SEMI-ANNUAL REPORT STUDY OF THE IMPACT OF THE ACA IMPLEMENTATION IN KENTUCKY Prepared for: Foundation for a Healthy Kentucky Prepared by: State Health Access Data Assistance Center (SHADAC) University of Minnesota 2221 University Ave, Suite 345 Minneapolis, MN September 2016 BRIEF SUMMARY OF PROGRESS ON THE STUDY The study began in March 2015. Since then, SHADAC’s main study accomplishments and deliverables include: • Study Methods and Plan • Data Scan and Gaps Analysis • Quarterly Snapshots: • January-March 2015 • April-June 2015 • July-September 2015 • October-December 2015 • January-March 2016 • Baseline data file • 2015 Semi-Annual Report • 2016 Annual Report • Special Report: Uninsurance Estimates from the American Community Survey (Memo), and KY Health Insurance Coverage 2014/Estimates from the American Community Survey (Infographic), September 2015 • Special Report: ACA Improves Health Insurance Coverage for Kentucky Children, October 2015 • Special Report: High-deductible Health Insurance in Kentucky, June 2016 • Special Issue Brief: Section 1115 Waivers and ACA Medicaid Expansions: A Review of Policies and Evidence from Five States, May 2016 • Quarterly Meetings with Oversight Committee to solicit ideas and feedback • Kentucky Health Reform Survey instrument development • Conducting the Kentucky Health Reform Survey (K-HRS), March-May 2016. In addition, SHADAC staff have tracked relevant developments in Kentucky, including newly-published studies, data updates, and state policy proposals and decisions. We track these ongoing developments through media coverage, peer-reviewed publications, and the grey literature. SHADAC has continued to analyze secondary federal and state data sources for the study’s ongoing indicators. SHADAC also has conducted the Kentucky Health Reform Survey and will continue to analyze the data. With in- put from the Foundation and Oversight Committee, SHADAC has begun planning for the qualitative component of the study and is on schedule to complete the focus groups and key stakeholder interview in Year 2 of the study. SEMI-ANNUAL REPORT STUDY OF THE IMPACT OF THE ACA IMPLEMENTATION IN KENTUCKY September 2016 TABLE OF CONTENTS I. Introduction 1 Overview 1 Purpose of Current Report 2 II. Study Findings: Data Update 3 Domain #1: Health Insurance Coverage 3 Domain #2: Access 4 Domain #3: Cost 13 Domain #4: Quality 19 Domain #5: Health Outcomes 21 III. Study Findings: Kentucky Health Reform Survey 24 IV. Discussion & Next Steps 36 V. Appendix: Data Sources, Methods and Indicators 37 VI. End Notes 39 INTRODUCTION Overview This report was produced by the State Health Access Data Assistance Center (SHADAC) at the University of Minnesota as part of a mixed-methods study, Study of the Impact of Implementation of the Affordable Care Act (ACA) in Kentucky, funded by the Foundation for a Healthy Kentucky (Foundation). The study evaluates Kentucky’s performance in five domains: coverage, access, cost, quality, and health outcomes. As part of this project, SHADAC uses semi-annual and annual reports to document the impact of the ACA in Kentucky using a set of indicators agreed upon by the Foundation and its ACA Impact Study Oversight Committee. These reports track change in the indicators throughout the duration of this 34-month study (March 2015 through January 2018), and include comparisons of Kentucky metrics with the United States and other states. This report includes data obtained from analysis of federal and state data resources. This report also presents preliminary findings from the Kentucky Health Reform Survey (K-HRS), which was conducted in spring 2016 by SHADAC and the University of Cincinnati’s Institute for Policy Research. Future reports will present additional analyses of K-HRS data, including examining K-HRS estimates in comparison to pre- and post-ACA trends from the Kentucky Health Issues Poll (KHIP). Additionally, future reports will present findings from the qualitative components of the study, including focus groups with Medicaid ben- eficiaries and interviews with key stakeholders in Kentucky. 1 1 | STATE HEALTH ACCESS DATA ASSISTANCE CENTER SEMI-ANNUAL REPORT Study of the Impact of the ACA Implementation in Kentucky INTRODUCTION Purpose and Layout of Current Report The main purposes of this report are to briefly provide an introduction and summary of progress on the study, as well as: 1) provide updates on key study indicators in Kentucky, and 2) present preliminary find- ings from the spring 2016 Kentucky Health Reform Survey. 1) Data update 2) Kentucky Health Reform Survey (K-HRS) findings This section provides a data update to the key The next section of the report presents prelimi- study indicators that were introduced in the nary findings from the Kentucky Health Reform study’s first semi-annual baseline report (August Survey (K-HRS), which was conducted from March 2015) and first annual report (February 2016). to May 2016. The survey asked non-elderly adults While these data include indicators from all five questions related to four primary domains: health study domains (coverage, access, cost, quality, and insurance coverage and experiences with Ken- health outcomes), they are focused more heavily tucky’s marketplace, kynect; affordability of care; on the access and cost domains, with an update to access to care; and health status. These domains 2014 estimates from the National Health Interview and the specific survey questions were selected in Survey (NHIS), the first year since ACA implemen- consultation with the Foundation and the study’s tation. This section also updates certain measures Oversight Committee with the goals of address- in other domains, such as employer-sponsored ing key study questions about the impacts of ACA insurance (coverage) and hospital admission rates implementation in Kentucky. We included several for diabetes, hypertension, and asthma (quali- questions to remain consistent with the existing ty), and it adds a new set of indicators that were ongoing Kentucky Health Issues Poll (KHIP) survey, not included in prior reports: cigarette smoking which will allow us to track trends pre- and post- among Kentucky adults and adolescents (health ACA across these surveys over time. This report outcomes). All of the updated data in this report focuses on six topics from the Kentucky Health represent the time period since implementation Reform Survey: uninsurance and coverage rates, of the ACA. Some of the updates were available experiences using kynect, concerns about losing for 2015, while others were only available for 2014 coverage, care delayed or forgone due to cost, at this time. As a baseline comparison, we use cal- dental coverage and care, and use of emergency endar year 2012 data for most indicators because departments. Future reports will examine trends it pre-dated the first ACA enrollment period; how- in indicators included in both the K-HRS and the ever, for certain indicators in which 2012 data KHIP. weren’t available, we use 2013 as a baseline. For selected indicators, we also compare Kentucky to national figures and nearby states for comparison (Arkansas, Indiana, Illinois, Ohio, Missouri, Tennes- see, Virginia, and West Virginia).1 SEPTEMBER 2016 | WWW.SHADAC.ORG 2 2 II. STUDY FINDINGS: DATA UPDATE DOMAIN #1: HEALTH INSURANCE COVERAGE Health insurance coverage is a critical component In this report, we present the most recent data of access to health care services. Having health available (2015) for the employer offer rate, com- insurance is associated with increased access pared to 2012 baseline data. to needed medical care, better health care out- “ comes, and improved health status.2 In this study, the metrics used to monitor health insurance In 2015, the employer coverage in Kentucky and over time include the offer rate fell significantly distribution of type of health insurance coverage (public, private, and uninsured); rates of under- from 54.4% to 47.8%, insurance; and the percentage of employers that offer health insurance coverage. Our data sources a 6.6 percentage point in this domain include federal surveys that provide drop compared to 2012. state-level estimates of health insurance coverage including the American Community Survey (ACS), This decrease was driven the Medical Expenditure Panel Survey-Insurance Component (MEPS-IC) and the Current Popula- primarily by a decline tion Survey (CPS). For this report, updated health in employer-based insurance data were only available for one met- ric – employer offer rate. Subsequent annual and semi-annual reports will provide updated data coverage offered by “ as they become available for several additional small employers. . metrics. COVERAGE MEASURES Decline in Employer Offers of Coverage, Most people in the U.S. with health insurance Driven by Small Employers coverage have health insurance that is sponsored Overall, Kentucky saw a statistically significant through an employer. Looking at employer offer decrease in employers offering health insurance rates, there has been a statistically significant since 2012. The employer offer rate fell from decrease from 2012-2015. Figure 1.1 shows that 54.4% in 2012 to 47.8% in 2015, a 6.6 percentage in 2015, 26.6% of Kentucky’s small employers (few- point drop. This decrease was driven primarily by er than 50 employees) and 98.3% of larger em- small employers; the offer rate for large employers ployers (50 or more employees) offered coverage. remained statistically steady at approximately 98% Offer rates among small employers experienced a in 2012 and 2015. statistically significant decrease from 2012-2015, with a decrease of nearly 10 percentage points over that time period. In contrast, offer rates among larger employers remained stable. This sig- nificant decrease in employers offering coverage is consistent with a long-term decline nationally, especially among smaller businesses, that preced- ed the ACA.3 3 3 | STATE HEALTH ACCESS DATA ASSISTANCE CENTER SEMI-ANNUAL REPORT Study of the Impact of the ACA Implementation in Kentucky DOMAIN #1: HEALTH INSURANCE COVERAGE 100% 2012 2015 98.7% 98.3% FIGURE 1.1: Employer Offer Rates by 90% Private Sector Employers 80% for Kentucky, 2012-2015 70% 60% 54.4% 50% 47.8%* 40% 36.4% 30% 26.6%* 20% 10% 0% All private Firms with less than Firms with 50 sector employers 50 employees or more employees *Difference is statistically significant at the 95% level. Source: 2012 and 2015 MEPS-IC. Estimates are for percent of private sector establishments that offer coverage by firm size. DOMAIN #2: ACCESS The U.S. Institute of Medicine defines health As new data become available, future reports care access as “the timely use of personal health will continue to provide more insight about any services to achieve the best health outcomes.”4 impacts of the ACA’s coverage expansions on Even among those with health insurance coverage, access to health care services. financial and non-financial access to care barriers “ can persist.5 We use 6 indicators to monitor health care access in this report.6 For the access Overall, the domain we obtained data for the indicators from indicators show that the National Health Interview Survey (NHIS). We include data for children under age 19 as well as health care access in non-elderly and elderly adults where data were available. In all cases, we present the most recent Kentucky has been data available (2014), compared to the 2012 baseline. relatively stable “ since 2012. Overall, the indicators show that health care access in Kentucky has been relatively stable since 2012. SEPTEMBER 2016 | WWW.SHADAC.ORG 4 4 ACCESS MEASURES Kentuckians Reporting a Usual Source of Care Figure 2.1 presents the percentage of Kentuckians Remains Stable reporting a usual source of care (not including an Having a usual source of care is “a summary mea- emergency department) by age and over time. sure of adequate access to primary care,”⁷ and Each age group showed higher reporting of a usu- some studies have found it to be even more im- al source of care between 2012 and 2014, but the portant for health outcomes than having health increase was only statistically significant for those insurance.⁸ The measure we use from the NHIS 65 years and older. Overall, 85.6% of Kentuckians asks, “Is there a place you usually go when you are reported having a usual source of care in 2014. sick or need advice about your health?” We also Elderly adults reported the highest usual source use responses to the follow up question: “What of care (98.3%), followed by children (94.8%) and kind of place is it?” to make sure that emergen- non-elderly adults (79.1%). cy department visits were not considered to be a usual source of care. FIGURE 2.1: 100% 2012 2014 98.3%* Usual Source of Care by 93.8% 94.8% 94.1% Age Category, Kentucky, 83.6% 85.6% 2012-2014 80% 77.6% 79.1% 60% 40% 20% 0% All 0-18 19-64 65+ *Difference is statistically significant at the 95% level. Source: SHADAC analysis of the civilian non-institutional population in the 2012 and 2014 NHIS using the SHADAC Data Center. For this measure, we also present comparisons source of care (Figure 2.3). Although there was a between Kentucky, neighboring states, and the relatively large difference between West Virginia's U.S. For these cross-state comparisons, we use 2012 and 2014 estimates for this indicator, the dif- two types of charts: The bar charts (e.g., Figure ference was not statistically significant.