" standard.
[6] The May 2025 enforcement pause applies to the provisions of the September 2024 Final Rule. The underlying MHPAEA statute (29 U.S.C. §1185a), the 2013 Final Rule (29 CFR §2590.712), and the CAA 2021 Section 203 requirements remain fully effective and enforceable.
[7] The six benefit classifications are defined at 29 CFR §2590.712(c)(2)(ii): (1) inpatient, in-network; (2) inpatient, out-of-network; (3) outpatient, in-network; (4) outpatient, out-of-network; (5) emergency care; and (6) prescription drugs.
[8] 29 CFR §2590.712(c)(4)(iii), requiring identification of the "processes, strategies, evidentiary standards, or other factors" used in applying the NQTL, and 29 U.S.C. §1185a(a)(8)(A)(iii)-(iv) (CAA 2021), requiring documentation of factors and evidentiary standards as distinct analytical elements.
[9] 29 U.S.C. §1185a(a)(8)(A)(iv)-(v) (CAA 2021), requiring the comparative analysis itself and findings/conclusions as separate documented elements. The "as written and in operation" standard at 29 CFR §2590.712(c)(4)(i).
[10] 29 U.S.C. §1185a(a)(8)(A)(vi) (CAA 2021), requiring that if the analysis identifies non-compliance, the plan must include "a description of the actions the plan or coverage has taken or intends to take to address the non-compliance."
[11] DOL EBSA, "FAQs About Mental Health and Substance Use Disorder Parity Implementation and the Consolidated Appropriations Act, 2021 Part 45," (2021). Contains illustrative examples of insufficient comparative analyses and the specific deficiencies identified. See also EBSA's comparative analysis review letters, published in redacted form in the Reports to Congress.
[12] 29 CFR §2590.712(c)(4)(iv), as clarified in the 2013 Final Rule preamble (78 Fed. Reg. at 68,252), establishing that the plan fiduciary bears responsibility for the comparative analysis regardless of benefit administration delegation.