Medicare's Hospital Value-Based Purchasing Program Values Quality over QALYs.

Medicare's Hospital Value-Based Purchasing Program (HVBP) is the first national pay-for-performance program to combine measures of quality of care with a measure of episode spending. We estimated the implicit tradeoffs between mortality reduction and spending reduction. To earn points in HVBP, a hospital can either lower mortality or reduce spending, creating a tradeoff between the 2 measures. We analyzed the quality performance and earned points of 2814 hospitals using publicly available data. We then quantified the tradeoffs between spending and mortality in terms of quality-adjusted life-years (QALYs). If incentives in the program were balanced, then the tradeoff between spending and QALYs should be comparable with those of high-value health interventions, roughly $50,000 to $200,000 per QALY. Instead, the tradeoff in HVBP was about $1.2 million per QALY. HVBP overvalues improvements in quality of care relative to spending reductions. We propose 2 possible policy adjustments that could improve incentives for hospitals to deliver high-value care.

Spending On Postacute Care After Hospitalization In Commercial Insurance And Medicare Around Age Sixty-Five.

Postacute care costs are the primary determinant of episode spending around hospitalization. Yet there is little evidence that greater spending on postacute care improves readmission rates or functional recovery. Recent Medicare payment reform evaluations have suggested that postacute care spending is responsive to episode-based incentives. However, it remains unknown whether Medicare payment policies are responsible for excess postacute care spending, compared with that of commercial payers. In a population-based, statewide collaborative of Michigan hospitals, we used regression discontinuity design among propensity-weighted, age-adjusted cohorts to compare postacute care spending between patients with commercial insurance and those with Medicare around age sixty-five. Spending was 68-230 percent greater among fee-for-service Medicare beneficiaries than among similar commercially insured people across varied medical and surgical conditions. Despite greater spending, there were no differences in readmission rates. These findings suggest that postacute care utilization is highly sensitive to payer influence, and there may be an opportunity for additional savings in Medicare without sacrificing quality.

The Five-Star Skilled Nursing Facility Rating System and Care of Disadvantaged Populations.

OBJECTIVES: To examine characteristics and locations of high- and low-quality skilled nursing facilities (SNFs) and whether certain vulnerable individuals were differentially discharged to facilities with lower quality ratings. DESIGN: Retrospective observational study. SETTING: Medicare-certified SNFs providing postacute care. PARTICIPANTS: SNF stays (N=1,195,166) of Medicare beneficiaries aged 65 and older admitted to 14,033 SNFs within 2 days of hospital discharge. MEASUREMENTS: We used Medicare claims from October 2013 to September 2014 and SNF 5-star ratings published on Nursing Home Compare. We describe the characteristics and populations of facilities according to quality, and the location of low (1 star) and high (5 stars) quality facilities. We used logistic regression models to estimate odds of admission to a low-quality facility after hospital discharge according to race, ethnicity, dual Medicare-Medicaid enrollment, functional status, discharge from a safety-net or low-quality hospital, and residence in a county with more low-quality SNFs. RESULTS: More than one-fifth (22.2%) of the facilities had a 5-star (high quality) rating, and 15.9% had a one-star (low quality) rating. Low-quality facilities were more likely to be in the south (44%), for profit (85%), and larger (>70 beds (86%)). Dual enrollment was the strongest predictor of admission to a 1-star facility (odds ratio (OR) = 1.53, 95% confidence interval (CI) = 1.51-1.55), although racial or ethnic minority status (black: OR = 1.25, 95% CI = 1.22-1.28; Hispanic: OR = 1.10, 95% CI = 1.06-1.14) and geographic prevalence of facilities (for a 10% increase in 1-star beds located in the county of individual's residence: OR = 1.27, 95% CI = 1.26-1.27) were also significant predictors. CONCLUSION: Vulnerable groups are more likely to be discharged to lower-quality facilities for postacute care. Policy-makers should monitor disparities in SNF quality. J Am Geriatr Soc 67:108-114, 2019.

Association of Medication Nonadherence Among Adult Survivors of Stroke After Implementation of the US Affordable Care Act.