⁹ 2.2) indicate whether Kentucky’s most recent estimates differ from the comparison states, while More Than 1 in 4 Reacting to Drug Costs the plot-charts (e.g., Figure 2.3) present baseline Another indicator of access is changes in prescrip- and most-recent estimates for Kentucky and com- tion drug usage due to cost. This is a summary mea- parison states, and they indicate whether those sure that includes: asking the doctor for cheaper within-state changes were statistically significant. medications, delaying refills, taking less medication than prescribed, skipping dosages, using alterna- Figure 2.2 shows that in 2014, the percentage tive therapies, and/or buying medications out of of Kentuckians with a usual source of care was the country. This measure indicates whether peo- significantly lower than that of four nearby states ple are making decisions based on cost that may (Arkansas, Indiana, Ohio, and Tennessee). Only negatively affect their health. For this indicator, one state – Ohio – and the U.S. had significant we present 2014 estimates for non-elderly adults increases in the overall rates of reported usual 5 5 | STATE HEALTH ACCESS DATA ASSISTANCE CENTER SEMI-ANNUAL REPORT Study of the Impact of the ACA Implementation in Kentucky DOMAIN #2: ACCESS 100% Kentucky United States Medicaid Expansion States Non-expansion States FIGURE 2.2: Usual Source of Care, 98% Kentucky Compared to 96% Neighboring States and 94% U.S. Rate, 2014 (all ages) 92% 91.3%* 90.5%* 90.8%* 90.4%* 90% 89.8% 89.3% 89.3% 88.8% 88.8% 88% 86% 85.6% 84% 82% 80% KY U.S. AR IL IN OH WV MO TN VA *Difference is statistically significant across states (e.g., Kentucky vs. Arkansas) at the 95% level. Source: SHADAC analysis of the civilian non-institutional population in the 2014 NHIS using the SHADAC Data Center. Note: While Indiana is a Medicaid expansion state, the state did not expand its Medicaid program until 2015. 100% 2012 2014 FIGURE 2.3: Usual Source of Care, Kentucky Compared to Neighboring States and 95% U.S. Rate, 2012-2014 91.3%* (all ages) 90.5% 90.8% 90.4% 89.9% 90% 89.3% 89.3% 88.8%* 88.8% 89.8% 89.3% 88.4% 88.1% 85.6% 87.7% 87.4% 87.4% 86.8% 85% 83.6% 82.4% 80% KY U.S. AR IL IN OH WV MO TN VA *Difference is statistically significant within the state (e.g., Arkansas 2012 estimate vs. Arkansas 2014 estimate) at the 95% level. Source: SHADAC analysis of the civilian non-institutional population in the 2012 and 2014 NHIS using the SHADAC Data Center. (ages 19-64) and elderly adults (ages 65 and old- tions due to cost between 2012 (26.2%) and 2014 er), as compared to the 2012 baseline (Figure 2.4). (22.8%), but this was not statistically significant. In 2014, 27.8% of non-elderly Kentuckians report- ed altering their prescription medications due to “ cost. Although this was a decrease of 6.3 percent- In 2014, 27.8% Kentuckians age points from the 2012 rate of 34.1%, the change was not statistically significant. There was also a 3.4 percentage point decrease for elderly adults, reported altering their prescrip- “ those ages 65 and older, who reported skipping, tion medications due cost. .. delaying, or altering their prescription medica- SEPTEMBER 2016 | WWW.SHADAC.ORG 6 6 FIGURE 2.4: 40% 2012 2014 Skipping, Delaying, or Altering Prescription 34.1% Drug Use Due to Cost, Kentucky, 2012-2014 30% 27.8% 26.2% (ages 19-64 & 65+) 22.8% 20% 10% 0% 19-64 65+ *Difference is statistically significant at the 95% level. Source: SHADAC analysis of the civilian non-institutional population in the 2012 and 2014 NHIS using the SHADAC Data Center. It will be important to follow this indicator as fu- ure 2.5). Figure 2.6 shows that only the U.S. and ture years of data become available to determine Indiana experienced significant increases in this whether statistically significant changes occur. measure between 2012 and 2014. We also present comparisons between Kentucky, Elderly Adults, Children Reported Higher nearby states, and the U.S. for this measure Shares Visiting a Health Care Provider (Figures 2.5 and 2.6). Indiana was the only neigh- Having a visit with a health care provider during boring state with a significantly higher rate than the past year is another way to gauge access to Kentucky of reporting skipping, delaying, or alter- health care. For this measure, we include visits ing prescription drug use due to cost in 2014 (Fig- FIGURE 2.5: 50% Kentucky United States Medicaid Expansion States Non-expansion States Skipping, Delaying, or Altering Prescription 41.0%* Drug Use Due to Cost, 40% Kentucky Compared to 34.7% 32.1% Neighboring States and 30.5% 29.7% 30.1% 30% 27.8% U.S. Rate, 2014 26.4% 27.4% 27.5% (ages 19-64) 20% 10% 0% KY U.S. AR IL IN OH WV MO TN VA *Difference is statistically significant across states (e.g., Kentucky vs. Arkansas) at the 95% level. Source: SHADAC analysis of the civilian non-institutional population in the 2014 NHIS using the SHADAC Data Center. Note: While Indiana is a Medicaid expansion state, the state did not expand its Medicaid program until 2015. 7 7 | STATE HEALTH ACCESS DATA ASSISTANCE CENTER SEMI-ANNUAL REPORT Study of the Impact of the ACA Implementation in Kentucky DOMAIN #2: ACCESS 50% FIGURE 2.6: 2012 2014 Skipping, Delaying, or 41.0%* Altering Prescription 40% 35.8% Drug Use Due to Cost, 34.1% 34.7% Kentucky Compared to 30.5%* 30.0% 29.7% 30.1% Neighboring States and 30% 27.4% 27.5% 31.9% 32.1% U.S. Rate, 2012-2014 30.0% 27.8% 27.9% (ages 19-64) 26.4% 25.4% 24.3% 24.9% 20% 22.1% 10% 0% KY U.S. AR IL IN OH WV MO TN VA *Difference is statistically significant within the state (e.g., Arkansas 2012 estimate vs. Arkansas 2014 estimate) at the 95% level. Source: SHADAC analysis of the civilian non-institutional population in the 2012 and 2014 NHIS using the SHADAC Data Center. 100% 2012 2014 FIGURE 2.7: 91.4% 91.8% Provider Visit in Past 85.8% Year by Age Category, 83.6% 80% 75.6% Kentucky, 2012-2014 73.8% 68.9% 65.6% 60% 40% 20% 0% All 0-18 19-64 65+ *Difference is statistically significant at the 95% level. Source: SHADAC analysis of the civilian non-institutional population in the 2012 and 2014 NHIS using the SHADAC Data Center. to a general provider in the 12 months preceding Kentuckians Using Emergency Department the survey. Results show that these levels held Stayed Stable in Past Year steady with no significant change from 2012-2014. We also examine the prevalence of visits to an As was the case at baseline, in 2014, children emergency department (ED) within the past year. and elderly adults continued to report the high- According to the Agency for Healthcare Research est percentages of seeing a provider, 83.6% and and Quality (AHRQ), “ED utilization reflects the 91.8%, respectively, compared to non-elderly greater health needs of the surrounding commu- adults (68.9%). Figure 2.7 presents the data by age nity and may provide the only readily available category. care for individuals who cannot obtain care else- where. Many ED visits are ‘resource sensitive’ and SEPTEMBER 2016 | WWW.SHADAC.ORG 8 8 potentially preventable, meaning that access to (25.4%), and children (19.7%). As with the other high-quality, community-based health care can access indicators, it will be important to follow ED prevent the need for a portion of ED visits.”10 use as future data become available to determine whether statistically significant changes emerge. Figure 2.8 presents the 2012-2014 ED visit data For this measure, we also present comparisons by age. While the estimates for ED use for Ken- between Kentucky and neighboring states (Fig- tuckians overall and for each age group were ures 2.9 and 2.10). In 2014, although Kentucky’s lower in 2014, this did not represent a statistically ED use was lower than 2012, it was still significant- significant change since the 2012 baseline. Elderly ly higher than the U.S. average (18.2%) and that Kentuckians continued to report the highest visits of two neighboring states (Missouri and Virginia). to an ED, at 32.0%, followed by non-elderly adults None of the comparison states had ED use that FIGURE 2.8: 40% 2012 2014 Emergency Department 35.3% Visits in the Past Year by 32.0% Age Category, Kentucky, 30.4% 30.2% 30% 2012-2014 27.8% 24.9% 25.4% 20% 19.7% 10% 0% All 0-18 19-64 65+ *Difference is statistically significant at the 95% level. Source: SHADAC analysis of the civilian non-institutional population in the 2012 and 2014 NHIS using the SHADAC Data Center. FIGURE 2.9: 30% Kentucky United States Medicaid Expansion States Non-expansion States Emergency Department Visits in the Past Year, 24.9% 25.2% Kentucky Compared to 23.7% 23.4% Neighboring States and 21.1% 20.9% 20% 19.8% U.S. Rate, 2014 (all ages) 18.2%* 18.7%* 17.8%* 10% 0% KY U.S. AR IL IN OH WV MO TN VA *Difference is statistically significant across states (e.g., Kentucky vs. Arkansas) at the 95% level. Source: SHADAC analysis of the civilian non-institutional population in the 2014 NHIS using the SHADAC Data Center. Note: While Indiana is a Medicaid expansion state, the state did not expand its Medicaid program until 2015. 9 9 | STATE HEALTH ACCESS DATA ASSISTANCE CENTER SEMI-ANNUAL REPORT Study of the Impact of the ACA Implementation in Kentucky DOMAIN #2: ACCESS 40% 2012 2014 FIGURE 2.10: Emergency Department Visits in the Past Year, 30.4% Kentucky Compared to 30% 26.5% Neighboring States and 23.7%* 24.3% U.S. Rate, 2012-2014 22.5% 21.5% 24.9% 19.8% 25.2% 20.9% 19.6% (all ages) 19.0% 23.4% 20% 21.1% 20.0% 18.2%* 18.7% 18.7% 17.6% 17.8% 10% 0% KY U.S. AR IL IN OH WV MO TN VA *Difference is statistically significant within the state (e.g., Arkansas 2012 estimate vs. Arkansas 2014 estimate) at the 95% level. Source: SHADAC analysis of the civilian non-institutional population in the 2012 and 2014 NHIS using the SHADAC Data Center. was significantly higher than Kentucky (six states More Than 9 in 10 Kentuckians Able to Find a had rates statistically indistinguishable from Doctor When Needed Kentucky). Being able to find a doctor when needed is an important component of health access. In 2014, Figure 2.10 shows how Kentucky’s ED use has 96.2% of Kentuckians surveyed said that they changed between 2012 and 2014 compared to were able to find a doctor when needed, and this changes in ED use in neighboring states. Like Ken- was not statistically different from 2012. Figure tucky, most states did not experience a statistically 2.11 shows provider availability across age cat- significant change in ED use; only the U.S. experi- egories; there were no significant changes from enced a significant decline in ED use, while Illinois 2012-2014. saw a significant increase in ED use. 2012 2014 FIGURE 2.11: 100% 96.2% 98.0% 98.4% 97.0% 97.4% 94.8% 93.1% 95.1% Found Doctor When Needed by Age 80% Category, Kentucky, 2012-2014 60% 40% 20% 0% All 0-18 19-64 65+ *Difference is statistically significant at the 95% level. Source: SHADAC analysis of the civilian non-institutional population in the 2012 and 2014 NHIS using the SHADAC Data Center. SEPTEMBER 2016 | WWW.SHADAC.ORG 1010 In addition to finding a doctor, some people also Kentucky of individuals reporting that providers face barriers with providers not accepting their accepted their insurance: Arkansas (98.6%) and insurance coverage. From 2012-2014, there was Missouri (98.7%). The remaining neighboring no significant change in Kentucky for the rate of states and the U.S. had rates that were not sta- people reporting that providers would accept their tistically different from Kentucky’s. From 2012 to coverage, with 96.4% of Kentuckians reporting 2014, only West Virginia and the U.S. had signif- acceptance in 2014 compared to 96.5% in 2012 icant increases in the rate of patients reporting (Figure 2.12). Figure 2.13 presents this informa- acceptance of their insurance coverage (Figure tion for neighboring states and the U.S. for all ages. 2.14). No states experienced significant declines. Two states had significantly higher rates than FIGURE 2.12: 2012 2014 Told Provider Accepts 98.5% 98.1% 100% 96.5% 96.4% 95.0% 97.3% 96.7% 95.6% Insurance by Age Category, Kentucky, 80% 2012-2014 60% 40% 20% 0% All 0-18 19-64 65+ *Difference is statistically significant at the 95% level. Source: SHADAC analysis of the civilian non-institutional population in the 2012 and 2014 NHIS using the SHADAC Data Center. FIGURE 2.13: 100% Kentucky United States Medicaid Expansion States Non-expansion States Told Provider Accepts Insurance, Kentucky 98.6%* 98.7%* 98.5% Compared to 98% 97.9% 97.5% 97.5% 97.6% Neighboring States and U.S. Rate, 2014 (all ages) 96.4% 96.4% 96.5% 96% 94% 92% 90% KY U.S. AR IL IN OH WV MO TN VA *Difference is statistically significant across states (e.g., Kentucky vs. Arkansas) at the 95% level. Source: SHADAC analysis of the civilian non-institutional population in the 2014 NHIS using the SHADAC Data Center. Note: While Indiana is a Medicaid expansion state, the state did not expand its Medicaid program until 2015. 