RESEARCH DOSSIER
Mental Health Benefit Parity and Insurance Coverage
Abstract. Mental health benefit parity concerns the relationship between insurance entitlements for mental health and substance use disorders and those for medical and surgical conditions. Its institutional content extends from formal cost-sharing and visit limits to medical-necessity criteria, provider admission, negotiated reimbursement, network accuracy, and the actual possibility of obtaining treatment. This monograph analyzes benefit parity as a comparative rule within heterogeneous health financing systems. It reconstructs the evolution of American parity legislation and its 2013, 2024, 2025, and 2026 enforcement regimes; evaluates coverage for psychotherapy, psychiatric services, medications, emergency and intensive treatment; critically appraises quantitative and nonquantitative restrictions; and compares insurance-based arrangements with publicly financed pathways in England, Australia, Canada, and European health systems. The empirical literature yields differentiated findings: certain commercially insured children received measurable out-of-pocket protection, several adult and substance-use studies detected modest or statistically indistinguishable utilization effects, and recent network audits show that nominal provider listings frequently fail to produce appointments. These findings answer different causal questions and cannot be combined as a universal effect of parity. The analysis separates legal obligation, contractual coverage, effective provider capacity, patient payment, utilization, and clinical outcome. Original analytical synthesis shows why each transition requires distinct evidence and why enforcement records, administrative claims, secret-shopper audits, and population surveys must retain their own denominators. A worked illustration demonstrates how equal copayments can coexist with divergent appointment probabilities and total patient costs. International comparison identifies different institutional routes to substantively comparable access while emphasizing the limits of transferring American legal classifications to tax-funded health systems. The conclusions establish source-grounded priorities for measurement, accountable network governance, procedural transparency, and regulation-sensitive evaluation without presenting a new theory or proprietary framework.
Keywords: mental health parity; health insurance; substance use disorder; mental health benefits; nonquantitative treatment limitations; medical necessity; provider networks; reimbursement; out-of-pocket expenditure; financial protection; treatment access; regulatory enforcement; comparative health policy; causal evaluation; health equity
Introduction
Benefit design is one of the points at which the economic promise of health insurance becomes a practical claim upon a health care system. Coverage can eliminate the price of a medical service from the patient's immediate decision while leaving eligibility conditions, authorization, available providers, pharmacy formularies, appointment supply, and clinical pathways intact. Mental health treatment makes these distinctions especially visible because treatment may be longitudinal, delivered outside hospitals, dependent on a continuing therapeutic relationship, and distributed across multiple professional disciplines. The primary scientific question is how a formally insured mental health need travels from recognition within a benefit package to accessible, appropriately financed care.
The scale of the problem is established by the World Health Organization's World Mental Health Today report (WHO, 2025), which estimates that more than a billion people live with mental disorders, and its Mental Health Atlas 2024 (WHO, 2025), based on reporting from 144 countries. These are population and systems descriptions, not tests of private insurance parity laws. The WHO reports illuminate why the legal entitlement to care cannot, by itself, explain service use in places where personnel, primary care integration, medicines, and public budgets are constrained.
This dossier is international in scientific scope and precise in institutional attribution. The United States supplies the most extensively codified example of insurer benefit parity, particularly under the Mental Health Parity and Addiction Equity Act (MHPAEA). England illustrates direct entitlement through publicly funded services, Australia illustrates a publicly financed benefit limited by session allowances and provider fees, Canada illustrates the boundary of statutory insured hospital and physician services, and European institutions illustrate the distinction between national benefit design and supranational coordination. These are comparative institutional cases, not a ranking of national health systems or an assertion that all jurisdictions employ equivalent legal duties.
The study uses documentary comparison of primary laws and official guidance; appraisal of original observational and quasi-experimental studies; explicit reconstruction of denominators; and hypothetical, reproducible insurance calculations. Empirical observations are dated to the source's collection period rather than the present research date. The source hierarchy prioritizes enforceable legislative and administrative materials for legal propositions, original study methods for empirical estimates, and WHO/OECD statistical documentation for cross-country patterns. A law's legal force, an agency's enforcement practice, the existence of a covered service, and an individual's ability to obtain it are evaluated as separate propositions. No patient-level data were accessed or reanalyzed and no original causal effect was estimated.
Foundations and Source-Critical History
From coverage equality to access as an empirical question
Parity first entered modern health insurance debates in response to benefit designs that treated mental illness as a separate, more tightly limited category of expenditure. Annual visits, lifetime inpatient day ceilings, differential coinsurance, and stand-alone dollar caps made the disparity visible in a benefits table. The historical policy response emphasized comparability across health conditions: where a plan offered benefits for mental health or substance use treatment, the applicable restrictions should be assessed against those governing general medical and surgical care. Yet the practical meaning of comparability changed as insurers moved from numerical benefit limits to managed-care decisions about provider panels, reimbursement, diagnosis, and medical necessity.
The central United States source is the federal MHPAEA law and regulations collection (U.S. Department of Labor, current agency compilation), documenting the 2008 statute and successive regulations. Its historical sequence matters. The federal 2008 law expanded parity across mental health and substance use benefits; the 2013 final regulations addressed quantitative and nonquantitative limitations under implementing federal rules; and the 2024 final regulations specified further requirements concerning network access, core benefits, discriminatory factors, outcomes data, and comparative analyses. A regulatory description dated 2024 does not establish the enforcement position in November 2026.
The Affordable Care Act's individual and small-group essential-health-benefit architecture addresses an important predecessor question: whether mental health and substance use care is included in the package. Marketplace coverage requirements establish covered categories including psychotherapy and counseling, inpatient behavioral treatment, and substance use treatment, with state- and plan-specific details. MHPAEA answers a separate comparative question regarding the restrictions imposed on the covered benefits. These complementary rules should not be collapsed into a promise that every intervention is automatically insured, clinically indicated, or available in a network.
Beyond the United States, the trajectory is more heterogeneous. England established a nationally financed pathway for psychological interventions, and NHS England's Talking Therapies service description identifies free patient access, professional referral, self-referral, and a range of delivery modes. Australia's Medicare mental health arrangements provide a benefit toward qualifying psychological services with specified annual session limits. Canada's Canada Health Act overview frames universal provincial obligations around medically necessary hospital and physician services. European Union mental health initiatives express shared policy objectives while actual reimbursement remains organized substantially through national social insurance and health systems. Each trajectory raises an analogous problem of equitable financing and usable services, though only some adopt an American-style MHPAEA comparison.
