RESEARCH DOSSIER
Capitation and Global Budget Payment in Health Insurance
Abstract. Capitation and global budget payment are prospective methods for financing health care providers within insurance and publicly organized coverage systems. Their economic effects follow from who bears variation in illness severity, service volume, unit cost, enrollment, and aggregate spending; what clinical activities the payment covers; and how payment is adjusted for need, quality, and extraordinary events. This monograph develops a comparative analysis of population-based capitation, hospital-specific fixed revenue, sector-wide expenditure ceilings, and hybrid fee-for-service systems under a floating point value. It distinguishes payer-to-insurer payments from insurer-to-provider contracts, accounting expenditure ceilings from fixed provider revenue, and formal incentive predictions from measured causal effects. It examines actuarial risk adjustment, attributed-population rules, selection, service displacement, quality monitoring, provider solvency, contract governance, and empirical identification. Primary evidence from the World Health Organization, U.S. Medicare and Medicaid authorities, Maryland's hospital payment reforms, Taiwan's National Health Insurance, European and U.S. primary-care studies, and selected international systems demonstrates that prospective payment can improve expenditure predictability while creating divergent incentives at the patient and provider levels. The synthesis identifies the measurement conditions under which expenditure restraint represents efficiency, the circumstances in which savings may reflect displaced costs or reduced appropriate care, and the safeguards needed for reliable scientific evaluation. No new proprietary theory, model, or doctrine is asserted.
Keywords: Health insurance; capitation; prospective payment; global budget; hospital global budgets; provider reimbursement; population attribution; risk adjustment; moral hazard; adverse selection; underprovision; service intensity; quality assurance; strategic purchasing; comparative health systems; causal inference; actuarial soundness; health expenditure.
Introduction
Health care payment establishes the rules under which insurers, public purchasers, and other financing institutions transform prepaid resources into clinically delivered services. Under prospective payment, a provider receives an amount specified before the full realization of clinical utilization or expenditure. Two arrangements deserve particular scrutiny. Capitation typically allocates a periodic payment for each attributed eligible person and defines services for which the recipient is accountable. A global budget establishes an aggregate spending or revenue envelope over a period for an institution, group, geographic area, or sector. Their effects depend on contract detail and institutional setting. An annual hospital revenue target may be enforced while unit prices and individual claims remain visible. A nationwide expenditure cap may coexist with fee-for-service claims that are retrospectively converted into payments using a variable monetary point value. Those arrangements cannot be evaluated as though they were identical.
This is an inquiry into insurance-financed purchasing, including compulsory public arrangements and private coverage that collect or pool funds against health costs. The WHO health financing framework distinguishes raising revenues, pooling funds, and purchasing health services. Payment design belongs principally to purchasing; its consequences extend to enrollment, risk pooling, access, fiscal transfers, and the provider labor market. The distinction is crucial because an insurer's per-member payment from government is an insurer financing mechanism, whereas a medical practice's per-member payment from an insurer is a provider reimbursement mechanism. A system may use both, one, or neither.
The central scientific problem concerns allocation of uncertainty. Clinical needs are heterogeneous, some expenditures are preventable, and providers hold information that purchasers cannot fully observe. A prospective payment makes some marginal spending the financial responsibility of providers. That can reward avoidance of low-value care and coordination. It can also reward selection of low-need patients, deferral of necessary treatment, cost shifting, or manipulation of recorded risk. A fixed budget can stabilize institutional revenue, but whether it improves the efficiency of the whole health system depends on spending outside its boundary. These predictions establish empirical questions, not sufficient evidence that any particular program succeeded or failed.
The inquiry separates five levels of claim. A contractual identity describes how payments are calculated. An economic proposition describes behavior under stated incentives. A legal or administrative statement describes rules applicable to a particular scheme and date. An observational finding describes measured differences without necessarily identifying a counterfactual. A causal evaluation estimates an intervention's effects under explicit assumptions. The WHO mixed-payment guide emphasizes that multiple payment streams coexist and may conflict; accordingly, every comparison here identifies the purchaser, provider, services, population, period, and surrounding payment methods.
This monograph undertakes structured source-critical synthesis rather than a new patient-level statistical analysis or systematic review with exhaustive database screening. Primary institutional descriptions establish currently documented rules; journal articles establish the questions, populations, designs, and results reported by their authors. Model equations below are explanatory accounting and incentive representations derived from established payment logic. They do not describe estimated structural parameters. The principal date of regulatory and program review is November 14, 2026. Findings from one jurisdiction are interpreted according to its own coverage, eligibility, pricing, and contracting institutions.
Foundations and Source-Critical History
The payment object and the source of its budget
Capitation is commonly formulated as a payment per person per unit of time for a specified bundle of care. Its financial meaning requires four elements: an attributable person-period, a defined service obligation, a payment rate, and an allocation of exceptional costs. An annual global hospital budget instead fixes or constrains aggregate revenue for a defined hospital service perimeter over a fiscal interval. A sector-wide global expenditure budget fixes the total available for a group of providers and may distribute the total in proportion to claims or service points. Although each arrangement reduces a provider's marginal reward for certain additional services, the strength and distribution of that reduction differ. The European Observatory's payment overview identifies capitation and budgets among several methods and places their incentives within wider health system financing.
A line-item appropriation for salaries and supplies belongs to the broader family of budgets, but its administrative constraints differ from flexible global revenue. Under line-item budgeting, savings on one input may not be transferable to another. Under a flexible institutional envelope, a hospital may reallocate spending among departments but must reconcile actual service obligations against the revenue ceiling. The WHO public-financial-management assessment explains how rigid budget classifications, fragmented funding streams, and late disbursements can weaken strategic purchasing. Thus, the nominal unit of payment is only one dimension of implementation.
Payment systems became analytically salient as purchasers sought mechanisms to contain costs without sacrificing clinical access. The causal mechanism is elementary: a fee-for-service marginal payment typically rises with an additional billable service; a fixed per-person or per-institution payment ordinarily does not. But providers operate under liquidity constraints, professional norms, legal obligations, quality contracts, and reputational incentives. The early Gosden et al. (2000) review compared capitation, salary, fee for service, and mixed approaches to primary-care remuneration. Its small evidence base and limited health-outcome evidence cautioned against strong predictions from payment labels alone.
Historical trajectories and institutional varieties
Modern capitation has appeared in managed-care organizations and primary-care programs in many jurisdictions. Some governments pay an insurer a risk-adjusted sum per covered person, after which the insurer contracts with doctors under different arrangements. Other systems pay primary-care clinicians directly through population lists. Mixed practice funding may combine weighted capitation, practice allowances, performance rewards, and payment for specific activities. The empirical relevance of these distinctions appears in England, where Opie-Martin et al. (2026) examined a targeted capitation supplement rather than a pure capitated contract. A supplementary payment can correct an existing rate's distribution without replacing the entire purchasing architecture.