1111 | STATE HEALTH ACCESS DATA ASSISTANCE CENTER SEMI-ANNUAL REPORT Study of the Impact of the ACA Implementation in Kentucky DOMAIN #2: ACCESS 100% FIGURE 2.14: 2012 2014 98.6% 98.7% Told Provider Accepts 98.5%* 97.9% Insurance, Kentucky 97.5%* 97.5% 97.6% 98% Compared to 98.2% 97.0% 96.5% 97.8% 96.4% Neighboring States and 97.4% 97.0% 96.9% 97.0% 97.1% U.S. Rate, 2012-2014 96% 96.5% 96.4% (all ages) 95.9% 94% 92% 90% KY U.S. AR IL IN OH WV MO TN VA *Difference is statistically significant within the state (e.g., Arkansas 2012 estimate vs. Arkansas 2014 estimate) at the 95% level. Source: SHADAC analysis of the civilian non-institutional population in the 2012 and 2014 NHIS using the SHADAC Data Center. SEPTEMBER 2016 | WWW.SHADAC.ORG 1212 DOMAIN #3: COST Health care costs are a topic of concern for many Because nearly half of Kentucky private-sector families, with out-of-pocket costs—including pre- employers offer health insurance, health care miums, co-pays, co-insurance, and deductibles— costs also are a concern for employers. We in- often varying substantially by benefit plan. This is clude five metrics related to cost in this report, especially important in Kentucky, where the most focusing primarily on individual or household recent estimates (2014) show that 19.1% of the spending. We also include a measure of premi- population is living below the poverty level (less ums for employer-sponsored insurance and one than 100% FPG), compared to 15.5% nationwide. health system indicator — a measure of hospital Kentucky’s poverty rate is the 5th highest in the uncompensated care that is aggregated across all country (see Figure 3.1 for additional detail on reporting hospitals in Kentucky. Data sources Federal Poverty Guidelines [FPG]).11,12 for the cost measures include the NHIS and the MEPS-IC. Our uncompensated care measure was FIGURE 3.1: Single Family obtained from the Kentucky Cabinet for Health Federal Poverty Person of Four and Family Services.13 The majority of the esti- Guidelines, 2014 100% FPG $11,670 $23,850 mates in the cost domain include all ages. 138% FPG $16,105 $32,913 Most of these indicators are only available 200% FPG $23,340 $47,700 through 2014. As new data become available, we Source: Office of the Assistant Secretary for Planning and Evaluation, U.S. Department of Health and Human Services, 2014 Poverty Guidelines. will continue to track any changes in this domain Accessible at: https://aspe.hhs.gov/2014-poverty-guidelines. in subsequent annual and semi-annual reports. COST MEASURES Kentuckians Reporting Trouble Paying not find a significant change for any of the age Medical Bills Dropped Between 2012 & 2014 subgroups (children, non-elderly adults, elderly Overall, 39.1% of Kentuckians in 2014 reported adults). This finding comes from SHADAC analysis that their families had trouble paying medical bills of the NHIS, which asks, “In the past 12 months — a statistically significant drop of 10.0 percentage did [you/anyone in the family] have problems points from 2012 (Figure 3.2). However, we did paying or were unable to pay any medical bills? FIGURE 3.2: 60% 2012 2014 Trouble Paying Medical 52.3% 52.7% Bills by Age Category, 49.1% Kentucky, 2012-2014 43.3% 42.1% 40% 39.1%* 22.9% 20% 17.1% 0% All 0-18 19-64 65+ *Difference is statistically significant at the 95% level. Source: SHADAC analysis of the civilian non-institutional population in the 2012 and 2014 NHIS using the SHADAC Data Center. The estimate reports the percentage of people who had trouble paying off medical bills in the last year or were currently paying off medical bills. 1313 | STATE HEALTH ACCESS DATA ASSISTANCE CENTER SEMI-ANNUAL REPORT Study of the Impact of the ACA Implementation in Kentucky DOMAIN #3: COST Include bills for doctors, dentists, hospitals, Figure 3.4 shows how Kentucky’s performance therapists, medication, equipment, nursing home, on this indicator has changed between 2012 and or home care.” We also compared this metric to 2014 compared to changes in neighboring states. the U.S. rate and to Kentucky’s neighboring states Only Kentucky and the U.S. had a significant in Figure 3.3. Kentucky had a significantly higher decrease in the percent of people reporting that percentage of people reporting trouble paying their families had trouble paying medical bills. medical bills compared to the U.S. Kentucky’s Kentucky’s 10.0 percentage point drop was nearly rate also was significantly higher than Illinois, 4 times as large as the U.S. drop of 2.6 percentage Missouri, Tennessee, and Virginia. No state had a points, but more Kentuckians still reported trouble rate that was significantly higher than Kentucky’s. paying medical bills in 2014. 50% Kentucky United States Medicaid Expansion States Non-expansion States FIGURE 3.3: Trouble Paying Medical Bills, 40% 39.1% 39.4% Kentucky Compared to 36.2% 34.1% 34.7% Neighboring States and 30.2%* 31.6%* U.S. Rate, 2014 (all ages) 30% 28.7%* 29.0%* 29.4%* 20% 10% 0% KY U.S. AR IL IN OH WV MO TN VA *Difference is statistically significant across states (e.g., Kentucky vs. Arkansas) at the 95% level. Source: SHADAC analysis of the civilian non-institutional population in the 2014 NHIS using the SHADAC Data Center. The estimate reports the percentage of people who had trouble paying off medical bills in the last year or were currently paying off medical bills. Note: While Indiana is a Medicaid expansion state, the state did not expand its Medicaid program until 2015. 60% FIGURE 3.4: 2012 2014 Trouble Paying 49.1% Medical Bills, 50% Kentucky Compared to 41.2% 40.3% 40.0% Neighboring States and 38.5% 40% 35.5% 36.5% U.S. Rate, 2012-2014 34.7% 39.1%* 39.4% 31.3% 30.8% (all ages) 36.2% 34.7% 30% 34.1% 30.2% 31.6% 28.7%* 29.0% 29.4% 20% 10% 0% KY U.S. AR IL IN OH WV MO TN VA *Difference is statistically significant within the state (e.g., Arkansas 2012 estimate vs. Arkansas 2014 estimate) at the 95% level. Source: SHADAC analysis of the civilian non-institutional population in the 2012 and 2014 NHIS using the SHADAC Data Center. SEPTEMBER 2016 | WWW.SHADAC.ORG 1414 No Significant Change in Delaying or delaying needed care due to cost (Figures 3.5), Skipping Care Due to Cost but this was not statistically different from 2012. Delaying or going without needed medical care Figure 3.6 shows how the rate of those who can be a major impediment to good health delayed needed care due to cost in Kentucky outcomes. It may cause serious conditions to compares to the rates in neighboring states and go undetected or become worse by being left the U.S. in 2014. The U.S. rate (7.1%) and the rates untreated, resulting in poorer health status and of four neighboring states—Arkansas, Illinois, higher treatment costs. Cost is a reason frequently Missouri, and Ohio—were significantly lower than cited for delaying or going without medical care. in Kentucky. The rates of the remaining four states In 2014, 10.1% of Kentuckians of all ages reported were not statistically different. FIGURE 3.5: 20% 2012 2014 Delayed or Went Without Needed Care Due to Cost, Kentucky, 2012-2014 (all ages) 15% 11.7% 10.1% 10.0% 10% 8.3% 5% 0% Delayed care Went without care *Difference is statistically significant at the 95% level. Source: SHADAC analysis of the civilian non-institutional population in the 2012 and 2014 NHIS using the SHADAC Data Center. These estimates report the percentage of people who delayed seeking medical care because of worry about the cost and the percentage of people who needed medical care but did not get it because they could not afford it in the last year. FIGURE 3.6: 15% Kentucky United States Medicaid Expansion States Non-expansion States Delayed Needed Care Due to Cost, Kentucky Compared to Neighboring States and 10.1% U.S. Rate, 2014 (all ages) 10% 8.9% 8.0% 7.8% 7.7% 7.1%* 7.2%* 7.0%* 6.8%* 6.8%* 5% 0% KY U.S. AR IL IN OH WV MO TN VA *Difference is statistically significant across states (e.g., Kentucky vs. Arkansas) at the 95% level. Source: SHADAC analysis of the civilian non-institutional population in the 2014 NHIS using the SHADAC Data Center. 1515 | STATE HEALTH ACCESS DATA ASSISTANCE CENTER SEMI-ANNUAL REPORT Study of the Impact of the ACA Implementation in Kentucky DOMAIN #3: COST Figure 3.7 shows how the rate of those who de- with delayed care, this was not a significant layed needed care due to cost changed between change from the 2012 baseline. Kentucky’s rate of 2012 and 2014 compared to changes in neighbor- forgone care was significantly higher than the U.S. ing states and the U.S. Only the U.S. and Missouri rate and the rates for four neighboring states; it experienced statistically significant declines from was not significantly different than Indiana, Virgin- 2012 to 2014. ia or West Virginia (Figure 3.8; 2014 data were not available for Tennessee). Figure 3.5 also displays results for going without needed care due to cost. In 2014, 8.3% of all Ken- tuckians reported going without needed care. As 15% FIGURE 3.7: 2012 2014 Delayed Needed Care Due to Cost, 11.7% 11.2% Kentucky Compared to Neighboring States and 9.5% 10% 8.7% 8.9% 8.9% 8.7% U.S. Rate, 2012-2014 10.1% 8.4% 7.7% 7.7% (all ages) 8.9% 8.0% 7.8% 7.1%* 7.2% 7.0% 7.1% 6.8% 6.8%* 5% 0% KY U.S. AR IL IN OH WV MO TN VA *Difference is statistically significant within the state (e.g., Arkansas 2012 estimate vs. Arkansas 2014 estimate) at the 95% level. Source: SHADAC analysis of the civilian non-institutional population in the 2012 and 2014 NHIS using the SHADAC Data Center. 10% Kentucky United States Medicaid Expansion States Non-expansion States FIGURE 3.8: Went Without Needed 8.3% Care Due to Cost, 7.4% Kentucky Compared to 7.0% Neighboring States and 6.1%* 6.1% U.S. Rate, 2014 (all ages) 5.3%* 5.1%* 5% 4.8%* 4.0%* 0% KY U.S. AR IL IN OH WV MO TN^ VA *Difference is statistically significant across states (e.g., Kentucky vs. Arkansas) at the 95% level. Source: SHADAC analysis of the civilian non-institutional population in the 2014 NHIS using the SHADAC Data Center. Note: While Indiana is a Medicaid expansion state, the state did not expand its Medicaid program until 2015. ^Data were not available from Tennessee for this indicator. SEPTEMBER 2016 | WWW.SHADAC.ORG 1616 FIGURE 3.9: 15% 2012 2014 Went Without Needed Care Due to Cost, Kentucky Compared to Neighboring States and 10.0% U.S. Rate, 2012-2014 10% (all ages) 7.4% 7.4% 7.7% 7.2% 7.0% 8.3% 6.7% 6.2% 5.2% 6.8% 5% 6.1% 6.1% 5.3%* 5.1% 5.3% 4.8%* 4.0%* 0% KY U.S. AR IL IN OH WV MO TN^ VA *Difference is statistically significant within the state (e.g., Arkansas 2012 estimate vs. Arkansas 2014 estimate) at the 95% level. ^Data were not available from Tennessee for this indicator. Source: SHADAC analysis of the civilian non-institutional population in the 2012 and 2014 NHIS using the SHADAC Data Center. Figure 3.9 shows how the rate of those who went After a slight increase from 2012 to 2013, char- without needed care due to cost changed between ity care and self-pay charges dropped substan- 2012 and 2014 compared to changes in neighbor- tially in 2014—the same year as Kentucky’s im- ing states and the U.S. Only the U.S. rate and those plementation of Medicaid expansion and kynect. for Missouri and Ohio decreased significantly. Between our baseline year of 2012 and 2015, Hospital Charity Care and Self-Pay Charges these uncompensated charges dropped 76.9%, Declined 77% from nearly $2.4 billion to $552 million. Due to the Figure 3.10 presents data on hospital charity care nature of these data, statistical significance testing and other care to uninsured Kentuckians (i.e., self- comparing Kentucky rates over time was not per- pay charges), which we use as a proxy for uncom- formed for this measure. pensated care. It is important to note that these Employer Coverage Premiums Increased for data do not include bad debt from people with Single, Statistically Stable for Family Coverage insurance, such as if a person with coverage does Figure 3.11 provides estimates of spending on not pay cost sharing (e.g., deductible) owed to health insurance premiums. In 2015, the aver- the hospital. Across the country, states that have age single premium for private-sector employer- expanded Medicaid have experienced larger de- sponsored insurance was $5,984, a statistically creases in charity care compared to states that significant increase of $587 since 2012. However, have not expanded Medicaid.14 This may be due in the average family premium for employer-based part to expansion states having a greater number coverage ($16,622) was not significantly different of people who now have more medical charges from 2012. It is important to note that, like with that are covered by insurance. long-term trends finding a decline in employers offering health insurance, premiums also have been increasing over the longer-term prior to the ACA.3 1717 | STATE HEALTH ACCESS DATA ASSISTANCE CENTER SEMI-ANNUAL REPORT Study of the Impact of the ACA Implementation in Kentucky DOMAIN #3: COST $3,000 FIGURE 3.10: Hospital Charity Care $2,570 $2,396 