Printer Friendly

Changes in hospital service demand, cost, and patient illness severity following health reform.

1 | BACKGROUND AND RESEARCH OBJECTIVES

Two major health insurance coverage provisions of the Patient Protection and Affordable Care Act (ACA) went into effect in 2014. First, Health Insurance Marketplaces ("insurance exchanges") for individuals to purchase private coverage were opened in all states. Income-based subsidies were distributed as an advance tax credit to individuals with income <400 percent of the federal poverty level (FPL). The second provision allowed states to opt to expand Medicaid eligibility ("Medicaid expansions") for most low-income adults to 138 percent of the FPL. As of December 2015, 30 states and the District of Columbia had adopted the expanded Medicaid eligibility criteria. Enrollment in Medicaid and the Children's Health Insurance Program grew by approximately 12.4 million individuals. (1) Provisions of the ACA resulted in gains in health insurance coverage for 14.5 million people by early 2016. (1)

Although it is clear that the coverage expansions were associated with sharp changes in payer mix, (2-4) less is known about changes in payer-specific utilization rates, how the magnitude of such changes varied with the size of the population targeted by the coverage expansions, or the impact of the expansions on utilization rates of various age/sex demographic groups.

The aim of this study was to analyze the relationship between the ACA insurance coverage provisions and hospital utilization rates, cost, and patient illness severity. We used two approaches in our assessment of the impact of coverage expansions on outcomes. The first approach estimated the average effect of the 2014 insurance expansions in (a) states expanding Medicaid, and (b) states not expanding Medicaid. The second approach leveraged variation across states in the "treatment intensity" associated with the major ACA coverage expansions. We used data on the size of the uninsured population eligible for exchange subsidies or Medicaid to estimate the elasticity of hospital utilization rates, cost, and patient illness severity with respect to the potential size of the coverage expansion. Evidence on the relationship between treatment intensity and payer-specific utilization rates may help policy makers anticipate the impact of future coverage policies.

1.1 | Research questions

1. How has the Affordable Care Act affected hospital inpatient and emergency department (ED) utilization rates, cost, and severity for different payer groups and for age/sex demographic groups (a) in states that implemented insurance exchanges only, and (b) in those that implemented insurance exchanges and expanded Medicaid as well?

2. To what extent is the treatment intensity (or coverage potential) of the expansion provisions associated with changes in hospital inpatient and ED utilization rates, cost, and illness severity?

1.2 | Review of recent literature

Data that capture the direct effects of the ACA insurance coverage expansions on hospital utilization have been emerging. Studies using administrative data from the Healthcare Cost and Utilization Project (HCUP) found that Medicaid expansion led to shifts in payer mix from uninsured toward Medicaid for inpatient discharges (2,5) and ED visits. (6,7) Survey data from the National Health Interview Survey suggest increases in all-payer inpatient utilization in Medicaid expansion states but no significant changes in ED utilization. (8,9) A survey of low-income adults in the 3 years following the ACA found better access to care, lower out-of-pocket spending, and improvements in preventive care for previously uninsured individuals gaining coverage. (10)

Numerous studies have leveraged earlier, pre-ACA coverage expansions to study the impact of policies that increased access to Medicaid or private insurance on hospital utilization. (11-15) For example, the Oregon Health Study found that low-income adults who gained Medicaid coverage through a lottery significantly increased both ED visits and hospital admissions. (16,17)

Other studies have used survey data to examine the association between the ACA insurance expansions and insurance access or take-up.(4,18-21) Frean, Gruber, and Sommers estimated the relative contributions of different ACA provisions to observed insurance changes, finding that Medicaid enrollment among previously eligible adults accounted for half of the reduction in uninsurance attributable to the Medicaid expansion. (22) Courtemanche and colleagues estimated that the full ACA implementation (including the Medicaid expansion) increased the proportion of residents with insurance by 9.5 percentage points compared with 6.2 percentage points in states that did not expand Medicaid. (4) As in their previous study, the identification strategy leveraged variation across geographic areas within states in the share of the population eligible for exchange subsidies and Medicaid expansion; (4,19) we adopted an analogous modeling strategy, using cross-state rather than within-state variation in treatment intensity.

1.3 | Study contribution

Extant literature has shown that recent insurance expansions have shifted hospital patient volumes from uninsured to various insured sources of payment, and the shift has been much larger in states that opted in to Medicaid expansion. What is not fully understood is how this shift has affected rates of hospital utilization, costs, and composition of discharges or visits in specific insurance coverage segments and age/sex demographic groups.

Our study adds to the literature by analyzing ED and inpatient utilization, cost, and patient illness severity using HCUP data from 36 states during the 2011-2015 time period. We also exploit state variation in pre-ACA coverage levels to examine the extent to which the size of the potentially affected population affects outcomes and how these estimated effects vary by payer. Unlike previous studies, we produced estimates of the coverage elasticity of utilization, or the extent to which utilization is responsive to the potential for additional insurance coverage provided by the policy changes. These estimates allow private and government payers to anticipate changes in per capita use, cost, and severity that are likely to occur in covered and uncovered populations and to assess the subpopulations and outcomes most responsive to changes in coverage in expansion and nonexpansion states.

2 | CONCEPTUAL FRAMEWORK

The goal of this study was to model the relationship between ACA coverage expansions and payer-specific changes in hospital utilization, cost, and patient illness severity. Insurance coverage can affect hospital utilization and cost through several channels, including price and access to services. Utilization may increase because acquisition of health insurance coverage lowers the out-of-pocket price. In addition, acquiring coverage may improve access to professional and organizational providers, which may result in higher utilization rates for referral-sensitive conditions. However, better access to primary care also may reduce the need for certain inpatient and ED services.

In addition to price and access effects, the decision to acquire coverage is likely to be associated with health status or demographic factors, implying that the health status composition of the uninsured and insured could change. Patients more likely to need care may be more likely to take up coverage, changing the health status and demographic patient mix within each payer type.

We analyzed changes in per capita rates of hospital utilization, costs

per visit, and patient illness severity (case mix) per visit. Higher rates of use, costs, and patient illness severity for a payer are consistent with adverse selection in take-up, pent-up demand for services, or a strong utilization response to reductions in the out-of-pocket price of care. Lower rates of use, costs, and patient illness severity could reflect favorable selection in take-up or substitution of ambulatory for hospital care.

We also analyzed the impact by age and sex groups by payer in expansion and nonexpansion states to further understand the impact of these policy initiatives. Because of differences in need and demand for health care and health insurance, we posited that large changes in utilization by age/sex group by payer are important to understand, as well as those age/sex/payer groups that do not appear to be affected by the coverage expansions. For example, although we expected to detect large decreases in uninsured use, a finding of no change in utilization within an uninsured age/sex/payer group would indicate that future policy initiatives may need to be modified to address uninsured use within this demographic group.

3 | METHODS

3.1 | Data

Our primary data source was the HCUP State Inpatient Databases (SID) and State Emergency Department Databases (SEDD) from 2011 through 2015. Hospital inpatient discharge and ED records were assembled from the HCUP SID and SEDD for 2011-2015. Our sample included 2011-2015 SID data from 36 states-18 expansion states and 18 non-expansion states. We included 2011-2015 SEDD data for 23 states-12 expansion states and 11 nonexpansion states (Table A3 in Appendix S1).