Global budgets are also heterogeneous. Maryland's 2014–2018 All-Payer Model applied regulated hospital revenue budgets within an all-payer hospital rate-setting structure. The subsequent Total Cost of Care arrangement expanded accountability beyond hospital services from 2019, and the state moved to the AHEAD Model on January 1, 2026, according to the Maryland payment chronology. Taiwan's National Health Insurance, by contrast, has used aggregate expenditure caps with fee-for-service claims converted through a budget-constrained point system. The Hsu (2014) hospital analysis describes Taiwan's regional capped claims architecture. The two cases imply different marginal incentives and different requirements for causal assessment.
The ongoing CMS AHEAD Model includes five participating states—Maryland, Connecticut, Hawaii, Rhode Island, and New York—and is described by CMS as running through December 31, 2035. AHEAD combines hospital global budgets with primary-care transformation, geographic accountability, and cost and quality objectives. It provides a current institutional example of simultaneous population and facility accountability. The model's announced aims are policy objectives; as of the review date, they cannot be treated as demonstrated long-term outcomes.
Methodological hierarchy for payment research
The strongest statements about program operation derive from governing rules and contracts, supplemented by the implementing authority's published methodology. The strongest statements about effects require defensible counterfactuals. Random assignment is rare for jurisdiction-wide payment reforms. Difference-in-differences studies demand credible parallel trends; interrupted time-series designs demand attention to preexisting trends, autocorrelation, concurrent reforms, and changes in measurement. Cross-sectional physician-payment comparisons are vulnerable to sorting by organizations and patients. Claims-based research can measure utilization and expenditures but may misclassify clinical appropriateness or fail to capture unbilled care.
An instructive warning comes from Zuvekas and Hill (2004). Using the 1996–1997 Medical Expenditure Panel Survey and generalized method-of-moments estimation, they reported that organizational characteristics of HMOs influenced access and quality measures more strongly than capitation itself; their capitation estimates were not statistically significant. This result illustrates why identifying the payment rate without observing the organization is insufficient. It does not show that capitation universally has zero effects.
A final historical point concerns terminology. “Global budget” may denote a public fiscal ceiling, a fixed hospital net-revenue arrangement, or a total spending cap that floats across competing providers. “Capitation” may denote an insurer's premium-like revenue from a public program or a clinician's reimbursement. Comparability requires identifying the economic holder of residual risk at every layer. Subsequent chapters preserve this distinction through explicit accounting boundaries.
Population Risk Assignment
Attribution, eligibility, and exposure time
Population-based payment begins with an attribution rule. Persons may be assigned to providers through voluntary registration, historical plurality of primary-care visits, geographic residence, claims algorithms, employer enrollment, or a combination. Enrollment and attribution have different meanings: enrollment establishes insurance eligibility, while attribution may determine which provider is responsible for care and which patient-period is included in a quality or spending benchmark. Attribution based on utilization can be endogenous: underuse itself changes who remains assigned. Prospective assignment limits this problem but can create data lags and errors for people who move or lose eligibility.
Let an attributed population comprise eligible persons indexed by i, with exposure fraction eᵢ during a given year. Let rᵢ denote a prospective annual rate after contractual risk adjustment. The prospective capitation entitlement is:
This identity is a payment accounting statement. If a rate is quoted per member per month, a comparable annual total uses the sum of covered member-months multiplied by the monthly rate. Midyear entrants, deaths, and dual coverage can change the denominator. The population ledger needs effective dates, payer identity, member identifier linkage, service coverage, and an explicit method for reconciling retrospective corrections. Without a reliable exposure ledger, apparent changes in spending per capita may merely reflect altered assignment.
Contract designers must also specify when assignment changes take financial effect. Retrospective reconciliation can transfer unexpected losses to purchasers, softening incentives to control costs. Pure prospective assignment leaves greater residual risk with providers but can produce sharp discontinuities. The financial issue is practical: if a person with unusually costly disease is assigned after the annual rate has been set, the physician organization may face expenses far above its per-member revenue. Reinsurance, risk corridors, high-cost carve-outs, and stop-loss provisions can share that exposure rather than forcing the practice to carry highly concentrated tail risk.
Risk adjustment and normative weights
A simple unadjusted capitation rate implicitly treats expected covered costs as equal across all assigned persons. That assumption is usually inappropriate where age, health status, disability, socioeconomic barriers, local wages, and geographic travel costs materially alter resources required for comparable access. Risk adjustment seeks to align prospective revenue with expected legitimate resource needs. Its technical form may be expressed as rᵢ = b × w(Xᵢ), where b is the base rate and w is a normalized weight constructed from approved covariates. The formula says nothing by itself about whether selected covariates are fair, immune to coding manipulation, or sufficient for unusually complex cases.
Panturu et al.'s provider-payment framework—authored by Panturu, van Kleef, Eijkenaar, and Cattel (2025)—distinguishes the choice of risk adjusters from the choice of payment weights and identifies trade-offs among selection, cost-control incentives, and feasibility. One may estimate future costs from clinical diagnoses, but observed costs are partly determined by historical access, treatment style, and reimbursement. Calibrating to actual historical spending can reproduce previous inequity if disadvantaged patients were systematically undertreated. Conversely, adjusting aggressively for all observable expenditure differences can reward avoidable cost patterns and weaken cost-control incentives.
A rate system therefore requires two distinct tests. Predictive calibration asks whether expected expenditures under the existing data-generating process match payments across groups. Normative adequacy asks whether rates permit the covered benefit to be supplied at acceptable quality for groups with legitimate health needs. The second question needs service-need and quality evidence as well as claims. Schokkaert and Van de Voorde (2004) examine how specifications of risk-adjustment formulas affect risk selection, underscoring that maximizing predictive fit is not equivalent to choosing socially appropriate transfer incentives.
Risk models can be measured through calibration by decile, absolute and proportional prediction error, subgroup residuals, temporal drift, and the distribution of provider-level financial margins. Evaluate population-wide results and tails: a model with small average error may materially underpay for high-complexity chronic disease or socially marginalized patients. The van Kleef and van Vliet (2022) analysis examines persistently high- and low-spending groups in health plan payment. Its direct setting is plan risk adjustment; the conceptual warning transfers to provider capitation, but the calibrated weights and institutional constraints do not transfer automatically.
Separating three risk-bearing layers
There are at least three distinct contracts in a financed health system. First, households, employers, and governments finance insurance or public coverage pools. Second, a program may pay an insurer or managed-care plan a risk-adjusted amount per member; the insurer bears risk for the agreed coverage scope, subject to public protections. Third, the insurer may compensate a provider organization by capitation, episodes, fee for service, or a salary arrangement. A government capitation transfer to a Medicaid managed-care organization does not establish that its individual physicians receive capitation. Similarly, an accountable care organization can share savings against a spending benchmark while claims remain fee-for-service.