and Self-Pay Charges $2,500 in Dollars (millions), Kentucky, 2012-2015 $2,000 $1,500 $1,000 $942 $552 $500 $0 2012 2013 2014 2015 Source: SHADAC analysis of 2012 to 2015 data from the Kentucky Cabinet for Health and Family Services’ Kentucky Outpatient Hospital Administrative Claims Data. $20,000 2012 2015 FIGURE 3.11: Average Premium per $16,622 Private Sector Employee $15,734 in Dollars, Kentucky, $15,000 2012-2015 $10,000 $5,984* $5,397 $5,000 $0 Family Premiums Single Premiums *Difference is statistically significant at the 95% level. Source: 2012 and 2015 MEPS-IC. These estimates represent the total annual premium cost. SEPTEMBER 2016 | WWW.SHADAC.ORG 1818 DOMAIN #4: QUALITY Achieving improvements in the quality of health cents). For the quality domain, our data sources care was a key goal of the ACA. There are a number for this update include the Agency for Healthcare of ways in which the law is focused on improving Research and Quality’s (AHRQ) Healthcare Cost the quality of care, including avoiding preventable and Utilization Project (HCUP) and the Youth Risk hospital readmissions, increasing the utilization of Behavior Surveillance System (YRBSS), which is a preventative care, and encouraging recommend- survey conducted in partnership between states ed health practices, such as breastfeeding for in- and the U.S. Centers for Disease Control and fants. For this data update, we include five metrics Prevention. Data in this domain are updated that relate to quality of care, with a focus on po- through 2014 for HCUP data and 2015 for YRBSS tentially preventable hospital admissions, hospital data. Generally, the indicators in this domain did mortality, and behaviors that may impact health not show much overall change between baseline outcomes (e.g. unprotected sex among adoles- levels and the most recent data years. QUALITY MEASURES Potentially Preventable Hospital Admissions Figure 4.1 presents data on potentially prevent- According to AHRQ, “one area where higher qual- able hospitalizations as the number of hospitaliza- ity and lower costs coincide is potentially prevent- tions per 100,000 adults. For diabetes short-term able hospital admissions—inpatient stays that complications, approximately 93 out of 100,000 could be prevented with high-quality primary and adults were admitted in 2014, an increase over preventive care. High rates of these potentially 84 in 2012. In contrast, both hypertension- and preventable hospital admissions identify areas asthma-related admissions decreased over the where possible improvements in the health care same period, with approximately 68 per 100,000 delivery system could be made to enhance patient in 2012 and 58 per 100,000 in 2014 for hyper- outcomes and decrease costs.”15 In this study, we tension, and 58 per 100,000 in 2012 and 44 per look at potentially avoidable hospitalizations for 100,000 in 2014 for asthma. three chronic conditions: diabetes, hypertension, and asthma. The data for these come from AHRQ’s HCUP dataset.16 FIGURE 4.1: 100 2012 2014 Diabetes (ages 18+), 92.8 90 Hypertension (ages 18+) 84.3 and Asthma (ages 18-39) 80 Hospital Admissions 70 67.9 (per 100,000), Kentucky, 60 58.4 57.9 2012-2014 50 44.0 40 30 20 10 0 Diabetes short-term admissions Hypertension admissions Asthma admissions Source: SHADAC analysis of 2012 and 2014 HCUP data. These estimates report the Diabetes Short-term Complications Admission Rate for adults (PQI 1), the Hypertension Admission rate for adults (PQI 7), and the Asthma in Younger Adults Admission Rate (PQI 15). 1919 | STATE HEALTH ACCESS DATA ASSISTANCE CENTER SEMI-ANNUAL REPORT Study of the Impact of the ACA Implementation in Kentucky DOMAIN #4: QUALITY Death Rate in Low Mortality Admissions Stable ures) are associated with low mortality rates and Figure 4.2 shows the number of deaths per 1,000 are used as one indicator of hospital quality; hos- patients of all ages who were hospitalized for pitals with high mortality rates associated with conditions that typically do not result in mortal- these low-mortality DRGs may provide lower ity. All cases treated in hospitals are classified quality care.18 The mortality rate presented here according to groups called diagnosis-related is risk-adjusted to take into account patients’ pri- groups (DRGs). DRGs are used to help determine or health status. Figure 4.2 shows that in 2014, how much a hospital gets paid for its services, Kentucky’s mortality rate for “low-mortality DRGs” adjusted for severity and other factors.17 Many was 0.325 per 1,000, slightly lower than the state’s DRGs (e.g., eye disorders, childbirth, knee proced- 2012 baseline rate of 0.330 per 1,000. 0.50 FIGURE 4.2: 2012 2014 Mortality Rate in Low Mortality DRGs (per 0.40 1,000 cases), Kentucky, 2012-2014 (all ages) 0.330 0.325 0.30 0.20 0.10 0.0 2012 2014 Source: SHADAC analysis of 2012 and 2014 HCUP data. The estimate reports the “Dying in the Hospital while Getting Care for a Condition that Rarely Results in Death Rate” cases. Unprotected Sex Among High School Students Figure 4.3 provides a snapshot of the 2013 base- Remains Statistically Unchanged line data and updated 2015 data for Kentucky. The 2015 YRBSS provides estimates of unprotect- Among high school students, 14.5% reported ed sex (i.e., no use of any birth control) among engaging in unprotected sex during their last high school students who reported that they were sexual intercourse in 2015, which was not statis- sexually active. This indicator was identified by the tically different from 2013. Female high school Foundation as an important part of the study’s students reported higher rates of unprotected population health and prevention measures. sex (17.5%) compared to males (11.6%) in 2015, Although the ACA includes certain provisions although neither of these were statistically differ- designed to increase access to contraception— ent than in 2013. such as requiring private health insurance plans to cover birth control prescribed by a health care provider with no cost-sharing—there are many factors that influence adolescents’ use of con- traception,19 so the law is not expected to have a strong effect on use of birth control by high school students. SEPTEMBER 2016 | WWW.SHADAC.ORG 2020 FIGURE 4.3: 20% 2013 2015 Unprotected Sex Among 18.6% 17.5% High School Students, Kentucky, 2013-2015 15.1% 15% 14.5% (grades 9-12) 11.2% 11.6% 10% 5% 0% All Female Male *Difference is statistically significant at the 95% level. Source: SHADAC analysis of 2013 and 2015 Youth Risk Behavior Surveillance System data. The estimate reports the percentage of high school students who did not use any method to prevent pregnancy during their last sexual intercourse. DOMAIN #5: HEALTH OUTCOMES Health outcomes are determined by a combi- Overall, the health outcomes measures were rel- nation of factors including genetics, behaviors, atively stable between 2012 and 2014. Kentucky environmental exposures, social factors, and continued to perform less favorably for these health care services and policies.20 Although the measures when compared to the U.S. overall. determinants are complex, the outcome mea- “ sures included in this report are at least partially influenced by access to high-quality care. While The health outcomes health outcomes are slow to change at a state or measures were relatively national level, monitoring them is key to under- standing the impacts of efforts to improve health in Kentucky. In this report, we provide updates to stable between “ two measures of health outcomes: self-reported 2012 and 2014. obesity and cigarette use. These measures are based on data from the Behavioral Risk Factor Sur- veillance System (BRFSS) and YRBSS. HEALTH OUTCOMES MEASURES Obesity Rate Remains Stable for Kentucky’s Nearly one in five high school students (18.5%) High School Students in Kentucky indicated they were obese in 2015, Obesity is associated with a range of chronic con- which was statistically unchanged from 2013. Al- ditions, including heart disease, high blood pres- though Kentucky’s adolescent obesity rates were sure, and diabetes.21 Obesity is prevalent among higher than the U.S., their statistical stability was adults and children in the U.S., though rates consistent with the U.S. While this measure could among children have stabilized in recent years.22 potentially be affected by improved health insur- Figure 5.1 shows estimates of the prevalence of ance coverage and access, it would likely take sub- self-reported obesity among high school students stantial time for this change to occur. in Kentucky from 2013-2015. 2121 | STATE HEALTH ACCESS DATA ASSISTANCE CENTER SEMI-ANNUAL REPORT Study of the Impact of the ACA Implementation in Kentucky DOMAIN #5: HEALTH OUTCOMES 20% 2013 2015 FIGURE 5.1: 18.5% Self-Reported Obesity, 18.0% Kentucky and U.S., 2013-2015 (high school 15% 13.7% 13.9% students, grades 9-12) 10% 5% 0% Kentucky U.S. *Difference is statistically significant at the 95% level. Source: SHADAC analysis of 2013 and 2015 Youth Risk Behavior Surveillance System data. Cigarette Use Declines Significantly Among benefits, with no cost-sharing.23 Since the most- Adults; Stable for High School Students recent study report, SHADAC has added two new The ACA incorporated certain policies to discour- indicators to measure tobacco use: Figures 5.2 age tobacco use and to provide people resources and 5.3 show estimates of the prevalence of cig- to quit. For example, the law allows insurers to arette use among high school students and adults charge higher premiums to people who use tobac- in both Kentucky and the U.S. Overall, Kentucky’s co, and the law also requires that private health rates of cigarette use among high school students insurance cover certain recommended preventive and adults remained higher than the U.S. rates. health care services, including tobacco-cessation 20% 2013 2015 FIGURE 5.2: 17.9% Cigarette Use, Kentucky 16.9% and U.S., 2013-2015, 15.7% (high school students, 15% grades 9-12) 10.8%* 10% 5% 0% Kentucky U.S. *Difference is statistically significant at the 95% level. Source: SHADAC analysis of 2013 and 2015 Youth Risk Behavior Surveillance System data. The estimate reports the percentage of high school students who currently smoked cigarettes, on at least 1 day during the 30 days before the survey. SEPTEMBER 2016 | WWW.SHADAC.ORG 2222 While U.S. rates of cigarette use among high use among adult Kentuckians decreased signifi- school students decreased significantly from cantly from 28.3% in 2012 to 26.2% in 2014; this is 15.7% in 2013 to 10.8% in 2015, Kentucky’s rates consistent with the significant decline seen in the were higher than the U.S. average and remained U.S. rate of cigarette use for adults (18.1% in 2012 statistically unchanged, at 16.9% in 2015 (Figure and 16.8% in 2014). 5.2). In contrast, Figure 5.3 shows that cigarette FIGURE 5.3: 30% 28.3% 2012 2014 Cigarette Use, Kentucky 26.2%* and U.S., 2012-2014, 25% (adults 18+) 20% 18.1% 16.8%* 15% 10% 5% 0% Kentucky U.S. *Difference is statistically significant at the 95% level. Source: SHADAC analysis of 2012 and 2014 Behavioral Risk Factor Surveillance System data. The estimate reports the percentage of adults who have smoked 100 or more cigarettes in their lifetime and who currently smoke some days or every day. 2323 | STATE HEALTH ACCESS DATA ASSISTANCE CENTER SEMI-ANNUAL REPORT Study of the Impact of the ACA Implementation in Kentucky III. STUDY FINDINGS: KENTUCKY HEALTH REFORM SURVEY As part of the Study of the Impact of Implementa- We analyze these outcomes using a set of vari- tion of the Affordable Care Act (ACA) in Kentucky, ables to explore differences for some sub-groups: we conducted a one-time survey of non-elderly coverage status and type, sex, race, education, adults in the Commonwealth to assess a set of household income, marital status, language spo- indicators post-implementation of the ACA in Ken- ken at home, health status, and geographical re- tucky. The Kentucky Health Reform Survey (K-HRS) gion. However, the discussion presented below addressed several key domains, including cov- is restricted to the variables that were associated erage status; experiences navigating Kentucky’s with the outcome; those that showed statistically state-based marketplace, kynect; affordability significant differences with respect to the overall of health care; access to health care; and health estimate. status. The K-HRS was a dual-framed landline and For ease of presentation, although the K-HRS find- cell phone survey implemented between March ings are limited to non-elderly adults, we do not 31 and May 3, 2016. The survey employed the specify this on each reference and instead refer to methodology used by the annual Kentucky Health “Kentuckians.” Issues Poll (KHIP) and included many of the same “ questions, which will allow us to compare K-HRS findings to findings from this annual Kentucky sur- The SHADAC 2016 vey and delve deeper into several policy-relevant areas. K-HRS found that only While future reports will present additional in- 8.9% of non-elderly depth analyses, including trend analyses of K-HRS and KHIP