Data from the single-year ACS Public Use Microdata Samples (PUMS) for 2011-2015 provided population denominators for utilization rates. Respondents to the ACS can be identified by health insurance coverage, state of residence, age category, sex, and household income bracket as a percentage of the FPL. This allowed us also to compute counts and rates of uninsurance by age, sex, and household income as a percentage of FPL. We used these measures to describe the size of the uninsured population in a state prior to the ACA coverage expansions.

We obtained quarterly state-level unemployment rates from the Bureau of Labor Statistics to control for time-varying economic conditions that may affect insurance status and health care utilization. (23)

State-level median household income data from Nielsen were included to control for changes in income levels over time that could affect Medicaid eligibility. (24)

3.1.1 | State-specific medicaid eligibility requirements

Medicaid income eligibility thresholds at the state level describe the context in which Medicaid expansions did (or did not) occur. We collected Medicaid program income eligibility limits for two categories of adults: (a) parents of dependent children, jobless and working, with full Medicaid benefits, and (b) other nondisabled adults, jobless and working, with full Medicaid benefits. (25) We used these income eligibility limits to estimate the size of the population that was eligible for but not enrolled in Medicaid in 2013.

3.2 | Sample selection

We included individuals in the largest patient segment likely to be affected by the Affordable Care Act insurance coverage expansion; adults aged 19-64 years. SID or SEDD states that submitted data for all years between 2011 and 2015 were included in the study. In addition, we included hospitals with data present in all quarters between 2011 and 2015.

3.3 | Units of analysis

The unit of analysis was a cell defined by state of patient residence, age category (19-34 years, 35-54 years, and 55-64 years), sex, year of discharge or visit, and expected source of payment.

3.4 | Outcome variables

3.4.1 | Utilization rates

Inpatient discharge and ED visit volumes were converted to per capita rates, by combining numerators from the HCUP and denominators from the ACS. HCUP SID and SEDD data were aggregated by state of patient residence, age category, sex, year of discharge or visit, and expected source of payment (Medicaid, uninsured, private insurance, and all payers). (26) ACS denominator data for rates were aggregated by patient residence state, age category, sex, year, and ACS health insurance coverage category.

3.4.2 | Cost and patient illness severity (case mix)

We used the cost-to-charge ratio (CCR) method to estimate costs and reflect the cost of service delivery by a given hospital. (27) We developed a case mix index by assigning discharge or visit records to a Clinical Classification Software (CCS) principal diagnosis category and then estimating average cost using 2013 HCUP SID and SEDD data. The case mix index reflects standardized costs, with higher values representing a costlier mix of patients.

3.5 | Predictor variables

Covariates in our models served as either variables quantifying the ACA coverage expansion policy or controls for time-varying state characteristics with the potential to affect hospital utilization.

3.5.1 | ACA Coverage expansion policy variables

To quantify effects of the ACA policy change on hospital utilization, we first used indicator variables that identified the state Medicaid expansion status in the post-ACA time period as a means of estimating the combined effect of all ACA coverage provisions implemented in 2014 in expansion and nonexpansion states.

Second, we defined a coverage expansion ratio (CER) for a given payer type as the ratio of the uninsured population eligible for entering that payer group under the coverage expansion to the population already in that payer group in 2013. CERs represent the potential payer-specific take-up induced by the ACA coverage expansions relative to the payer-specific population in 2013. Variation across states in these coverage expansion ratios allows us to assess the elasticity of outcomes with respect to the treatment intensity of the coverage expansion. See the Appendix S1 for calculation of CER variables.

3.5.2 | Other covariates

In addition to the variables characterizing the ACA coverage expansion, our models included fixed effects for demographic groups and states. We included state-specific linear time trends to model counterfactual trends in both expansion and nonexpansion states. Unemployment rates and income also were included as controls because they can affect insurance coverage and demand for hospital care independently of the ACA coverage expansion.

3.6 | Statistical specifications

We estimated two sets of models with separate treatment parameters for expansion and nonexpansion states in the post-ACA time period. The first model (indicator model) used information on the post-ACA Medicaid expansion status of states to estimate policy effects on outcomes:

E[[y.sub.sdt]|x*]= f([[micro].sub.s] + [[lambda].sub.d] + [[psi].sub.s][l.sub.t] + [[tau].sub.1d][P.sub.t]*|{[ME.sub.s]=0} + [[tau].sub.2d][P.sub.t]*|{[ME.sub.s] = 1} + [x'.sub.sdt][beta]), (1)

for a given insurance type I, where: y = outcome; x = covariate; I = time (continuous); P = post-health insurance exchange (HIX)/Medicaid expansion (ME) implementation indicator; |{*} = indicator function, equal to 1 if its argument is true and 0 otherwise; ME = Medicaid expansion state indicator; i = insurance type (all payer, uninsured, Medicaid, private); s = state; d = demographic category [sex (M, F) x age (19-34 years, 35-54 years, 55-64 years)]; t = year.

The second model (CER model) substituted the CER for information on post-ACA state Medicaid expansion status to capture ACA policy effects specific to insurance coverage type i:

E[[y.sub.sdt]|*]= f([[micro].sub.s] + [[lambda].sub.d] + [[psi].sub.s][l.sub.t] + [[theta].sub.1d][P.sub.t]*l{[ME.sub.s]=0}*[UT.sub.sd] + [[theta].sub.2d][P.sub.t]*|{[ME.sub.s]=1}*[UT.sub.sd] + [x'.sub.sdt] [beta]), (2)

where UT = coverage expansion ratio (CER).

In Equation 1, the effects of the ACA coverage expansions are identified off of a before-after comparison between 2014 and 2015 and earlier years (2011-2013) within each state-demographic-payer cell, after adjusting for a linear time trend. [[tau].sub.1d] captures the average regression-adjusted deviation from the trend in 2014-2015 observed in the nonexpansion states, and [[tau].sub.2d] captures the deviation observed in the expansion states.

In contrast, the CER model presented in Equation 2 exploits cross-sectional variation in treatment intensity among states with the same Medicaid expansion status. Both models rely on the assumption that a state-specific linear trend appropriately models the counterfactual evolution of hospital utilization within each demographic cell (after controlling for fixed effects and other covariates). In Equation 2, this assumption is made implicitly by omitting any intercept shift associated with 2014-2015 so that a state with a zero CER would not have any deviation from the state-specific trend.

We implemented Poisson pseudo-maximum likelihood regressions--for outcomes expressed as population rates, observations were weighted by the corresponding population size, and for encounter-level means, observations were weighted by the corresponding encounter volume. Modeling rates with denominator weights is equivalent to modeling the numerator with the denominator as the exposure term. Standard errors were adjusted to accommodate state-level clustering. For all policy variables, we report semi-elasticities associated with the coefficients [exp(coefficient estimate) - 1].

4 | RESULTS

4.1 | Study sample descriptive statistics

4.1.1 | Demographic and insurance coverage profiles

Table 1 contains population demographic profiles in expansion and nonexpansion states prior to the Medicaid expansion and lth Insurance Marketplace rollouts in 2014. Age and sex profiles were similar across states for the inpatient and ED study samples, although expansion states had larger total populations.