Importance: Among adults with chronic disease, survivors of stroke have high out-of-pocket financial burdens. The US government enacted the Affordable Care Act (ACA) in 2010 and implemented the law in 2014 to provide more low-income adults with health insurance coverage. Objective: To assess whether ACA implementation is associated with cost-related nonadherence (CRN) to medication among adult survivors of stroke. Design, Setting, and Participants: This study analyzed data from the 2000 to 2016 National Health Interview Survey, an in-person household survey of the noninstitutionalized US population conducted annually by the National Center for Health Statistics. Conducted at the University of Michigan Medical School, Ann Arbor, from July 24, 2017, to February 28, 2018, the study had a sample of 13 930 survivors of stroke. Analyses were stratified by age (45-64 years vs ≥65 years). Time was treated as a continuous variable and as a categorical variable across 4 periods (2000-2005, historical control; 2006-2010, economic recession and peak unemployment; 2011-2013, before ACA implementation; and 2014-2016, after ACA implementation). Percentages are weighted to reflect US population estimates. Main Outcomes and Measures: The primary outcome was the self-report of CRN, defined as the inability to afford prescribed medications within the past 12 months. Results: Among the 13 930 total survivors of stroke, 38.1% were aged 45 to 64 years (50.5% were female and 49.5% were male, with a mean [SE] age of 56.0 [0.10] years), and 61.9% were aged 65 years or older (54.9% were female and 45.1% were male, with a mean [SE] age of 76.2 [0.09] years). From 2011 to 2013 through 2014 to 2016, Medicaid increased (from 24.0% [95% CI, 21.0%-27.2%] in 2011-2013 to 30.8% [95% CI, 27.3%-34.6%] in 2014-2016; P < .001) and uninsurance decreased (from 13.7% [95% CI, 11.3%-16.4%] to 6.8% [95% CI, 5.3%-8.8%]; P < .001) among survivors of stroke aged 45 to 64 years. Among survivors aged 45 to 64 years, CRN increased over time before ACA implementation (from 18.6% [95% CI, 16.5%-20.9%] in 2000-2005, to 22.6% [95% CI, 19.7%-25.9%] in 2006-2010, to 23.8% [95% CI, 20.7%-27.3%] in 2011-2013) and decreased after ACA implementation to 18.1% (95% CI, 15.4%-21.3%; P = .01) in 2014 to 2016. The period after ACA implementation was associated with lower odds of CRN after adjustment for sociodemographics, year, and clinical factors (odds ratio [OR], 0.63; 95% CI, 0.47-0.85). The difference was attenuated after further adjustment for health insurance coverage (OR, 0.76; 95% CI, 0.56-1.03). Conclusions and Relevance: After the ACA implementation, health insurance coverage increased and CRN decreased among adult survivors of stroke, suggesting that further expansion of Medicaid coverage is likely to be advantageous for survivors.

Performance in the Medicare Shared Savings Program by Accountable Care Organizations Disproportionately Serving Dual and Disabled Populations.

BACKGROUND: The growth of accountable care organizations (ACOs) and other alternative payment models has prompted concern about whether these models will disadvantage providers who serve vulnerable populations, particularly those living in poverty or with a disability. OBJECTIVE: To examine performance by ACOs in the top quintile of their proportion of beneficiaries dually enrolled in Medicare and Medicaid (high-dual) and the top quintile of disabled beneficiaries (high-disabled). RESEARCH DESIGN: This is a retrospective cohort study. SUBJECTS: The 333 ACOs in the Medicare Shared Savings Program in 2014, followed through 2016. MEASURES: Quality scores, savings per beneficiary, whether or not the ACO shared savings, and amount of shared savings. RESULTS: High-dual and high-disabled ACOs had slightly lower quality and similar or higher baseline spending than other ACOs, but achieved greater savings per beneficiary than other ACOs ($212 vs. $51 for high-dual ACOs, P=0.04; $241 vs. $44 for high-disabled ACOs, P=0.012). Further, these ACOs were equally or more likely to earn shared savings; just over 30% of high-dual ACOs earned shared savings compared with 25% of non-high-dual ACOs (P=0.35) and 38% of high-disabled ACOs earned shared savings compared with 23% of non-high-disabled ACOs (P=0.013). In longitudinal analyses, we found a decrease in the differences in quality between high-social risk and other ACOs over time. Savings remained higher for high-dual and high-disabled ACOs relative to other ACOs over 2014-2016 though the gap narrowed over time. CONCLUSIONS: High-dual and high-disabled ACOs had similar or higher spending than other ACOs at baseline, but achieved greater savings and were equally or more likely to earn shared savings, suggesting that alternative payment models can have positive financial outcomes for providers who serve vulnerable populations.