findings, this report provides a discus- adult Kentuckians were sion around the several timely findings that re- uninsured, a historic late to Kentucky’s ongoing health care reforms: uninsurance, consumer experience navigating the marketplace, concern about losing coverage, for- low for uninsurance in “ gone or delayed care due to cost, dental coverage the Commonwealth...... and care, and emergency department use. KENTUCKY HEALTH REFORM SURVEY PRELIMINARY FINDINGS Uninsurance Rate Falls to Historic Low of 8.9% through the individual market), public health Since implementation of the ACA’s main coverage insurance programs, and the uninsured.27 The expansions in 2014, Kentucky has experienced largest source of coverage in 2016 was group one of the largest gains in health insurance cov- coverage, with 44% of Kentuckians reporting erage in the nation. The KHIP found that between health insurance through their own or a family 2013 and 2014, the uninsurance rate among non- member’s employer. The second most common elderly adults in Kentucky fell from 25% to 12%,24 source of coverage in Kentucky was public health and it remained statistically stable in 2015.25 insurance programs (e.g., Medicaid, Medicare for The SHADAC 2016 K-HRS found that only 8.9% the disabled, military coverage), which was report- of non-elderly adult Kentuckians were unin- ed by 41% of Kentuckians. Only 4% of Kentuckians sured, a historic low for uninsurance in the purchased non-group individual or family plans, Commonwealth.26 either through kynect or the off-exchange market. Figure 6.1 shows the distribution of Kentucky’s non-elderly adult population by coverage type: group coverage (i.e., employer-based insurance), non-group coverage (i.e., coverage purchased SEPTEMBER 2016 | WWW.SHADAC.ORG 2424 FIGURE 6.1: 50% Sources of Health 44.2% Insurance Coverage, 41.0% 2016 40% 30% 20% 10% 8.9% 3.9% 2.1% 0% Group Non-group Public Uninsured Other Private Source: 2016 Kentucky Health Reform Survey. Note: The category of "other" represents people who reported having health insurance but didn't specify the type of coverage. The analyses presented in the rest of the report Individual characteristics associated with a re- combine the group and non-group coverage duced likelihood of being uninsured include being categories into a larger “private coverage” catego- female, being married, having attended or gradu- ry. Two exceptions are made for coverage through ated from college, speaking only English at home, kynect and ED visits, where non-group coverage and having a family income of 201% of the Federal will be reported independently due to the rel- Poverty Guideline (FPG) or more. As presented in evance of the analysis and policy issues.28 We Figure 6.2, the only characteristic associated with exclude the “other” coverage type due to its small a higher uninsurance rate was having less than a sample size and limited policy relevance. high school degree (25.5%). FIGURE 6.2: 30% Uninsurance Rate by Education, 2016 25.5%* 20% 11.0% 10% 8.9% 4.1%* 1.0%* 0% Total Less Than High School Some College College High School Graduate Graduate *Difference is statistically significant at the 95% level between the specific estimate and the total estimate for the overall non-elderly population. Source: 2016 Kentucky Health Reform Survey. 2525 | STATE HEALTH ACCESS DATA ASSISTANCE CENTER SEMI-ANNUAL REPORT Study of the Impact of the ACA Implementation in Kentucky K-HRS PRELIMINARY FINDINGS FIGURE 6.3: Reason Percent Main Reason Cited for Being Uninsured, 2016 Too expensive/ could not afford 56.2% Do not need health insurance 10.0% Unemployed/ not working 8.7% Will get insurance soon 6.9% Don’t want government involved in their health care 5.3% Not eligible/ health condition 5.3% Spiritual/ natural healing 2.4% Do not know how 2.3% Never looked into it 2.0% Other 1.0% Source: 2016 Kentucky Health Reform Survey. Cost is Main Reason for Being Uninsured were more likely to have private coverage than Figure 6.3 shows the reasons that the uninsured the average Kentuckian. Figure 6.4 shows that gave for being uninsured. Over half (56%) of those residents of Northern Kentucky and the Lexington who were uninsured reported unaffordability of Area were significantly more likely to have private health insurance as the main reason to remain health insurance coverage, while residents of East- uninsured. Almost one in five uninsured Kentuck- ern Kentucky were significantly less likely to have ians reported reasons that suggest they were not private coverage. interested in obtaining coverage: 10% reported not needing health insurance, 5% did not want Public Health Insurance Coverage Plays a the government involved in their healthcare, and Large Role in Kentucky 2% preferred spiritual and natural healing. A quar- Public health insurance programs cover 41.0% ter of the uninsured reported reasons indicating of non-elderly adults in Kentucky. As presented they are not opposed to obtaining coverage: 9% on the following page in Figure 6.5, non-whites reported being unemployed or not working as the were more likely to be enrolled in public pro- reason for being uninsured, 7% had already initi- grams, as well as individuals with incomes below ated an application (i.e., will get insurance soon), 138% of the FPG—those newly eligible for Med- 5% believed their health condition made them icaid since implementation of the ACA Medicaid ineligible or they were not eligible to get public expansion in 2014—whereas those with higher programs or employer-based insurance, and 4% incomes (201+% FPG) were significantly less likely did not know how to apply for health insurance (17.9%).29 To see the dollar values associated with coverage or had not looked into it. these FPG percentages, please see Figure 3.1 in Section II. Almost Half of Non-Elderly Adults Get Health Those more likely to be covered by public pro- Insurance through the Private Sector grams include: individuals who had not graduated Almost half (48.1%) of non-elderly adults in Ken- from high school or who only had a high school tucky were covered by either health insurance diploma, who were unemployed or full-time stu- through their (or a relative’s) employer or health dents, were divorced or widowed, who reported insurance they purchased in the individual market. having fair or poor health, or who lived in Eastern Kentuckians who were white, married, employed, Kentucky. or who had incomes of 201+% of the FPG, had attended some college or had a college degree, or reported having very good or excellent health SEPTEMBER 2016 | WWW.SHADAC.ORG 2626 FIGURE 6.4: 70% Private Coverage by 61.5%* Geographic Region, 60% 56.1%* 2016 50% 48.1% 47.6% 48.8% 40% 35.5%* 30% 20% 10% 0% Total Western KY Louisville Northern KY Lexington Eastern KY Area Area *Difference is statistically significant at the 95% level between the specific estimate and the total estimate for the overall non-elderly population. Source: 2016 Kentucky Health Reform Survey FIGURE 6.5: 80% Public Coverage by Race 70.9%* and Income, 2016 70% 60% 53.6%* 50% 46.7% 41.0% 39.3%* 40% 30% 20% 17.9%* 10% 0% Total White Non-white <138% FPG 138-200% 201+% FPG FPG *Difference is statistically significant at the 95% level between the specific estimate and the total estimate for the overall non-elderly population. Source: 2016 Kentucky Health Reform Survey 1 in 4 Insured Kentuckians Found Their Current Under the ACA, states were given the option of cre- Health Insurance Through kynect ating their own state-based marketplaces (SBMs), One of the key provisions of the ACA was the cre- developing a partnership with the federal govern- ation of health insurance marketplaces. These ment, or allowing the federal government to run marketplaces provide a place for people to shop a marketplace for them. In 2014 Kentucky imple- for private individual market health (and dental) mented an SBM, called kynect. Since his election insurance and obtain income-based financial as- in November 2015, newly elected Governor Matt sistance. Additionally, marketplaces can also serve Bevin announced plans to transition from the Com- as a place where people apply for public coverage, monwealth’s state-based marketplace, kynect, to such as Medicaid. 2727 | STATE HEALTH ACCESS DATA ASSISTANCE CENTER SEMI-ANNUAL REPORT Study of the Impact of the ACA Implementation in Kentucky K-HRS PRELIMINARY FINDINGS a model using the federal Healthcare.gov health college graduates were less likely to use kynect. insurance marketplace platform.30 Non-whites also showed a higher rate of using kynect’s resources. In 2016, one third of the non-elderly, adult “ population in Kentucky had shopped for cover- age through kynect since it was created. Among Among those who individuals enrolled in public programs, 60.4% looked for coverage through kynect, whereas only looked for coverage 7.3% of those with employer-based insurance through kynect, 71% did so. Although no statistically significant dif- ference was found with respect to the total esti- considered the overall mate, 38.7% of uninsured Kentuckians and 39.7% of those with non-group coverage used kynect experience to be “ to shop for health insurance coverage. Figure excellent or good. 6.6 shows that while individuals with no college education had a higher rate of using kynect, 60% FIGURE 6.6: 52.5%* Use of kynect by Education and Race, 50% 46.7%* 2016 40% 38.3%* 32.9% 30.8%* 30.7% 30% 20% 14.2%* 10% 0% Total White Non-white Less Than High School Some College High School Graduate College Graduate *Difference is statistically significant at the 95% level between the specific estimate and the total estimate for the overall non-elderly population. Source: 2016 Kentucky Health Reform Survey Among those who looked for coverage through Finally, among those who shopped for health kynect, 71% considered the overall experience insurance coverage through kynect, in-person to be excellent or good. However, not everyone assistance was the most used resource (52%) who shopped through kynect obtained coverage while the call center and the website were used through the marketplace. In 2016, among those only by 42% and 38%, respectively.31 As presented insured, 25.8% reported they obtained their health in Figure 6.8, the choice of assistance resource was insurance coverage through kynect. As presented associated with income. Low income individuals in Figure 6.7, the marketplace was an important (<138% FPG) had a higher rate of use of in-person gateway to public health insurance programs; assistance, whereas they were less likely to use the over half of those with public coverage found their website than the average Kentuckian. In contrast, coverage through kynect (54.3%). Consistent with higher-income individuals (201+% FPG) used the this result, Figure 6.7 also shows that those eligible website more often and in-person assistance less for public programs due to their income (<138% regularly. FPG) were more likely to obtain health insurance coverage through this portal (56.6%). SEPTEMBER 2016 | WWW.SHADAC.ORG 2828 FIGURE 6.7: 60% 56.6%* Coverage Through 54.3%* kynect by Insurance 50% Type and Income, 2016 40% 32.1% 30% 29.2% 25.8% 20% 10% 5.0%* 0% Total Public Non-group <138% FPG 138-200% FPG 201+% FPG *Difference is statistically significant at the 95% level between the specific estimate and the total estimate for the overall non-elderly population. Source: 2016 Kentucky Health Reform Survey FIGURE 6.8: 70% Total <138% FPG 138-200% FPG 201+% FPG Use of kynect Resources 62.6%* 60.7%* by Income, 2016 60% 52.3% 49.5% 50% 45.4% 43.7% 41.7% 40% 38.0% 37.2%* 34.1% 33.9% 30.4%* 30% 20% 10% 0% Web Call Center In-person Assistance *Difference is statistically significant at the 95% level between the specific estimate and the total estimate for the overall non-elderly population. Source: 2016 Kentucky Health Reform Survey “ 1 in 10 Insured Kentuckians is Concerned about Losing Current Coverage Twice as many people with Governor Bevin's administration has developed plans to transition kynect to a federally support- incomes below 138% FPG ed SBM using the Healthcare.gov marketplace were concerned about losing platform.32 He also announced intentions to shift away from Kentucky’s traditional ACA-based Med- coverage compared to those “ icaid expansion to an alternative expansion based with incomes between on a Section 1115 demonstration waiver. 