Expansion states had a higher percentage of population covered by Medicaid and a lower percentage of population uninsured in 2013. Although overall levels of uninsurance differed, the distribution of the uninsured population by FPL was similar in expansion and nonexpansion states during 2013. Expansion states had slightly higher proportions of those previously eligible for Medicaid but not enrolled. The proportions of those newly eligible for the health insurance exchanges tended to be slightly higher in nonexpansion states. Coverage expansion ratios were higher in expansion states for Medicaid beneficiaries and uninsured individuals, whereas the coverage expansion ratio for private insurance was higher in nonexpansion states.

4.1.2 | Observed outcome trends

Discharge and visit rates

Inpatient discharge rates declined between 2011-2013 and 2014-2015 for all payers in expansion and nonexpansion states (Table 2). For the uninsured, there were large percentage increases in discharge rates in nonexpansion states (14.0 percent) and even larger decreases in expansion states (-36.2 percent). All-payer ED visit rates increased by roughly equivalent percentages in expansion and nonexpansion states. However, ED visit rates for uninsured individuals declined in expansion states (-6.7 percent) but increased at a faster pace than for all payers in nonexpansion states (18.4 percent). A similar pattern, of smaller magnitude, was observed for private insurance, with a small decrease in expansion states (-0.2 percent) and an increase of 7.9 percent in non-expansion states. Medicaid ED visit rates increased in expansion states (7.0 percent) and remained unchanged in nonexpansion states.

Case mix

The inpatient case mix index had the largest percentage changes in expansion states for the uninsured (-3.3 percent) and Medicaid (7.3 percent). Changes in case mix index for other payers, and the ED in general, were smaller in magnitude.

Costs

Inpatient costs increased for all insurance coverage classes in expansion and nonexpansion states. Percentage increases in costs were smallest for uninsured patients in expansion states (0.2 percent) and largest for Medicaid patients in expansion states (11.0 percent). Compared with nonexpansion states, the change in estimated all-payer costs in expansion states was somewhat higher for inpatient care (8.6 percent in 2014 vs 6.1 percent for nonexpansion states) and somewhat lower for ED visits (15.0 percent vs 16.8 percent for non-expansion states).

4.2 | Model results for inpatient and emergency department utilization

4.2.1 | Indicator model results

Percent change estimates associated with the [tau] coefficients for the indicator model (Equation 1) are reported for utilization rates, cost, and the case mix index in Tables 3 and 4 (also reported graphically in Figure A1a-c in Appendix S1).

Discharge and visit rates

The largest ACA coverage expansion effects were on inpatient discharge rates for the uninsured in expansion states (Table 3). Discharge rates for all demographic groups except young females decreased by 39 percent or more per capita on average. Medicaid discharge rates increased by more than 14 percent in expansion states for males (aged 19-34 and 35-54 years), whereas discharge rates decreased for all female age groups. For Medicaid enrollees in nonexpansion states, discharge rates decreased for those aged 55-64 years. ACA effect estimates for the privately insured and all payers, when significant, were smaller.

Uninsured ED visit rates decreased by more than 10 percent for females (aged 19-34 years and 35-54 years) and males (aged 35-54 years) (Table 4). Total visit rates increased among Medicaid enrollees in expansion states for males in all age groups. Notably, Medicaid visit rates decreased for older females (aged 55-64) and males in all age groups in nonexpansion states. All-payer ED visit rates for males and females aged 55-64 years in expansion states increased.

Costs

The ACA effects on inpatient cost were mostly negative and significant. The ACA effects on ED cost for all insurance coverage categories and most demographics groups in non expension states were positive and significant. However, in expansion states, they were negative and significant (Table 4).

Case mix

Affordable Care Act effects on the case mix index generally were much smaller than effects on utilization rates or costs. Only two effects exceeded 5 percent in magnitude, both for inpatient discharges among females aged 19-34 years in expansion states: Medicaid (6.2 percent) and uninsured (-14.8 percent).

Coverage expansion ratio model results

Coefficient estimates for the CER model are reported as semi-elasticities in Table 5, interpreted as the average percent change in outcome rates in 2014-2015, compared with 2011-2013, and associated with a unit increase in the CER. For example, the CER coefficient for uninsured discharge rates among females aged 35-54 years in expansion states was -0.0063, or a roughly 0.63 percent decrease in discharge rate for a unit increase in the CER (which represents a 1 percent increase in policy-related coverage in the population).

Significant CER model coefficients closely followed patterns observed for the indicator model effects. A comparison of the indicator (Tables 3 and 4) and CER estimates (Table 5) reveals a strong correlation between the indicator and CER model coefficients. Although some CER coefficients failed to achieve significance in cases in which indicator model effects were significant, there was a strong concordance in the estimates for all outcomes, demographic groups, payers, and settings.

5 | DISCUSSION

States adopting Medicaid expansion reduced post-ACA inpatient and ED per capita utilization among the uninsured, whereas inpatient and ED visit rates increased for Medicaid enrollees. There were, however, some exceptions to this generalization. Medicaid discharge rates in expansion states for the youngest women fell, whereas those for males in all age groups increased, likely because of health differences between pre-existing and new Medicaid enrollees in these demographic groups. There were no large significant inpatient effects estimated for those with private insurance or for all payers in either expansion or nonexpansion states.

Although we cannot be certain given the cross-sectional nature of our data, dramatic decreases observed in the Medicaid expansion states in utilization rates among the uninsured may suggest that those exiting uninsurance are less healthy than those who remain uninsured. Acknowledging that those who were uninsured in 2014 or 2015 include individuals who lost coverage between 2013 and 2015, a large majority also were uninsured in 2013 (about 77 percent, Table A2), so the dominant change in the uninsured is likely due to coverage acquisition. Similarly, increases in Medicaid utilization rates suggest that newly enrolled members, the majority of whom were previously uninsured (Table A2), are less healthy than those currently enrolled, generating increases in utilization.

If the least healthy uninsured individuals were gaining Medicaid coverage, we would expect to find positively correlated changes in costs and the case mix index. For the uninsured in expansion states, we found negative effects for inpatient and ED utilization rates, cost, and the case mix index, as expected. For Medicaid, however, the sign of utilization effects tended to be opposite to that of cost and case mix effects. For example, the Medicaid inpatient discharge effect was negative for females aged 19-34 years in expansion states (-17.7 percent), whereas the cost and case mix index effects were positive (2.8 and 6.2 percent, respectively; see Table 3). One explanation consistent with this finding is a possible difference in the Medicaid population pre- and post-ACA expansion. In the ED setting, costs increased consistently in nonexpansion states for all age groups and payers, suggesting that this effect was common to nonexpansion states. Taken together, significant changes in uninsured utilization rates, case mix index, and cost lend support to the hypothesis of selective take-up of coverage among the previously uninsured on the basis of health status.

In aggregate, the ED results generally mirrored trends reported by Garthwaite et al (28) in a study of visits to 126 investor-owned hospitals in the United States where Medicaid ED visits rose and uninsured ED visits fell in expansion states. The results were less pronounced in nonexpansion states.

