There is growing interest among insurers and policy makers in the United States in primary care playing an expanded role to improve quality and lower cost. These activities often are not costless for primary care practices or clinics (PCCs) to implement, and the costs are not always reimbursed. Thus, a PCC facing expectations to provide additional, unreimbursed services is likely to consider their return on investment (ROI). Additionally, when private businesses are used to implement government policy, the objectives, resources, and constraints of the businesses need to be carefully considered. We develop a general model for the PCC’s ROI in a tiered cost-sharing health insurance benefit design in which PCCs with lower total risk-adjusted annual per capita cost of care (TCOC) are assigned to a tier with lower consumer cost-sharing, leading to an increase in patient volume. We focus on price discounts as one way to lower the PCC’s TCOC, but the same analysis would apply to any unreimbursed effort to make the PCC more attractive to prospective patients, e.g., in a capitation system with regulated fees. We use data from a large state employee insurance program that uses such a TCOC design. The study finds that discounts of 10–20 percent can have a positive ROI for the PCC, meaning a PCC may generate more income through a gain of patient volume than is lost through lowering their unit prices. These results are sensitive to the model’s parameters. The study finds that primary care clinics offering 10–20 percent price discounts within a tiered cost-sharing insurance model can achieve a positive return on investment. Increased patient volume from lower cost-sharing tiers can offset the revenue loss from discounts. However, the profitability of these strategies depends strongly on specific model parameters and assumptions.

The term "upcoding" has several interpretations. In a recent scoping review, the RAND Corporation defines upcoding as, "…the coding of a patient to a higher complexity level than they would be if payment were unrelated to complexity." [1] In some contexts, upcoding implies fraudulent behavior. For example, Coustasse defines upcoding as submission of payment codes for more severe and expensive diagnoses or procedures than the provider actually diagnosed or performed [2]. The Centers for Medicare and Medicaid Services (CMS) provides the following example of upcoding: billing a follow-up visit using a higher-level evaluation and management code, such as a comprehensive new-patient office visit [3]. Coding diagnoses for medical conditions that the patient never had certainly constitutes fraud, but CMS already allows coding additional diagnoses that increase the cost of treating a patient for a specific condition. The adjustment is made through interaction terms among diagnoses in the computation of risk scores. For example, capitation payment for a diabetic patient is higher if the patient also has congestive heart failure [4]. The following analysis assumes there is universal agreement that fraudulent upcoding is rightly illegal and focuses on a suggestion to improve the accuracy of legal upcoding. There is an adage, "Never assume malevolence when stupidity works just as well." A modification might read, "Never assume malevolence when Econ 101 works just as well." A useful guide to understanding human behavior is that people generally do what they are paid to do. In traditional fee-for-service (FFS) Medicare (TM), health care providers are paid to provide services to patients. In contrast, MA plans are paid a capitation amount, and the capitation amount is higher for patients with more medical conditions. Thus, MA plans are rewarded for coding more diagnoses. If we rule out fraudulent coding then the additional diagnoses coded by the MA plan compared to a similar patient in TM represent diagnoses that are observed and documented, but not necessarily diagnoses that increased the cost of treating the patient over the time period used to compute the capitation payment. CMS understands the problem and has responded by reducing the capitation payments to MA plans by 5.9%–an amount that CMS is required to apply, but less than CMS is authorized to use [5]. This uniform adjustment has been criticized, because upcoding varies from MA one plan to another [6]. Upcoding is far from an MA-only problem, although MA accounts for the greatest economic impact of upcoding [7]. In its Accountable Care Organization (ACO) alternative payment models, CMS pays ACOs on a FFS basis, but ACOs are subject to rewards and penalties based on their performance relative to an administratively determined benchmark. The benchmark, in turn, is computed using diagnosis-based risk adjustment. The more diagnoses that are coded for the patient, the higher the benchmark, and the greater the probability that the ACO will be rewarded with shared savings, rather than penalized. As a result, ACOs that face the risk of a financial penalty for spending above the benchmark have a similar incentive to upcode as MA plans, and they respond accordingly [8]. In response, CMS has found it necessary to impose upcoding ceilings on ACOs in TM, similar to those imposed on MA plans. A recent study also found upcoding by hospitals, whose prospective payments are a function of diagnosis-based risk payments [9]. Upcoding is symptomatic of larger issues in provider payments. Should providers be paid for services or for diagnoses? FFS payments or risk-adjusted capitation? Each payment system has advantages and disadvantages. FFS payment provides a claims-based record of what actually was done to the patient, but FFS payment also encourages the prescription of low-value and wasteful services [10]. Capitation payment, adjusted for the enrollee's diagnosis-based "risk" discourages overuse, but encourages upcoding and skimping on services unless health outcomes are closely monitored. To balance these incentives, Newhouse proposed a blended payment system [11]. CMS's efforts to improve risk adjustment in its V28 initiative involve eliminating some of its Hierarchical Condition Categories (HCCs), and altering the mapping of diagnoses into the remaining HCCs [12]. Lieberman and Ginsburg [6] and Jung, Carlin, and Feldman [13] propose upcoding adjustments that are tied to the MA plan's observed level of upcoding. Other proposals include limiting the sources of diagnosis codes and developing diagnosis weights specific to the MA sector [13]. This approach will not solve all the problems of the Medicare program. For example, outlays will still exceed revenue [14], and TM's level of resource use and fee schedule will continue to set the payment rate for diagnoses in the capitation system [13, 15]. Any payment system will require monitoring and some further diagnosis interactions may need to be added to CMS's current list. But MedPAC estimates that losses to upcoding reached $50 billion in 2024 [5], and thus a sizable investment in addressing the issue is justified. The author declares no conflicts of interest. The data that support the findings of this study are found in the references.