U.S. state Medicaid managed-care rate setting is a concrete illustration. The CMS 2026–2027 rate guidance covers managed-care capitation rates for specified rating periods beginning between July 1, 2026, and June 30, 2027, and emphasizes actuarial soundness in state rate certification. Its direct regulatory object is state-to-plan payment. Providers beneath those plans may be reimbursed differently. The Medicare Shared Savings Program description similarly illustrates shared population accountability that is not automatically full prospective physician capitation.
A fourth layer arises when hospitals are paid under global budgets while doctors, pharmacies, post-acute providers, and insurers remain on alternative contracts. A hospital can gain from reducing an avoidable admission within its fixed budget, yet the community may see increased spending elsewhere. The appropriate accounting unit for social evaluation is total covered expenditure for a defined population, adjusted for outcomes and cost displacement—not solely the global-budget provider's revenue line.
Rate adequacy and financial sustainability
The appropriate per-member rate is related to expected benefit obligations and administrative capacity, not simply the purchaser's preferred cash ceiling. Rate formation may start from prior eligible cost experience, adjust for inflation, technology, practice transformation, epidemiology, and changing covered benefits, and then reconcile available financing. Costing must distinguish salaries funded elsewhere, capital depreciation, delegated pharmacy costs, emergency transport, and unpaid work. A provider may appear profitable because another institution pays its staff or supplies.
The WHO Ukraine costing report, based on 2021 expenses of 100 selected primary-care providers and published in 2024, estimated an inflation-adjusted 2023 rate of 878.42 hryvnias compared with an actual 786.65 hryvnias. These are historical report estimates, not 2026 Ukrainian prices or a nationally representative causal finding. They illustrate the need to examine inputs, setting, and methodology when judging a nominal capitation rate. A nationally uniform rate may undercompensate small rural facilities facing sparse populations and fixed staffing requirements; targeted, transparent supplements can address these differences.
Financial soundness is different for a small clinic and a large diversified health system. If average claims cost per patient has variance σ² and independent patient risks, the standard deviation of a group mean declines proportionally with the square root of population size. Actual health care risks are correlated through outbreaks, local supply prices, and demographic concentrations; therefore the independence approximation can understate aggregate variance. Working capital, reserves, counterparty solvency, risk-sharing contracts, and prompt adjustment for extraordinary events are part of payment adequacy.
Service Incentives and Underservice Risk
Marginal revenue, marginal cost, and clinical choice
For a fee-for-service provider receiving price p for an additional billable service with incremental cost c, direct short-run financial contribution is p − c, before fixed costs, risk and any quality-related payments. Under fixed comprehensive capitation, the direct contractual payment for one additional covered encounter is ordinarily zero, so direct contribution becomes −c. Under a true hospital-specific fixed revenue budget the same approximation holds within the covered services and contract period; services outside the budget may have different prices. These are marginal accounting incentives, not full descriptions of medical decision-making. Professional responsibility, accreditation, continuity, patient demand, legal duty, and reputational considerations can alter actual care.
The shift generates multiple possible behaviors. The provider may invest in prevention and care coordination that avert expensive downstream treatment while preserving health. It may substitute telephone care or multidisciplinary work for unnecessary visits. It may refer complex patients elsewhere, reduce visit duration, delay access, lower staffing ratios, ration diagnostic services, or avoid accepting highly complex cases. Mixed payment may yield additional fee-for-service revenue for carved-out procedures, encouraging substitution toward activities excluded from capitation. All such responses are contingent on contract boundaries and opportunities for monitoring.
A national outpatient physician study by Balkrishnan et al. (2002) found shorter visits on average among capitated physicians but more preventive counseling, using a cross-sectional analysis of 46,320 U.S. ambulatory visits. Specifically, visit length was estimated to be 5.6% lower, whereas counseling was 17% more likely in capitated plans. These associations support the possibility of simultaneous reductions in face-to-face time and growth in certain preventive activities. The design does not establish a universal causal mechanism or determine whether shorter visits were clinically appropriate.
Payment boundaries and cost displacement
A prospective payment creates incentives precisely at the edge of its covered benefit. Full primary-care capitation may leave specialty care, hospitalizations, diagnostics, and medicines financed separately. If practice A must pay for a referral from its own budget, its marginal referral calculus differs from practice B that passes specialty costs to the insurer. A hospital global budget may exclude professional fees, pharmaceuticals dispensed outside hospital, rehabilitation, or long-term care. Savings to hospital accounts can accompany increased outpatient, home-care, insurer, or household spending. Research must follow the entire cost trajectory for the same population wherever data permit.
The arrangement also affects prices. Under a fixed revenue budget, a facility may have permission to adjust billed unit rates in response to changing volume, while aggregate revenue is constrained. Under a sector cap with fee-for-service claims, each additional claim may increase a provider's share of a fixed pool, even if aggregate system reimbursement stays constant. If total eligible service points are Q and a fixed sector budget is G, the ex post price per point can be v = G/Q. Provider j with qⱼ points is then paid Gqⱼ/Q. Its marginal revenue from additional points is positive when it gains share from rivals, even though the collective price per point falls as total claims rise. Thus, the label “global budget” can coexist with an incentive to produce more billable volume.
This logic helps interpret Taiwan. Hsu (2014) analyzed medical claims under regional fee-for-service budgets and reported early increases in hospital claims after implementation. Cheng et al. (2009) compared 2002 and 2004 hospital behavior and reported a 7% rise in length of stay and a 15% increase in procedures and medications per admission following the program's introduction. Such before-and-after changes require caution regarding simultaneous clinical and institutional changes, but directly contradict a simplistic assumption that a global budget necessarily reduces all service volume.
Selection and patient-level service allocation
Underservice risk and selection risk are related but analytically distinct. Underservice involves insufficient appropriate treatment for someone already assigned. Selection involves changing who is enrolled, attributed, retained, or encouraged to remain within the provider's practice. Risk adjustment may reduce predictable gains from selection while leaving incentives to avoid unmeasured complexity, nonadherent patients, people requiring intensive communication, or patients whose costs are hard to code. Quality monitoring of current patients cannot fully detect people deterred from entry in the first place.
Selective practice closure, long waiting lists, inconvenient appointment times, exclusion of costly specialists, and strategic referrals may all alter patient composition. Appropriate empirical indicators include acceptance rates by severity, attributed-population churn, transfers out, specialist referral timing, emergency department use, disease-specific access, and external complaint or appeal records. An analyst must also consider lawful service-boundary exclusions and baseline supply shortages. A country with few clinicians may exhibit waiting times under multiple payment systems; attributing those delays to capitation without credible comparison confuses financing incentives with physical capacity.
Mixed payment and comparative evidence
Most real systems combine methods. The WHO five-step assessment begins by mapping payment flows and progresses through incentive assessment, broader effects, governance, and policy options. This provides a practical research structure: identify which provider receives salary, capitation, activity payments, capital grants, risk corridors, and performance bonuses; establish whether payments reward the same outcomes; then examine actual responses under each combination.