138% and 200% FPG. 2929 | STATE HEALTH ACCESS DATA ASSISTANCE CENTER SEMI-ANNUAL REPORT Study of the Impact of the ACA Implementation in Kentucky K-HRS PRELIMINARY FINDINGS 30% FIGURE 6.9: Concerned About Losing 25.8%* Current Coverage by Insurance Type and Income, 2016 20% 18.5%* 12.6% 11.4% 10% 5.8%* 3.9%* 0% Total Public Private <138% FPG 138-200% FPG 201+% FPG *Difference is statistically significant at the 95% level between the specific estimate and the total estimate for the overall non-elderly population. Source: 2016 Kentucky Health Reform Survey Due to the uncertainty about the future of Ken- average Kentuckian. Employed and married indi- tucky’s implementation of the ACA, our survey viduals also expressed less concern. asked whether respondents were concerned about losing their health insurance coverage in Nearly Half of Uninsured Kentuckians Delayed the next 12 months. or Went Without Medical Care Due to Cost Compared to other states, Kentucky had one of Overall, at the time of the survey in spring 2016, the largest gains in health insurance since imple- 11.4% of insured Kentuckians reported being mentation of the ACA’s main coverage expansions. concerned about losing their coverage. This varied As the rate of uninsurance falls, health care access based on characteristics of the population, mainly and affordability become important indicators of by type of coverage and household income. Figure whether coverage is allowing people to use health 6.9 shows that 18.5% of Kentuckians enrolled in care services and protecting them from financial public health insurance programs indicated this risk. This section provides a current picture of concern, which was significantly more than the healthcare affordability in Kentucky. overall insured population. By comparison, only 5.8% of privately insured Kentuckians were con- Although almost 21% of Kentuckians delayed or cerned about losing their coverage. More than went without medical care due to cost in the 12 one in four Kentuckians with a household income months prior to the time of the survey, the unin- below 138% of FPG also worried about losing sured were more than twice as likely to delay or their coverage, which was significantly higher than forgo care (48%).33 Kentuckians who described the overall population. Twice as many lower-in- their health as “fair” or “poor” also were more come people (<138% FPG) were concerned about likely to say they delayed or went without care due losing coverage compared to those with incomes to cost (28% and 43%, respectively), while those between 138 and 200% of FPG, and 6.5 times with “excellent” or “very good” health status were higher than those with higher incomes (201+% less likely to say they delayed or went without care FPG). Individuals who had never been married, (7% and 14%, respectively)(not shown in figure). reported having fair or poor health status, or lived Figure 6.10, shows lower-income Kentuckians in Eastern Kentucky also showed a significant- (<138% of FPG) were significantly more likely to ly higher concern. In contrast, individuals with report delayed or forgone care due to cost com- private coverage, household income of 201% of pared to Kentuckians overall. the FPG or higher, a college degree, and good health status were less concerned than the SEPTEMBER 2016 | WWW.SHADAC.ORG 3030 FIGURE 6.10: 40% Percent Who Delayed or Went Without Medical Care Due to Cost by 29.4%* 30% Income, 2016 26.0% 20.5% 20% 13.3%* 10% 0% Total <138% FPG 138-200% FPG 201+% FPG *Difference is statistically significant at the 95% level between the specific estimate and the total estimate for the overall non-elderly population. Source: 2016 Kentucky Health Reform Survey Nearly 1 in 4 Went Without Dental Care effective treatment of medical conditions. Almost The K-HRS also asked respondents whether they 14% reported going without routine, or preventive had gone without specific types of care (i.e., pre- care, which the ACA requires that private health scription drugs, dental care, routine care, men- insurance covers with no cost-sharing. tal health care, and specialist care) due to cost.34 Figure 6.11 also shows that 37% of Kentuckians Figure 6.11 shows that almost one in four Kentuck- went without at least one of these five types of ians went without dental care due to cost in the care due to cost.35 This result suggests that despite prior 12 months. Additionally, nearly one in five the significant gains in coverage post-ACA, strong Kentuckians reported going without prescribed barriers to accessing some types of health care re- drugs due to cost, which could serve as a barrier to main in Kentucky. FIGURE 6.11: 40% 36.8% Forgone Care Due to Cost by Type of Care, 2016 30% 24.6% 18.9% 20% 13.8% 11.6% 10% 6.3% 0% Any Care Dental Care Prescription Routine Care Specialist Mental Health Drugs Care Care Source: 2016 Kentucky Health Reform Survey 3131 | STATE HEALTH ACCESS DATA ASSISTANCE CENTER SEMI-ANNUAL REPORT Study of the Impact of the ACA Implementation in Kentucky K-HRS PRELIMINARY FINDINGS 70% FIGURE 6.12: 63.9%* Forgone Care Due to 60% Cost by Health Status, 2016 50.2%* 50% 40% 36.8% 36.2% 30% 26.9%* 21.3%* 20% 10% 0% Total Excellent Very Good Good Fair Poor *Difference is statistically significant at the 95% level between the specific estimate and the total estimate for the overall non-elderly population. Source: 2016 Kentucky Health Reform Survey Similar to delayed and forgone medical care, 2 in 3 Kentuckians Have Dental Coverage three characteristics were highly correlated with Access to dental care is a concern because research forgoing these types of care: insurance status, has found that poor oral health is associated with income, and health status. Almost three out of other medical conditions, such as cardiovascular five uninsured Kentuckians (59%) went without disease, diabetes, and microbial infections.36 Ad- at least one of these types of care in the prior 12 ditionally, in the U.S. hospital emergency de-part- months, compared to 40% for those enrolled in ment (ED) visits for dental conditions have contin- public coverage and to 31% for those with private ued to rise and the share of ED visits paid for by coverage. Additionally, half of lower-income Medicaid has increased.37 Kentuckians (<138% FPG) went without at least Two thirds of non-elderly, adult Kentuckians re- one type of care; one in three higher-income ported having some form of dental insurance cov- (201+% FPG) individuals reported this experience. erage. In addition, Figure 6.13 shows that dental Figure 6.12 presents the association between care coverage is highly associated with health health status and forgone care due to cost. The insurance coverage: Those without health insur- percentage of Kentuckians who reported going ance were significantly less likely to have dental without any type of care among those with poor insurance (10.2%). The type of health coverage an health was triple that of Kentuckians with excel- individual has was also associated with dental in- lent health (64% and 21%, respectively). While surance; 85.0% of Kentuckians with private health the previous results suggested that having lower insurance report having dental coverage, while income or being uninsured would increase one’s only 56.5% of those with public health insurance chances of forgoing care, the results shown in Fig- reported having dental coverage. ure 6.12 may suggest that forgoing care could af- Dental coverage was also associated with house- fect individuals’ health. hold income, education, marital status, employ- ment, and health status. Kentuckians were less likely to have dental coverage if they were older, had lower incomes, had less education, were not employed, or reported worse health status. SEPTEMBER 2016 | WWW.SHADAC.ORG 3232 A substantial portion of Kentuckians do not receive dentist within the past 12 months, only 38% of dental care as frequently as recommended by lower-income individuals (138% of FPG or less) the American Dental Association,38 as only 60% did. The uninsured were least likely to have visited had a visit within the prior 12 months and 17% a dentist within the past 12 months (31%) and had not visited the dentist in over 5 years (Figure the most likely not to have received dental care 6.14). Frequency of dental visits is associated with in more than 5 years (41%). Slightly over half of health insurance status and type, sex, education, those enrolled in public programs had their last household income, employment, marital status, dental care visit in the last 12 months, but 20% and health status. While 78% of higher-income had not received dental care in the last 5 years. Kentuckians (201% of FPG or more) visited the FIGURE 6.13: 90% 85.0%* Dental Coverage by 80% Insurance Status and Type, 2016 70% 66.0% 60% 56.5%* 50% 40% 30% 20% 10.2%* 10% 0% Total Uninsured Public Private *Difference is statistically significant at the 95% level between the specific estimate and the total estimate for the overall non-elderly population. Source: 2016 Kentucky Health Reform Survey FIGURE 6.14: 90% Total <138% FPG 138-200% FPG 201+% FPG Last Dental Care Visit by 80% 77.6%* Income, 2016 70% 59.9% 60% 53.9% 50% 40% 38.4%* 30% 28.0%* 19.8% 20% 16.7% 17.0%* 17.0% 13.0% 13.4% 13.0% 9.6%* 10.0% 10% 7.7%* 5.2%* 0% <1 year 1-2 years 2-5 years 5+ years *Difference is statistically significant at the 95% level between the specific estimate and the total estimate for the overall non-elderly population. Source: 2016 Kentucky Health Reform Survey 3333 | STATE HEALTH ACCESS DATA ASSISTANCE CENTER SEMI-ANNUAL REPORT Study of the Impact of the ACA Implementation in Kentucky K-HRS PRELIMINARY FINDINGS As noted previously in Figure 6.11, one out of Forgoing dental care is also associated with health four non-elderly adults in Kentucky went without status, marital status, employment status, and dental care in the 12 months prior to the survey household income. Unemployed, disabled, and due to cost. Figure 6.15 shows that Kentuckians retired non-elderly adults were twice as likely with no dental coverage were more likely to go to report forgone dental care in comparison to without dental care services (39%). These results employed individuals. Higher-income (201% of suggest that being without health insurance FPG or more), married, and employed Kentuckians is a stronger barrier to accessing dental care were less likely to report forgoing dental care due services than not having dental coverage, as 45% to cost; whereas lower-income (<138% FPG), of uninsured Kentuckians went without dental divorced/separated, unemployed/disabled/retired care (compared to 39% of those without dental individuals were more likely to report this barrier insurance). Individuals with dental coverage and than the average Kentuckian. Lastly, individuals private health insurance coverage were less likely with fair or poor health also reported higher to forgo dental care than Kentuckians on average, percentages of forgone dental care due to cost. at 17% and 19%, respectively. 50% FIGURE 6.15: 45.0%* Forgone Dental Care by Health Insurance Status 40% 38.8%* and Dental Coverage, 2016 30% 28.0% 24.6% 20% 18.9%* 17.2%* 10% 0% Total Uninsured Public Private No Dental Dental Coverage Coverage *Difference is statistically significant at the 95% level between the specific estimate and the total estimate for the overall non-elderly population. Source: 2016 Kentucky Health Reform Survey 4 in 10 Kentuckians Visited the Emergency health insurance coverage (not shown in figure), Department in the Last 12 Months either in the marketplace or not, also sought this Visits to the Emergency Department (ED) are emergency care. “ costly due to the amount of resources required to provide care to each one of these visits.39 However inevitable in some cases, they are The most common preventable and unnecessary in others.40 Figure reason reported for 6.16 shows that 40.5% of Kentuckians visited the ED in the prior 12 months for themselves or using the Emergency a family member.41 Among the uninsured, 42% Department was that sought this care, which was not a statistically significant difference from the overall rate. How- ever, almost half of those enrolled in public health they experienced a “ insurance programs visited the ED (48%), while medical emergency...... only 19% of those who purchased their own SEPTEMBER 2016 | WWW.SHADAC.ORG 3434 Figure 6.16 also shows that while 39% of Ken- For example, the most common reason was that tuckians who only speak English at home they experienced a medical emergency (29.1%), received care in the ED, 58% of those who speak and 7.1% said their doctors directed them to the a different language at home also visited the emergency department. However, some other ED, which could be due to language barriers or answers suggested that respondents faced bar- differences in health literacy. Other characteristics riers to obtaining care in other settings, such as negatively associated with visiting the ED to seek a regular health clinic, indicating the emergency care were health status and household income. department visit may have been avoidable or care could have been obtained in another setting. For Among K-HRS respondents who reported using example, 28.7% said they used the emergency the ED, certain responses suggested the emergen- department because other facilities were not cy department visit was necessary (Figure 6.17). open when they needed care. FIGURE 6.16: 70% Emergency Department Visits by Health 60% 58.2%* Insurance Type and Language Spoken at 50% 48.2%* Home, 2016 42.4% 40.5% 40% 38.7%* 34.1%* 30% 20% 10% 0% Total Uninsured Public Private Only English Other Language *Difference is statistically significant at the 95% level between the specific estimate and the total estimate for the overall non-elderly population. Source: 2016 Kentucky Health Reform Survey FIGURE 6.17: Main Reason Cited for Reason Percent Visit to the Emergency There was a medical emergency 29.1% Department, 2016 Other facilities weren’t open when needed care 28.7% Named a specific health condition only 18.2% Your doctor directed you to go there 7.1% Do not have a regular doctor 3.0% Proximity to Emergency Department 2.9% Transferred/ taken by ambulance 2.6% Other 8.4% Source: 2016 Kentucky Health Reform Survey. 