Purpose – The paper aims to explore the use of dashboards in healthcare organizations and their potential for evaluating financial, operational, clinical, and patient centric metrics. Design/methodology/approach – In this study, we scanned articles in the databases PubMed, Cochrane Library and Ovid MEDLINE. In the review process, the articles included comprised of two categories (information needs and dashboard use in healthcare). In the first phase, the needs were identified for management, providers, consumers, and employers and in the subsequent phase, the role of dashboards was evaluated to serve the needs of the identified groups. Findings – The paper provides insights into how dashboards should be designed to meet the information needs of the end user. Further, the paper identifies key information needs of healthcare organization’s management, providers, consumers and employers and the role of dashboards in improving quality, effective care. Research limitations/implications – Because of the chosen research approach, the research results may lack comprehensiveness. Practical implications – The paper informs the design of dashboards in terms of information and indicators in the context of healthcare organizations. Originality/value – Previous research discussing dashboards in healthcare reports the use of different types of dashboards in healthcare. While much research focuses on synthesizing existing dashboard use, literature is scarce in identifying how dashboards can be used in meeting the information needs of end users. Keywords: Primary care clinic (PCC), sponsor, total cost of care (TCC), cost, tiered health plan, return on investment (ROI), electronic health records (EHR) The paper aims to explore the use of dashboards in healthcare organizations and their potential for evaluating financial, operational, clinical, and patient centric metrics. In this study, we scanned articles in the databases PubMed, Cochrane Library and Ovid MEDLINE. In the review process, the articles included comprised of two categories (information needs and dashboard use in healthcare). In the first phase, the needs were identified for management, providers, consumers, and employers, and in the subsequent phase, the role of dashboards was evaluated to serve the needs of the identified groups The paper provides insights into how dashboards should be designed to meet the information needs of the end user. Further, the paper identifies key information needs of healthcare organization’s management, providers, consumers and employers and the role of dashboards in improving quality, effective care. The design of dashboards should align with the end goal and needs of the user. While middle and senior management would be interested in tactical and strategic dashboards (operational and financial), PCC staff would monitor operational and clinical metrics using an operational dashboard. PCC providers are likely to use a clinical and patient-centric dashboard to ensure efficient care delivery. The goal of a patient-centric dashboard would be to capture information about patient satisfaction and help them schedule, communicate and exchange info with specialists. An effective user-centric dashboard design would be made possible by keeping in mind of the objectives of the end user.