We also found that the magnitude of change in outcomes was associated with the size of the population eligible for insurance coverage take-up, as measured by the CER. The CER coefficients demonstrate how the potential for coverage translated into utilization. For example, one of the larger impacts was for uninsured discharge rates among males aged 35-54 years in expansion states of -0.0080, or a roughly 0.793 percent decrease in discharge rate for a unit increase in the CER. Conversely, cases in which the CER response is limited (eg, few significant coefficients such as males 19-34, see Table 5) can inform policy makers that coverage expansions did not bring about major compositional changes in the insured population.

Per capita utilization and cost have implications for private and government payer budgeting and planning. In particular, levels of use, and therefore aggregate expense, are likely to increase for Medicaid in expansion states. Uncompensated care delivery costs for hospital inpatient and ED services will decline on the basis of our findings, consistent with Dranove and colleagues. (29)

There are limitations that may affect our findings, including possible synchronization issues with the numerators based on encounter data and denominators based on the ACS. We also note that, although the SID and SEDD data could support more geographically disaggregated analyses that leverage variation in CER across hospital markets within states, we conducted our analysis at the state level to avoid the challenges involved in measurement within sociodemographic cells at substate levels of geography.

Three recent studies focused on the impact of the ACA using substate geographic areas. Courtemanche and colleagues (4,19) employed variation in substate coverage rates to isolate the impact of the ACA coverage expansions on insurance coverage and health on the basis of survey responses (ACS (19) and Behavioral Risk Factor Surveillance System [BRFSS] (4)). They found that the largest effects generally were concentrated in substate geographic areas with higher rates of uninsurance. Also using the BRFSS, Benitez and colleagues (21) examined impacts on access to care in Kentucky in high- and low-poverty areas based on ZIP Code and found greater improvements in access to care in high-poverty areas. In this study, we focused on estimating the impact of the ACA within age and sex demographic groups by payer, where data are best available at the state level. As data on insurance rates at lower levels of geography become available, future investigations may focus on these geographic variations.

Our model includes fixed effects for state, demographic groups, and time. A limitation of the fixed-effects approach is the potential for time-varying confounders that are not included in the model. Another limitation is that we did not follow individual patients over time as they changed insurance status. As such, we were not able to pinpoint sources of changes in utilization, such as moral hazard or improved access to ambulatory care. We used the CCR method to generate cost estimates in this study. CCR cost estimates are known to have biases when used in small subsets of patients but generally agree well with cost-accounting-based estimates at the hospital level. Also, our cost estimates were not adjusted for labor costs, which could affect differences observed between expansion and nonexpansion states. However, the use of state fixed effects and state-specific time trends should address this issue.

In this study, we assumed that a coverage expansion ratio based on the size of the uninsured population represents the treatment intensity of the ACA and estimated the reduced-form relationship between this treatment intensity and changes in hospital outcomes. We believe that this makes intuitive sense, and our empirical results support this assumption. There may be, however, other cross-state differences correlated with the ACA that modify the effect of insurance coverage on utilization. Even without taking a stand on whether changes in the insured population are the only causal mechanism linking CER to hospital outcomes, estimates of the CER model may be useful for examining this mechanism while potentially facilitating prediction of Medicaid expansion impacts in nonexpansion states. Future investigation within a structural model framework and collection of additional details on state policy environments would help determine the importance of these alternative mechanisms.

It is interesting to note what our study did not find. When assessing all-payer ACA effects on use of hospital services by adults younger than age 65 years, we found that inpatient and ED outcomes did not change appreciably from historical trends after the health insurance exchange and Medicaid expansion implementations. Thus, our findings do not offer evidence of substantive changes in consumer behavior with respect to hospital use in response to insurance acquisition.

6 | CONCLUSION

We analyzed HCUP SID data from 36 states and SEDD data from 23 states to provide new evidence on how the insurance coverage provisions of the ACA affected hospital utilization rates, costs, and patient illness severity by payer in 2014 and 2015, the first 2 years after implementation of the ACA major coverage provisions. We found that variation across states in the size of the potential expansion population can help explain our results. Our findings may be of interest to state and federal policy makers as well as hospitals and payers to improve and predict the impact of future coverage interventions on hospital services.

ACKNOWLEDGMENTS

Joint Acknowledgment/Disclosure Statement: This study was funded by the Agency for Healthcare Research and Quality under contract HHSA-290-2013-00002-C. The views expressed herein are those of the authors and do not necessarily reflect those of the Agency for Healthcare Research and Quality or the U.S. Department of Department of Health and Human Services. No official endorsement by any agency of the federal or state governments, Truven Health Analytics or RAND Corporation is intended or should be inferred.

DISCLOSURES

None.

ORCID

Gory Pickens https://orcid.org/0000-0003-4689-185X

Zeynal Karaca https://orcid.org/0000-0001-7744-4476

Teresa B. Gibson https://orcid.org/0000-0002-5853-7447

REFERENCES

(1.) US Department of Health and Human Services. Medicaid & CHIP: December 2015 Monthly Applications, Eligibility Determinations and Enrollment Report. https://www.medicaid.gov/medicaid/program-information/downloads/December-2015-enrollment-report.pdf. February 29, 2016. Accessed November 18, 2018.

(2.) Nikpay S, Buchmueller T, Levy H. Affordable Care Act Medicaid expansion reduced uninsured hospital stays in 2014. Health Aff (Millwood). 2016;35(1):106-110.

(3.) Freedman S, Nikpay S, Carroll A, Simon K. Changes in inpatient payer-mix and hospitalizations following Medicaid expansion: evidence from all-capture hospital discharge data. PLoS ONE. 2017;12(9):1-9.

(4.) Courtemanche C, Marton J, Ukert B, Yelowitz A, Zapata D. Effects of the Affordable Care Act on health care access and self-assessed health after 3 years. Inquiry. 2018;55:1-10.

(5.) Hellinger FJ. In four ACA expansion states, the percentage of uninsured hospitalizations for people with HIV declined, 2012-14. Health Aff (Mil/wood). 2015;34(12):2061-2068.

(6.) Pines JM, Zocchi J, Moghtaderi A, et al. Medicaid expansion in 2014 did not increase emergency department use but did change insurance payer mix. Health Aff (Millwood). 2016;35(8):1480-1486.

(7.) Nikpay S, Freedman S, Levy H, Buchmueller T. Effect of the Affordable Care Act Medicaid expansion on emergency department visits: evidence from state-level emergency department databases. Ann Emerg Med. 2017;70(2):215-225.e6.

(8.) Wherry LR, Miller S. Early coverage, access, utilization, and health effects associated with the Affordable Care Act Medicaid expansions: a quasi-experimental study. Ann Intern Med. 2016;164(12):795-803.

(9.) Chen J, Vargas-Bustamante A, Mortensen K, Ortega AN. Racial and ethnic disparities in health care access and utilization under the Affordable Care Act. Med Care. 2016;54(2):140-146.

(10.) Sommers BD, Maylone B, Blendon RJ, Orav EJ, Epstein AJ. Three-year impacts of the Affordable Care Act: improved medical care and health among low-income adults. Health Aff (Millwood). 2017;36(6):1119-1128.