Racial Differences in Long-Term Outcomes Among Older Survivors of In-Hospital Cardiac Arrest.

BACKGROUND: Black patients have worse in-hospital survival than white patients after in-hospital cardiac arrest (IHCA), but less is known about long-term outcomes. We sought to assess among IHCA survivors whether there are additional racial differences in survival after hospital discharge and to explore potential reasons for differences. METHODS: This was alongitudinal study of patients ≥65 years of age who had an IHCA and survived until hospital discharge between 2000 and 2011 from the national Get With The Guidelines-Resuscitation registry whose data could be linked to Medicare claims data. Sequential hierarchical modified Poisson regression models evaluated the proportion of racial differences explained by patient, hospital, and unmeasured factors. Our exposure was black or white race. Our outcome was survival at 1, 3, and 5 years. RESULTS: Among 8764 patients who survived to discharge, 7652 (87.3%) were white and 1112 (12.7%) were black. Black patients with IHCA were younger, more frequently female, sicker with more comorbidities, less likely to have a shockable initial cardiac arrest rhythm, and less likely to be evaluated with coronary angiography after initial resuscitation. At discharge, black patients were also more likely to have at least moderate neurological disability and less likely to be discharged home. Compared with white patients and after adjustment only for hospital site, black patients had lower 1-year (43.6% versus 60.2%; relative risk [RR], 0.72), 3-year (31.6% versus 45.3%; RR, 0.71), and 5-year (23.5% versus 35.4%; RR, 0.67; all P<0.001) survival. Adjustment for patient factors explained 29% of racial differences in 1-year survival (RR, 0.80; 95% confidence interval, 0.75-0.86), and further adjustment for hospital treatment factors explained an additional 17% of racial differences (RR, 0.85; 95% confidence interval, 0.80-0.92). Approximately half of the racial difference in 1-year survival remained unexplained, and the degree to which patient and hospital factors explained racial differences in 3-year and 5-year survival was similar. CONCLUSIONS: Black survivors of IHCA have lower long-term survival compared with white patients, and about half of this difference is not explained by patient factors or treatments after IHCA. Further investigation is warranted to better understand to what degree unmeasured but modifiable factors such as postdischarge care account for unexplained disparities.

Readmission Rates and Skilled Nursing Facility Utilization After Major Inpatient Surgery.

BACKGROUND: There is widespread interest in reducing use of postacute care (ie, care after hospital discharge) following major surgery, provided that such reductions do not worsen quality outcomes such as readmission rates. OBJECTIVES: To describe the association between changes in skilled nursing facility (SNF) use and changes in readmission rates after surgery. RESEARCH DESIGN: This was a observational study. SUBJECTS: Fee-for-service Medicare beneficiaries undergoing coronary artery bypass grafting (CABG) or total hip replacement (THR) from 2008 to 2013. MEASURES: Primary exposure was risk-adjusted SNF use initiated 0-2 days after hospital discharge, and the primary outcome was risk-adjusted readmission rates from 3 to 30 days after discharge. RESULTS: Among 176,994 patients who underwent CABG at 804 hospitals and 233,955 patients who underwent THR at 1220 hospitals, hospital-level SNF utilization increased after CABG (16.4%-19.0%, P=0.001) and THR (40.8%-45.5%, P<0.001), from 2008 to 2013. Hospital readmission rates decreased for CABG (14.7%-12.7%, P<0.001) but did not change for THR (4.9%-4.8%, P=0.55), from 2008 to 2013. However, there was wide variation in hospital-level change in readmission rates. After adjusting for hospital characteristics and baseline readmission rates, there was no statistically significant association between change in SNF use and change in readmission rates (0.017 and 0.011 percentage point increase in SNF use for every one percentage point increase in readmission rates for CABG and THR respectively, P=0.58 and 0.32). CONCLUSIONS: Changes in use of postacute care after THR and CABG have not been associated with changes in readmission rates.

Relationship between Hospital 30-Day Mortality Rates for Heart Failure and Patterns of Early Inpatient Comfort Care.