The Next Generation Accountable Care Organization (NGACO) model (active during 2016-21) tested the effects of high financial risk, payment mechanisms, and flexible care delivery on health care spending and value for fee-for-service Medicare beneficiaries. We used quasi-experimental methods to examine the model's effects on Medicare Parts A and B spending. Sixty-two ACOs with more than 4.2 million beneficiaries and more than 91,000 practitioners participated in the model. The model was associated with a $270 per beneficiary per year, or approximately $1.7 billion, decline in Medicare spending. After shared savings payments to ACOs were included, the model increased net Medicare spending by $56 per beneficiary per year, or $96.7 million. Annual declines in spending for the model grew over time, reflecting exit by poorer-performing NGACOs, improvement among the remaining NGACOs, and the COVID-19 pandemic. Larger declines in spending occurred among physician practice ACOs and ACOs that elected population-based payments and risk caps greater than 5 percent.

OBJECTIVE:To assess key birth outcomes in an alternative maternity care model, midwifery-based birth center care. DATA SOURCES:The American Association of Birth Centers Perinatal Data Registry and birth certificate files, using national data collected from 2009 to 2019. STUDY DESIGN:This observational cohort study compared key clinical birth outcomes of women at low risk for perinatal complications, comparing those who received care in the midwifery-based birth center model versus hospital-based usual care. Linear regression analysis was used to assess key clinical outcomes in the midwifery-based group as compared with hospital-based usual care. The hospital-based group was selected using nearest neighbor matching, and the primary linear regressions were weighted using propensity score weights (PSWs). The key clinical outcomes considered were cesarean delivery, low birth weight, neonatal intensive care unit admission, breastfeeding, and neonatal death. We performed sensitivity analyses using inverse probability weights and entropy balancing weights. We also assessed the remaining role of omitted variable bias using a bounding methodology. DATA COLLECTION:Women aged 16-45 with low-risk pregnancies, defined as a singleton fetus and no record of hypertension or cesarean section, were included. The sample was selected for records that overlapped in each year and state. Counties were included if there were at least 50 midwifery-based birth center births and 300 total births. After matching, the sample size of the birth center cohort was 85,842 and the hospital-based cohort was 261,439. PRINCIPAL FINDINGS:Women receiving midwifery-based birth center care experienced lower rates of cesarean section (-12.2 percentage points, p < 0.001), low birth weight (-3.2 percentage points, p < 0.001), NICU admission (-5.5 percentage points, p < 0.001), neonatal death (-0.1 percentage points, p < 0.001), and higher rates of breastfeeding (9.3 percentage points, p < 0.001). CONCLUSIONS:This analysis supports midwifery-based birth center care as a high-quality model that delivers optimal outcomes for low-risk maternal/newborn dyads.

OBJECTIVES:To explain key challenges to evaluating Center for Medicare and Medicaid Innovation (CMMI) accountable care organization (ACO) models and ways to address those challenges. STUDY DESIGN:We enumerate the challenges, beginning with the conception of the alternative payment model and extending through the decision to scale up the model should the initial evaluation suggest that the model is successful. The challenges include churn at the provider and ACO levels, beneficiary leakage and spillover, participation in prior payment models, and determinants of shared savings and penalties. METHODS:We explain challenges posed in evaluations of voluntary ACO models vs models in which ACOs are randomly assigned to the treatment group. We also note the relationship between the design used in an evaluation and subsequent plans for scaling up successful models. RESULTS:The optimal research design is inextricably tied to the plans for scaling up a successful model. Decisions regarding churn, leakage, spillover, and participating in past payment models can alter the estimated effects of the intervention on participants in the model. CONCLUSIONS:If CMMI intends to offer the model to a larger, but similar, group of volunteers, then the estimated treatment effect based on voluntary participants may be the most policy-relevant parameter. However, if the scaled-up population has different characteristics than the evaluation sample, perhaps due to mandatory participation, then the evaluator will need to employ pseudo-randomization appropriate for observational data.