Evidence types and the history of evaluation
Coverage mandates can be studied through contract audits, claims, population surveys, administrative appeals, and provider availability exercises. Their estimands differ. A contract audit asks whether written terms comply with a rule. A difference-in-differences study estimates a change in specified outcomes among exposed and comparison populations, conditional on assumptions about untreated trends and other reforms. A secret-shopper survey estimates the probability of contacting a listed provider and obtaining an appointment under the study's call protocol. None directly measures every dimension of clinical benefit.
Early state-level studies already demonstrated uncertainty in utilization effects. Harris and colleagues' state parity law evaluation (2006) employed repeated cross-sections and state changes in policy. Subsequent work on children found limited or statistically indistinguishable treatment-use effects in some samples, whereas Li and Ma's child and adolescent analysis (2020) estimated a 2.80-percentage-point relative utilization increase in a particular middle-income subgroup of states without prior parity protections. The subgroup, survey waves, policy exposure definition, and comparison states are essential to interpretation. A numerical result that is correct for that sample cannot stand for the average effect of all parity regimes on all insured persons.
The chronology therefore motivates a disciplined interpretation: statutory design, insurer response, provider supply, and consumer behavior change over different intervals. Initial compliance work may occur before the service supply changes that would permit a sizable utilization response. Changes in utilization may also reflect diagnosis, reporting, insurance enrollment, and medical practice shifts. These mechanisms remain empirical questions rather than presumed consequences of the statute.
Mental health as an insured risk and a public health obligation
Psychiatric conditions affect service use through fluctuating need, recurrence, comorbidity, and variation in care intensity. Private insurers have conventional expenditure-control reasons to assess necessity and manage provider contracts, while patients need reliable access to evidence-based treatment during episodes of illness. The economic conflict arises when generalized utilization management is applied under conditions of asymmetric information and uneven provider supply. A parity rule attempts to discipline the comparison of restrictions; it does not determine the clinically optimal treatment plan for each individual.
WHO's mental health systems atlas documents the global variation in resources, service configuration, and information capacity. Its survey design also demonstrates a core constraint on international parity research: jurisdictions do not consistently observe the same patient-level coverage variables, diagnostic classifications, provider payment arrangements, or unmet need measures. International empirical generalization must respect those data-generation differences.
Parity Definition and Legal Scope
The object of legal comparison
Under the American parity framework, the reference is usually the financial requirements and treatment limitations applicable to medical/surgical benefits within benefit classifications. Relevant classifications include inpatient and outpatient services, emergency care, and prescription drugs, with specified in-network and out-of-network subdivisions under the governing rules. Financial requirements include deductibles, coinsurance, copayments, and out-of-pocket rules; quantitative treatment limitations include numerical visit or day limits. Nonquantitative treatment limitations (NQTLs) include prior authorization, concurrent review, reimbursement methodologies, network admission standards, and medical-necessity processes. The Department of Labor's parity law collection identifies these distinctions in the statute and implementing framework.
Formal comparability cannot be evaluated by comparing the average cost-sharing rate for mental health with the average cost-sharing rate for all physical conditions. The statutory tests are classification- and limitation-specific. For financial requirements and quantitative limits, the 2013 framework considers whether a specified requirement applies to substantially all medical/surgical benefits and is the predominant level in that classification. The correct comparator is the applicable medical/surgical distribution under the governing rule, not the cheapest medical benefit available. In practical audits, the analysis requires reliable plan benefit data, anticipated or actual allowed amounts as appropriate to the rule, and stable service classification.
NQTL analysis has a different evidentiary center. The evaluation asks whether the processes, strategies, evidentiary standards, and factors used in designing and applying the restriction to mental health and substance use benefits are comparable to those applied to medical/surgical benefits, and no more stringently applied. A plan could have equal nominal copays while using authorization criteria and contracting practices that burden mental health services more severely. Conversely, unequal observed utilization does not automatically prove a parity violation: prevalence, provider availability, preferences, clinical need, and geography may differ. Demonstrating regulatory noncompliance requires investigation of the relevant plan restriction and comparator.
Statutory coverage and material limits
The Marketplace essential-benefits explanation confirms that U.S. Marketplace plans include defined categories of mental health and substance use treatment. Other plan types vary in legal coverage source and applicable exemptions. Thus the claim that parity requires every American insurer to cover every mental health service misstates the object of the law. Parity is primarily comparative where relevant benefits are offered; specific coverage mandates arise from complementary law or plan obligations. For Medicaid managed care and related programs, the Medicaid parity guidance identifies a distinct federal regulatory structure, including Medicaid alternative benefit plans and the Children's Health Insurance Program. Benefit classification, delivery arrangement, and the responsible authority must be identified before legal conclusions are drawn.
Medical necessity illustrates the boundary between legitimate clinical review and unequal restriction. Insurers may define covered levels of care and apply utilization review. A parity inquiry examines how standards and decision processes were developed and applied across mental health and medical/surgical domains, and whether restrictions produce a pattern warranting scrutiny. The precise clinical guideline used matters, but a reviewer must also inspect staffing, appeal timing, documentation demands, reviewer qualifications, and authorization thresholds. Comparing a written manual while ignoring its implementation produces a materially incomplete account.
The 2024 rule and its post-2024 enforcement regime
A particularly important source-critical issue is the status of the 2024 final MHPAEA rule (U.S. Departments of Labor, HHS, and Treasury, 2024). The rule was finalized in September 2024, became effective in November 2024, and assigned staggered 2025 and 2026 applicability dates. It addressed meaningful core benefits, discriminatory factors and evidentiary standards, access-related outcomes information, and expanded NQTL comparative analysis requirements. Describing these provisions as unqualified operative federal enforcement requirements today would omit the subsequent agency position.
Following litigation filed in January 2025, the Departments' May 2025 enforcement statement declared that they would not enforce requirements introduced by the 2024 rule beyond the 2013 rule for failures occurring before the litigation's final decision plus an additional 18 months. The Departments expressly preserved the underlying statutory MHPAEA obligations, including the Consolidated Appropriations Act, 2021 duty to perform and document NQTL comparative analyses. Nonenforcement of new regulatory provisions is a policy about agency enforcement; it is neither repeal of the statute nor confirmation that the 2024 text has been judicially invalidated.