3535 | STATE HEALTH ACCESS DATA ASSISTANCE CENTER SEMI-ANNUAL REPORT Study of the Impact of the ACA Implementation in Kentucky IV. DISCUSSION & NEXT STEPS Discussion This study will document gains in health insurance pre-ACA trend of increasing over time. Quality in- coverage, monitor whether those gains are dicators also have shown mixed results, with de- maintained over time, and examine the impact of creases in certain preventable hospitalizations and the ACA. Recent evidence shows that measures increases in others. Finally, while one health out- intended to assess the impact of implementation comes indicator—adult cigarette use—has shown of Medicaid expansion under the ACA may improvement, two other indicators have remained underestimate the results in the short-term, unchanged: adolescent obesity and cigarette use. indicating the need for studies that evaluate the impact over time.42 We also presented initial findings from our quantitative primary data collection activities— Implementation of the ACA in Kentucky has initial results for the Kentucky Health Reform significantly increased health coverage in the Survey—in this report. Although nearly half Commonwealth.43 The preliminary results of our of Kentuckians reported having private health Kentucky Health Reform Survey, which found that insurance coverage, public health insurance only 8.9% of non-elderly adult Kentuckians were coverage also plays a large role in Kentucky, uninsured and are consistent with other surveys covering nearly as many Kentuckians as private that have found significant drops in Kentucky’s coverage. We also found that kynect was an uninsurance rate. important avenue for people to obtain health In this update of our study, we found that access insurance, with the majority reporting positive to health care services remained largely stable be- experiences with the marketplace. Despite the tween our baseline year of 2012 and 2014, the first recent strides in health insurance coverage since year of ACA implementation. Cost indicators have the implementation of the ACA, more than 10% been mixed: While hospital charity care and self- of Kentuckians reported being concerned about pay charges have dropped by more than 75% since losing their coverage, and 1 in 5 Kentuckians 2012, and a smaller share of Kentuckians report reported delaying or going without medical care trouble paying medical bills, the percentages of Ken- due to cost. We will soon begin interviews with tuckians delaying or forgoing care due to cost has Kentucky stakeholders and focus groups with not changed, and premiums for some employer- Medicaid beneficiaries to further explore findings sponsored insurance plans have continued their through the qualitative component of the study. Next Steps This semi-annual report provides an updated further developments in the health care environ- data comparison to the baseline assessment of ment in Kentucky. SHADAC will collaborate with the health care environment in Kentucky across the Foundation and the ACA Impact Study Over- the domains of coverage, access, cost, quality, sight Committee on key next steps for the study, and health outcomes. In combination with prior which include: reports, these indicators demonstrated improve- • Conducting focus groups and key informant ments in certain areas, although many of the in- interviews in summer and fall 2016. dicators have so far seen limited or no significant • Continued production of Quarterly Snapshots changes. We will continue to track the study in- to show change in indicators that have more dicators and use the other components of our frequent data availability. study to inform these findings. This report also presents preliminary findings from the Kentucky • Production of a second special report on sub- stance use in 2016. Health Reform Survey (K-HRS). Future reports will • The Year 2 annual report, with data updates present additional analyses of the K-HRS data that (as available) to the standard metrics pre- will allow comparisons to the KHIP as well as more sented in previous annual and semi-annual in depth analysis in several policy-relevant areas reports, as well as additional findings from than possible in the KHIP. the Kentucky Health Reform Survey, to be As the study proceeds, SHADAC will use future submitted in February 2017. annual and semi-annual reports to document SEPTEMBER 2016 | WWW.SHADAC.ORG 3636 V. APPENDIX: DATA SOURCES, METHODS, & INDICATORS In this Appendix, we describe our data collection procedures and methods for the study. The Appendix is organized by data source, and it includes a brief data source description, a discussion on how the estimates were obtained, and some notes about specific indicators where relevant. Medical Expenditure Panel Survey – Insurance year, emergency department visit in the last year, found Component (2012, 2013, 2014, 2015) doctor when needed, told provider accepts insurance, The Medical Expenditure Panel Survey – Insurance and changes to medical drug use due to cost. Component (MEPS-IC) is a federal survey sponsored The changes to drugs due to cost measure includes by the U.S. Department of Health and Human Ser- asking the doctor for cheaper medications, delaying vices, Agency for Healthcare Research and Quality. The refills, taking less medication than prescribed, skip- MEPS-IC collects information from public and private ping dosages, using alternative therapies, or buying employers about the health insurance plans they of- medications out of the country within the past year. fer to employees, including benefits, costs, and other The trouble paying off medical bills measure includes characteristics. The sample size in 2015 was over 39,000 people who are paying off medical bills within the past businesses at the national level. Summary reports with year. detailed state-level tables for private sector employers are released in July of each year following the survey Behavioral Risk Factor Surveillance System year. Unlike with the ACS and CPS, a public use data file (2011, 2012, 2013, 2014) is not available from the MEPS-IC. The Behavioral Risk Factor Surveillance System (BRFSS) For this report, SHADAC used data from the MEPS-IC to is a state-based survey sponsored by the CDC and the estimate private-sector employer offer rates and premi- Kentucky Cabinet for Health and Family Services. The ums. We accessed these estimates from the MEPS-IC BRFSS survey asks about health conditions, risk behav- web site. iors, preventive health practices, access to health care, and health insurance coverage. State-level results are National Health Interview Survey available from the CDC for all states. Kentucky BRFSS (2012, 2013, 2014) data are analyzed at the Area Development District The National Health Interview Survey (NHIS) is a feder- (ADD) level for the state’s 15 ADDs. The sample size for al survey sponsored by the Centers for Disease Control each ADD is 500 completed surveys, to ensure an ade- & Prevention (CDC) and the National Center for Health quate sample size for analysis. Statistics (NCHS). The NHIS asks about health insurance SHADAC has changed the way we obtained the BRFSS coverage, health care utilization and access, health con- since the baseline report, opting to access and analyze ditions and behaviors, and general health status, as well the public use data for all estimates. To maintain consis- as many demographic and socioeconomic characteris- tency and comparability, we have updated our baseline tics. It has a total sample of more than 112,000 in 2014 estimates, as well. This report includes a new indicator (the NHIS does not release state-level sample sizes). with the estimate of cigarette use in adults, added to Summary reports, with state estimates for the 43 larg- the health outcomes domain. This estimate reports the est states of types of coverage (including Kentucky) percentage of adults who have smoked 100 or more cig- are released six months after data collection. Data files arettes in their lifetime and who currently smoke some with state-level and other geographic identifiers can be days or every day. accessed only through a Census Research Data Cen- ter (RDC). Access to data in Research Data Centers is Healthcare Cost and Utilization Project (2012, only allowed after a proposal has been submitted and 2013, 2014) approved by NCHS and only to researchers who have The Healthcare Cost and Utilization Project (HCUP), is Special Sworn Status. SHADAC has an approved project sponsored by the U.S. Department of Health and Hu- for accessing this restricted data in the RDC for the pur- man Services’ Agency for Healthcare Research and pose of posting estimates on our Data Center. Changing Quality (AHRQ) and provides data on health statistics variable definitions or adding variables means amend- and information on hospital inpatient and emergency ing our annual proposal to the RDC. SHADAC used data department utilization. from the NHIS to estimate nine different measures in We use HCUP data for estimates in the quality domain, the cost and access domains. Measures within the cost including diabetes short-term admissions, hypertension domain include trouble paying medical bills, delayed admissions, asthma admissions, and death rate in needed care due to cost, and went without needed low-mortality DRGs. These indicators were previously care due to cost. For the access domain, the measures reported with data from a different source and due include: usual source of care, provider visit in the last to potential differences in the methodology, these 3737 | STATE HEALTH ACCESS DATA ASSISTANCE CENTER SEMI-ANNUAL REPORT Study of the Impact of the ACA Implementation in Kentucky APPENDIX data may not match similar data in prior reports. The Kentucky Health Reform Survey (2016) diabetes admission estimate reports the diabetes The Kentucky Health Reform Survey (K-HRS) was short-term complications admission rate for adults. conducted by SHADAC and the University of Cincinnati The hypertension estimate reports the hypertension Institute for Policy Research from March-May 2016. admission rate for adults. The asthma estimate reports The methodology and a substantial part of the survey asthma in younger adults’ admission rate for adults ages instrument were based on the existing Kentucky Health 18 to 39. The death rate estimate reports those dying in Issues Poll (KHIP), allowing for comparisons of the the hospital while getting care for a condition that rarely estimates from the K-HRS to prior KHIP estimates and results in death rate cases. Because these administrative potentially future KHIP estimates, yet the K-HRS included data are not based on a sample, there was no need for more depth in several policy-relevant areas than statistical testing of differences. possible in the KHIP. Survey questions were selected in consultation with the Foundation and study Oversight Youth Risk Behavior Surveillance System (2013, Committee, with overarching goals of maintaining 2015) consistency with the KHIP to allow trend analyses and The Youth Risk Behavior Surveillance System (YRBSS) investigating key components of ACA implementation survey asks students in grades 9-12 about tobacco in Kentucky, such as the Commonwealth’s kynect state- use, sexual behaviors, alcohol and drug use, diet and based marketplace. The dual-frame (landline and cell exercise, obesity, asthma, and behaviors related to phone) survey sampled non-elderly adult Kentuckians violence and injury. Kentucky also administers a middle- for a total of 1639 interviews. The measures in this school version for grades 6-8. The YRBSS is given to a report include data on uninsurance and coverage types, sample of students, and is a bi-annual survey conducted concern about losing coverage, forgone or delayed care in odd-numbered years, with results released the year due to cost, dental coverage and care, and emergency following the survey. In 2015, the Kentucky sample department use. from the YRBSS included more than 2,500 students. The source for the indicators obtained for this source is Report Estimate Considerations online data from the CDC. Suppression rules depended on the source of the data We include the following three measures from the and the availability of measures of uncertainty and/or survey: unprotected sex among high school students sample sizes. Estimates from the NHIS are suppressed if in the quality domain, as well as obesity rates and either the number of sample cases was too small or the cigarette use in the health outcomes domain. The relative standard error was greater than 30%. In cases estimate on unprotected sex reports the percentage where standard errors were not available, we