In Oregon, Ukhanova et al. (2021) compared 27 community health centers adopting an alternative per-member payment methodology with six nonadopting centers, using retrospective electronic health records and difference-in-differences methods. The reported effects on preventive care were mixed across service measures rather than a uniform loss of provision. In national U.S. ambulatory data, Morenz et al. (2023) found associations between capitation and some preventive services but not most measured targets. These findings distinguish the risk of underprovision from evidence that underprovision actually occurred on every clinical dimension.
The empirical inference is bounded. A payer could improve practice access by financing stable teams, yet the same nominal capitation arrangement could become restrictive if rates are low and services are broad. Both outcomes are consistent with the economic mechanism. A scientific account must measure baseline resources, adjustment variables, patient health, treatment appropriateness, and organizational response rather than converting the payment rule itself into a conclusion about quality.
Quality Safeguards and Evaluation
From accounting savings to clinical value
A payer's expenditure reduction has at least three possible explanations. The same beneficial clinical outcomes may have been produced with fewer resources, indicating improved technical efficiency. Expenditures may have shifted to other payers, facilities, future years, or patients, indicating a transfer of accounting responsibility. Appropriate care may have been delayed or withheld, producing adverse outcomes or unmeasured losses in health. A fourth possibility is that recorded spending changed through revised prices, coding, or claims eligibility while real service delivery changed little. It follows that expenditure growth alone cannot establish welfare improvement.
Quality assessment must cover processes, outcomes, access, and patient experience. Process measures include indicated testing, medicines, prevention, safe discharge, referral completion, and timely follow-up. Outcomes include mortality and morbidity for clinically meaningful risk-standardized conditions, preventable complications, readmissions, functional status, and disease control. Access measures include appointment times, network travel, emergency visits for ambulatory-care-sensitive conditions, patient retention, refusal of new patients, and inequities across risk groups. Cost measures must include total payer expenditure, net provider revenue, capital and infrastructure, uncompensated services, and out-of-pocket spending. None of these measures alone constitutes a comprehensive quality assessment.
Payment-linked quality bonuses can mitigate incentives to withhold services, but they introduce measurement trade-offs. If quality is assessed only on patients for whom a test was recorded, providers can improve apparent compliance by altering documentation or patient attribution. If mortality is used without sufficient adjustment for case severity and volume, comparisons may penalize hospitals that care for highly complex patients. If bonuses are too small relative to the savings from withholding costly care, their financial effect may be weak. Conversely, poorly targeted quality incentives can produce excessive low-value measurement activity. Effective governance treats measured outcomes as a validation system subject to audit rather than a complete proxy for clinical value.
Evidence from hospital global budgets
Maryland's 2014 All-Payer Model provides one of the most extensively evaluated contemporary examples of hospital-specific global budgeting under regulated rates. The CMS historical model description explains the policy's all-payer approach to inpatient and outpatient hospital payments. A different source, the HSCRC chronology of reforms, documents the transition from the 2014–2018 All-Payer period to the 2019–2025 Total Cost of Care period and the AHEAD transition in 2026. Separating these phases prevents the attribution of outcomes measured before 2026 to the newer program.
Morrison et al. (2021), in their Medicare–commercial difference-in-differences evaluation, compared Maryland outcomes with a hospital-based comparison group. The authors reported slower growth in Medicare hospital expenditures 4.5 years after implementation and in commercial hospital spending after four years. Medicare admissions declined relative to comparison; commercial admissions did not, although inpatient spending declined. Relative changes in emergency and other hospital outpatient spending produced an estimated $796 million reduction in Medicare hospital spending in the study's framework. That estimate is a modeled comparative reduction in specific Medicare hospital expenditure categories, not a national budget saving, a proven change in total health spending, or the policy's prospective savings target.
The study's payer-specific contrasts matter. If the same hospital budget is associated with fewer admissions among Medicare beneficiaries but unchanged admissions among commercial-plan patients, one cannot treat the revenue cap as producing a uniform demand response. The populations differ in age, morbidity, contracts, utilization, and exposure to other programs. The authors also emphasized that aligning hospital and nonhospital incentives may be necessary to reduce total spending. This is an empirical limitation on a hospital-only financial assessment.
Maryland's experience also presents a governance lesson. Under the current AHEAD hospital payment design, CMS describes prospective Medicare fee-for-service inpatient and outpatient global budgets adjusted for quality and health outcomes. State implementation occurs within a longer model, with other payer arrangements and primary-care reforms. The HSCRC AHEAD program materials provide contemporaneous implementation detail. Because reform components change over time, evaluation of post-2026 outcomes must identify performance years, population exposure, implementation phases, and national payment changes. An institutional promise to measure equity or population health is a safeguard specification, not evidence that disparities have already narrowed.
Counterevidence from Taiwan
Taiwan's sector-wide budgeting offers a different set of observed responses. Using municipal-hospital panel data, Chang and Hung (2008) reported cost containment accompanied by adverse quality indicators in their sampled facilities. Their conclusion is evidence for a possible trade-off under the observed conditions, not proof of an inevitable national outcome. The narrower sampling frame and potential contemporaneous changes limit extrapolation. It is important to consider that Taiwan's global budget combines spending caps with service-point incentives: under that arrangement, higher service volumes can dilute the value of each point without increasing the total envelope.
In national claims research covering cardio-cerebrovascular inpatient admissions, Kan et al. (2014) reported higher treatment intensity in for-profit and private nonprofit hospitals after the 2002 reform, without corresponding improvements in measured mortality. The result suggests that budget ceilings and claim competition can coexist with more treatments and questionable marginal value. An observed rise in treatment intensity is not, in itself, a positive quality result; equally, the absence of mortality improvement is not proof that every additional intervention was inappropriate. Other relevant outcomes and patient composition require evaluation.
The underprovision concern appears in stroke outcomes. Tung and Chang (2010) analyzed 258,167 ischemic stroke admissions from 1998–2007 and associated reimbursement reductions with increased adjusted 30-day mortality following the global budget's implementation. Residual confounding, secular clinical change, and concurrent reforms remain potential limitations of the observational design. The analysis nevertheless establishes that safety risks deserve direct measurement rather than being inferred away from aggregate spending control. Later interrupted time-series research on Taiwan's stroke care reported both favorable and unfavorable changes in treatment processes following fee-for-service reimbursement cuts; its segmented-trend study reinforces the need to avoid single-endpoint judgments.
Together, these findings reject a one-directional empirical story. Maryland, Taiwan, and primary-care systems differ in unit of payment, spending perimeter, rate setting, provider competition, and data capture. Any synthesis that treats an economy-wide cap on reimbursement points as equivalent to a fixed hospital revenue entitlement erases the mechanisms required to explain their outcomes. The stronger conclusion is conditional: global budgeting can limit the aggregate payer obligation while producing varying provider-level incentives, and the clinical effects must be established in each contract environment.