Health Services Research encourages authors to report marginal effects instead of odds ratios for logistic regression with a binary outcome. Specifically, in the instructions for authors, Manuscript Formatting and Submission Requirements, section 2.4.2.2 Structured abstract and keywords, it reads "Reporting of odds ratios is discouraged (marginal effects preferred) except in case-control studies" (see the HSR website https://www.hsr.org/authors/manuscript-formatting-submission-requirements). We applaud this decision. We also encourage other journals to make the same decision. It is time to end the reporting of odds ratios in the scientific literature for most research studies, except for case–control studies with matched samples. HSR's decision is due to increasing recognition that odds ratios are not only confusing to non-researchers,1, 2 but that researchers themselves often misinterpret them.3, 4 Odds ratios are also of limited utility in meta-analyses. Marginal effects, which represent the difference in the probability of a binary outcome between comparison groups, are more straightforward to interpret and compare. Below, we illustrate the difficulties in interpreting odds ratios, outline the conditions that must be met for odds ratios to be compared directly, and explain how marginal effects overcome these difficulties. Consider a hypothetical prospective cohort study of whether a new hospital-based discharge program affects the 30-day readmission rate, a binary outcome, observed for each patient who is discharged alive. The program's goal is to help eligible patients avoid unnecessary readmissions, and patients are randomized into participating in the program or not. Suppose that a carefully designed study estimates the logistic regression coefficient (the log odds) on the discharge program to be − 0.2 $$ -0.2 $$ , indicating that readmission rates are lower for patients who participate in the discharge program than patients who do not. When writing about the results, the researcher must decide how to report the magnitude of the change and has several choices for how to do so. One option is to report the odds ratio, which in this case is 0.82 = exp − 0.2 $$ 0.82=\exp \left(-0.2\right) $$ , and then compare it with other published odds ratios in the literature. However, this estimated odds ratio of 0.82 depends on an unobservable scaling factor that makes its interpretation conditional on the data and on the model specification.3, 5 As odds ratios are scaled by different unobservable factors and are conditional on different model specifications, the estimated odds ratio cannot be compared with any other odds ratio.6, 7 Even within a single study, odds ratios based on models including different sets of covariates cannot be compared. It would be more accurate to report that, "The estimated odds ratio is 0.82, conditional on the covariates included in the regression, but a different odds ratio would be found if the model included a different set of explanatory variables." Due to an unobserved scaling factor that is included in every estimated odds ratio, odds ratios are not generalizable. Odds ratios from different covariate specifications within the same study or between different studies can almost never be compared directly. The explanation for this requires an understanding of how logistic regression differs from linear regression.3 In least squares regression, adding covariates that predict the outcome—but are independent of other covariates (and are therefore not mediators or confounders)—does not change either the estimated parameters or the marginal effects. Adding more independent covariates to a linear regression just reduces the amount of unexplained variation, which reduces the error variance ( σ 2 $$ {\sigma}^2 $$ ), and results in smaller standard errors for each parameter or marginal effect because of improved precision. For example, in a perfectly executed randomized controlled trial (RCT), the assignment to treatment is independent of all covariates, and the covariates are balanced in the treatment and comparison groups. In a perfectly executed RCT, the estimated treatment effect from a least squares regression should be the same whether covariates are included or not. The only difference in the estimated treatment effect with or without covariate adjustment is the standard errors. Including covariates corrects for any imbalance in the covariates resulting from sampling variation. Adding covariates thus improves statistical significance while leaving the expected value of the estimated treatment effects unchanged. This result does not carry over to logistic regression (or to probit regression). In contrast to linear regression applied to the RCT, adding covariates will change the estimated coefficients in a logistic regression of a binary outcome from the same RCT, even when those added covariates are not confounders.3-7 Therefore, the estimated odds ratios also change unlike the linear regression where the estimated coefficients do not change. The reason that the odds ratios change is because the estimated coefficients in a logistic regression are scaled by an arbitrary factor equal to the square root of the variance of the unexplained part of binary outcome, or σ $$ \sigma $$ . That is, logistic regressions estimate β / σ $$ \beta /\sigma $$ , not β $$ \beta $$ (for the full mathematical derivation, see Norton and Dowd3). Furthermore and more problematic, σ $$ \sigma $$ is unknown to the researcher. Because the estimated coefficients in a logistic regression are scaled by an arbitrary factor σ $$ \sigma $$ , the odds ratios are also scaled by an arbitrary factor (odds ratio = exp β / σ $$ \exp \left(\beta /\sigma \right) $$ ). Ideally, this arbitrary scaling factor σ $$ \sigma $$ would be invariant to changes in covariate specification, but it is not. In fact, this scaling factor changes when more explanatory variables are added to the logistic regression model, because the added variables explain more of the total variation and reduce the unexplained variance and reduce σ $$ \sigma $$ . Therefore, adding more independent explanatory variables to the model will increase the odds ratio of the variable of interest (e.g., treatment) due to dividing by a smaller scaling factor (σ), which does not occur when representing the strength of association via relative risks or absolute risks. In the same perfectly executed RCT, including additional covariates to a logistic regression on a binary outcome would change the magnitude of the estimated treatment effect (log odds, β / σ $$ \beta /\sigma $$ ) and the corresponding odds ratio ( exp β / σ $$ \exp \left(\beta /\sigma \right) $$ ). As a result, the interpretation of the odds ratio depends on the covariates included in the model. A comparison of ORs from prior literature is not meaningful if either the covariate specification is different or if the sample is different because the unknown σ $$ \sigma $$ is different for each study. In the readmission example above, a clearer option would be to report marginal effects in terms of a percentage point change in the probability of readmission, along with the base readmission rate for context.8 In health services research, the most common way of reporting marginal effects is through average marginal effects—the average of the marginal effects computed for each person. These are interpreted as the mean percentage point difference—not the percent difference—in outcome probabilities that accompany a change in the treatment variable's value. For binary treatments, an alternative is to present the predicted probabilities of the outcome when the treatment variable equals 0 and 1. Marginal effects are much less sensitive to the unknown scaling factor and exhibit little change when independent covariates are added to the logistic regression model. When averaged, many of these small changes cancel out.3 The magnitude of average marginal effects can be compared across different studies, whereas the magnitude of odds ratios cannot. For this reason, marginal effects are preferable to report from logistic regression from RCTs and nonrandomized studies. By extension from odds ratios not being comparable across studies due to unknown scaling factors being different, they have limited utility in systematic reviews and meta-analyses. Marginal effects overcome these difficulties. Similarly, marginal effects are preferable to odds ratios or coefficients when using logistic regression to generate predictive models that will be applied to other populations. The magnitude of the unknown scaling factor in odds ratios or log odds will differ across populations, limiting the generalizability of a predictive model to a population other than the one in which it is tested and trained. The choice of how to report results from a logistic regression is important because logistic regression is one of the most common statistical tools in the health services research toolkit. It is also important that researchers—especially researchers who study public policies and quality of care—communicate their results and conclusions clearly to other researchers, policymakers, and the public. Therefore, HSR's stand on odds ratios will help improve interpretation and communication. We commend Health Services Research for deciding to discourage the reporting of odds ratios in most studies. We agree wholeheartedly with this decision, which keeps Health Services Research at the forefront of best practices. Dr. Maciejewski was also supported by a Research Career Scientist award from the Department of Veterans Affairs (RCS 10-391).