BACKGROUND: The Centers for Medicare & Medicaid Services rewards hospitals that have low 30-day riskstandardized mortality rates (RSMR) for heart failure (HF). OBJECTIVE: To describe the use of early comfort care for patients with HF, and whether hospitals that more commonly initiate comfort care have higher 30-day mortality rates. DESIGN: A retrospective, observational study. SETTING: Acute care hospitals in the United States. PATIENTS: A total of 93,920 fee-for-service Medicare beneficiaries admitted with HF from January 2008 to December 2014 to 272 hospitals participating in the Get With The Guidelines-Heart Failure registry. EXPOSURE: Early comfort care (defined as comfort care within 48 hours of hospitalization) rate. MEASUREMENTS: A 30-day RSMR. RESULTS: Hospitals' early comfort care rates were low for patients admitted for HF, with no change over time (2.5% to 2.6%, from 2008 to 2014, P = .56). Rates varied widely (0% to 40%), with 14.3% of hospitals not initiating comfort care for any patients during the first 2 days of hospitalization. Risk-standardized early comfort care rates were not correlated with RSMR (median RSMR = 10.9%, 25th to 75th percentile = 10.1% to 12.0%; Spearman's rank correlation = 0.13; P = .66). CONCLUSIONS: Hospital use of early comfort care for HF varies, has not increased over time, and on average, is not correlated with 30-day RSMR. This suggests that current efforts to lower mortality rates have not had unintended consequences for hospitals that institute early comfort care more commonly than their peers.

Dually Enrolled Beneficiaries Have Higher Episode Costs On The Medicare Spending Per Beneficiary Measure.

Cost measures are a growing part of Medicare's value-based payment programs. Medicare Spending per Beneficiary (MSPB) is the cost measure included in Medicare's Hospital Value-Based Purchasing (VBP) Program. Beneficiaries who are dually enrolled in Medicare and Medicaid are known to have higher spending on care, but it is unknown whether spending on the MSPB measure varies based on dual enrollment and whether this has implications for the performance of safety-net hospitals. We found that after adjustment for comorbidities, dually enrolled beneficiaries had 4.3 percent higher spending, which was primarily driven by higher costs in the postacute setting associated with use of institutional postacute care. Hospitals in the highest quintile of the disproportionate share hospital index had poorer performance on the MSPB measure, and were more likely to be penalized under VBP. After adjustment for dual status, differences in MSPB performance between safety-net and non-safety-net hospitals were no longer significant. This suggests that differences in performance between the two types of hospitals were driven at least in part by differences in their patient populations. However, overall VBP payment impacts were largely unchanged after the MSPB measure was adjusted for dual-enrollment status.

Medicare's Acute Care Episode Demonstration: Effects of Bundled Payments on Costs and Quality of Surgical Care.

OBJECTIVE: To evaluate whether participation in Medicare's Acute Care Episode (ACE) Demonstration Program-an early, small, voluntary episode-based payment program-was associated with a change in expenditures or quality of care. DATA SOURCES/STUDY SETTING: Medicare claims for patients who underwent cardiac or orthopedic surgery from 2007 to 2012 at ACE or control hospitals. STUDY DESIGN: We used a difference-in-differences approach, matching on baseline and pre-enrollment volume, risk-adjusted Medicare payments, and clinical outcomes to identify controls. PRINCIPAL FINDINGS: Participation in the ACE Demonstration was not significantly associated with 30-day Medicare payments (for orthopedic surgery: -$358 with 95 percent CI: -$894, +$178; for cardiac surgery: +$514 with 95 percent CI: -$1,517, +$2,545), or 30-day mortality (for orthopedic surgery: -0.10 with 95 percent CI: -0.50, 0.31; for cardiac surgery: -0.27 with 95 percent CI: -1.25, 0.72). Program participation was associated with a decrease in total 30-day post-acute care payments (for cardiac surgery: -$718; 95 percent CI: -$1,431, -$6; and for orthopedic surgery: -$591; 95 percent CI: $-$1,161, -$22). CONCLUSIONS: Participation in Medicare's ACE Demonstration Program was not associated with a change in 30-day episode-based Medicare payments or 30-day mortality for cardiac or orthopedic surgery, but it was associated with lower total 30-day post-acute care payments.

Moneyball in Medicare.