Clinical quality contracts and public accountability
Quality contracts should specify the covered cohort and denominator for every metric, mandatory minimum standards, independent clinical review, data provenance, missing-data treatment, and response to violations. Risk adjustment for performance must be distinct from payment risk adjustment where their goals differ. Clinical review can examine cases where major outcomes deteriorate or referral patterns shift. Patient complaints, appeals, and access surveys should complement administrative records. The CMS AHEAD model requirements place quality adjustment inside hospital budgets; these provisions demonstrate the institutional relevance of quality safeguards, while their eventual effectiveness remains an evaluative question.
An enforcement framework must also detect exclusion. Monitoring only assigned patients may conceal the fact that patients with severe disease never secured attribution. Useful indicators are new-patient acceptance by morbidity, geographic gaps, attrition following serious diagnoses, and unplanned transfers. Providers need clinically justified exceptions when measured indicators are inappropriate for individual patients. Otherwise, a target to increase testing can itself induce unnecessary testing. Meaningful monitoring therefore combines quantitative thresholds with case-level validation and transparent correction processes.
Distributional equity is integral to quality rather than an appendix to it. Equal capitated revenue for unequal clinical need can deepen existing disparities. Even sophisticated diagnosis-based adjustment can miss housing instability, limited transport, language support needs, disability accommodations, and long travel distances. The English equity-targeted capitation evaluation provides a relevant recent case: the 2021 funding top-up was evaluated in 62 intervention practices and 62 matched comparison practices. Investigators reported a 3.2-percentage-point higher Quality and Outcomes Framework score (95% confidence interval 0.5–5.9) and no statistically significant differences in evaluated staffing or patient-experience measures. Its design supports a measured association under adjustment assumptions, not a claim that every equity objective was achieved.
Data, Measurement, and Validation
A reproducible unit of analysis
A payment study requires a dated, version-controlled data dictionary. At minimum, define the purchaser, beneficiary or member, provider entity, contract type, payer program, service category, covered period, geographic region, and applicable rate schedule. For capitated arrangements record member-months, risk weights, enrollment gaps, retroactive eligibility, carve-outs, stop-loss provisions, reconciliation transfers, and changes in attributed provider. For global budgets record the annual ceiling, revenue within scope, exclusions, reconciliation rules, adjustment formulas, volume accounting, and cross-payer allocations. If the system uses claims points, preserve both the number of points and the eventual monetary conversion factor.
The denominator should match the inferential question. Hospital spending divided by total residents is not equivalent to spending per attributed insured person. Admissions per thousand eligible member-years differ from admissions per thousand active patients. A quality indicator calculated only from completed visits conditions on contact and can conceal a drop in access. A fixed hospital budget may refer to net patient revenue rather than recognized insurer expenditure, particularly where accounting adjustments and out-of-area services are important. Each numerator and denominator must be traceable to primary record categories.
An audit-ready minimum record contains: a contract identifier and effective date; provider and payer entity identifiers; member identifier linkage; contract inclusion and exclusion lists; risk-adjustment version; rate or budget version; payment and encounter dates; realized payments by category; clinically validated quality measures; record-extraction date; data revisions; and explicit treatment of missing records. Ideally the analyst obtains provider cost information as well as claims; a payment reduction that erodes essential staffing capacity may be invisible in payer data before access deteriorates.
Estimands and identification
For any proposed empirical evaluation define the treatment, comparison, population, observation window, outcome, and estimand before selecting a regression. An intention-to-treat estimate includes all people or facilities assigned to the reform, including those who do not fully comply. A treatment-on-treated estimate captures actual participation but can be biased by self-selection. For staggered state or hospital adoption, conventional two-way fixed effects may mix heterogeneous treatment effects across cohorts; cohort-specific estimands and event-study diagnostics are preferable when implementation timing differs materially.
A conventional difference-in-differences estimand is:
Its causal interpretation requires assumptions about counterfactual trends, composition, spillovers, and measurement. A placebo test using pretreatment periods can reveal gross instability but cannot prove the future parallel-trend assumption. Administrative policy changes may alter how conditions are coded, making apparent risk-adjusted mortality changes partly a product of measurement drift. Multiple-comparison correction and prespecified primary endpoints help avoid interpreting chance findings among dozens of quality metrics.
In interrupted time-series analysis, estimate pre-intervention slope, immediate level change, post-intervention slope change, seasonal patterns, and autocorrelation. A post-intervention jump in visits may reflect delayed coding completion or an influenza epidemic rather than the payment rule. Interrupted time series improves on simple before-after comparisons but remains vulnerable to co-occurring reforms. Cross-sectional analyses of physician remuneration require particularly careful control for practice ownership, HMO organization, geographic selection, referral networks, and patient sorting. The contrast between Balkrishnan et al.'s national visit study and Zuvekas and Hill's endogeneity analysis illustrates these identification issues.
A primary-care capitation evaluation should include practice-level and member-level outcomes. The first measures labor, fixed costs, operating margins, and overall capacity; the second measures diagnoses, missed appointments, preventive care, consultations, hospitalization, mortality, and equity. Cluster-robust inference should account for treatment assignment at the practice or organizational level; a large number of encounters does not create a large number of independently treated providers. Where the intervention affects only 27 centers, effective statistical precision is more constrained than the number of recorded visits might suggest. The Oregon community-clinic evaluation makes the distinction between provider-level intervention and encounter-level data particularly relevant.
Sensitivity analysis and validation
Every evaluation should test the consequences of attribution changes, historical coding intensity, alternative comparison groups, reform launch dates, and excluded costs. Hospital findings require separate results for inpatient, emergency, outpatient, physician, pharmacy, post-acute, and total payments where feasible. Cost outcomes should be reported in constant-price terms and, for multi-payer comparisons, under clearly specified rate and benefit equivalence assumptions. Absolute dollar changes and percentage changes should be reported together; percentages may be large when initial levels are small.
Risk-adjustment performance should be tested both in-sample and on independent time periods. A model that predicts historic claims well may degrade after providers begin documenting additional diagnoses in response to a higher payment weight. Testing must distinguish genuine morbidity change from coding response. The Panturu et al. (2025) assessment is especially pertinent because it places the choice of adjusters and weights within explicit incentive trade-offs. Cost-based calibration without diagnostic and clinical audits risks rewarding the production of more expensive-looking cases rather than better care.
Quality endpoints require formal case-mix assessment and numerator integrity checks. For example, a recorded reduction in avoidable admissions may mean better community management, restricted access, migration to a different hospital, or administrative recoding. Linking longitudinal member histories, death records, emergency events, and transfer data helps distinguish these explanations. If no cross-provider linkage exists, the publication must identify that limitation rather than treating a hospital's discharge file as the entire patient pathway.
Subgroup evaluation is essential. Examine morbidity, age, disability, rurality, income or area deprivation, and continuity of insurance separately, with appropriate privacy and statistical safeguards. Extreme-cost cases deserve explicit analysis because the mean can mask a small number of patients carrying exceptional clinical and financial risk. Subgroup effect heterogeneity must be tested with interaction estimates and uncertainty intervals; separate statistically significant and insignificant subgroup results do not automatically establish a significant difference between groups.