Many interventions are based on voluntary participation in the treatment group and difference-in-differences (DID) models frequently are used to estimate the effect of the treatment on treatment group versus the untreated control group. Expected gains in the form of resolve or capacity to adhere to the intervention are likely to be unobserved by the analyst and affect outcomes only after the subject learns the actual content of the intervention effect. When an omitted variable is both time-varying and subject-varying, it will not be undetectable by all the usual DID specification tests, including tests of the parallel trends assumption, and will not be corrected by the standard two-way fixed effect model. Both the internal and external validity of estimated treatment effect can be threatened, whether the estimates are biased from a policy standpoint depends on how the intervention will be expanded if it proves to be successful. When the analyst suspects that unobserved expected gains are a source of bias in a DID model, there are a number of appropriate econometric methods available that double as specification tests. We provide a simulation example to show how the problem arises, and how it can be addressed.

OBJECTIVES:Private managed care plans in the Medicare Advantage (MA) program have been gaining market share relative to traditional fee-for-service Medicare (TM), yet there are no obvious structural changes to Medicare that would explain this growth. Our goal is to explain the growth in MA market share during a period when it increased dramatically.STUDY DESIGN:Data are drawn from a representative sample of the Medicare population from 2007 to 2018.METHODS:We decomposed MA growth into changes in the values of explanatory variables that influence MA enrollment (eg, income and payment rate) and changes in preferences for MA vs TM (estimated coefficients) using a nonlinear version of the Blinder-Oaxaca decomposition to distinguish the sources of MA growth. We find that the relatively smooth growth in MA market share masks 2 distinct growth periods.RESULTS:From 2007 to 2012, 73% of the increase was due to changes in the values of the explanatory variables, and only 27% was due to changes in coefficients. In contrast, from 2012 to 2018, changes in explanatory variables, particularly MA payment levels, would have led to a decline in MA market share if that effect had not been offset by changes in the coefficients.CONCLUSIONS:Overall, we find that MA is becoming more appealing to more educated and nonminority beneficiaries than in the past, although minority and lower-income beneficiaries are still more likely to pick the program. Over time, if preferences continue to shift, the nature of the MA program will change as it moves more toward the middle of the Medicare distribution.