(11.) Nikpay S, Buchmueller T, Levy H. Early Medicaid expansion in Connecticut stemmed the growth in hospital uncompensated care. Health Aff (Millwood). 2015;34(7):1170-1179.

(12.) Lo N, Roby DH, Padilla J, et al. Increased service use following Medicaid expansion is mostly temporary: evidence from California's low income health program. Health Policy Brief, UCLA Center for Health Policy Research. 2014. http://healthpolicy.ucla.edu/publications/Documents/PDF/2014/Demand_PB_FINAL_10-8-14.pdf. Accessed October 17, 2016.

(13.) Miller S. The effect of insurance on emergency room visits: an analysis of the 2006 Massachusetts health reform. J Pub Econ. 2012:96(11-12):893-908.

(14.) Ellimoottil C, Miller S, Ayanian J, Miller C. Effect of insurance expansion on utilization of inpatient surgery. JAMA Surg. 2014;149(8):829-836.

(15.) Sommers BD, Chua KP, Kenney GM, Long SK, McMorrow S. California's early coverage expansion under the Affordable Care Act: a county-level analysis. Health Serv Res. 2016:51(3): 825-845.

(16.) Finkelstein A, Taubman S, Wright B, et al. The Oregon Health Insurance Experiment: evidence from the first year. Q J Econ. 2012;127(3):1057-1106.

(17.) Taubman S, Allen H, Wright B, Baicker K, Finkelstein A, Oregon Health Study Group. Medicaid increases emergency department use: evidence from Oregon's Health Insurance Experiment. Science. 2014:343(6168):263-268.

(18.) Courtemanche C, Marton J, Yelowitz A. Who gained insurance coverage in 2014, the first year of full ACA implementation? Health Econ. 2016;25:778-784.

(19.) Courtemanche C, Marton J, Ukert B, Yelowitz A, Zapata D. Impacts of the Affordable Care Act on health insurance coverage in Medicaid expansion and non-expansion states. J Policy Anal Manage. 2017;36(1):178-210.

(20.) Sommers BD, Blendon RJ, Orav EJ. Both the 'private option' and traditional Medicaid expansions improved access to care for low-income adults. Health Aff (Millwood). 2016;35(1):96-105.

(21.) Benitez JA, Creel L, Jennings J. Kentucky's Medicaid expansion showing early promise on coverage and access to care. Health Aff (Millwood). 2016;35(3):528-534.

(22.) Frean M, Gruber J, Sommers BD. Premium subsidies, the Mandate, and Medicaid expansion: coverage effects of the Affordable Care Act. J Health Econ. 2016;53:72-86.

(23.) US Department of Labor, Bureau of Labor Statistics. Local area unemployment statistics. http://www.bls.gov/lau/. 2015. Accessed October 17, 2016.

(24.) The Nielsen Company. Nielsen demographic data. http://www.tetrad.com/demographics/usa/nielsen. Accessed October 17, 2016.

(25.) Kaiser Family Foundation. Medicaid income eligibility limits for adults as a percent of the Federal Poverty Level. http://kff.org/health-reform/state-indicator/medicaid-income-eligibility-limits-for-adults-as-a-percent-of-the-federal-poverty-level/. Accessed October 17, 2016.

(26.) Barrett M, Lopez-Gonzalez L, Hines A, Andrews R, Jiang J. An examination of expected payer coding in HCUP Databases. HCUP Methods Series Report #2014-03. Agency for Healthcare Research and Quality. http://www.hcup-us.ahrq.gov/reports/methods/2014-03.pdf. 2014. Accessed October 17, 2016.

(27.) Agency for Healthcare Research and Quality. Cost-to-charge ratio files. https://www.hcup-us.ahrq.gov/db/state/costtocharge.jsp. 2015. Accessed October 7, 2016.

(28.) Garthwaite C, Gross T, Notowidigdo M, Graves JA. Insurance expansion and hospital emergency department access: evidence from the Affordable Care Act. Ann Intern Med. 2017;166(3):172-180.

(29.) Dranove D, Garthwaite C, Ody C. Uncompensated care decreased at hospitals in Medicaid expansion states but not at hospitals in nonexpansion states. Health Aff (Millwood). 2016;35(8):1471-1479.

SUPPORTING INFORMATION

Additional supporting information may be found online in the Supporting Information section at the end of the article.

Gary Pickens PhD (1) | Zeynal Karaca PhD (2) | Teresa B. Gibson PhD (3) | Eli Cutler PhD (4) | Michael Dworsky PhD (5) | Brian Moore PhD (3) | Herbert S. Wong PhD (2)

(1) IBM Watson Health, Wilmette, Illinois

(2) U.S. Department of Health and Human Services, Agency for Healthcare Research and Quality, Rockville, Maryland

(3) IBM Watson Health, Ann Arbor, Michigan

(4) Qventis (Formerly of IBM Watson Health), Mountain View, California

(5) RAND Corporation, Santa Monica, California

Correspondence

Zeynal Karaca, PhD, U.S. Department of Health and Human Services, Agency for Healthcare Research and Quality, 5600 Fishers Lane, Office # 07N39, Rockville, MD 20857.

Email: zeynal.karaca@ahrq.hhs.gov

DOI: 10.1111/1475-6773.13165
TABLE 1 Demographic profiles of inpatient and emergency department
study samples, nonexpansion and expansion states, 2013, patients aged
19-64 y

                                 Emergency department
                                 analysis state sample
Characteristic                   Nonexpansion states   Expansion states

                                         11                    12
Number of states                 45 130 832            72 012 682
Total population
Female, by age in y (%)                  50.6                  50.2
  All                                     9.0                   9.0
  19-26                                   8.4                   8.8
  27-34                                  10.7                  10.6
  35-44                                  11.6                  11.5
  45-54                                  10.9                  10.3
  55-64
Male, by age in y (%)                    49.4                  49.8
  All                                     9.4                   9.5
  19-26                                   8.4                   8.9
  27-34                                  10.5                  10.5
  35-44                                  11.2                  11.2
  45-54                                  10.0                   9.6
  55-64
Insurance (%)                             7.8                  11.2
  Medicaid                                4.4                   3.3
  Medicare                                2.7                   1.3
  Other                                  62.0                  64.7
  Private insurance                      23.1                  19.4
  Uninsured
    <100% FPL                             7.9                   5.6
    10096-137% FPL                        2.8                   2.3
    138%-249% FPL                         6.5                   5.5
    250%-399% FPL                         3.7                   3.4
    >400% FPL                             2.3                   2.6
Eligibility 2013 (%)
  Newly eligible for Medicaid            NA                     2.7
   but not enrolled, 2013
  Previously eligible for                 1.7                   2.6
   Medicaid but not enrolled,
   2013
  Uninsured and newly eligible            3.0                   8.9
   exchange, for health
   insurance 2013
  Uninsured and newly eligible            9.3                   5.5
   for health insurance
   exchange subsidy, 2013
2013 CER (x100)
  Medicaid                               22.1                  75.2
  Private                                21.0                  13.7
  Uninsured                              63.1                  86.4
  All payers                             14.6                  16.8

                                 Inpatient analysis state sample
Characteristic                   Nonexpansionstates   Expansion states