US policymakers place high priority on tying Medicare payments to the value of care delivered. A critical part of this effort is the Hospital Value-based Purchasing Program (HVBP), which rewards or penalizes hospitals based on their quality and episode-based costs of care and incentivizes integration between hospitals and post-acute care providers. Within HVBP, each patient affects hospital performance on a variety of quality and spending measures, and performance translates directly to changes in program points and ultimately dollars. In short, hospital revenue from a patient consists not only of the DRG payment, but also of that patient's marginal future reimbursement. We estimate the magnitude of the marginal future reimbursement for individual patients across each type of quality and performance measure. We describe how those incentives differ across hospitals, including integrated and safety-net hospitals. We find evidence that hospitals improved their performance over time in the areas where they have the highest marginal incentives to improve care, and that integrated hospitals responded more than non-integrated hospitals.

Association Between Race, Neighborhood, and Medicaid Enrollment and Outcomes in Medicare Home Health Care.

BACKGROUND/OBJECTIVES: More than 3 million Medicare beneficiaries use home health care annually, yet little is known about how vulnerable beneficiaries fare in the home health setting. This is particularly important given the recent launch of Medicare's Home Health Value-Based Purchasing model. The objective of this study was to determine odds of adverse clinical outcomes associated with dual enrollment in Medicaid and Medicare as a marker of individual poverty, residence in a low-income ZIP code tabulation area (ZCTA), and black race. DESIGN: Retrospective observational study using individuals-level logistic regression. SETTING: Home health care. PARTICIPANTS: Fee-for-service Medicare beneficiaries from 2012 to 2014. MEASUREMENTS: Thirty- and 60-day clinical outcomes, including readmissions, admissions, and emergency department (ED) use. RESULTS: Home health agencies serving a high proportion of dually enrolled, low-income ZCTA, or black beneficiaries were less often high-quality. Dually-enrolled, low-income ZCTA, and Black beneficiaries receiving home health care after hospitalization had higher risk-adjusted odds of 30-day readmission (odds ratio [OR] = 1.08, P < 0.001; OR = 1.03, P < 0.001; and OR = 1.02, P = 0.002 respectively) and 30-day ED use (OR = 1.20, 1.07, and 1.15, P < 0.001 for each). Those receiving home health care without preceding hospitalization had higher 60-day admission (OR = 1.06, P < 0.001; OR = 1.01, P = 0.002; and OR = 1.05, P < 0.001), and 60-day ED use (OR = 1.16, 1.03, and 1.19, P < 0.001 for each). Differences were primarily within agencies rather than between the agencies where these beneficiaries sought care. CONCLUSION: Medicare beneficiaries receiving home health services who are dually enrolled, live in a low-income neighborhood, or are black have higher rates of adverse clinical outcomes. These populations may be an important target for quality improvement under Home Health Value-Based Purchasing.

Performance And Participation Of Physicians In Year One Of Medicare's Value-Based Payment Modifier Program.

In 2015 Medicare launched the Physician Value-Based Payment Modifier program, the largest US ambulatory care pay-for-performance program to date and a precursor to the forthcoming Merit-based Incentive Payment System. In its first year, the program included practices with a hundred or more clinicians. We found that 1,010 practices met this criterion, 899 of which had at least one attributed beneficiary. Of these latter practices, 263 (29.3 percent) failed to report performance data and received a 1 percent reporting-based penalty. Of the 636 practices that reported performance data, those that elected quality tiering-voluntarily receiving performance-based penalties or bonuses-and those with high use of electronic health records had better performance on quality and costs than other practices. Practices with a primary care focus had better quality than other practices but similar costs. These findings translated into differences in the receipt of penalties and bonuses and may have implications for performance patterns under the Merit-based Incentive Payment System.

Risk Adjustment May Lessen Penalties On Hospitals Treating Complex Cardiac Patients Under Medicare's Bundled Payments.

To reduce variation in spending, Medicare has considered implementing a cardiac bundled payment program for acute myocardial infarction and coronary artery bypass graft. Because the proposed program does not account for patient risk factors when calculating hospital penalties or rewards ("reconciliation payments"), it might unfairly penalize certain hospitals. We estimated the impact of adjusting for patients' medical complexity and social risk on reconciliation payments for Medicare beneficiaries hospitalized for the two conditions in the period 2011-13. Average spending per episode was $29,394. Accounting for medical complexity substantially narrowed the gap in reconciliation payments between hospitals with high medical severity (from a penalty of $1,809 to one of $820, or a net reduction of $989), safety-net hospitals (from a penalty of $217 to one of $87, a reduction of $130), and minority-serving hospitals (from a penalty of $70 to a reward of $56, an improvement of $126) and their counterparts. Accounting for social risk alone narrowed these gaps but had minimal incremental effects after medical complexity was accounted for. Risk adjustment may preserve incentives to care for patients with complex conditions under Medicare bundled payment programs.