did not of sexually active high school students who did not suppress any estimates. use any method to prevent pregnancy during their It should be noted that we lacked the necessary last sexual intercourse. The obesity measure reports information to perform an “overlap adjustment” to the percentage of students who were above the 95th our statistical tests. Since we are comparing Kentucky’s percentile for Body Mass Index based on gender-and- estimates to national estimates (which include age-specific reference data from the 2000 CDC growth Kentuckians), the proportion of Kentuckians in the charts. The cigarette measure reports the percentage of population considered in the estimate should be taken high school students who currently smoked cigarettes, into account. However, this specific information was on at least 1 day during the 30 days before the survey. not available for most estimates. By not conducting an overlap adjustment we are slightly less likely to report Kentucky Outpatient Hospital Administrative that a difference is statistically significant. Claims Data (2012, 2013, 2014, 2015) The Kentucky Outpatient Hospital Administrative Claims Data were provided by the Kentucky Cabinet for Health and Family Services. These data were updated to include 2012 data (previously unavailable) and the full year of 2013 data (previous data only included the first three quarters of the year), and include charges for self-pay and charity care. In these data we are not able to discern between paid and unpaid charges. Since hospitals are likely to receive some payment for at least of portion of self-pay charges, we acknowledge that not all self-pay charges become “uncompensated”. For the purposes of estimating uncompensated care, we assume that the majority of the self-pay charges are not paid in full. SEPTEMBER 2016 | WWW.SHADAC.ORG 3838 VI. ENDNOTES 1 Though Arkansas is not technically a border state, we include it because it is often compared to Kentucky due to similarities in health status, demographics, and state policies. 2 Institute of Medicine. (2009). America’s uninsured crisis: Consequences for health and health care. (Report Brief). Available at: http://www. nationalacademies.org/hmd/~/media/Files/Report%20Files/2009/Americas-Uninsured-Crisis-Consequences-for-Health-and-Health-Care/ Americas%20Uninsured%20Crisis%202009%20Report%20Brief.pdf 3 Planalp, C., Sonier, J., & Fried, B. (2015). State-Level Trends in Employer-Sponsored Health Insurance: A State-by-State Analysis. State HeathAccess Data Assistance Center. Available at: http://www.shadac.org/publications/ state-level-trends-employer-sponsored-health-insurance-2016-report 4 Institute of Medicine, Committee on Monitoring Access to Personal Health Care Services. (1993). Access to Health Care in America. Washington, DC. National Academy Press. 5 Call, K., McAlpine, D., Garcia, C., Shippee, N., Beebe, T., Adeniyi, T., & Shippee, T. (2014). Barriers to Care in an Ethnically Diverse Publicly Insured Population: Is Health Care Reform Enough? Medical Care, 52(8) 6 SHADAC is investigating the feasibility of including another access indicator, wait time to see a primary care provider, in future reports. For more information on this, see the section on the National Health Interview Survey in the Appendix. 7 U.S. Department of Health and Human Services, Assistant Secretary for Planning and Evaluation. (Undated). Health System Measurement Project: Percentage of People Who Have a Specific Source of Ongoing Medical Care. Available at: https:// webcache.googleusercontent.com/search?q=cache:1aPHByoep8QJ:https://healthmeasures.aspe.hhs.gov/ percentage-people-who-have-specific-source-ongoing-medical-care+&cd=2&hl=en&ct=clnk&gl=us 8 DeVoe, J., Tillotson, C., Wallace, L., Lesko, S., & Pandhi, N. (2012). Is health insurance enough? A usual source of care may be more important to ensure a child receives preventive health counseling. Maternal and Child Health Journal, 16(2), 306-315 9 For some of the NHIS indicators, the difference between 2012 and 2014 U.S. estimates is statistically significant even though a similar or larger difference was not statistically significant for Kentucky; this may be because the U.S. sample size is larger than for individual states, making it more likely to find that a difference is statistically significant. 10 Agency for Health care Research and Quality. (2014). HCUP Statistical Brief: Overview of Emergency Department Visits in the United States, 2011. Available at: https://www.hcup-us.ahrq.gov/reports/statbriefs/sb174-Emergency-Department-Visits-Overview.pdf 11 Office of the Assistant Secretary for Planning and Evaluation (2014). Poverty Guidelines. Available at: https://aspe.hhs. gov/2014-poverty-guidelines 12 U.S. Census Bureau (2014). Percent of people below poverty level in the past 12 months. American Fact Finder. 2014 American Community Survey 1-year Estimates. Available at: http://factfinder.census.gov/faces/tableservices/jsf/pages/productview.xhtml?src=bkmk 13 Charity care and self-pay charges are used as a proxy for uncompensated care. 14 Cunningham, P., Garfield, R., & Rudowitz, R. (2015). How Are Hospitals Faring Under the Affordable Care Act? Early Experiences from Ascension Health. The Henry J. Kaiser Family Foundation. Available at: http://kff.org/health-reform/issue-brief/ how-are-hospitals-faring-under-the-affordable-care-act-early-experiences-from-ascension-health/ 15 Torio, C., Elixhauser, A., & Andrews, R. (2013). Trends in Potentially Preventable Hospital Admissions among Adults and Children, 2005– 2010. Agency for Healthcare Research and Quality. Healthcare Cost and Utilization Project. Statistical Brief #151. Available at: https://www. hcup-us.ahrq.gov/reports/statbriefs/sb151.pdf 16 Commonwealth of Kentucky, Cabinet for Health and Family Services. (2012). Prevention Quality Indicators. Available at: http://chfs.ky.gov/ ohp/healthdata/pqis.htm 17 Centers for Medicare and Medicaid Services. (2014). Acute Inpatient PPS. Available at: https://www.cms.gov/Medicare/Medicare-Fee-for- Service-Payment/AcuteInpatientPPS/index.html?redirect=/acuteinpatientpps/ 18 Agency for Healthcare Research and Quality. (2009). Patient Safety Indicators Technical Specifications PSI #2 Death in Low-mortality DRGs. Available at: http://www.qualityindicators.ahrq.gov/Downloads/Modules/PSI/V41/TechSpecs/PSI%2002%20Death%20in%20Low- mortality%20DRGs.pdf 19 Klein, J., Barratt, M., Blythe, M., Braverman, P., Diaz, A., Rosen, D., & Wibbelsman, C. (2007). "Contraception and Adolescents." Pediatrics 120, 5. Available at: http://pediatrics.aappublications.org/content/120/5/1135 20 U.S. Department of Health and Human Services, Office of Disease Prevention and Health Promotion. (2015). Healthy People 2020: Foundational Health Measures: Determinants of Health. Available at: http://www.healthypeople.gov/2020/about/ foundation-health-measures/Determinants-of-Health 21 Boban, M., Francis, L., Kayalar, A., & Cone, J. (2008). Obesity: Effects of Cardiovascular Disease and its Diagnosis. Journal of the American Board of Family Medicine, 21(6). Available at http://www.jabfm.org/content/21/6/562.full National Institutes of Health, National Heart, Lung, and Blood Institute. (2013). Why Obesity is a Health Problem. Available at: https:// 22 www.nhlbi.nih.gov/health/educational/wecan/healthy-weight-basics/obesity.htm 3939 | STATE HEALTH ACCESS DATA ASSISTANCE CENTER SEMI-ANNUAL REPORT Study of the Impact of the ACA Implementation in Kentucky END NOTES 23 The Centers for Disease Control and Prevention. (Undated) Coverage for Tobacco Use Cessation Treatments. Available at: https://www.cdc. gov/tobacco/quit_smoking/cessation/pdfs/coverage.pdf 24 Foundation for a Healthy Kentucky. (2015). Employer-provided insurance rises; uninsured rate drops. Available at: http://www.healthy-ky. org/sites/default/files/KHIPInsuranceCoverageFINAL.pdf 25 Foundation for a Healthy Kentucky. (2016). State’s uninsured rate at 13%; unstable insurance declines. Available at: http://www.healthy-ky. org/sites/default/files/KHIP%20health%20insurance%20FINAL.pdf 26 Other surveys, such as Gallup, have found different estimates. However, this figure is the lowest ever estimated using the same or equivalent methodology. 27 Group coverage includes health insurance through an employer. Non-group coverage includes self-purchased health insurance for an individual or family and health insurance purchased through the marketplace for which a premium is paid. Public coverage includes Medicaid (including Kentucky Health Choices, Anthem, Passport, CoventryCares, Humana/CareSource, and WellCare), Medicare, Military care (including TRICARE and CHAMP-VA), and health insurance obtained through the marketplace for which a premium is not paid. Other coverage is health insurance for which type could not be determined due to limited information provided by the respondent. 28 Only 70 individuals reported having non-group coverage in the KHRS. Although a small sample size, it is still valid to obtain estimates for some indicators for this group. However, the reader needs to consider this indicator with caution, as the standard errors of tend to be high and confidence intervals wide. 29 An annual income of $16,394 for a single person or $33,534 for a family of four is equivalent to 138% of the Federal Poverty Guidelines in 2016. (Source: Office of the Assistant Secretary for Planning and Evaluation. Available at: https://aspe.hhs.gov/poverty-guidelines). Similarly, 200% of the FPG is equivalent to incomes of $23,760 for a single person and $48,600 for a family of four. 30 Letter from Governor Matthew Bevin to U.S. Department of Health & Human Services Secretary Sylvia Mathews Burwell. (December 30, 2015). Available at: http://www.xcenda.com/Documents/HPW1-15-16_Bevin-Burwell.pdf 31 Note that individuals may have used more than one resource, which implies that the sum of these estimates surpasses 100%. 32 Letter from U.S. Department of Health and Human Services Secretary Sylvia Burwell to Governor Matt Bevin (July 20, 2016). Available at: http://static1.squarespace.com/static/56992609c647ad16c04a9386/t/5792792d1b631b28a684b6aa/1469217069843/ Secretary+Burwell+response+to+KY+Governor+Bevin+Letter+re+KY+Kynect+Tran...+%281%29.pdf 33 This estimate is higher than estimates from the 2014 NHIS, which found that 14.5% of non-elderly Kentuckians delayed and 11.9% went without medical care, but the K-HRS estimates combine these two barriers. 34 The reader should note that these indicators are slightly different from that of delayed or forgone medical care, as they only refer to forgone prescription drugs, dental care, routine care, mental care, and specialist care. 35 As these categories are not mutually exclusive, forgoing any care is defined as forgoing at least one of the five types of care: dental, routine, mental, or specialist care, or prescription drugs. However, it does not include forgoing medical care; the difference between the question for medical care and those for these other types of care make them not comparable. 36 Institute of Medicine, Committee on an Oral Health Initiative. (2011). Advancing Oral Health in America. The National Academies Press. Available at: http://www.hrsa.gov/publichealth/clinical/oralhealth/advancingoralhealth.pdf Wall, T., & Vujicic, M. (2015). Emergency Department Use for Dental Conditions Continues to Increase. Available at: http://www.ada.org/~/ 37 media/ADA/Science%20and%20Research/HPI/Files/HPIBrief_0415_2.ashx 38 American Dental Association. (2013). American Dental Association Statement on Regular Dental Visits. Available at: http://www.ada.org/ en/press-room/news-releases/2013-archive/june/american-dental-association-statement-on-regular-dental-visits 39 Blue Cross Blue Shield of Massachusetts. (Undated). Typical Costs for Common Medical Services. Available at: http://www.bluecrossma. com/blue-iq/pdfs/TypicalCosts_89717_042709.pdf 40 Caldwell, N., Srebotnjak, T., Wang, T., & Hsia, R. (2013). “How Much Will I Get Charged for This?” Patient Charges for Top Ten Diagnoses in the Emergency Department. PLOS ONE, 8(2) 1-6. Available at: http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0055491 41 Although they are similar, the K-HRS and NHIS questions on Emergency Department use are not comparable. The 2016 K-HRS asks whether respondents used the ED for themselves or a family member (40.5%), while the NHIS asks whether respondents used the ED for themselves only (24.9% in 2014). 42 Sommers, B., Blendon, R., Orav, J., & Epstein, A. (2016). Changes in Utilization and Health Among Low-Income Adults After Medicaid Expansion or Expanded Private Insurance. JAMA Internal Medicine. Available at: http://static.politico.com/9b/ ca/6d295d7d4975a810287854c70efc/jama-study-medicaid-expansion-two-years-in.pdf 43 State Health Access Data Assistance Center. (2016). Annual Report: Study of the Impact of the ACA Implementation in Kentucky. Available at: http://www.healthy-ky.org/sites/default/files/REVISED%20FINAL%20FULL%20Annual%20Report%203.21.pdf SEPTEMBER 2016 | WWW.SHADAC.ORG 4040 4141 | STATE HEALTH ACCESS DATA ASSISTANCE CENTER