                                         18                   19
Number of states                 80 244 758           92 690 496
Total population
Female, by age in y (%)                  50.4                 50.1
  All                                     9.1                  8.9
  19-26                                   8.6                  8.8
  27-34                                  10.7                 10.6
  35-44                                  11.5                 11.4
  45-54                                  10.6                 10.4
  55-64
Male, by age in y (%)                    49.6                 49.9
  All                                     9.5                  9.5
  19-26                                   8.7                  9.0
  27-34                                  10.5                 10.6
  35-44                                  11.1                 11.1
  45-54                                   9.8                  9.7
  55-64
Insurance (%)                             7.4                 10.6
  Medicaid                                4.1                  3.5
  Medicare                                2.6                  1.6
  Other                                  62.3                 64.5
  Private insurance                      23.5                 19.8
  Uninsured
    <100% FPL                             7.8                  5.8
    10096-137% FPL                        2.9                  2.3
    138%-249% FPL                         6.5                  5.6
    250%-399% FPL                         3.8                  3.5
    >400% FPL                             2.5                  2.6
Eligibility 2013 (%)
  Newly eligible for Medicaid             NA                   2.4
   but not enrolled, 2013
  Previously eligible for                 1.7                  2.4
   Medicaid but not enrolled,
   2013
  Uninsured and newly eligible           13.2                  9.0
   exchange, for health
   insurance 2013
  Uninsured and newly eligible            9.3                  5.6
   for health insurance
   exchange subsidy, 2013
2013 CER (x100)
  Medicaid                               23.1                 81.9
  Private                                21.3                 14.0
  Uninsured                              63.1                 86.7
  All payers                             14.8                 17.1

Note: Percentages may not sum to 100 percent because of rounding
issues. Abbreviations: CER, coverage expansion ratio; FPL, federal
poverty level; NA, not applicable.
Source: American Community Survey Public Use Microdata Samples,
2011-2015.

TABLE 2 Observed inpatient discharge and emergency department visit
outcome rates for expansion and nonexpansion states by insurance type,
2011-2013 and 2014-2015

Outcomes                              Setting

Medicaid
  Discharge or visit rate per 1000    Inpatient
                                      Emergency department
  Case mix index                      Inpatient
                                      Emergency department
  Total cost ($)                      Inpatient
                                      Emergency department
Private insurance
  Discharge or visit rate per 1000    Inpatient
                                      Emergency department
  Case mix index                      Inpatient
                                      Emergency department
  Total cost ($)                      Inpatient
                                      Emergency department
Uninsured
  Discharge or visit rate per 1000    Inpatient
                                      Emergency department
  Case mix index                      Inpatient
                                      Emergency department
  Total cost ($)                      Inpatient
                                      Emergency department
All payer
  Discharge or visit rate per 1000    Inpatient
                                      Emergency department
  Case mix index                      Inpatient
                                      Emergency department
  Total cost ($)                      Inpatient
                                      Emergency department

Outcomes               Expansion   2011-2013   2014-2015    % Change
                       state
  Medicaid
  Discharge or visit   No             276         246       -11.1
  rate per 1000        Yes            228         193       -15.4
                       No            1027        1028         0.0
                       Yes            778         832         7.0
  Case mix index       No               0.79        0.80      1.1
                       Yes              0.82        0.87      7.3
                       No               1.01        1.02      1.0
                       Yes              1.01        1.02      1.0
  Total cost ($)       No            8254        8559         3.7
                       Yes           9875      10 966        11.0
                       No             502         571        13.8
                       Yes            588         665        13.2
Private insurance
  Discharge or visit   No              59          55        -7.5
  rate per 1000        Yes             60          53       -10.9
                       No             166         180         7.9
                       Yes            166         166        -0.2
  Case mix index       No               0.98        0.99      0.5
                       Yes              0.98        0.98      0.4
                       No               1.11        1.12      1.2
                       Yes              1.08        1.10      2.1
  Total cost ($)       No          10 023      10 588         5.6
                       Yes         11 135      11 944         7.3
                       No             654         745        13.9
                       Yes            681         786        15.4
Uninsured
  Discharge or visit   No              46          52        14.0
  rate per 1000        Yes             42          27       -36.2
                       No             487         577        18.4
                       Yes            358         334        -6.7
  Case mix index       No               0.99        0.99      0.7
                       Yes              0.98        0.95     -3.3
                       No               0.96        0.97      1.1
                       Yes              0.98        0.98      0.6
  Total cost ($)       No            8977           9448      5.2
                       Yes         10 174      10 197         0.2
                       No             485         552        13.8
                       Yes            569         653        14.8
All payer
  Discharge or visit   No              87          85        -2.2
  rate per 1000
                       Yes             87          83        -4.6
                       No             350         372         4.5
                       Yes            308         326         4.2
  Case mix index       No               0.96        0.97      0.6
                       Yes              0.95        0.97      1.3
                       No               1.03        1.05      1.4
                       Yes              1.03        1.05      1.5
  Total cost ($)       No            9817      10 309         6.1
                       Yes         11 080      11 957         8.6
                       No             555         640        16.8
                       Yes            623         716        15.0

TABLE 3 Adjusted percent changes in inpatient discharges in Medicaid
expansion and nonexpansion states in 2014-2015 (after affordable care
act implementation) compared with 2011-2013, by age, sex, and insurance
type-indicator model

                     State expan-   Demographic group (y)
Insurance coverage   sion status    Female 19-34   Female 35-54

Discharge rates
  Medicaid           No               3.9           -2.5
                     Yes            -17.7 (*)       -4.5 (*)
  Private            No               2.8           -0.7
                     Yes              2.7            1.2
  Uninsured          No               5.4            6.1
                     Yes            -12.3          -42.2 (*)
  All payer          No               1.7           -0.2
                     Yes              2.5            1.9 (*)
Case mix index
  Medicaid           No              -1.4           -0.4
                     Yes              6.2 (*)        1.7 (*)
  Private            No              -1.7            0.4
                     Yes             -1.5 (*)       -2.4 (*)
  Uninsured          No               2.4 (*)        1.5
                     Yes            -14.8 (*)       -4.7
  All payer          No              -1.0            0.7
                     Yes              0.0           -1.0
Estimated cost
  Medicaid           No              -5.5           -3.2
                     Yes              2.8           -5.1 (*)
  Private            No             -10.0 (*)       -4.9 (*)
                     Yes             -3.7 (*)       -6.7 (*)
  Uninsured          No              -4.9 (*)       -4.0 (*)
                     Yes            -15.6 (*)       -9.1 (*)
  All payer          No              -7.4 (*)       -3.5 (*)
                     Yes             -2.3           -5.8 (*)