The evidence record and its limitations
This monograph's evidence register pairs each external research link with the specific subject or claim checked on the official institutional page or indexed scholarly record. Scientific articles are interpreted at the level of verified publication metadata and available abstracts or accessible full text; inaccessible paywalled details are not represented as independently examined. The register distinguishes institutional policy descriptions, method guidance, systematic reviews, and empirical designs. Bibliographic completeness is pursued using original publisher metadata, PubMed records, or institutional publication pages. The findings do not constitute a formal evidence synthesis with a reproducible comprehensive search strategy or a pooled meta-analysis.
The absence of original data restricts inference in important ways. This inquiry cannot recompute Maryland counterfactual spending, re-estimate Taiwan's mortality associations, verify raw patient-level claims, or adjudicate disputed causal identification assumptions from original microdata. It can compare the reported findings, state their denominators and caveats, and test logical consistency between payment structures and empirical interpretations. These are substantive scholarly tasks, but their evidentiary level must be represented accurately.
A further limitation arises from rapid institutional change. Maryland's move to AHEAD in January 2026 means a 2014–2018 evaluation and the model's 2026 website describe different regulatory moments. The Maryland Department of Health timeline corroborates the transition, while CMS describes the national model's current participating states and intended time horizon. Likewise, a February 2026 Medicaid guide is appropriate for its specified state-to-plan rating periods, not for all capitation contracts at every time. Source dates and scope are conditions of valid comparison.
International Synthesis and Scientific Conclusions
United States: regulated hospital revenue and payer alignment
Maryland demonstrates what happens when hospital-level revenue responsibility is created within a rate-regulated all-payer environment. Its distinctive feature is that a hospital has an incentive to contain spending within a defined fixed revenue entitlement while managing its services and community relationships. The Morrison et al. (2021) evaluation provides support for slower hospital spending growth and selected utilization reductions under the All-Payer Model, particularly for Medicare. The interpretation requires attention to the comparison design, payer differences, and spending outside hospitals. Stronger conclusions about systemwide efficiency require linking hospital payments to physician, pharmacy, rehabilitation, home-care, and patient costs.
The 2026 AHEAD transition expands the policy horizon. CMS presents a structure in which participating states coordinate hospital budget incentives, primary-care transformation, multi-payer alignment, and population health targets. The current CMS program description supplies the institutional design and participating jurisdictions. At the review date, AHEAD's published goals and payments cannot be substituted for independent long-term effectiveness estimates; reforms starting in 2026 require future causal assessment using actual implementation data. Moreover, participating states have different starting institutions. A statewide budget model pioneered through decades of Maryland rate regulation is not an immediately transferable contract template for every other state.
Medicaid provides a separate lesson. State-to-plan capitation obliges actuaries to estimate the insurer's cost of furnishing covered benefits to an eligible population, but the plan may then use fee for service or value-based arrangements downstream. The CMS rate-review instructions make clear the regulatory importance of certified actuarially sound plan rates. Analyses that merge state insurer payments and provider capitation into one undifferentiated “capitation effect” risk double-counting exposure and misidentifying the entity that bears marginal cost.
Taiwan: aggregate control and relative-volume competition
Taiwan demonstrates the difference between aggregate expenditure control and uniform restraint at the facility level. A sector ceiling combined with fee-for-service point claims determines a common financial pool while leaving relative claim volume strategically valuable to individual providers. The Cheng et al. (2009) findings of increased resource intensity after implementation, and the Kan et al. (2014) findings of higher treatment intensity without measured mortality benefit in some ownership groups, are coherent with this institutional design. They do not prove that the budget alone caused every reported difference, but they make an unqualified assumption of falling service volume untenable.
Tung and Chang's stroke mortality assessment adds a different form of counterevidence. Even when aggregate spending growth is controlled, reimbursement pressure can be associated with worse outcomes among clinically vulnerable inpatients. A scientific evaluation therefore needs both financial and clinical denominators, with attention to whether hospitals face uniform conversion-factor reductions or differentiated financial pressures. Taiwan's experiences show why quality safeguards should be assessed at the disease and patient levels rather than inferred from macro-level budget adherence.
England, Oregon, and Ukraine: primary-care payment adequacy
England's primary-care system illustrates the importance of weighted allocation and supplementary funding. The Opie-Martin et al. (2026) study evaluates a targeted health-equity top-up rather than a wholesale conversion from fee for service. Its measured improvement in the quality indicator and absence of significant effects on other reported endpoints imply that an additional allocation can change some outcomes while leaving others empirically unresolved. The relevant scientific unit is the marginal supplement layered over an existing practice funding scheme.
Oregon's community health centers provide a different experiment in prospective primary-care payment. In the Ukhanova et al. (2021) comparison, preventive measures did not move uniformly. This pattern is compatible with heterogeneity in practice behavior and patient services, and it limits any claim that one payment type determines all aspects of preventive care. More generally, quality evaluation must consider the specific services within the provider's contractual responsibility, baseline service access, and support for delivery transformation.
Ukraine's costing exercise raises the question that precedes incentive design: whether the prospective amount is adequate to purchase the promised service. The WHO/Europe technical report records a historical cost-based gap in sampled primary-care providers. The figures refer to 2023 estimates developed from 2021 data, and cannot be extrapolated to current prices without re-estimation. Their scientific significance lies in the structure of a valid pricing inquiry: identify costs actually paid by providers, distinguish supplementary budget financing, and test whether a uniform capitation rate covers fixed costs across rural and urban settings.
Comparative design matrix
| Payment arrangement | Main financial unit | Marginal activity incentive | Principal exposure | Priority measurement |
|---|---|---|---|---|
| Primary-care capitation | Attributed member-month | Usually no direct incremental payment for covered visit | Morbidity mix, enrollment, covered-services cost | Appropriateness, access, prevention, referrals, churn |
| State-to-plan managed-care capitation | Eligible member-month at plan level | Depends on downstream contracts | Benefit claims, provider prices, member risk | Plan-level cost, access, reserves, actuarial soundness |
| Hospital-specific global revenue budget | Hospital annual revenue perimeter | Revenue largely constrained within scope | Service costs, volume mix, quality adjustments | Hospital and system costs, admissions, patient outcomes |
| Sector budget with variable service points | Aggregate sector pool plus relative claims | Positive reward for increasing share of service points | Diluted point value, claim competition | Total points, unit value, unnecessary intensity, mortality |
| Fixed line-item budget | Appropriated input categories | Constrained by category allocations | Input insufficiency and fiscal rigidity | Supplies, staffing, budget execution, waiting times |
| Blended prospective and activity payment | Multiple contract components | Depends on carve-outs and bonus weights | Misaligned margins, gaming, transfer costs | Cross-setting episode cost and quality |
This matrix is an original analytical classification of existing institutional mechanisms, not a new formal theory. It arranges comparable accounting attributes to expose differences that are obscured by broad reimbursement labels. The categories are derived from the WHO provider-payment framework, current government model descriptions, and reviewed empirical cases. Actual countries may combine several rows and modify them with risk corridors, supplemental funding, and quality requirements.