A March 30, 2026, joint litigation status report states that the Departments intended to propose significant regulatory revisions by December 31, 2026. This is a documented litigant statement of intended administrative action, not evidence that a future proposal was adopted. The latest directly verified agency operational guidance in this study, Field Assistance Bulletin 2026-03 (September 8, 2026), concentrates EBSA enforcement on treatment exclusions, medical-necessity review, and network admission/reimbursement. It reiterates continued statutory responsibilities while narrowing enforcement focus. These time-specific qualifications are necessary in any account purporting to describe parity regulation as of November 16, 2026.
The bulletin also distinguishes enforcement priority from the total field of legal compliance: a supervisory focus on three categories does not make every other potentially discriminatory NQTL permissible. State insurance authorities may also exercise separate responsibilities. The relevant analysis should therefore record which law, regulator, plan category, date, and administrative enforcement posture each claim concerns.
Principles of legal interpretation for comparative researchers
A defensible comparison follows four steps. First identify who is covered: an employee group, an individual-market enrollee, a Medicaid beneficiary, or a person eligible for public services. Second identify the legal source of the benefit and the enforcement institution. Third isolate the actual treatment or payment restriction and its proper comparator. Fourth examine both documents and operational evidence, specifying whether the question concerns formal compliance, financial affordability, available appointments, or health outcomes. The difference between these inquiries is substantive. For example, a denial appeal establishes an adverse decision and possibly its correction, while the absence of a psychiatrist from a network may prevent a claim from ever being submitted. Regulatory case files and claims files therefore observe different parts of access.
Access to Therapy and Medication
The benefit pathway from indication to delivered care
Effective insurance coverage is a sequence of potential transitions: a person experiences a need, recognizes it, seeks help, satisfies benefit eligibility, obtains a clinically appropriate referral or authorization where required, locates an available practitioner or facility, accepts an appointment, receives service, and settles the residual patient payment. Insurance design influences every transition, but by different instruments. Copays change an immediate price; prior authorization changes time and administrative burden; provider reimbursement changes which clinicians can afford to participate; coverage exclusions limit the range of reimbursable interventions. A failure upstream can generate zero observed claims, even when the relevant benefit is nominally generous.
Psychotherapy is an instructive case because the same service can appear as physician care, psychologist care, licensed clinical social-work treatment, group treatment, or remote treatment depending on jurisdiction and insurer. Accurate measurement must specify which professional and code are covered, whether direct patient billing is allowed, and whether the available provider accepts the actual plan. A provider directory that lists a psychiatrist without an available appointment cannot establish timely access. A consumer's out-of-network choice may restore access but transfer cost and claim-submission work to the patient.
The 2026 commercial network audit by Haeder, Lou, Xu, and Miller illustrates this distinction. Researchers made 5,822 secret-shopper calls to mental health providers listed in major commercial plans in five U.S. states between October 2025 and March 2026. They classified 61.3% of calls as nonevaluable, commonly because no definitive response could be reached; among evaluable calls, the modeled probability of scheduling an appointment was 23.6%, and 55.6% encountered directory errors. A single call produced an appointment with any available clinician in approximately 12.4% of cases. These percentages have different denominators and outcomes and must not be added or treated as national population access rates. Their evidentiary value is direct observation of a limited network search protocol, not proof that each unsuccessful call independently violated MHPAEA.
The Pennsylvania Marketplace directory study (2025) similarly compared regulatory submissions for 8,306 counselor listings with consumer-facing directories and appointment inquiries. Researchers found that 19.9% of filed listings were absent from the consumer directory and only 35.3% fully matched directory entries. These are proportions of filed listings under the study's protocol, not of all insured individuals who sought treatment. Read together, the two studies establish that nominal network size and actionable provider capacity are distinct observables and that administrative directories require validation.
Reimbursement, contracting, and provider behavior
Provider reimbursement is a central mediator between benefit entitlement and appointment availability. A therapist or psychiatrist decides whether to participate in an insurer's network partly on the relationship among the allowed payment, billing effort, appointment length, schedule control, patient demand, and delays in reimbursement. An insurance plan can meet formal cost-sharing parity while retaining a provider network whose reimbursement and administrative burdens undermine the practical value of coverage. A parity review therefore requires comparing the factors and processes used to set network admission and payment methodologies, not merely counting providers.
At the same time, observed out-of-network care can reflect several mechanisms. Patients may choose a practitioner for therapeutic continuity, specialization, privacy, location, or appointment availability. The correct measurement distinguishes patient preference from lack of in-network supply and records whether an exception or network-gap arrangement was requested. Empirical association between a high out-of-network share and mental health treatment does not by itself establish the intent or legal sufficiency of reimbursement policy. It provides a basis for an appropriately targeted operational audit.
Provider availability also differs geographically. Rural distance, telehealth eligibility, language match, child psychiatry capacity, and the intensity of required treatment shape the relevant network denominator. A national provider-to-population ratio obscures the range of appointment times within a person's clinically acceptable travel and timing threshold. The natural unit of access measurement is frequently a person with specified clinical need in a defined service area and time window, rather than a plan's aggregate number of contracted professionals.
Medical necessity and treatment continuity
Medical necessity review can operate at admission, continuation, discharge, and medication approval. For severe mental illness or substance use disorders, continuity is itself an important characteristic of treatment. A utilization-management process that repeatedly restarts authorization or imposes inconsistent documentary requests may create missed sessions and premature transitions in intensity of care. The proper inquiry compares the relevant review processes with medically analogous management of complex, relapsing medical conditions and assesses the actual application of clinical criteria.
The Department of Labor's September 2026 enforcement bulletin explicitly treats prior authorization, concurrent review, retrospective review, treatment exclusions, and network reimbursement as high-priority NQTL subjects. It also explains that proprietary guidelines are not categorically prohibited, provided the comparison and application satisfy continuing statutory standards; access to the criteria remains important for investigations and beneficiaries. This is a nuanced operational position: the existence of a clinical guideline cannot substitute for testing how that guideline is used.
For medication, parity operates in conjunction with coverage rules on formularies, step therapy, prior authorization, quantity limits, and specialty dispensing. The comparator must reflect the governing benefit classification and the clinical role of the medicine, while a separate clinical evaluation assesses the justification and consequences of switching or delay. Prescription coverage can be formally present yet practically constrained by high pharmacy costs, inconsistent supply, or inability to access a prescribing clinician. The entire treatment pathway should be considered when attributing medication use changes to a parity intervention.
Distinctive public financing pathways
England's NHS Talking Therapies offers psychological treatment without a patient charge through a publicly financed service and permits self-referral for eligible conditions. The access bottleneck can instead be waiting time, local professional supply, suitability of the service, or referral to specialist care. The NHS service access guidance distinguishes services accessible directly from those requiring a general practitioner. Comparing an NHS appointment with an American reimbursed psychotherapy claim therefore requires observing waiting times and service eligibility, as well as patient fees.
Australia's Medicare benefit arrangements provide, for qualifying patients, up to ten individual and ten group sessions in a calendar year and a referral process, with providers free to charge above the reimbursed amount. The professional eligibility rules specify the clinical and referral conditions. The legal design produces a measurable limit on publicly reimbursed psychotherapy sessions and possible gap payments. This is a benefit quantity and financial access question within Medicare, not an automatic example of an American-type parity violation.
Canada's public insurance coverage explanation places the legally insured-services core around medically necessary hospital, physician, and certain surgical-dental services, with provinces and territories determining service coverage within their obligations. Mental health care delivered by physicians or hospitals and psychotherapy delivered by independent nonphysician professionals can consequently fall within different financing arrangements. A comparative research file must identify province, professional class, delivery site, and any supplemental coverage. The broad label 'universal health coverage' cannot establish that all community counseling has identical public funding.
Access and quality cannot be inferred from spending alone
A rise in reimbursed treatment expenditure can represent improved access, higher fees, more intensive need, or changes in coding. A fall can represent successful preventive care, effective treatment, denial of services, or migration to out-of-network care. Likewise, no change in aggregate utilization can coexist with meaningful redistribution between high-need and low-need groups. The empirical variable must match the proposed causal mechanism. For treatment quality, clinical outcome measures, continuity, shared decision-making, and patient-reported functioning should be evaluated where collected; insurance administrative data alone rarely identify them comprehensively.
Outcomes and Enforcement
Evaluating the financial protection effect
One objective of parity is to reduce the financial burden borne by people requiring mental health treatment. Kennedy-Hendricks and colleagues' Federal Parity and Spending study (2018) compared commercially insured children newly exposed to parity with children whose plans were never subject to that law, using claims from three national insurers for 2008–2012. Among children with mental health conditions, the authors estimated an average annual out-of-pocket spending difference of minus $140 relative to the comparison trend, with a 95% confidence interval from minus $196 to minus $84. Among the high-spending subgroup, the estimate was minus $234, with a confidence interval from minus $391 to minus $76. These are difference-in-differences estimates for defined insured populations, not estimates of an equal reduction in national household mental health spending.
This finding is economically important even when changes in service use are small. A benefit may improve insurance protection by shifting a given treatment expense from household to payer without bringing a new patient into treatment. The converse is equally possible: an individual can obtain more covered services while still incurring substantial indirect costs for transportation, lost work, repeated authorization, and the therapeutic relationship. Financial protection and utilization are therefore separate outcomes.
Why utilization studies disagree
Drake, Busch, and Golberstein's Medical Expenditure Panel Survey analysis (2019) used 66,602 person-year observations for continuously employer-insured adults aged 26–64 and a difference-in-differences design covering 2005–2013. The study found no statistically significant parity-associated changes in the overall probability, quantity, or spending on mental health services, while identifying a marginally significant shift among users toward specialty rather than primary care settings. It estimates a different population and measurement pathway from the child claims analysis.
A separate study of high utilizers of mental health services (2019) matched adults in newly exposed commercial plans with a comparison group and found modest increases in outpatient visits during the two years after parity, with larger estimates among high baseline users, without significant corresponding reductions in emergency visits or hospitalizations. Busch and colleagues' substance use treatment evaluation (2014) found no change in the proportion using substance use disorder treatment in its first-year postlaw claims comparison, although spending per enrollee increased modestly. The specialty outpatient behavioral study (2018/2019) likewise reported no substantial average utilization or out-of-pocket spending response under its controls and sample.
These results can coexist. Population age, prior state parity exposure, employer plan structure, baseline utilization, benefit service category, observation period, and the distinction between changes in patient payment and changes in provider capacity alter the estimand. Studies with null results cannot establish that parity never affects care; positive subgroup estimates cannot establish an average effect for every insurance market. Nor should an observational estimate automatically be interpreted as a fully causal effect. The identifying assumptions, parallel pretrends, measurement of treatment need, and simultaneous reforms remain decisive.
Li and Ma's child utilization study (2020) reports a 2.80-percentage-point increase in a defined middle-income subgroup, based on the 2007 and 2011–2012 National Survey of Children's Health. The study illustrates heterogeneity across income strata and state policy history; the estimate should retain that subgroup definition every time it is reported. The contrast with other child studies also warns against replacing a study's actual comparison group with a generalized narrative about national reform.
Regulatory outputs as process evidence
The Department of Labor's 2025 MHPAEA Report to Congress (issued February 20, 2026) covers comparative analysis review activity from August 1, 2023, through July 31, 2025. It reports that EBSA closed 89 investigations involving NQTL reviews during that reporting period and issued 25 initial determination letters concerning 43 limitations. Among categories cited were preauthorization, provider network admission standards, and out-of-network reimbursement. The report also records corrective action plans and a set of specific claim reprocessing remedies. These quantities describe selected enforcement operations; the investigations are not a random sample of all health plans, and their results cannot be used as a national violation prevalence estimate.
Detailed enforcement records are valuable because they connect abstract rules to documented insurer behavior. They can identify missing NQTL analyses, noncomparable review procedures, inaccessible treatment categories, or failures to specify the evidence behind restrictions. They also reveal the administrative work required to remedy a violation, including new policy documents, changed utilization-management systems, and retrospective identification of affected claims. The 2025 report's case accounts show why a corrective letter is an intermediate output, whereas confirmed reprocessing and revised operations constitute more direct evidence of remedial action.
An enforcement measure needs an explicit denominator and temporal label. Counting 43 challenged limitations, 25 initial determinations, 89 closed EBSA investigations, and all persons potentially affected measures four different things. Even the count of members affected by insurer-wide corrective practices is a coverage footprint, not a causal count of patients who obtained extra clinical treatment. A program evaluation should document whether the corrective action was completed, whether the old restriction disappeared in operation, and how beneficiaries' subsequent claims or appointments changed.
Changing enforcement discretion and legal interpretation
The May 2025 federal statement and the September 2026 bulletin require evaluations after 2024 to distinguish nonenforcement of newly introduced regulatory provisions from continued statutory obligations. An apparent reduction in enforcement actions may result from an agency's decision to prioritize different categories, changed complaint flows, staff resources, settlement dynamics, or genuine improvement in compliance. Simple before-and-after comparisons of investigations are therefore insufficient to infer insurer behavior.
The research record should retain both the rule text and the agency posture in force during a plan's relevant period. A file analyzed in 2026 may concern an authorization decision taken in 2022; the legal tests and agency actions relevant to that decision are historically situated. This prevents current policy guidance from being projected backward into the original administrative decision.
Patient outcomes and distribution
Clinical improvement is the final public health concern, but it is rarely observed directly in insurer benefit audits. More appointments can improve outcomes if the care is indicated, evidence-based, timely, and sustained. An increase in treatment claims may also be a response to greater distress or improved detection of previously unmet need. The 2024 National Survey on Drug Use and Health (SAMHSA, 2025) estimates that 22.9% of U.S. adults received any listed form of mental health treatment in 2024, and 70.8% of adults meeting its serious-mental-illness definition received treatment. These estimates use a survey treatment definition including medication and telehealth, and are not direct evidence of parity compliance or of adequate treatment quality.
WHO's 2025 mental health data synthesis indicates that globally fewer than one in ten persons with major depressive disorder are estimated to receive minimally adequate treatment. This proportion refers to a particular adequacy threshold, condition, and international estimate. It is not directly comparable to the U.S. proportion of all adults receiving any mental health treatment. Their juxtaposition demonstrates the need for denominators matching severity and adequacy definitions before national comparisons are interpreted.
Parity also affects distributional concerns. High-need users face larger cumulative exposure to coinsurance and care management, while people with low incomes are especially sensitive to fixed copays, transport, and time costs. Children, rural patients, people seeking culturally and linguistically appropriate care, and those with complex co-occurring conditions may confront access barriers that average benefit measures conceal. Distributional assessment should specify the need strata and the relevant service type, rather than assuming a single national insured average is representative.
Data, Measurement, and Validation
A hierarchy of research units
The correct unit depends on the question. Legal interpretation may use a statutory section, plan year, product, insurer, and benefit classification. Actuarial comparison may use allowed charges, covered lives, or claims exposure. Provider-network adequacy may use geographic service areas, clinicians who accept new patients, and contact attempts. Clinical assessment may use patient episodes, validated symptom scales, treatment adherence, or remission. Without separating these units, a technically correct number can be applied to an inappropriate proposition.
A minimum comparative data record should include jurisdiction, law and guidance version, plan or public-payer type, coverage period, eligible population, clinical condition definition, covered service, place of service, network status, authorization requirement, provider type, approved benefit, payment source, patient liability, outcome definition, study design, and data collection dates. For linked records, unique but privacy-preserving person and encounter identifiers allow repeated observations to be handled without confusing a single patient with several claims.
Contract, claims, and directory data
A contract benefit schedule describes what a payer promises. A paid claim establishes that one service was successfully billed and processed under identified rules; it does not establish the number of patients who sought care unsuccessfully. A denial file provides the opposite administrative observation, but usually excludes patients who abandoned the search before submitting a claim. A network roster counts listed relationships, which may include inaccurate contact details and nonparticipating practitioners. Thus contract, claims, denial, and directory datasets should be linked conceptually without being substituted for one another.
The Haeder et al. commercial secret-shopper protocol offers a concrete validation design: calls follow a defined simulated patient request, in named states and plans, and outcomes separate unreachable listings, erroneous directories, capacity shortages, and successful scheduling. Its cross-sectional design cannot identify why the network looks as it does or estimate an intervention's future effect. It does, however, supply a more direct test of actionable access than the number of entries in a network directory.
The 2025 Pennsylvania directory audit highlights the further difference between data submitted to a regulator and listings displayed to consumers. A robust network assessment should reconcile the plan's regulatory filing, consumer directory, and provider confirmation. For a statistically interpretable estimate, it must report the sampling frame, response disposition, repeated call policy, and rate among all attempted calls as well as among definitive outcomes.
Population surveys and cross-national comparability
Population surveys observe treated and untreated people but depend on diagnosis, recall, language, question design, household coverage, and response. SAMHSA's 2024 NSDUH methodology and results includes a specified civilian, noninstitutionalized population and revised mental health treatment questions. The source notes breaks in question comparability, particularly after 2022 and in 2024. This prevents a simple multiyear series from being treated as a uniform indicator without survey-method review.
The OECD–European Commission unmet-care analysis (2024) uses EU-SILC and documents cost, distance, and waiting-time reasons for unmet medical care. Its 2023 EU figure of 2.4% concerns reported unmet medical care in the whole surveyed population under that measure, not mental health-specific unmet need or a rate among people seeking psychiatric care. The report expressly notes variation in question wording and the substantially different rates obtained when the denominator is restricted to people with care needs. It is useful for methodological triangulation, not for claiming an international mental health parity success rate.
WHO's Mental Health Atlas survey draws from 144 reporting countries on policy, funding, workforce, and services. The coverage and representativeness of its country responses and availability of particular indicators must be inspected before computing world-level percentages. An unweighted average of country rates would answer a different question from a population-weighted global proportion. A summary of health financing should similarly differentiate domestic government expenditure, compulsory insurance, voluntary insurance, and out-of-pocket payment when country systems report them.
Identification strategies and counterfactual limits
A causal estimate of parity requires a credible counterfactual: what would have happened to the same eligible population, insurer, or service without the policy? A common empirical design compares more and less exposed plan groups before and after legal implementation. Its key assumption is that, absent reform, the outcome trends would have followed sufficiently similar paths conditional on adjustment. Preexisting state parity laws, plan exemption rules, employer insurance choices, economic shocks, Medicaid expansion, payment reforms, and diagnostic changes can violate that assumption or change the treatment contrast.
Difference-in-differences studies should report the preperiod, comparison population, policy adoption timing, standard error clustering level, outcome baseline, attrition, treatment intensity, and sensitivity to alternative policy dates. Where staggered reforms or heterogeneous effects are present, a simple two-way fixed-effects coefficient can be misleading; a design must be chosen for the specific variation and cohort timing. Observational estimates should be interpreted as conditional evidence unless the identification assumptions are independently persuasive.
Another strategy tests a narrower mechanism. A plan might remove an asymmetrical prior authorization rule, allowing evaluation of its immediate effect on denial rates, service initiation, and time to first appointment. Such work needs a comparison service and operational evidence that authorization behavior truly changed. An improvement in nominal compliance without new clinician capacity may primarily alter claims processing rather than appointments. The resulting estimate should be described as an effect of that procedural change, not the entire parity law.