Insurance coverage   Female 55-64      Male 19-34      Male 35-54

Discharge rates
  Medicaid           -12.7 (*)          -4.5            -7.0
                      -8.7 (*)          16.0 (*)        14.3 (*)
  Private              2.9               0.8             5.1 (*)
                      -1.8              -3.3            -4.1 (*)
  Uninsured            4.7               7.8             6.1
                     -39.6 (*)         -44.7 (*)       -49.0 (*)
  All payer            2.2 (*)           0.8 (*)         1.5 (*)
                      -0.5               4.8 (*)         1.8
Case mix index
  Medicaid             0.1              -0.4             0.6
                       1.3              -1.1            -1.7 (*)
  Private              0.1              -1.3            -0.9
                       1.9 (*)           0.4             1.0
  Uninsured            1.2              -0.3             0.0
                       2.1 (*)           0.2            -0.2
  All payer            0.2              -0.9            -0.2
                       1.8 (*)           0.1             0.1
Estimated cost
  Medicaid            -4.8 (*)          -5.7            -3.6
                      -7.5 (*)         -12.1 (*)        -9.8 (*)
  Private             -6.6 (*)          -4.5 (*)        -5.2 (*)
                      -4.3 (*)          -3.6 (*)        -4.0 (*)
  Uninsured           -4.5 (*)          -3.7 (*)        -4.9 (*)
                       0.3              -7.0 (*)        -5.0 (*)
  All payer           -5.6 (*)          -5.3 (*)        -4.5 (*)
                      -4.7 (*)          -2.6 (*)        -4.2 (*)

Insurance coverage   Male 55-64

Discharge rates
  Medicaid           -10.4 (*)
                       5.7
  Private              5.2 (*)
                       0.0
  Uninsured            9.2 (*)
                     -39.2 (*)
  All payer            5.0 (*)
                       4.4 (*)
Case mix index
  Medicaid            -0.5
                      -0.2
  Private             -0.5
                       1.1 (*)
  Uninsured           -0.6
                       0.0
  All payer           -0.5
                       0.4 (*)
Estimated cost
  Medicaid            -4.9 (*)
                      -6.9 (*)
  Private             -6.2 (*)
                      -4.5 (*)
  Uninsured           -5.7 (*)
                      -4.4
  All payer           -5.3 (*)
                      -4.7 (*)

(*) P < 0.05.

TABLE 4 Adjusted percent changes in emergency department visits in
Medicaid expansion and nonexpansion states in 2014-2015 (after
affordable care act implementation) compared with 2011-2013, by age,
sex, and insurance type-indicator model

                 State expansion   Demographic group (y)
Insurance        status            Female 19-34   Female 35-54
coverage

Visit rates
  Medicaid       No                  6.3            1.3
                 Yes                -6.0            5.3
  Private        No                  7.8 (*)       12.3 (*)
                 Yes                -3.2            2.1
  Uninsured      No                  5.1 (*)        7.6 (*)
                 Yes               -14.6 (*)      -19.9 (*)
  All payer      No                  2.7 (*)        3.9 (*)
                 Yes                -4.6 (*)        4.1 (*)
Case mix index
  Medicaid       No                 -0.7           -0.1
                 Yes                -0.5           -0.7 (*)
  Private        No                 -1.4 (*)       -0.4
                 Yes                -0.5           -0.8 (*)
  Uninsured      No                 -0.7            0.3
                 Yes                -0.2           -1.5 (*)
  All payer      No                 -0.6            0.5
                 Yes                -0.2           -0.7 (*)
Estimated cost
  Medicaid       No                  2.1            6.0 (*)
                 Yes                -5.5 (*)       -8.2 (*)
  Private        No                 -2.1            2.3
                 Yes                -1.0           -3.1 (*)
  Uninsured      No                  2.8            6.8 (*)
                 Yes                -3.3           -4.9
  All payer      No                  0.4            5.4 (*)
                 Yes                -2.4           -4.7 (*)

Insurance        Female 55-64   Male 19-34   Male 35-54   Male 55-64
coverage

Visit rates
  Medicaid       -13.3 (*)      -10.4        -12.4        -14.3 (*)
                  10.4           25.2 (*)     30.2 (*)     23.3 (*)
  Private          7.9 (*)       -0.4          7.9 (*)      6.7 (*)
                   8.9 (*)       -4.0 (*)      2.3          9.4
  Uninsured       -1.0            1.7          2.9          0.2
                  -0.2          -10.4        -14.8 (*)      0.4
  All payer       -0.1           -4.4 (*)     -1.8          0.3
                  12.4 (*)       -0.6          6.1 (*)     16.6 (*)
Case mix index
  Medicaid         1.0            0.0          1.1          2.6
                  -2.1 (*)       -0.4         -1.9 (*)     -2.5 (*)
  Private          0.2           -0.5          0.3          1.1
                  -1.1 (*)       -0.1         -0.8 (*)     -1.0 (*)
  Uninsured        0.8            0.2          0.8 (*)      1.9 (*)
                  -3.5 (*)       -0.4         -2.3 (*)     -3.4 (*)
  All payer        1.3 (*)        0.1          1.1 (*)      2.2 (*)
                  -1.6 (*)        0.4 (*)     -0.8 (*)     -1.6 (*)
Estimated cost
  Medicaid         8.1 (*)        3.7          7.7 (*)     11.0 (*)
                  -9.3 (*)       -7.6 (*)     -9.8 (*)     -6.4
  Private          6.1 (*)        0.2          3.5 (*)      6.9 (*)
                  -2.3 (*)       -1.2         -4.2 (*)     -3.5 (*)
  Uninsured        6.1 (*)        3.3          6.8 (*)      7.6 (*)
                  -8.7 (*)       -3.9         -8.7 (*)    -11.4 (*)
  All payer        8.8 (*)        2.1          6.3 (*)      8.8 (*)
                  -5.3 (*)       -2.6         -6.1 (*)     -6.6 (*)

(*) P < 0.05.

TABLE 5 Semi-elasticity estimates for inpatient discharges and
emergency department visits in Medicaid expansion and nonexpansion
states, by age, sex, and insurance type--coverage expansion ratio model

                                       State
Outcome                    Insurance   expansion  Female 19-34
                           coverage    status

Inpatient discharges
  Discharge rate           Medicaid    No          0.0039
                                       Yes        -0.0020
                           Private     No         -0.0003
                                       Yes        -0.0019
                           Uninsured   No          0.0006
                                       Yes        -0.0016
                           All Payer   No          0.0001
                                       Yes         0.0039
  Case mix                 Medicaid    No         -0.0022 (*)
   index                               Yes         0.0002
                           Private     No         -0.0011
                                       Yes        -0.0012
                           Uninsured   No          0.0004
                                       Yes        -0.0018 (*)
                           All Payer   No         -0.0020 (*)
                                       Yes        -0.0014 (*)
  Estimated                Medicaid    No         -0.0045 (*)
   cost                                Yes        -0.0003
                           Private     No         -0.0053 (*)
                                       Yes        -0.0039 (*)
                           Uninsured   No         -0.0008 (*)
                                       Yes        -0.0020 (*)
                           All Payer   No         -0.0070 (*)
                                       Yes        -0.0035 (*)
Emergency department visits
  Discharge rate           Medicaid    No          0.0027
                                       Yes        -0.0007
                           Private     No          0.0034 (*)
                                       Yes        -0.0012
                           Uninsured   No          0.0007
                                       Yes        -0.0018 (*)
                           All Payer   No          0.0024
                                       Yes        -0.0018
  Case mix                 Medicaid    No         -0.0008 (*)
   index                               Yes        -0.0001
                           Private     No         -0.0008 (*)
                                       Yes        -0.0010 (*)
                           Uninsured   No         -0.0002 (*)
                                       Yes         0.0000
                           All Payer   No         -0.0009 (*)
                                       Yes        -0.0006 (*)
  Estimated                Medicaid    No         -0.0006
   cost                                Yes        -0.0010 (*)
                           Private     No         -0.0019 (*)
                                       Yes        -0.0020 (*)
                           Uninsured   No          0.0003
                                       Yes        -0.0004
                           All Payer   No         -0.0017
                                       Yes        -0.0026 (*)