Primary care clinics are a frequent focus of policy initiatives to improve the value of health care; yet, it is unclear whether they have the ability or incentive to take on the additional tasks that these initiatives ask of them. This paper reports on a qualitative study assessing barriers that clinic leaders face to reducing cost within a tiered cost-sharing commercial health insurance benefit design that gives both consumers and clinics a strong incentive to reduce cost. We conducted semi-structured interviews of clinical and operational leaders at a diverse set of 12 Minnesota primary care clinics and identified 6 barriers: insufficient information on drivers of cost; clinics controlling a portion of spending; patient preference for higher cost specialists; administrative challenges; limited resources; and misalignment of incentives. We discuss approaches to reducing these barriers and opportunities to implement them.

This JAMA Guide to Statistics and Methods discusses instrumental variable analysis, a method designed to reduce or eliminate unobserved confounding in observational studies, with the goal of achieving unbiased estimation of treatment effects.

Consumers in private health insurance markets are highly inertial. The literature has repeatedly found consumers are willing to pay thousands of dollars to keep their health plan. However, the causes of inertia are not well understood, despite their importance in determining whether welfare can be improved by reducing inertia and which types of policies would be effective in doing so. Using administrative data from California's Health Insurance Marketplace, we separately identify three sources of inertia-tastes for provider continuity, inattention, and hassle costs-using two-stage models of inattention and health plan choice. We find that eliminating inattention and hassle costs would reduce repeated health plan choice by 53 percentage points and that interventions to reduce inattention and hassle costs are complements. Inattention and hassle costs cost consumers over a billion dollars in foregone consumer surplus in 2018, roughly $1,790 per household per year or half the annual premium paid by the median household, with inattention accounting for the largest source of forgone surplus. We conclude that interventions to reduce inertial plan choice should jointly focus on hassle costs and particularly inattention, but not tastes for provider continuity. (C) 2022 Elsevier B.V. All rights reserved.

Background: Lack of affordable health care affects the uninsured, commercially insured, and Medicare beneficiaries. Yet, the wide variation in providers’ prices and practice styles suggests that more affordable care already may be available and data on low value and wasteful care suggest that lower cost care need not come at the expense of better quality. Although price variation has received the most attention in the literature and legislation, total cost of care is a function of both unit prices (fees) and the quantity of services. Objective: To partition provider-specific variation in total annual risk-adjusted per capita expenditures on health care services into variation in unit prices (fees) versus quantities of services, and to explore the relationship between low value, avoidable, discretionary, and recommended care to total health expenditures. The analysis is important because both prices and quantities of services can affect affordability and reductions in prices versus quantities have very different effects on providers’ profits. Setting: 2018 data from the Minnesota State Employees Group Insurance Program (SEGIP) that offers a tiered cost-sharing health insurance benefit design to 130,000 State employees and their dependents (SEGIP “members”). Exposure: Each year during open enrollment, SEGIP members choose a primary care clinic (PCC). The PCC can make decisions regarding both unit prices and prescribed services. PCCs are placed in one of four cost-sharing tiers based on the total annual risk-adjusted per capita health expenditures for the SEGIP members who choose their clinic. Members choosing higher cost PCCs face higher deductibles, copayments, and maximum out-of-pocket spending limits. Measures: Overall prices and use of inpatient, outpatient hospital, professional, and pharmaceutical services, total and avoidable use of emergency department visits and inpatient admissions, low value care, testing for patients with pneumonia, and recommended preventive care. Results: Differences in total risk-adjusted annual per capita health expenditures across the care systems were substantial. Higher cost providers had both higher unit prices and higher use of services. Variation in the quantity of health care services explained more of the variance in total spending than variation in prices. Prices for professional services and use of inpatient, outpatient hospital, and pharmaceutical services, and ambulatory care sensitive admissions, contributed significantly to high total expenditures. Lower cost PCCs in the lowest cost-sharing tier had higher rates of low value care and lower emergency department visits per capita. Neither the number of investigations for patients with pneumonia nor the receipt of recommended mammography screening varied systematically by tier. Conclusions: Efforts to identify and expand sources of affordable care, including improved information and incentives for consumers, need to account for variation in both prices and quantities of services. Efforts to encourage more efficient use of health care services by providers need to consider the effect of those efforts on the provider’s internal costs and thus their profits.