Interdependence in decision-making by medical consultants: implications for improving the efficiency of inpatient physician services.

OBJECTIVES: Hospital administrators are seeking to improve efficiency in medical consultation services, yet whether consultants make decisions to provide more or less care is unknown. We examined how medical consultants account for prior consultants' care when determining whether to provide intensive consulting care or sign off in the treatment of complex surgical inpatients. We applied three distinct theoretical frameworks in the interpretation of our results. METHODS: We performed a retrospective cohort study of consultants' care intensity, measured alternately using a dummy variable for providing two or more days consulting (versus one) and a continuous measure of total days consulting, with 100% Medicare claims data from 2007-2010. Our analytic samples included consults for beneficiaries undergoing coronary artery bypass grafting (n = 61,785) or colectomy (n = 33,460) in general acute care hospitals. We compared the care intensity of consultants who observed different patterns of consulting care before their initial consults using ordinary least squares regression models at the patient-physician dyad level, controlling for patient comorbidity and many other patient- and physician-level factors as well as hospital region and year fixed effects. RESULTS: Consultants were less likely to provide intensive consulting care with each additional prior consultant on the case (1.2-1.7 percent) or if a prior consultant rendered intensive consulting care (20.6-21.5 percent) but more likely when prior consults were more concentrated across consultants (2.9-3.1 percent). Effects on consultants' total days consulting were similar. CONCLUSION: On average, consultants appeared to calibrate their care intensity for individual patients to maximize their value to all patients. Interventions for improving consulting care efficiency should seek to facilitate (not constrain) consultants' decision-making processes.

Effect of a Hospital-wide Measure on the Readmissions Reduction Program.

BACKGROUND: The Hospital Readmissions Reduction Program penalizes hospitals that have high 30-day readmission rates across specific conditions. There is support for changing to a hospital-wide readmission measure to broaden hospital eligibility and provide incentives for improvement across more conditions. METHODS: We used Medicare claims from 2011 through 2013 to evaluate the number of hospitals that were eligible for penalties, in that they met a volume threshold of 25 admissions over a 3-year period for a specific condition or 25 admissions over a 1-year period for the cohorts included in the hospital-wide measure. We estimated the expected effects that changing from the condition-specific readmission measures to a hospital-wide measure would have on average penalties for safety-net hospitals (i.e., hospitals that treat a large proportion of low-income patients) and other hospitals. RESULTS: Our sample included 6,807,899 admissions for the hospital-wide measure and 4,392,658 admissions for the condition-specific measures. Of 3443 hospitals, 688 were considered to be safety-net hospitals. Changing to the hospital-wide measure would result in 76 more hospitals being eligible to receive penalties. The hospital-wide measure would increase penalties (mean [±SE] Medicare payment reductions across all hospitals) from 0.42±0.01% to 0.89±0.01% of Medicare base diagnosis-related-group payments. It would also increase the disparity in penalties between safety-net hospitals and other hospitals from -0.03±0.02 to 0.41±0.06 percentage points. CONCLUSIONS: A transition to a hospital-wide readmission measure would only modestly increase the number of hospitals eligible for penalties and would substantially increase the penalties for safety-net hospitals.

Most Marketplace Plans Included At Least 25 Percent Of Local-Area Physicians, But Enrollment Disparities Remained.

The Affordable Care Act allows commercial insurers participating in the Marketplaces to vary the size of their provider networks as long as the providers are "sufficient" in numbers and types. Concerns have been growing over the increasing use of restricted-provider or narrow networks in Marketplace plans because of their implications for reduced access to care, but little is known about the breadth and stability of these networks over time or what types of enrollees choose such plans. Using national data, we found that in 2016, 60 percent of provider networks in plans offered in the federally facilitated Marketplaces included at least one-quarter of local-area physicians, and that consumers' access to broad-network plans remained stable between 2015 and 2016. Hispanic and low-income people made up a disproportionate share of enrollees in smaller-network plans (those with fewer than one-quarter of local-area physicians). It will be important to monitor the impact of narrow networks on access to and quality of care as well as on health outcomes.