A reproducible measurement illustration
Consider two outpatient benefit categories, mental health and medical/surgical, each with a $25 copayment and a $150 payer-allowed charge per visit. Formal copayment equality is satisfied in this simple hypothetical comparison. Suppose the probability that a randomly selected in-network provider directory contact leads to an appointment is 0.15 for mental health and 0.60 for a medical/surgical comparator under identical search procedures. The ratio of appointment probabilities is 0.25 and the difference is minus 45 percentage points. This observed disparity would warrant investigation of network processes and underlying service supply; by itself it does not prove a legal violation.
If a patient must make repeated, independent attempts until the first success, the expected number of attempts is 1/p: approximately 6.67 for mental health and 1.67 for the comparator. The independence assumption is unlikely to hold exactly in real networks because directories share outdated information and providers' availability is correlated. The example demonstrates the sensitivity of consumer search burden to genuine capacity and the necessity of reporting a search protocol. Even when the contractual copayment remains equal, time costs, delay, and the risk of costly out-of-network care can differ sharply.
For an illustrative eight-session course of therapy, a $25 copay yields $200 in direct in-network copays. If a network gap forces a patient to pay an assumed $160 per self-pay session with no reimbursement, the same course costs $1,280, a difference of $1,080. These are declared hypothetical amounts, not market averages. The correct observational study would record patient payment after reimbursements, plan exceptions, deductibles, and whether each session was clinically indicated and delivered.
Validation, uncertainty, and data governance
A full validation protocol uses document-level legal version checks; random and risk-based verification of directories; subgroup and geographic stratification; uncertainty intervals for sample estimates; independent reproduction of arithmetic; and qualitative review of authorization rules. The review should distinguish measured statistical uncertainty from uncertainty caused by missing provider dispositions, unobserved unmet need, and changing enforcement policy. Numerical confidence intervals do not cure sample selection or incorrect legal attribution.
Mental health information is sensitive. Studies using claims, clinical outcomes, or inferred diagnoses should minimize data fields, apply appropriate privacy and access controls, and avoid releasing identifiable treatment histories. Network searches can often be evaluated with simulated inquiries, which reduce exposure of personal patient data, though they still need attention to ethical research practice and the burden placed on practitioners. Transparency in aggregate findings can coexist with protection of individual records.
International Synthesis and Scientific Conclusions
Comparing functions rather than importing identical legal categories
The international comparative object is the insured person's route to clinically appropriate mental health care, together with the allocation of payment responsibility and the authority to challenge unjustified barriers. The American system makes parity a comparative obligation across specified categories of health plan benefit and restriction. England finances many services directly through the NHS and evaluates access through service eligibility, waits, referrals, workforce, and quality. Australia's Medicare subsidizes specified mental health services subject to professional eligibility, plans, visit allowances, and possible patient gaps. Canada's federal statute guarantees a public insurance core of hospital and physician care while provinces, territories, and supplemental arrangements shape the availability of other mental health professionals. These arrangements pursue related access and affordability aims through different legal objects.
The European Commission's comprehensive mental health approach (2023) supports national action across prevention, treatment, discrimination, and social participation, but a Commission strategy is not a uniform enforceable EU insurance-parity code. The European Health Insurance Card rules coordinate medically necessary state-provided care during temporary stays under host-country conditions; they do not replace national benefit coverage, guarantee free treatment, or independently settle the scope of psychotherapy benefits. The distinction matters because international portability is often mistaken for equivalence of national clinical entitlements.
For comparative analysis, functional questions can be posed consistently without declaring the legal categories identical: Is indicated psychotherapy included? Which providers qualify? What does the patient pay at point of service? How long is the wait? What review is required for continuation of care? Is urgent psychiatric care available? Can a patient appeal a restriction and obtain a remedy in time? How are medication and co-occurring substance use services handled? Who verifies that provider directories and reimbursement actually support access? Each answer should identify its own statute, benefit program, observation period, and population.
The interaction between insurance and treatment capacity
Parity instruments principally regulate the distribution of covered financial and administrative restrictions. Their effect on treatment access depends on the existence of clinicians, facilities, medicines, referral pathways, and informational systems. The WHO 2025 systems evidence describes large cross-country differences in workforce and funding, making this dependence especially visible in low-resource settings. Accordingly, insurance design is a component of mental health access, while effective capacity comes from broader health system institutions and investment.
The U.S. network audit evidence shows that resource constraints and administrative inaccuracy also operate inside high-income private insurance markets. The Haeder et al. 2026 appointment evidence should therefore inform regulatory measurement, while remaining bounded to the carriers, states, and study protocol investigated. It would be unsound to infer from an international workforce shortage that every specific insurer violates parity, just as it would be unsound to infer from a parity-compliant contract that patients have access to clinically appropriate care.
What empirical research can presently establish
Across the original studies reviewed, the defensible conclusions are differentiated. Federal parity can redistribute out-of-pocket financial risk in some insured populations; evidence of aggregate increases in treatment use is mixed and frequently modest; certain subgroups may experience more pronounced changes; contract limits and provider availability can keep practical access below the potential implied by formal benefits. This is supported by the children's spending estimate, the adult survey evaluation, the high-utilizer analysis, and the 2014 substance use evaluation. Their distinct designs and periods prevent pooled generalization without a prespecified synthesis method and compatible outcomes.
Regulatory records document specific patterns of noncompliance and corrective action; they do not quantify population-level causal benefit. The 2025 congressional report is especially useful for identifying the operational categories that agencies scrutinized and the documentary gaps they encountered. Evaluations after September 2026 must record the agency's revised enforcement priorities and continuing statutory baseline. A change in enforcement scope should be modeled as a change in the measurement process as well as a possible influence on insurer behavior.
Research priorities grounded in existing evidence
The strongest next studies would link plan-level benefit and NQTL characteristics to verified provider availability and longitudinal patient outcomes while retaining appropriate privacy controls. For a provider network, collect confirmed new-patient acceptance, time to appointment, distance, language, specialization, and out-of-network exceptions. For utilization management, record the time from request to decision, denial and overturn rates, interruptions of treatment, diagnostic group, and comparator services. For financial protection, capture deductible status, copays, coinsurance, nonreimbursed care, and indirect patient cost. For clinical outcomes, use validated disorder-specific measures and appropriate follow-up rather than interpreting paid claims as recovery.
Jurisdictional comparisons require an institution-by-institution mapping: private insurer, compulsory social fund, tax-funded provider, or provincial plan; the entitlement's legal source; budget and benefit limits; practical network or workforce barriers; and the appeal mechanism. Such a map is an analytical description of existing systems, not a new doctrine. Multi-country research should prioritize harmonized definitions of treatment need and receipt over headline comparisons built from unaligned survey questions.