Conditions for interpreting cost containment
The international evidence supports a sequential analytical test. First, identify the perimeter of covered services and whether the budget controls provider revenue, insurer expenditure, or sector reimbursement. Second, identify whether marginal activity generates a payment and where that marginal payment is funded. Third, establish who bears demographic and clinical risk, and how prospective weights and exceptional-cost protection modify that burden. Fourth, measure complete patient pathways and clinical results, accounting for spending displaced to institutions outside the formal budget. Fifth, compare observed changes with a credible counterfactual and document uncertainty and institutional context.
This procedure exposes three recurrent sources of mistaken inference. A hospital's falling claims do not establish falling total spending if care shifts elsewhere. A sector-wide expenditure cap does not establish a fall in service volume if individual providers gain from producing more service points. A statistically significant rise in recorded quality indicators does not establish improved health if incentives have changed who is counted. These are analytical deductions from the payment identities and the design limitations documented in the research literature, not independently estimated universal effect sizes.
Governance, regulation, and future inquiry
Government and private purchasers should specify adjustment rules prospectively, publish the contract's included service categories and beneficiary attribution, maintain systems for independent data validation, and define enforceable quality floors. The contract must identify recourse when necessary care is withheld or a provider cannot sustain the level of service funded. Patient appeal rights and audit access are especially important when savings from refusing care accrue to the decision maker. For smaller providers, financial protection against high-cost cases can improve continuity without eliminating incentives for ordinary cost management.
The WHO strategic-purchasing discussion links payment to performance and population needs, while its mixed-provider-payment analysis highlights governance across overlapping purchasers. These frameworks align with the comparative findings: neither a capitation formula nor a global-budget number specifies the full institutional incentive environment. Decisions about benefits, attribution, rate adequacy, information rights, quality enforcement, and risk-sharing are constitutive parts of the arrangement.
Future empirical research should prioritize patient-linked, multi-payer studies covering the shift between hospital and community settings; prospective evaluation of AHEAD jurisdictions after implementation; separate analysis of Taiwan's floating point values and clinical intensity; provider-level financial resilience under high-risk patient concentrations; patient access effects of capitation adjustment formulas; and long-term clinical and financial changes after targeted primary-care supplements. These are research directions implied by evidence gaps, not claims of established new theoretical systems.
Conclusions
Capitation transfers specified service-cost risk from the purchaser to a provider in relation to an attributed population and time period. Its economic effect depends on covered benefits, enrollment stability, patient need, risk adjustment, provider size, and protection against extraordinary cost. A flat rate can create valuable incentives for prevention and efficient coordination, while inadequate adjustment and weak monitoring can increase selection and underprovision incentives. Evidence from primary-care studies identifies heterogeneous associations and effects rather than a single universal clinical outcome.
Global budgets have two particularly important forms. A hospital-specific revenue envelope constrains institutional payments over an agreed period. A sector-wide cap implemented through relative service points controls collective spending while permitting each provider to gain by increasing its share of claims. Their mathematical accounting differs, and so do their incentives. Maryland and Taiwan document why the words “global budget” alone cannot define a treatment for comparative research.
Expenditure predictability, clinical efficiency, and population health are distinct outcomes. A decline in hospital spending may reflect useful prevention, displacement of services, coding change, or reduced appropriate treatment. Financial evaluation must therefore be integrated with longitudinal patient outcomes, risk-standardized quality, access, equity, and full-pathway costs. Published observational findings warrant interpretation within the specific comparison, population, period, and institutional contract studied.
The central synthesis is that prospective payment allocates residual risk through the boundary of the covered service or spending perimeter. That allocation generates incentives whose consequences depend on risk adjustment, competing payment streams, provider capabilities, and enforceable patient protections. Effective capitation and global budgeting consequently require verifiable population attribution, actuarially adequate rates, a transparent budget reconciliation method, accessible clinical-quality safeguards, and causal assessment of systemwide—not merely account-level—outcomes.
No numerical estimates beyond those attributed to cited studies are claimed as original empirical results. No external peer review or independent patient-level reanalysis is claimed. The monograph's original contribution is its comparative source-critical synthesis of distinct payment identities, risk allocation, adverse incentives, and evidence constraints across jurisdictions.
Technical Appendix: Reproducible Definitions and Tests
Accounting identities and boundaries
Define for each contract: N attributed patients, eᵢ eligible time fraction, rᵢ annual adjusted capitation entitlement, L eligible provider service costs, S direct supplemental revenue, and T reconciliation and stop-loss transfers. Define provider financial margin from the covered arrangement as M = Σeᵢrᵢ + S + T − L. This is an accounting identity only; it omits unmeasured capital and costs unless included in L. If quality bonuses are separately funded, classify them within S and identify whether they enter the capitation base or the subsequent annual reconciliation.
For a hospital global budget, let B denote agreed annual net revenue, A permitted adjustments for population shifts, inflation, quality, and other contract terms, and R realized covered receipts. A simple budget compliance gap is D = R − (B + A). The financial consequences of a gap depend on formal reconciliation and carry-forward provisions; D is not necessarily hospital profit or insurer savings. Hospital operating surplus requires subtracting actual operating costs from recognized revenues and tracing all other payments and obligations.
For a point-based sector cap, let G denote the common financing pool and qⱼ each provider's recorded eligible claims points. Then Q = Σqⱼ and realized provider revenue is Pⱼ = Gqⱼ/Q. Holding all other provider points constant, the derivative with respect to the provider's own points is ∂Pⱼ/∂qⱼ = G(Q−qⱼ)/Q², which is positive when other providers claim any points. Holding the total Q mechanically fixed is not appropriate when changing one provider's qⱼ; the denominator changes endogenously. The derivative demonstrates the incentive for relative claim competition under a sector ceiling and does not estimate its behavioral elasticity.
Minimal preregistered empirical reporting template
For each outcome report the jurisdiction and reform phase, unit of observation, payer and provider population, original and comparison assignment rules, treatment start, observation dates, eligibility exposure denominators, contract coverage, potential spillovers, missingness, inflation adjustment, effect estimator, confidence interval, and a sensitivity test for alternative control groups or intervention dates. At least one clinical outcome and one access outcome should accompany any primary financial outcome when the proposed claim concerns health system performance. When clinical data are unavailable, state that the financial estimand cannot establish quality or welfare.
Predefine how to assess residual risk: compare actual covered costs with risk-adjusted revenue for age, clinical burden, rurality, deprivation, and exceptional costs; report distributions and tail outcomes rather than averages only. A payment reform can reduce aggregate variance for the payer while transferring concentrated volatility to smaller providers. The contract should specify both financing adequacy and provider solvency protection.