OBJECTIVESTo understand responses of primary care clinics to inclusion in a tiered total cost of care insurance benefit design.STUDY DESIGNWe used a qualitative design beginning with longitudinal analysis of administrative data on consumer clinic choice, clinic tier placement, and clinic actions, followed by in-depth interviews with key informants from clinics, administering health plans, and program administrators.METHODSWe collected data via semistructured interviews with purposively sampled key informants selected from clinics that prospectively reduced prices to move to, or remain in, a tier with lower cost sharing. Data from interview transcripts were coded using qualitative coding software and analyzed for thematic responses.RESULTSOur findings suggest that clinics respond to the incentives in the tiered cost-sharing benefit design. Two motivations cited by clinics are (1) concern over developing a reputation as a high-cost clinic and (2) concern about the possible loss of patients due to higher cost sharing. Some clinics have agreed to price reductions or risk-sharing arrangements to move to, or remain in, a tier with lower cost sharing. Clinic informants reported that price reductions alone are not scalable. They sought greater transparency in tier assignment and increased data sharing to help them reduce costly or unnecessary utilization.CONCLUSIONSManagers of primary care clinics respond to a tiered benefit design that holds them accountable for total cost of care. They respond by offering price discounts and expressing interest in reducing costly referrals and unnecessary use of services.

OBJECTIVE To describe physicians' variation in de-adopting concurrent statin and fibrate therapy for type 2 diabetic patients following a reversal in clinical evidence. DATA SOURCES We analyzed 2007-2015 claims data from OptumLabs® Data Warehouse, a longitudinal, real-world data asset with de-identified administrative claims and electronic health record data. STUDY DESIGN We modeled fibrate use among Medicare Advantage and commercially insured type 2 diabetic statin users before and after the publication of the ACCORD lipid trial, which found statins and fibrates were no more effective than statins alone in reducing cardiovascular events among type 2 diabetic patients. We modeled fibrate use trends with physician random effects and physician characteristics such as age and specialty. DATA EXTRACTION We identified patient-year-quarters with one year of continuous insurance enrollment, type 2 diabetes diagnoses, and fibrate use. We designated the physician most responsible for patients' diabetes care based on evaluation and management visits and prescriptions of glucose-lowering drugs. PRINCIPAL FINDINGS Fibrate use increased by 0.12 percentage points per quarter among commercial patients (95% CI, 0.10 to 0.14) and 0.17 percentage points per quarter among Medicare Advantage patients (95% CI, 0.13 to 0.20) before the trial and then decreased by 0.16 percentage points per quarter among commercial patients (95% CI, -0.18 to -0.15) and 0.05 percentage points per quarter among Medicare Advantage patients (95% CI, -0.06 to -0.03) after the trial. However, 45% of physicians treating commercial patients and 48% of physicians treating Medicare Advantage patients had positive trends in prescribing following the trial. Physicians' characteristics did not explain their variation (pseudo R2  = 0.000). CONCLUSION On average, physicians decreased fibrate prescribing following the ACCORD lipid trial. However, many physicians increased prescribing following the trial. Observable physician characteristics did not explain variations in prescribing. Future research should examine whether physicians vary similarly in other de-adoption settings.