The international scientific contribution of this dossier is the carefully sourced integration of legal, contractual, economic, administrative, and clinical evidence into a single comparative account of mental health insurance coverage. The synthesis shows how a formally equal benefit can yield unequal patient access, why that observation demands specific causal and legal investigation, and how diverse public and private financing systems can be compared without forcing them into one country's statutory terminology. The argument rests on verified legislation, published empirical designs, and explicit illustrative assumptions, with no claim to a novel law of insurance behavior.
Conclusion
Mental health benefit parity is a consequential institutional method for limiting discriminatory insurance restrictions and improving the conditions under which covered psychiatric and substance use care can be obtained. Its practical effect is mediated by benefit inclusion, classification, medical necessity, provider contracting, reimbursement, usable networks, patient cost, and treatment capacity. These mechanisms operate across different organizational levels and generate distinct forms of evidence.
American MHPAEA provides a particularly developed comparative legal system, but its current interpretation requires an accurate regulatory chronology: the 2013 framework and 2021 statutory NQTL duties continue to matter; the 2024 regulations introduced additional requirements; the 2025 nonenforcement statement limited federal enforcement of new provisions; and the September 2026 EBSA bulletin specified enforcement priorities. A scientific account must preserve those distinctions rather than treating an original promulgation date as a complete description of law in operation.
The best-supported empirical result is qualified heterogeneity. Financial protection can improve for affected users without a large aggregate rise in utilization. Some subgroups exhibit measurable increases in care, while other studies report small or statistically indistinguishable effects. Provider network investigations document a separate obstacle between covered benefits and available appointments. These findings call for precise denominators, justified counterfactuals, reliable directories, scrutiny of operational review practices, and accountability for remedies actually delivered.
Across public and private health systems, effective parity is evaluated most meaningfully through concrete access and financing functions: which needed services are included, how decisions are made, what patients pay, whether providers are reachable, and whether appropriate care occurs. The comparative task is to trace these functions with accurate legal and statistical evidence. Mental health insurance research thereby contributes directly to the study of financial protection, health system performance, and distributive fairness.
Technical Appendix: Comparative Measures and Worked Calculations
Benefit restriction classification
A researcher should assign each observed plan rule to the correct institutional category before evaluating its legal significance. An additional $15 mental health copay is a financial requirement. A six-visit ceiling is a quantitative treatment limit. Prior authorization based on a severity guideline, an exclusion applied by facility type, a provider reimbursement schedule, and network admission criteria are nonquantitative instruments. Each should be mapped to the appropriate benefit classification and medical/surgical comparator. Cross-category raw averages are descriptive and cannot replace the governing legal test.
Appointment probability and search burden
Let p denote the conditional probability of a successful appointment in one standardized contact attempt. Under identical and independent contact outcomes, the expected number of attempts to first appointment equals 1/p. At p = 0.15, E[N] = 6.6667. At p = 0.60, E[N] = 1.6667. Their ratio is 4. The appointment probability ratio is 0.15/0.60 = 0.25; the difference is 0.15 − 0.60 = −0.45, or −45 percentage points. In practice, contact outcomes are correlated, provider directories can contain duplicate entries, and patient willingness to continue searching is endogenous. The example is mathematical illustration, not an empirical projection.
Cost-of-care illustration
Assume eight covered sessions, an in-network copay of $25 and an out-of-network or self-pay charge of $160 with no reimbursement. In-network cost equals 8 × $25 = $200; self-pay equals 8 × $160 = $1,280; the difference equals $1,080. These amounts ignore deductibles, allowed-charge balance billing rules, network-gap exceptions, and all indirect costs by explicit assumption. The calculation illustrates how nominal cost sharing under a covered option differs from realized household liability if that option is unavailable.
Preferred reporting ledger
For each empirical estimate, preserve numerator, denominator, unit, period, jurisdiction, sample selection, comparison group, confidence interval, source version, and interpretation restrictions. For each legal proposition, preserve the authority, regulated population, effective and applicability dates, contemporaneous enforcement guidance, and any documented later revision. For each observed access failure, retain whether it was a stale directory record, a provider not accepting patients, a delayed appointment, an authorization rejection, or an out-of-pocket barrier. This separation makes review reproducible and prevents a single outcome indicator from acquiring legal meaning that its source does not support.
References
Busch, Susan H., Andrew J. Epstein, Michael O. Harhay, David A. Fiellin, Hyong Un, Deane Leader Jr., and Colleen L. Barry. 2014. “The Effects of Federal Parity on Substance Use Disorder Treatment.” American Journal of Managed Care 20(1):76–82. https://pubmed.ncbi.nlm.nih.gov/24512166/
Drake, Coleman, Susan H. Busch, and Ezra Golberstein. 2019. “The Effects of Federal Parity on Mental Health Services Use and Spending: Evidence From the Medical Expenditure Panel Survey.” Psychiatric Services 70(4):287–293. https://pubmed.ncbi.nlm.nih.gov/30691381/
European Commission. 2023. “A Comprehensive Approach to Mental Health.” Directorate-General for Health and Food Safety, policy and communication overview. https://health.ec.europa.eu/non-communicable-diseases/mental-health_en
European Commission. Current policy page. “European Health Insurance Card.” Employment, Social Affairs and Inclusion. https://employment-social-affairs.ec.europa.eu/policies-and-activities/moving-working-europe/eu-social-security-coordination/european-health-insurance-card_en
Haeder, Simon, Yuqin Lou, Wendy Xu, and Benjamin F. Miller. 2026. “Mental Health Ghost Networks in Commercial Insurance Plans.” JAMA Health Forum 7(10):e263746. https://jamanetwork.com/journals/jama-health-forum/fullarticle/2855163
Harris, Kathleen M., and colleagues. 2006. “The Effects of State Parity Laws on the Use of Mental Health Care.” Medical Care. https://pubmed.ncbi.nlm.nih.gov/16707997/
Health Canada. Current guidance. “About the Canada Health Act.” https://www.canada.ca/en/health-canada/services/health-care-system/canada-health-care-system-medicare/canada-health-act.html
Health Canada. Current guidance. “How Publicly Funded Health Care Coverage Works.” https://www.canada.ca/en/health-canada/services/health-care-system/canada-health-care-system-medicare/canada-health-act/how-publicly-funded-coverage-works.html
HealthCare.gov. Current consumer guidance. “Mental Health & Substance Abuse Coverage.” U.S. Department of Health and Human Services. https://www.healthcare.gov/coverage/mental-health-substance-abuse-coverage/
Kennedy-Hendricks, Alene, Andrew J. Epstein, Elizabeth A. Stuart, Rebecca L. Haffajee, Emma E. McGinty, Alisa B. Busch, Haiden A. Huskamp, and Colleen L. Barry. 2018. “Federal Parity and Spending for Mental Illness.” Pediatrics 142(2):e20172618. https://pubmed.ncbi.nlm.nih.gov/30037977/
Li, Xiaoxue, and Jie Ma. 2020. “Does Mental Health Parity Encourage Mental Health Utilization Among Children and Adolescents? Evidence From the 2008 Mental Health Parity and Addiction Equity Act (MHPAEA).” Journal of Behavioral Health Services & Research 47:38–53. https://pubmed.ncbi.nlm.nih.gov/31037480/
National Health Service. Current guidance. “How to Access Mental Health Services.” NHS England. https://www.nhs.uk/mental-health/social-care-and-your-rights/how-to-access-mental-health-services/
NHS England. Current service guidance. “NHS Talking Therapies for Anxiety and Depression.” https://www.england.nhs.uk/mental-health/adults/nhs-talking-therapies/
OECD and European Commission. 2024. “Unmet Healthcare Needs.” Health at a Glance: Europe 2024, OECD Publishing. https://www.oecd.org/en/publications/2024/11/health-at-a-glance-europe-2024_bb301b77/full-report/component-54.html
Services Australia. Updated 2026. “Mental Health Care and Medicare.” https://www.servicesaustralia.gov.au/mental-health-care-and-medicare?context=60092
Services Australia. Current provider guidance. “Eligibility for a Mental Health Treatment Plan.” https://www.servicesaustralia.gov.au/eligibility-for-mental-health-treatment-plan?context=20
Substance Abuse and Mental Health Services Administration (SAMHSA). 2025. Key Substance Use and Mental Health Indicators in the United States: Results From the 2024 National Survey on Drug Use and Health. https://www.samhsa.gov/data/sites/default/files/reports/rpt56287/2024-nsduh-annual-national/2024-nsduh-annual-national-html-071425-edited/2024-nsduh-annual-national.htm
Substance Abuse and Mental Health Services Administration (SAMHSA). 2025. “2024 NSDUH Annual National Report.” https://www.samhsa.gov/data/report/2024-nsduh-annual-national-report
U.S. Centers for Medicare & Medicaid Services. Current guidance. “Medicaid and CHIP Parity.” https://www.medicaid.gov/medicaid/benefits/behavioral-health-services/parity
U.S. Department of Labor. Current compilation. “Law and Regulations for Mental Health and Substance Use Disorder Parity.” https://www.dol.gov/agencies/ebsa/laws-and-regulations/laws/mental-health-and-substance-use-disorder-parity/law-and-regulations
U.S. Departments of Labor, Health and Human Services, and the Treasury. 2013. “Final Rules Under the Paul Wellstone and Pete Domenici Mental Health Parity and Addiction Equity Act of 2008.” Federal Register 78 (November 13). https://www.govinfo.gov/content/pkg/FR-2013-11-13/pdf/2013-27086.pdf
U.S. Departments of Labor, Health and Human Services, and the Treasury. 2024. “Requirements Related to the Mental Health Parity and Addiction Equity Act.” Federal Register 89 (September 23):77586 ff. https://www.govinfo.gov/content/pkg/FR-2024-09-23/html/2024-20612.htm
U.S. Departments of Labor, Health and Human Services, and the Treasury. 2025. “Statement Regarding Enforcement of the Final Rule on Requirements Related to the Mental Health Parity and Addiction Equity Act.” May 15. https://www.dol.gov/index.php/agencies/ebsa/laws-and-regulations/laws/mental-health-parity/statement-regarding-enforcement-of-the-final-rule-on-requirements-related-to-mhpaea
U.S. Department of Labor, Employee Benefits Security Administration. 2026. 2025 MHPAEA Report to Congress. February 20. https://beta.dol.gov/research-data/surveys-reports-publications/2025-mhpaea-report-congress
U.S. Department of Labor, Employee Benefits Security Administration. 2026. “Field Assistance Bulletin No. 2026-03: Guiding Principles for EBSA's Enforcement of MHPAEA's Nonquantitative Treatment Limitation Requirements.” September 8. https://beta.dol.gov/policy-regulations/reference-materials/guidance-searches/guidance/ebsa-field-assistance-bulletin-no-2026-03
U.S. District Court for the District of Columbia. 2026. The ERISA Industry Committee v. Department of Health and Human Services et al., Case No. 1:25-cv-00136-TJK, Joint Status Report, March 30, 2026, Document 18. https://litigationtracker.law.georgetown.edu/wp-content/uploads/2025/01/ERISA-INDUSTRY-COMMITTEE-v.-DEPARTMENT-OF-HEALTH-AND-HUMAN-SERVICES-et-al_2026.03.30_JOINT-STATUS-REPORT.pdf
Haeder, Simon F., and Jane M. Zhu. 2025. “ACA Network Regulatory Filings Are Inaccurate, Poorly Match Provider Directories.” American Journal of Managed Care 31(9):e265–e269. DOI 10.37765/ajmc.2025.89791. https://pubmed.ncbi.nlm.nih.gov/40966601/
Haffajee, Rebecca L., Michelle M. Mello, Fang Zhang, Alisa B. Busch, Alan M. Zaslavsky, and J. Frank Wharam. 2019. “Association of Federal Mental Health Parity Legislation With Health Care Use and Spending Among High Utilizers of Services.” Medical Care 57(4):245–255. DOI 10.1097/MLR.0000000000001076. https://pubmed.ncbi.nlm.nih.gov/30807450/
Gertner, Alex K., Jason Rotter, and Gracelyn Cruden. 2018. “Effects of the Mental Health Parity and Addiction Equity Act on Specialty Outpatient Behavioral Health Spending and Utilization.” Journal of Mental Health Policy and Economics 21(3):91–103. https://pubmed.ncbi.nlm.nih.gov/30530870/
World Health Organization. 2025. World Mental Health Today: Latest Data. Geneva: WHO. ISBN 9789240113817. https://www.who.int/publications/i/item/9789240113817
World Health Organization. 2025. Mental Health Atlas 2024. Geneva: WHO. ISBN 9789240114487. https://www.who.int/publications/i/item/9789240114487