Source-critical decision rules
Evidence is strongest when a claim matches the source's actual object. Institutional pages establish current descriptions, not proof of program success. Peer-reviewed observational work establishes associations and qualified counterfactual estimates, not automatic external validity. A direct mathematical identity establishes payment mechanics under defined assumptions, not observed behavior. Each cited result is assigned its own source record with a date and scope statement in the accompanying EVIDENCE.json. Source-linked numerical and program facts in this text are restricted to claims supported by those records.
References
Balkrishnan, R., Hall, M. A., Mehrabi, D., Chen, G. J., Feldman, S. R., & Fleischer, A. B., Jr. (2002). Capitation payment, length of visit, and preventive services: Evidence from a national sample of outpatient physicians. American Journal of Managed Care, 8(4), 332–340. PubMed record.
Centers for Medicare & Medicaid Services. (n.d.). AHEAD (Achieving Healthcare Efficiency through Accountable Design) Model. CMS program page. Accessed November 14, 2026.
Centers for Medicare & Medicaid Services. (n.d.). Maryland All-Payer Model. CMS program record. Accessed November 14, 2026.
Centers for Medicare & Medicaid Services. (n.d.). Medicare Shared Savings Program: About the Program. CMS program information. Accessed November 14, 2026.
Centers for Medicare & Medicaid Services. (2026). Rate Review and Rate Guides: 2026–2027 Medicaid Managed Care Rate Development Guide. Medicaid program guidance. Accessed November 14, 2026.
Chang, L., & Hung, J.-H. (2008). The effects of the global budget system on cost containment and the quality of care: Experience in Taiwan. Health Services Management Research, 21(2), 106–116. PubMed and DOI record. DOI: 10.1258/hsmr.2008.007026.
Cheng, S.-H., Chen, C.-C., & Chang, W.-L. (2009). Hospital response to a global budget program under universal health insurance in Taiwan. Health Policy, 92(2–3), 158–164. PubMed and DOI record. DOI: 10.1016/j.healthpol.2009.03.008.
European Observatory on Health Systems and Policies. (n.d.). Payment mechanisms. WHO/Observatory thematic page. Accessed November 14, 2026.
Gosden, T., Forland, F., Kristiansen, I. S., Sutton, M., Leese, B., Giuffrida, A., Sergison, M., & Pedersen, L. (2000). Capitation, salary, fee-for-service and mixed systems of payment: Effects on the behaviour of primary care physicians. Cochrane Database of Systematic Reviews, (3), CD002215. PubMed and DOI record. DOI: 10.1002/14651858.CD002215.
Health Services Cost Review Commission, Maryland. (n.d.). AHEAD Model. HSCRC program information. Accessed November 14, 2026.
Health Services Cost Review Commission, Maryland. (n.d.). Maryland’s Total Cost of Care Model. HSCRC model chronology. Accessed November 14, 2026.
Hsu, P.-F. (2014). Does a global budget superimposed on fee-for-service payments mitigate hospitals' medical claims in Taiwan? International Journal of Health Care Finance and Economics, 14(4), 369–384. PubMed and DOI record. DOI: 10.1007/s10754-014-9149-6.
Kan, K., Li, S.-F., & Tsai, W.-D. (2014). The impact of global budgeting on treatment intensity and outcomes. International Journal of Health Care Finance and Economics, 14(4), 311–337. PubMed and DOI record. DOI: 10.1007/s10754-014-9150-0.
Maryland Department of Health. (n.d.). Maryland's AHEAD Primary Care Programs. Maryland AHEAD transition page. Accessed November 14, 2026.
Morenz, A. M., Zhou, L., Wong, E. S., & Liao, J. M. (2023). Association between capitated payments and preventive care among U.S. adults. AJPM Focus, 2(3), 100116. PubMed and DOI record. DOI: 10.1016/j.focus.2023.100116.
Morrison, M., Haber, S., Beil, H., Giuriceo, K., & Sapra, K. (2021). Impacts of Maryland's global budgets on Medicare and commercial spending and utilization. Medical Care Research and Review, 78(6), 725–735. PubMed and DOI record. DOI: 10.1177/1077558720954693.
Opie-Martin, S., Tracey, F., Whitfield, E., Co, M., Beech, J., Pettigrew, L. M., Clarke, J. M., & Lloyd, T. (2026). Evaluating the impact of capitation funding top-up payments in primary care. BMC Primary Care, 27, article 48. PubMed and DOI record. DOI: 10.1186/s12875-025-03136-x.
Panturu, A., van Kleef, R., Eijkenaar, F., & Cattel, D. (2025). A framework for the design of risk-adjustment models in healthcare provider payment systems. Medical Care Research and Review, 82(1), 43–57. PubMed and DOI record. DOI: 10.1177/10775587241273355.
Schokkaert, E., & Van de Voorde, C. (2004). Risk selection and the specification of the conventional risk adjustment formula. Journal of Health Economics, 23(6), 1237–1259. PubMed and DOI record. DOI: 10.1016/S0167-6296(03)00040-7.
Tung, Y.-C., & Chang, G.-M. (2010). The effect of cuts in reimbursement on stroke outcome: A nationwide population-based study during the period 1998 to 2007. Stroke, 41(3), 504–509. PubMed and DOI record. DOI: 10.1161/STROKEAHA.109.568956.
Tung, Y.-C., Chang, G.-M., & Cheng, S.-H. (2015). Long-term effect of fee-for-service-based reimbursement cuts on processes and outcomes of care for stroke: Interrupted time-series study from Taiwan. Circulation: Cardiovascular Quality and Outcomes, 8(1), 30–37. PubMed and DOI record. DOI: 10.1161/CIRCOUTCOMES.114.001086.
Ukhanova, M., Marino, M., Angier, H., Jacob, L., O’Malley, J., Cottrell, E. K., Dambrun, K., & Heintzman, J. (2021). The impact of capitated payment on preventive care utilization in community health clinics. Preventive Medicine, 145, 106405. PubMed and DOI record. DOI: 10.1016/j.ypmed.2020.106405.
van Kleef, R. C., & van Vliet, R. C. J. A. (2022). How to deal with persistently low/high spenders in health plan payment systems? Health Economics, 31(5), 784–805. PubMed and DOI record. DOI: 10.1002/hec.4477.
World Health Organization. (2019). Analytical guide to assess a mixed provider payment system. WHO/UHC/HGF/Guidance/19.5. Geneva: WHO. ISBN 978-92-4-151533-7. WHO publication record.
World Health Organization. (n.d.). Health financing. WHO thematic overview. Accessed November 14, 2026.
World Health Organization. (n.d.). Health Financing and Economics: Public Financial Management. WHO technical overview. Accessed November 14, 2026.
World Health Organization Regional Office for Europe. (2024). Results of costing for provider payments in primary health care in Ukraine: Technical report. WHO/EURO:2024-8277-48049-71211. Copenhagen: WHO Regional Office for Europe. WHO/Europe publication record.
Zuvekas, S. H., & Hill, S. C. (2004). Does capitation matter? Impacts on access, use, and quality. Inquiry, 41(3), 316–335. PubMed publication record.