Background The Kidney Allocation System (KAS) includes a scoring system to match transplant candidate life expectancy with expected longevity of the donor kidney, and a backdating policy that gives waitlist time credit to patients waitlisted after starting dialysis treatment (post-dialysis). We estimated the effect of the KAS on employment among patient subgroups targeted by the policy. Methods We used a sample selection model to compare employment after transplant before and after KAS implementation among patients on the kidney-only transplant waitlist between December 4, 2011 and December 31, 2017. Results Post-dialysis transplant recipients aged 18-49 were significantly more likely to be employed 1-year post transplant in the post-KAS era compared to the pre-KAS era. Transplant recipients aged 35-64 with no dialysis treatment were significantly less likely to be employed 1 year after transplant in the post-KAS era compared to the pre-KAS era. Conclusions This study provides the first assessment of employment after DDKT under the KAS and provides important information about both the methods used to measure employment after transplant and the outcome under the KAS. Changes in employment after DDKT among various patient subgroups have important implications for assessing long-term patient and societal effects of the KAS and organ allocation policy.

Publicly subsidized private health insurance markets in the United States were created under the assumption that competition would maximize consumer welfare. However, consumers in these markets are willing to pay thousands of dollars to stay in the same health plan, even after accounting for premiums and coverage generosity. While this inertial behavior can undermine welfare gains from competition, it is unclear whether policymakers should attempt to reduce it. One reason not to reduce inertia is that it may result from consumers’ desire to maintain continuity of care with their health care providers, which has strong positive effects on health. On the other hand, inertia may result from inattention and hassle costs. Using administrative data from California’s Health Insurance Marketplace, we present descriptive evidence of inertial plan choice and inattention. We then separately identify three sources of inertia—tastes for provider continuity, inattention, and hassle costs—using two-stage models of inattention and plan choice. We find that nearly all inertial plan choice results from inattention and hassle costs, the former more so than the latter. As a result of these two sources of inertia, consumers lost over a billion dollars in forgone surplus in 2018, or $1,440 to $1,584 per household per year—roughly half the annual premium paid by the median household in California’s Marketplace. We conclude that inertia is amenable to policy remediation, and that interventions to reduce hassle costs and especially inattention can improve consumer welfare and efficiency in private health insurance markets.

Efforts to improve the efficiency of the US health-care system involve both provider payment reform and efforts to give consumers the information they need to choose efficient providers and a financial incentive to do so. An example of the latter type of initiative is tiered cost-sharing. We analyze data from a long-standing tiered cost-sharing system for primary care gatekeeper clinics. These clinics control access to specialists and hospitals and are held accountable for their patients' total annual risk-adjusted spending on covered health-care services. Consumers choosing higher cost clinics face higher levels of deductibles, copayments, and out-of-pocket maximums. We find that when choosing a primary care clinic, consumers are responsive to the clinic's tier. Consumers exhibit a high level of inertia, but nonetheless, many clinics voluntarily reduce their fees to move to, or retain placement in, lower cost tiers.

High-quality health care not only includes timely access to effective new therapies but timely abandonment of therapies when they are found to be ineffective or unsafe. Little is known about changes in use of medications after they are shown to be ineffective or unsafe. In this study, we examine changes in use of two medications: fenofibrate, which was found to be ineffective when used with statins among patients with Type 2 diabetes (ACCORD lipid trial); and dronedarone, which was found to be unsafe in patients with permanent atrial fibrillation (PALLAS trial). We examine the patient and provider characteristics associated with a decline in use of these medications. Using Medicare fee-for-service claims from 2008 to 2013, we identified two cohorts: patients with Type 2 diabetes using statins (7 million patient-quarters), and patients with permanent atrial fibrillation (83 thousand patient-quarters). We used interrupted time-series regression models to identify the patient- and provider-level characteristics associated with changes in medication use after new evidence emerged for each case. After new evidence of ineffectiveness emerged, fenofibrate use declined by 0.01 percentage points per quarter (95% CI − 0.02 to − 0.01) from a baseline of 6.9 percent of all diabetes patients receiving fenofibrate; dronedarone use declined by 0.13 percentage points per quarter (95% CI − 0.15 to − 0.10) from a baseline of 3.8 percent of permanent atrial fibrillation patients receiving dronedarone. For dronedarone, use declined more quickly among patients dually-enrolled in Medicare and Medicaid compared to Medicare-only patients ( P < 0.001), among patients seen by male providers compared to female providers ( P = 0.01), and among patients seen by cardiologists compared to primary care providers ( P < 0.001).