                           Demographic group (y)
Outcome                    Female 35-54   Female 55-64    Male 19-34

Inpatient discharges
  Discharge rate           -0.0008        -0.0159         -0.0021
                           -0.0002        -0.0008          0.0008
                           -0.0002         0.0012         -0.0005
                           -0.0008        -0.0052 (*)     -0.0024 (*)
                            0.0007         0.0005          0.0010
                           -0.0063 (*)    -0.0061 (*)     -0.0070 (*)
                           -0.0015 (*)     0.0005         -0.0005
                           -0.0008        -0.0159         -0.0021
  Case mix                 -0.0041 (*)    -0.0034          0.0000
   index                   -0.0001        -0.0001         -0.0001
                           -0.0020 (*)    -0.0020 (*)     -0.0006 (*)
                           -0.0044 (*)     0.0006         -0.0001
                            0.0002         0.0002         -0.0001
                           -0.0006         0.0002 (*)      0.0000
                           -0.0029 (*)    -0.0028 (*)     -0.0012 (*)
                           -0.0036 (*)    -0.0009 (*)     -0.0007 (*)
  Estimated                -0.0055 (*)    -0.0111 (*)     -0.0007
   cost                    -0.0009 (*)    -0.0011 (*)     -0.0008 (*)
                           -0.0054 (*)    -0.0068 (*)     -0.0017 (*)
                           -0.0089 (*)    -0.0059 (*)     -0.0023 (*)
                           -0.0008 (*)    -0.0008 (*)     -0.0006 (*)
                           -0.0012 (*)     0.0000 (*)     -0.0009 (*)
                           -0.0069 (*)    -0.0094 (*)     -0.0036 (*)
                           -0.0072 (*)    -0.0066 (*)     -0.0020 (*)
Emergency department visits
  Discharge rate            0.0031 (*)    -0.0314         -0.0029 (*)
                            0.0011 (*)     0.0014          0.0016 (*)
                            0.0069 (*)     0.0056 (*)      0.0000
                            0.0033 (*)     0.0122 (*)     -0.0017 (*)
                            0.0012 (*)    -0.0003          0.0002
                           -0.0025 (*)    -0.0001         -0.0013
                            0.0043 (*)     0.0011         -0.0017
                            0.0046 (*)     0.0138 (*)      0.0000
  Case mix                 -0.0008 (*)     0.0018         -0.0001
   index                   -0.0001        -0.0001         -0.0001
                           -0.0008 (*)    -0.0006         -0.0004 (*)
                           -0.0015 (*)    -0.0019 (*)     -0.0004 (*)
                            0.0000         0.0001          0.0000
                           -0.0002 (*)    -0.0004 (*)     -0.0001
                           -0.0007 (*)     0.0000         -0.0004 (*)
                           -0.0012 (*)    -0.0022 (*)     -0.0002
  Estimated                -0.0002         0.0094          0.0000
   cost                    -0.0011 (*)    -0.0010         -0.0006 (*)
                           -0.0005         0.0018         -0.0005
                           -0.0058 (*)    -0.0059 (*)     -0.0009
                            0.0008 (*)     0.0008 (*)      0.0004
                           -0.0006 (*)    -0.0011 (*)     -0.0005
                            0.0012         0.0057 (*)     -0.0003
                           -0.0053 (*)    -0.0071 (*)     -0.0017 (*)

                           Demographic group (y)
Outcome                    Male 35-54       Male 55-64

Inpatient discharges
  Discharge rate           -0.0055 (*)      -0.0078 (*)
                            0.0006           0.0008
                           -0.0009           0.0025 (*)
                           -0.0085 (*)      -0.0041
                            0.0006           0.0012 (*)
                           -0.0080 (*)      -0.0062 (*)
                           -0.0030           0.0036 (*)
                           -0.0055 (*)      -0.0078 (*)
  Case mix                 -0.0002          -0.0014
   index                   -0.0002          -0.0002
                           -0.0018 (*)      -0.0022 (*)
                           -0.0005           0.0002
                            0.0000          -0.0001
                            0.0000           0.0000
                           -0.0022 (*)      -0.0034 (*)
                           -0.0016 (*)      -0.0018 (*)
  Estimated                -0.0008          -0.0038 (*)
   cost                    -0.0008 (*)      -0.0009 (*)
                           -0.0041 (*)      -0.0065 (*)
                           -0.0043 (*)      -0.0056 (*)
                           -0.0009 (*)      -0.0010 (*)
                           -0.0007 (*)      -0.0006 (*)
                           -0.0056 (*)      -0.0089 (*)
                           -0.0042 (*)      -0.0062 (*)
Emergency department visits
  Discharge rate           -0.0049 (*)      -0.0103 (*)
                            0.0023 (*)       0.0025 (*)
                            0.0033 (*)       0.0043
                            0.0015           0.0119 (*)
                            0.0004           0.0000
                           -0.0019 (*)       0.0000
                           -0.0006           0.0014
                            0.0038 (*)       0.0164 (*)
  Case mix                 -0.0002           0.0014 (*)
   index                   -0.0002 (*)      -0.0002
                           -0.0006 (*)       0.0000
                           -0.0015 (*)      -0.0020 (*)
                            0.0000           0.0003
                           -0.0003 (*)      -0.0004 (*)
                           -0.0006 (*)       0.0008
                           -0.0013 (*)      -0.0023 (*)
  Estimated                 0.0008           0.0051 (*)
   cost                    -0.0009 (*)      -0.0004
                           -0.0005           0.0024
                           -0.0053 (*)      -0.0070 (*)
                            0.0008 (*)       0.0011 (*)
                           -0.0011 (*)      -0.0015 (*)
                            0.0009           0.0058 (*)
                           -0.0052 (*)      -0.0078 (*)

Notes: Semi-elasticity is calculated as exp(theta) - 1. Semi-elasticity
is interpreted as the average percent change in outcome rates in
2014-2015 (after Affordable Care Act implementation), compared with
2011-2013, associated with a unit increase in the coverage expansion
ratio.
(*) P<0.05.
COPYRIGHT 2019 Health Research and Educational Trust
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2019 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:RESEARCH ARTICLE
Author:Pickens, Gary; Karaca, Zeynal; Gibson, Teresa B.; Cutler, Eli; Dworsky, Michael; Moore, Brian; Wong,
Publication:Health Services Research
Date:Aug 1, 2019
Words:8541
Previous Article:Individual market health plan affordability after cost-sharing reduction subsidy cuts.
Next Article:Decomposing changes in the growth of U.S. prescription drug use and expenditures, 1999-2016.
Topics:

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters