REPORT DECEMBER 2020 STATE ALL-PAYER CLAIMS DATABASES Tools for Improving Health Care Value Part 2: The Uses and Benefits of State APCDs Douglas McCarthy Senior Research Advisor The Commonwealth Fund ABSTRACT TOPLINES ISSUE: Many states and stakeholders are seeking to control the rising All-payer claims databases can cost of health care and increase its value. All-payer claims databases help state health care purchasers “buy smart,” raise awareness (APCDs) facilitate such efforts by aggregating data on health care of the need for health system services paid for by health insurers and public programs, thereby change, and fuel data-informed offering a broad perspective on cost, utilization, and quality of care. policymaking. GOALS: Describe the uses and benefits of state-level APCDs as well as challenges to realizing their value, including data limitations and To reap the benefits of an all- antitrust concerns. payer claims database, states need to cultivate relationships METHODS: Interviews with staff and stakeholders of eight diverse state- with stakeholders and learn how level APCDs, supplemented by a review of documentary evidence. to use data to meet their needs. KEY FINDINGS AND CONCLUSIONS: APCDs are used to: 1) report on health system spending, utilization, and performance; 2) enhance state policy and regulatory analysis; 3) inform the public about health care prices and quality; 4) enable value-based purchasing and health care improvement; 5) support public health monitoring and improvement; and 6) provide reliable data for health care research and evaluation. The benefits of state APCDs include raising awareness of the need for change; fueling data-informed policymaking; and generating knowledge for improvement. Fulfilling the purposes of an APCD requires cultivating relationships with stakeholders and learning how to effectively use data to meet their needs. State All-Payer Claims Databases: Tools for Improving Health Care Value PART 22 INTRODUCTION approaches and contexts for implementing an APCD as States and health care stakeholders that wish to take well as to highlight relatively advanced uses of data (see effective action to control health care spending and ensure the section, “How This Study Was Conducted”). The first report describes how the states established their APCDs. its value require systemic information on costs, utilization, This report describes the uses of APCDs, their benefits as and quality of services. To support this objective, 21 perceived by stakeholders (Exhibit 2), and challenges that states have created or are implementing all-payer claims must be overcome to realize their value. It concludes with databases (APCDs). These aggregate health care payment lessons learned, which could be useful for other states. data for state residents from commercial health insurers, some employee benefit plans, and the Medicaid and Medicare programs (Exhibit 1). Eleven other states have STATE APCD USE CASES: A GROWING indicated strong interest in establishing APCDs, while USER BASE voluntary efforts serve specific geographic areas or The collective uses of state APCDs have increased as the purposes in at least five states.1 usability and integrity of the data have improved and as stakeholders recognize more opportunities to apply the This report, the second in a two-part series, synthesizes data.2 States typically require that use of their APCDs will the experiences of eight state-level APCDs. The purpose benefit residents of the state, as well as meeting other is to inform states about what to consider when creating objectives specific to each state. A synthesis of the many an APCD and to help them realize the potential of their applications of state APCD data by stakeholders in the APCD. Study sites were selected to exemplify diverse study states suggests six overarching use cases (Exhibit 2). Exhibit 1 State Exhibit Activity on All-Payer 1. State Activity Claims on All-Payer Claims Databases Databases Existing In Implementation Strong Interest Existing Voluntary Effort Study Sites State Agency -Maine -Minnesota -New Hampshire -Utah Administrator Under State Authority -Arkansas -Colorado -Virginia Voluntary Collaborative -Wisconsin Source: Adapted from The APCD Council with permission. © 2009-2020 University of New Hampshire, The APCD Council, National Association of Health Data Organizations. All Rights Reserved. commonwealthfund.org Report, December 2020 Source: Adapted from The APCD Council (permission forthcoming). © 2009-2020 University of New Hampshire, The APCD Council, National Association of Exhibit 2 APCD Use Cases and Users of APCD Data and Analyses State All-Payer Claims Databases: Tools for Improving Health Care Value PART 23 Exhibit 2. APCD Use Cases and Users of APCD Data and Analyses Use Cases Users Policymakers Researchers Consultants Consumers Purchasers Providers Insurers Reporting on health care spending, utilization, and performance P P P P S P P Enhancing state policy and regulatory analysis P S S S S S Informing the public about health care prices and quality S P S S P S S Enabling value-based purchasing and health care improvement P P P P S P Supporting public health monitoring and improvement P S S S S S Providing reliable data for health care research and evaluation S S S S P S Source: Author’s analysis. Note: P = Primary user. S = Secondary user. Policymakers includes state legislators, state agencies, and local officials. Purchasers includes public and private employers and coalitions. Providers includes hospitals, health systems, and health care practitioners. Insurers includes health and dental insurers. Consultants category may also include brokers and vendors. Note: P = Primary user. S = Secondary user. Policymakers includes state legislators, state agencies, and local officials. Purchasers includes public and private employers and coalitions. Providers includes hospitals, health systems, and health care practitioners. Insurers includes health and dental insurers. Consultants category may also include brokers and vendors. Colorado and Maine attend to all six use cases; other states cost of care and total resource use for commercially give priority to a subset ofSource: use cases and Author’s users, depending analysis. insured populations.3 By highlighting differences in the on their statutory or organizational mandate and relative level and drivers of spending among states and resources (Appendix Exhibit A). Minnesota focuses on use regions within states (Exhibit 3), such an analysis can cases one, two, and five; New Hampshire emphasizes use help policymakers and other stakeholders target their cases two, three, five, and six; and Wisconsin prioritizes use cost control efforts.4 Both states distributed practice-level cases four and six. Each use case is described in turn below. reports identifying actionable opportunities for physicians to improve their performance relative to their peers.5 Reporting on Health Care Spending, Utilization, Preventable hospital use. The Minnesota Department of and Performance Health used its APCD to identify, for a one-year period, 1.3 Study sites conduct analyses or sponsor research (to million hospital and emergency department (ED) visits varying degrees depending on their budgets), examining that were potentially preventable, and which represented patterns and trends in health care spending, utilization, $1.9 billion or 4.8 percent of total health care spending.6 and performance overall and by age group, type of Low-value care. Five study sites are using Milliman’s insurance coverage, and geographic area. Recent public Health Waste Calculator to quantify the frequency of reports have examined topics such as price increases for — and potential savings from reducing — services that prescription drugs, trends in prescription opioid drug use, provide little or no clinical benefit to patients (Exhibit 4). and the use and distribution of telemedicine visits. In Virginia, low-cost services were delivered much more Total cost of care. Colorado and Utah participated in frequently and accounted for almost twice the aggregate a multistate project led by the Network for Regional cost of unnecessary high-cost services.7 These insights are Health Improvement that calculated risk-adjusted total being used in collaborative efforts to guide improvement.8 commonwealthfund.org Report, December 2020 Total Median Risk-Adjusted Per Member Per Month Commercial Cost State All-Payer Claims Databases: Tools for Improving Health Care Value PART 24 by Colorado Region Exhibit 3. Total Median Risk-Adjusted Per Member Per Month Commercial Cost by Colorado Region Cost Utilization Compared to the CO Price Compared to the CO PMPM Statewide Median* Statewide Median* West $584 7% 29% East $551 8% 6% Greeley $492 3% 22% Fort Collins $453 1% 7% Grand Junction $449 13% 7% Denver $444 Statewide 5% 6% Median 6% Boulder $412 $424 8% Pueblo $378 4% 6% Colorado Springs $335 10% 10% Statewide Statewide Median Median Exhibit 4 Center for Improving Value in Health Care. Notes: Total includes inpatient, outpatient, professional, and pharmacy. Source: Statewide medians only reflect results for 163 adult primary care practices included in the 2016 Colorado APCD study. Utilization and Spending for Low-Value Services by State in 2017 Notes: Total includes inpatient, outpatient, professional, and pharmacy. Statewide medians only reflect results for 163 adult primary care practices included in the 2016 Colorado APCD study. Exhibit 4. UtilizationSource: andCenter Spending for Improvingfor ValueLow-Value in Health Care. Services by Payer and State in 2017 Commercial Medicaid Low-Value Total Low-Value Low-Value Total Low-Value Services per Low-Value Spending Services per Low-Value Spending 1,000 Spending PMPM 1,000 Spending PMPM Maine 322 $54,356,000 $10.38 317 $9,630,000 $4.36 Colorado 419 $150,576,000 $10.39 339 $69,052,000 $4.98 Virginia 477 $219,343,000 $6.16 106 $45,055,000 $3.11 Source: VBID Health, Utilization and Spending on Low-Value Medical Care Across Four States, May 2020. Notes: Washington State was included in the analysis but is not shown. Total Low-Value Spending includes both payer costs and member out-of-pocket costs. Low-Value Services per 1,000 = number of low-value serviced provided per 1,000 members in that sector. Low-Value Spending Per Member Per Month (PMPM) = total low-value spending divided by total member months of enrollment for that state. Claims from each APCD were run through the Milliman MedInsight Health Waste Calculator, a proprietary, algorithm-based software program designed to quantify low-value care use and spending by differentiating whether the use of a specific medical service was clinically necessary, likely low-value, or low-value. Low-value care was quantified by analyzing 47 clinical services deemed as low- value by sources such as the United States Preventive Services Task Force and the Choosing Wisely campaign. Prescription Drug Spending. To help policymakers and that drugs administered in physicians’ offices (and paid payers identify targets for cost control, Maine recently for under a medical benefit) tend to be high cost, high enacted legislation requiring its APCD to report, by Notes: Washington State was included in the analysis but is not shown. Total Low-Value Spending in cost includes bothgrowth, and payer costs and account member for nearly out-of-pocket as much costs. Low-Value in total Services per 1,000 = number of low-value serviced provided per 1,000 members in that sector. Low-Value Spending Per Member Per Month (PMPM) = total low-value spending divided by total member months of typefor enrollment ofthat payer, the costliest state. Claims from each APCDdrugs, the were run most through thefrequently health Milliman MedInsight Health Waste careaspending Calculator, as prescription proprietary, algorithm-based software drugs filled at program designed the low- to quantify value care use and spending by differentiating whether the use of a specific medical service was clinically necessary, likely low-value, or low-value. Low-value care was quantified by analyzing 47 prescribed clinical services deemeddrugs, andbythe as low-value drugs sources with such as the States the United highest year- Preventive Services Taskpharmacy (and paid Force and the Choosing for under a pharmacy benefit). Wisely campaign. over-year cost increases. A Minnesota analysis found Source: VBID Health, Utilization and Spending on Low-Value Medical Care Across Four States, May 2020. commonwealthfund.org Report, December 2020 State All-Payer Claims Databases: Tools for Improving Health Care Value PART 25 Enhancing State Health Policy and Regulatory APCD data to answer questions about, for example, the Analysis frequency and cost of so-called “surprise” medical bills for State insurance departments, Medicaid agencies, attorneys out-of-network services. APCD data also can be used in a general, and legislators are making increasing use of APCD policy intervention. Legislation enacted in Colorado, for data to enhance policymaking, regulatory oversight, and example, limits out-of-network billing to a percentile of planning functions (Exhibit 5). Leaders in several states the in-network allowed amount in the same geographic say that legislators are routinely requesting analyses of area based on commercial claims for the prior year.9 Exhibit 5. Example Uses of APCD Data for State Health Policy and Regulatory Analysis • Review the adequacy, fairness, and affordability of insurers’ premium rates11 • Assess health plan network adequacy (out-of-network use) • Determine reserve requirements for new insurance products • Analyze the feasibility of state risk-adjustment and reinsurance programs Insurance • Estimate the impact of changes in geographic rating areas (e.g., combining higher- and lower-cost rating regulators10 areas or creating a single statewide rating area) • Evaluate how patient cost-sharing effects utilization of services • Quantify the timeliness of insurer claims payment • Implement “surprise-billing” legislation limiting out-of-network charges • Understand cost impact of COVID-related moratorium on elective services • Conduct comparative studies of utilization, cost, and quality of care • Determine the adequacy of reimbursement rates for attracting and retaining providers by way of comparison to commercial insurance rates • Assess expected and actual utilization and costs for expansion populations Medicaid • Quantify the opportunity to impact spending for high-cost beneficiaries agencies • Examine how a Medicaid accountable care program reduced rehospitalizations • Inform the creation of a “centers of excellence” program • Describe the frequency and total cost of low-value care • Identify opportunities to reduce pharmacy spending by examining top drugs by spending, volume, and specialty usage as well as generic substitution rates • Conduct competitive market analysis for mergers and potential anticompetitive agreements in health care Attorney • Study competition between hospitals by assessing how far patients are willing to travel to receive inpatient General’s and outpatient care office • Assess competition between physician groups and between hospital outpatient departments and freestanding facilities • Evaluate options for implementing a public insurance option • Assess the impact of out-of-network billing practices • Calculate the percent of health care spending devoted to primary care Legislative • Quantify costs of emergency department visits for mental health conditions to demonstrate the opportunity requests, for improving access to services mandates, or attention • Measure trends in the costs of treating firearm-related injuries in the state • Assess potential savings from bulk purchases of vaccines • Identify possible violations of state regulation of pharmacy benefit managers by comparing payments reported by pharmacies to payments reported in claims Source: Author’s analysis of APCD websites and reports. commonwealthfund.org Report, December 2020 State All-Payer Claims Databases: Tools for Improving Health Care Value PART 26 Informing the Public About Health Care Prices and Quality CASE EXAMPLE: NEW HAMPSHIRE’S Colorado, Maine, and New Hampshire are among nine HEALTHCOST TRANSPARENCY WEBSITE states that have created websites displaying average or In 2007, the New Hampshire Insurance Department median prices for common elective procedures, along launched a HealthCost website, which uses data with providers’ quality of care.12 These sites report on from the state’s APCD to estimate prices paid to bundled prices for common episodes of care including health care facilities for common medical tests services obtained before, during, and after a procedure.13 and procedures. Users can learn the total cost Other study states offer more limited information. For of a procedure — including physician, lab, and example, Virginia publishes statewide and regional prices facility fees — based on their insurance coverage, according to care setting, whether a hospital, ambulatory deductible, and co-insurance. The website now surgical center, or physician’s office.14 covers more than 100 medical tests and procedures One study found that New Hampshire’s consumer and two dozen dental procedures. It also displays shopping website (see Case Example: New Hampshire’s quality measures for the state’s hospitals, such HealthCost Transparency Website) had a procompetitive as patient experience and infection rates. The effect on prices of medical imaging procedures, as some website’s use has been growing (it reached up consumers selected lower-cost providers, which led to to 30,000 visitors in one recent month) through providers reducing their prices.15 Assuming an annual cost outreach to employers and links from social media of $1 million to develop and operate the state’s APCD and and Google searches, according to state officials. website, the estimated $44 million saved over five years by consumers and insurers on the cost of imaging procedures The existence of the website appears to have led represents a substantial return on the state’s investment. to changes in market behaviors.22 To estimate Additional research is needed to determine whether and market effects, a researcher compared changes in to what degree this experience may be generalizable to the price of medical imaging procedures that were other procedures and states.16 and were not posted on the website during the five years after its launch. Results imply that the website While the benefits of health care quality ratings are well led to reductions of 4 percent in patients’ out-of- known,17 the efficacy of publishing health care prices pocket costs and 5 percent in insurers’ total costs remains uncertain.18 One plausible concern is that for X-rays, computed tomography, and magnetic disclosing prices in markets where competition is weak resonance imaging services. Savings increased over may lead lower-priced providers to raise their prices, an effect that has been observed in other sectors.19 Another time while price variation narrowed. Savings were concern is that only a subset of health care services are twice as great for patients responsible for the full “shoppable” and few consumers shop for those services, cost of the procedure under their deductible. These although a majority say they would like to do so.20 Experts cost reductions translated to estimated savings have recommended ways of increasing the likelihood that of approximately $7.9 million for patients and $36 price transparency will have desired effects on consumer million for insurers over the five-year period.23 and provider behavior.21 commonwealthfund.org Report, December 2020 State All-Payer Claims Databases: Tools for Improving Health Care Value PART 27 Having an APCD is a huge resource for us. It helps us know where to look and what to do. Having good data allows us to avoid unintended consequences [of shifting costs from one sector to another]. …Regional comparisons give us a treasure map on where to start in identifying market distortions…If we can make commercial insurance more affordable by removing market distortions, then there are better incentives for people on Medicaid when a pay raise puts them off enrollment. John Bartholomew Colorado Department of Health Care Policy and Financing APCDs also can be used to help consumers make insurance purchasing decisions. Colorado’s state CASE EXAMPLE: ANALYZING DELIVERY insurance marketplace uses information about total SYSTEM PERFORMANCE IN WISCONSIN annual health care costs and out-of-pocket expenses from the state’s APCD in its Quick Cost and Plan Finder tool. It The Wisconsin Health Information Organization has helps customers compare health plans for people with honed its APCD data analytic tools over the years similar demographic characteristics, expected health care to meet the needs of its stakeholders, many of use, and prescription medication use. which are integrated delivery systems. Subscribers Enabling Value-Based Care Delivery and Health can benchmark the quality and efficiency of Care Improvement health care providers to identify opportunities for In Maine and Wisconsin, hospitals and health systems improving health system performance and market seeking to improve access to and quality of care and agility (prices are normalized to mask negotiated monitor their competitive market position are regular fee schedules). They can use the tools to answer users of APCDs. Officials report that such monitoring has caused some hospitals to modify their referral patterns. questions such as: How does this system stack up against competitors? What is causing variation in In several states, interest in using APCD data also has been growing among purchasers seeking to promote quality of care and resource use? What doctors do I value-based care and form purchasing alliances; insurers need to work with to improve quality and efficiency? evaluating opportunities to develop products that may Sophisticated data users can access the tools on a promote more responsive or competitive markets; and portal or download de-identified data directly into consultants, brokers, and vendors for testing new payment models and developing services to improve delivery their own IT systems for custom analyses. system performance and meet market needs (Exhibit 6). commonwealthfund.org Report, December 2020 State All-Payer Claims Databases: Tools for Improving Health Care Value PART 28 One of my constituents had to have a colonoscopy, and they had to pay for it out of their pocket, and I told them to go on to CompareMaine to see what facilities get paid what, because maybe there’s a cheaper option for them if they travel. And they found one, and it saved them thousands of dollars. Maine State Legislator (as related by Karynlee Harrington, executive director, Maine Health Data Organization) Exhibit 6. Example Uses of APCD Data for Value-Based Purchasing and Improvement • Assess cost drivers, provider prices, and out-migration of community services to support efforts to negotiate discounts or shift care to lower-cost local providers • Examine episode-of-care data to identify higher-quality and lower-cost providers for “centers of excellence” programs Purchaser & • Analyze pharmacy data to define the scope of cost reduction strategies such as generic drug substitution and Stakeholder specialty drug options Coalitions • Create primary care practice reports that allow providers to benchmark the utilization and quality of care they provide in comparison to peers • Identify collaborative opportunities to reduce the provision of low-value care through benefit design, education, and incentives • Examine statewide medical cost structure, distribution of services, and utilization patterns to guide product and benefit designs to lower costs and meet needs Health • Assess the opportunity for entering the individual insurance market with a value-based reimbursement model and Dental • Research utilization of dental services for potential enhancement of access to services by the underserved Insurers population • Build networks of high-value providers by comparing providers’ clinical performance relative to health care spending to achieve that performance • Assess out-migration of rural hospital services to determine the opportunity for adding local services so that residents can receive care in the community. • Develop pricing bundles for an episode of care to lower health care costs without sacrificing quality of service. • Evaluate the benefits and risks of participating in an accountable care organization (ACO) and commercial Health Care payer contracts Providers • Link Medicaid ACO enrollment data with APCD data (with patient consent) to identify those in need of and System outreach services that can help prevent ED visits • Provide data to support certificate-of-need applications, community health needs assessments, medical staff development, and facility and strategic planning • Assess the performance of potential referral providers to ensure that they offer high-quality care at a reasonable price • Understand the insurance and provider composition of various geographies to promote employer-provider connections and encourage a competitive market Consultants • Build high-value, low-cost physician networks by creating provider performance benchmarks to identify best Brokers performers and assess potential quality improvements and cost savings if providers can achieve the best Vendors performance • Create population-level benchmarks for cohorts based on geography, age, gender, and previous medical conditions for use in a mobile app that would allow users to know what they can expect to spend on health care Source: Author’s analysis of APCD websites and reports. commonwealthfund.org Report, December 2020 State All-Payer Claims Databases: Tools for Improving Health Care Value PART 29 CASE EXAMPLE: USING REFERENCE-BASED PRICES IN COLORADO The Colorado APCD contributed data to a multistate study conducted by the RAND Corporation that compared payments to hospitals by private health plans to the rates that Medicare would pay for the same services at those facilities.24 A county-level analysis by the Center for Improving Value in Health Care found that hospital payments ranged from just above Medicare rates (115%) to nearly six times Medicare payments (576%) across the state.25 Purchasing coalitions are using these data to assess the reasonableness of prices paid to hospitals and develop strategies to increase affordability. In western Colorado, the Peak Health Alliance negotiated a local hospital fee schedule based on the premise that lower prices would allow residents to receive more of their care in the community. Health insurers used the negotiated fees in customized insurance plans offering average savings of 20 percent for residents of Summit County in 2020.26 This data-driven approach would not have been feasible without the data available through the APCD. Supporting Public Health Monitoring and Improvement The Colorado APCD has been Several states’ health departments are active users of APCDs, for purposes such as: an essential tool for Peak Health Alliance, providing custom analyses • Estimating the prevalence and cost of care which guide Peak’s negotiations attributable to treated chronic conditions and the potential impact of preventing or delaying the and ultimately lead to Peak’s onset of chronic disease (Exhibit 7); achievement in reducing the price • Measuring rates of chronic care management of health insurance premiums for and cancer screenings (overall and by payer) and thousands of Coloradans. identifying providers with low rates for inclusion in quality improvement programs; Tamara Pogue, • Assessing the completeness of a state’s immunization registry and obtaining an accurate CEO, Peak Health Alliance count of the number of immunization providers in the state. Colorado and Wisconsin tapped their APCDs to report Local health departments are also potential users of on populations at risk of serious illness from COVID-19, APCDs, for community health needs assessments.27 which may help officials address the current pandemic Colorado’s Center for Improving Value in Health Care and plan for future outbreaks. Wisconsin’s Medicaid developed a template that local public health departments agency used this information to conduct outreach to can use to create county health profiles that are based on high-risk beneficiaries. Some states are reporting on public data, including cost and quality of care and chronic the epidemiology of COVID-19 testing and treatment condition data from the APCD. and associated changes in the use of health care and Several study sites are stepping up attention to public telemedicine during the pandemic.29 health in response to the COVID-19 pandemic.28 commonwealthfund.org Report, December 2020 State ExhibitAll-Payer 7 Claims Databases: Tools for Improving Health Care Value PART 210 Diabetes-Attributable Health Care Spending in Minnesota Exhibit 7. Diabetes-Attributable Health Care Spending in Minnesota Per-Person Health Care Spending Among All Per-Person Spending Minnesotans and People with Diabetes Associated with Diabetes, 2009 2015 by Type of Spending $20,000 $19,510 $4,000 $17,620 Rx $750 Rx Health Care Spending in $ $16,000 $3,000 $1,180 $12,000 $2,000 $8,000 Medical $6,120 $3,050 Medical $5,060 $2,570 $3,800 $3,750 $1,000 $4,000 $0 $0 Per person total Per person total Per person per year 2009 2015 spending, all people spending, among spending associated people with diabetes with diabetes, people with diabetes Source: Minnesota Department of Health, Treated Chronic Disease Prevalence and Costs in Minnesota. Mathematica Policy Research analysis of the Minnesota APCD and other data. Source: Minnesota Department of Health, Treated Chronic Disease Prevalence and Costs in Minnesota. Mathematica Policy Research analysis of the Minnesota APCD and other data. Providing Reliable Data For Health Care Research BENEFITS OF STATE APCDS and Evaluation Stakeholders report that they find value in using APCD State APCDs offer an attractive source of multipayer data data and analyses for a range of purposes and needs that can ultimately help drive health system change. Looking for studies of changes in health care coverage or financing across these varied use cases, there are three overarching when it is important to place findings in the context of the benefits of state APCDs: state policy or market environment (Appendix Exhibit 1. Raising stakeholder awareness and engagement: B). APCD data are useful to compare the differential Public reports and initiatives supported by state effects of policies or interventions across payer types, APCD data can help educate key stakeholders about such as evaluations of the impact of multipayer primary the needs and opportunities to improve health care care medical home and accountable care initiatives. 30 system performance. Access to health care cost and Researchers also may benefit from the contextual quality information can empower consumers and knowledge that APCD staff can offer about payers employers to play a more active role in purchasing coverage and choosing where to get care. Public and providers in a state. A project led by the National attention to variations in care and spending within a Bureau of Economic Research developed a Uniform Data state can change stakeholder behavior. Structure file format to allow standardized measurements 2. Fueling data-informed policymaking: State officials and produce comparable findings across state APCDs. 31 say that APCD data can help answer policy questions. commonwealthfund.org Report, December 2020 State All-Payer Claims Databases: Tools for Improving Health Care Value PART 211 Analyses using APCD data have helped officials To support the use of APCD data by external stakeholders, and legislators and their constituents understand some states sponsor user groups and maintain a log of cost drivers, access patterns, and quality of care so data issues and resolutions.32 Less experienced users that they can make fact-based decisions to improve typically prefer to receive standard data tables or reports, coverage, financing, and regulation of the health care which an APCD administrator must have the resources to system across the public and private sectors. produce even if paid for through user fees. 3. Generating knowledge for improvement: Research Four of five consumers surveyed think it is important using state APCD data creates new knowledge about for states to provide comparative information on health specific treatments, provider practices, facilities, care prices, but only one of five residents of a state with delivery systems or networks, payers, and geographic a medical shopping website knew of it.33 States such as regions. Evaluations of interventions and programs New Hampshire are learning how to make their websites using APCD data provide an evidence-based feedback user friendly and encourage their use.34 Maine’s “Right to loop for improving and refining approaches. Shop” law requires health insurers to offer incentives for members to select lower-cost, higher-quality providers, CHALLENGES TO REALIZING THE VALUE OF which may be facilitated by — and lead to greater use of — STATE APCDS the state’s transparency website.35 If We Build It, Will They Come? Data Limitations and Completeness The uptake of an APCD involves a convergence of factors: APCDs are complex databases that require a learning a perceived need for data by stakeholders coupled with curve to master. Health care claims data originate in an awareness of the APCD as a possible data source, the payment for services and so have limitations when fitness of the data to meet the particular need, and the used for research and analysis.36 Several APCD leaders ability to analyze or interpret and effectively apply the noted the importance of efforts to ensure the integrity data. Study sites have learned that stakeholders benefit and credibility of their analytic efforts (see Part 1 for from education on the uses to which APCD data can be the role of advisory committees in this regard).37 State put and how it can augment their existing sources of data. officials caution that reports using APCD data should This effort may be more effectual when timed to coincide document data limitations to help assure appropriate with an initiative or opportunity to act on the data. interpretations. The [APCD] has really helped our research team. In fact, we just got a highly competitive grant to help 125 primary care practices throughout Virginia do a better job of screening and counseling for unhealthy alcohol use. This study could help up to 1.25 million Virginians. We are using the APCD as a counseling and tracking tool for practices. The grant reviewers specifically commented on the value of having this data. … Alex Krist, MD, MPH Professor, Family Medicine and Population Health, Virginia Commonwealth University commonwealthfund.org Report, December 2020 State All-Payer Claims Databases: Tools for Improving Health Care Value PART 212 One key limitation is missing data. For example, federal Antitrust and Market Competition regulations restrict the submission of some substance A review by scholars at the University of California abuse treatment data to APCDs.38 Another is the Supreme Hastings College of Law found that APCDs generally fall Court’s decision exempting self-insured ERISA plans from within a “safe harbor” under federal antitrust guidelines state APCD submission requirements.39 However, APCDs when they report the average or median price paid for a can still lawfully receive claims data from some self- service by multiple payers.44 Most states require public insured plans. Administrators estimate that their APCDs 40 disclosure of APCD data uses while offering stakeholders represent the majority (50% to 90%) of their states’ insured equal opportunity to make approved use of the data. residents. While state APCDs can report on health care APCD administrators rely on advisory bodies to help spending for commercially insured residents, some states determine appropriate data uses that may advance a data collect additional data to report on total statewide health requester’s interests while also improving the functioning care spending trends.41 Policymakers have proposed of the health care system or market. Ultimately, broader creating a national or federated approach to address data regulatory rules determine the nature and benefits of gaps for all states, but such a project would likely face its market competition in a state. own limitations.42 INSIGHTS AND LESSONS LEARNED CASE EXAMPLE: MINNESOTA’S TOTAL Fulfilling the purposes of an APCD requires continuous HEALTH CARE SPENDING ESTIMATES learning and adaptive management. Study states have The Minnesota Department of Health periodically taken a range of approaches to meet stakeholders’ synthesizes summary data collected from a variety information needs. While states have learned from one of health care payers (including those that do not another, shared knowledge is more easily translated in contribute claims data to the APCD) to report on the technical than the policy context.45 States’ varied total health care spending trends in the state.43 experiences in pursuing a transparency agenda, for Because these summary data are inadequate example, shaped distinct courses even among geographic for in-depth analyses, the state uses its APCD in neighbors, sometimes requiring that they change direction a complementary fashion to examine spending in response to changes in the policy environment.46 drivers by type of service, geographic area, or Realizing the potential of an APCD requires building demographics. “The APCD is both a tool to dive relationships to understand and meet needs. Several down much deeper, but also to help triangulate leaders emphasized the importance of spending time whether aggregated numbers are precise enough with new and existing state legislators and legislative to tell the story,” says Stefan Gildemeister, the staff, particularly those serving on committees with Department of Health’s chief economist. jurisdiction over health and budget matters. This interaction provides an opportunity to learn about their It has been surprising just how much of a need there is for data from the APCD for a wide variety of projects, even after explaining all of the potential limitations. Michael Lundberg CEO, Virginia Health Information commonwealthfund.org Report, December 2020 State All-Payer Claims Databases: Tools for Improving Health Care Value PART 213 needs and interests and to describe how the APCD can be CONCLUSION used to answer specific policy questions. In many states, State APCDs continue to demonstrate their relevance. APCD staff or officials regularly interact with stakeholders Data from APCDs can inform state policy and support health care stakeholders as they try to “buy smart” and and organizations that have an interest in using APCD improve health system performance. States can use APCD data.47 APCD staff also need to maintain good working data and analyses to meet the need for fiscally responsible relationships with insurers that submit data and with spending on public programs and to guide changes in the providers that are the subject of public disclosure. 48 health care system brought on by the COVID-19 pandemic. HOW THIS STUDY WAS CONDUCTED Data Collection and Analysis: We conducted semistructured interviews with APCD leaders in each state and with select stakeholders (e.g., legislator, employer, Medicaid official) in some states. Interviews were recorded (with permission) and transcribed. Data derived from interviews and documentary sources were organized in cross-case displays for topical content analysis.49 Findings were validated and refined based on a comparison with other published literature and through review by interviewees. Site Selection: Based on a literature scan and expert advice, we selected eight U.S. states (Exhibit 1) whose APCDs are characterized by diverse approaches and contexts. The APCDs, which have been in operation for four to 17 years, were also selected to highlight relatively advanced uses of data. We excluded some states that are the subject of other research (Massachusetts, Rhode Island), that only recently implemented an APCD (Delaware), or that have a unique policy context (all-payer rate setting in Maryland). The states we chose represent New England, the Midwest, the South, and the West. Contextual Environments: Study states represent a variety of markets and public policies. Collectively, they tend to perform better than average among all states on rankings of health system performance (median 12; range 3 to 47), small group insurance market competition (median 16; range 1 to 36), and ensuring that information is available to the public (median 13; range 1 to 37), as well as on an assessment of health care price transparency laws (median grade C; range A to F). All but Wisconsin have expanded Medicaid under the Affordable Care Act. These factors suggest that most study states are amenable to adopting health reforms and policies to promote health system improvement, which may have influenced the creation of an APCD. State Ranking or Grade Ark. Colo. Maine Minn. N.H. Utah Va. Wisc. Average Median Health System Performance (1) 47 9 12 3 10 11 29 12 17 12 Insurance Market Competition (2) 26 12 17 14 32 36 5 1 18 16 Ensuring Data Is Available for Use (3) 17 1 4 10 37 11 14 35 16 13 Healthcare Price Transparency (4) D B A C A D C F C C Expanded Medicaid Under ACA (5) Y Y Y Y Y Y Y N Y Y Sources: (1) The Commonwealth Fund, Scorecard on State Health System Performance (2019) (1=highest performing state). (2) Kaiser Family Foundation, State Health Facts: Small Group Insurance Market Competition, Rank on Herfindahl-Hirschman Index (1=most competitive market). (3) Center for Data Innovation, The Best States for Data Innovation (2017). The rank is a composite of 9 indicators (1=best at making data available for public use). (4) Catalyst for Payment Reform and the Source on Healthcare Price and Competition, “2020 Report Card on State Price Transparency Laws,” 2020. (5) The Commonwealth Fund, Medicaid Expansion Status, 2019. commonwealthfund.org Report, December 2020 The Commonwealth State All-Payer ClaimsFund Databases: Tools for Improving Health Care How Value PART High 214 Is America’s Health Care Cost Burden? 14 APPENDIX EXHIBIT A STATE DATA RELEASES USERS OF APCD DATA OR ANALYSES Public Fee-Based Government External Consumer Price Transparency Website Custom Reports/Analysis/Data Mart Purchasers/Stakeholder Coalitions Standard Analytic Reports/Tables State and Local Health Agencies Consultants/Brokers/Vendors Public Use Files (Deidentified) Research Reports/Snapshots State Employee Benefit Plan State Insurance Agency State Attorney General State Medicaid Agency External Researchers Payers and Providers Research Datasets State Legislature Consumers Arkansas * X X X X X X X X X X X X Colorado X X X X X X X X X X X X X X X X Maine X X X X X X X X X X X X X X X X X Minnesota X X X X New Hampshire X X X X X X X X X X X X Utah * X X X X X X X X X X X X Virginia * X X X X X X X X X X X X X Wisconsin * X X X X X X X X X COUNT 3 8 4 4 6 6 8 6 7 5 2 7 6 6 6 4 7 Source: Author’s analysis. * Notes on Consumer Price Transparency: Arkansas pricing data are available from a third party (mymedicalshopper.com); Utah’s website is under development; Virginia publishes median prices by care setting only (not by provider); Wisconsin recently discontinued its consumer medical shopping website. commonwealthfund.org Report, December 2020 The Commonwealth State All-Payer ClaimsFund Databases: Tools for Improving Health Care How Value PART High 215 Is America’s Health Care Cost Burden? 15 APPENDIX EXHIBIT B Topic Example Findings of Research and Evaluation Using APCD Data Staff of the Federal Trade Commission used Colorado APCD data to analyze variation in complexity-adjusted hospital prices paid by commercial health insurers for five procedures. They uncovered significant variation Hospital Price in the prices paid by different insurers for the same service in the same hospital, and even greater variation in Variation prices across hospitals. The researchers estimated that insurers would spend 10 percent to 20 percent less for these procedures if they paid the lowest price that each hospital received.50 Using data from the Colorado and Utah APCDs, researchers compared the experiences of Medicaid beneficiaries in Colorado, which expanded Medicaid under the Affordable Care Act, and Utah, which did Effect of not. They found that beneficiaries in Colorado gained an additional two months of coverage and were 16 Medicaid percentage points less likely to experience a coverage disruption in a given year compared to beneficiaries Expansion in Utah.51 During this period, new mothers in Utah were more likely to lose their Medicaid coverage and they had fewer outpatient visits paid by Medicaid during the six months after childbirth compared to new mothers in Colorado.52 Researchers examined Maine APCD data for commercially insured and Medicare beneficiaries aged 50 to 75 Coverage of before and after passage of the Affordable Care Act, which eliminated patient cost-sharing for preventive Preventive care. This change in coverage reduced median out-of-pocket payments by $94 in rural areas and $63 in Care urban areas, which led to a 40 percent relative reduction in the disparity in colonoscopy rates between urban and rural areas.53 Using the Virginia APCD, researchers found that new long-term opioid prescription-filling behavior is common after orthopedic surgical procedures in patients who were not taking opioids preoperatively.54 Opioid Use Research using New Hampshire APCD data found that, among adults with office visits for noncancer low-back pain, those who visited a chiropractor were 55 percent less likely to fill a prescription for an opioid analgesic compared with those who did not receive chiropractic services.55 Researchers used data from the Maine APCD to examine the cost and outcomes of total hip replacement. Clinical Cost They found that newer muscle-sparing and minimally invasive surgical techniques were more effective than and Outcomes traditional methods and can reduce costs in combination with centralization or regionalization of services.56 Analysis of medically tailored meal delivery using data from the Colorado APCD showed reductions of 13 Social Service percent in 30-day hospital readmissions and 24 percent in total medical costs for those with three chronic Interventions conditions.57 Using Minnesota APCD data, researchers found that telemedicine visits increased more than sevenfold from Use of 2010 to 2015. In metropolitan areas, telemedicine visits were primarily direct-to-consumer services provided Telemedicine by midlevel providers and covered by commercial insurance. In nonmetropolitan areas, telemedicine was used primarily for services delivered by physicians to publicly insured populations.58 commonwealthfund.org Report, December 2020 State All-Payer Claims Databases: Tools for Improving Health Care Value PART 216 NOTES Database Frequently Asked Questions (CIVHC, Feb. 1. APCD Council, Interactive State Report Map, accessed 2020). May 1, 2020. 10. Julia Lerche and Ross Winkelman, Applicability of 2. APCD Council, APCD Showcase Case Studies All-Payer Claims Databases for Rate Review and Other (University of New Hampshire, n.d.). Regulatory Functions (State Health Reform Assistance Network, June 2014). 3. Funding was provided by the Robert Wood Johnson Foundation; see: Network for Regional Healthcare 11. Many states review insurer premium rate filings to Improvement, “Healthcare Affordability: Data Is the ensure they are adequate to meet the need for services Spark, Collaboration Is the Fuel,” Nov. 8, 2018. and are not unfairly discriminatory in application. Colorado recently enacted legislation to include 4. For example, the analysis showed that Colorado could affordability as a criterion in rate reviews, although save $141 million annually if total costs were reduced its implementation has been delayed during the to the multistate average; see: Center for Improving COVID-19 pandemic. Value in Health Care, “Colorado’s Health Care Costs Continue to Rise Above Other States (CIVHC, Nov. 8, 12. See National Conference of State Legislatures, 2018). Transparency of Health Costs: State Actions (NCSL, n.d.). The state price transparency websites described 5. Utah’s reports were disseminated by the Utah here should not be conflated with a recent federal Partnership for Value-Driven Healthcare, which regulation requiring hospitals to post their billed offered clinics practical guidance on how to use the charges, or with President Trump’s Executive Order information for improvement. (not yet implemented) that would require hospitals to 6. Minnesota Department of Health, Novel MDH disclose their negotiated payment rates — a step that Study Yields First Statewide Estimate of Potentially the health care industry opposes. Preventable Health Care Events (MDH, July 23, 2015). 13. Bundled prices for episodes of care are reported for a 7. John N. Mafi et al., “Low-Cost, High-Volume Health subset of procedures on these websites. For example, Services Contribute the Most to Unnecessary Health the Colorado Shop-for-Care website reports episode Spending,” Health Affairs 36, no. 10 (2017): 1701–4. prices for 22 relatively expensive procedures while 8. For example, the state of Virginia created a task force reporting facility-only prices for imaging procedures; of employers to act on the data as part of an initiative the CompareMaine website reports on episodes of led by the Virginia Center for Health Innovation in care for 12 surgical procedures. collaboration with health systems and networks; 14. In response to a legislative mandate, Utah’s state see: Virginia Center for Health Innovation, Employer auditor is developing a transparency website using Task Force Launched to Focus on Reducing Low-Value data from the state’s APCD. The Wisconsin Health Health Services (VCHI, July 9, 2019); Virginia Center Information Organization recently discontinued its for Health Innovation, Virginia Receives a $2.2M consumer shopping website due to lack of use and Grant to Tackle the Overuse of Unnecessary Health loss of state funding. The group continues to publish Care (VCHI, March 13, 2019). ratings of the quality and efficiency of primary care 9. Center for Improving Value in Health Care, HB 19-1174 practices — an application that is more aligned with Out-of-Network Bill: Colorado All Payer Claims the organization’s future direction. commonwealthfund.org Report, December 2020 State All-Payer Claims Databases: Tools for Improving Health Care Value PART 217 15. Zach Y. Brown, “Equilibrium Effects of Health Care 19. David Cutler and Leemore Dafny, “Designing Price Information,” Review of Economics and Statistics Transparency Systems for Medical Care Prices,” New 101, no. 4 (Oct. 2019): 699–712. In a related analysis, England Journal of Medicine 364, no. 10 (2011): 894-5; the author estimated that wider uptake of a public D. Andrew Austin and Jane G. Gravelle, CRS Report for website would produce greater “supply-side” effects Congress: Does Price Transparency Improve Market whereby providers are induced to reduce their prices Efficiency? Implications of Empirical Evidence in Other over time. See: Zach Y. Brown, An Empirical Model Markets for the Health Sector (Congressional Research of Price Transparency and Markups in Health Care Service, April 29, 2008). (University of Michigan, 2018). 20. Amanda Frost and David Newman, Spending on 16. The effects of a public website may differ from the Shoppable Services in Health Care (Health Care Cost effects of employers’ price transparency tools. Studies Institute, 2016); Ateev Mehrotra et al., “Americans of the latter generally find that payments are lower Support Price Shopping for Health Care, But Few for some services received by those who used a price Actually Seek Out Price Information,” Health Affairs comparison tool compared to those who did not. 36, no. 8 (2017): 1392–1400. However, few people used the tools and there was no observed reduction in overall spending in the 21. Ateev Mehrotra et al., “Consumers’ and Providers’ time periods measured (which may not have been Responses to Public Cost Reports, and How to Raise long enough to detect supply-side effects on prices). the Likelihood of Achieving Desired Results,” Health See: Christopher Whaley et al., “Association Between Affairs 31, no. 4 (2012): 843–51; Lovisa Gustafsson and Availability of Health Service Prices and Payments Shawn Bishop, Hospital Price Transparency: Making for These Services,” JAMA 312, no. 16 (2014): 1670–76; It Useful for Patients (The Commonwealth Fund, Feb. Sunita Desai et al., “Association Between Availability 12, 2019). of a Price Transparency Tool and Outpatient Spending,” JAMA 315, no. 17 (2016): 1874–81; 22. In 2010, based in part on the cost of procedures Anna D. Sinaiko, Karen E. Joynt, and Meredith B. revealed by the HealthCost website, the health Rosenthal, “Association Between Viewing Health insurer Anthem engaged in hardline negotiations Care Price Information and Choice of Health Care with a high-cost hospital. Media attention to price Facility,” JAMA Internal Medicine 176, no. 12 (Dec. variations reportedly led to public and employer 2016): 1868–70; Sunita Desai et al., “Offering a Price support for Anthem’s position, leading the hospital Transparency Tool Did Not Reduce Overall Spending to agree to a reduction in its rates. See: Ha Tu and Among California Public Employees and Retirees,” Rebecca Gourevitch, Moving Markets: Lessons from Health Affairs 36, no. 8 (Aug. 2017): 1401–7. New Hampshire’s Health Care Price Transparency 17. Judith H. Hibbard, Jean Stockard, and Martin Tusler, Experiment (California Health Care Foundation, April “Hospital Performance Reports: Impact on Quality, 7, 2014). Market Share, and Reputation,” Health Affairs 24, no. 4 (2005): 1150–60. 23. Zach Y. Brown, “Equilibrium Effects of Health Care Price Information,” Review of Economics and Statistics 18. Ateev Mehrotra, Michael E. Chernew, and Anna D. (Sept. 2019). Sinaiko, “Promise and Reality of Price Transparency,” New England Journal of Medicine 378, no. 14 (April 5, 24. Chapin White and Christopher Whaley, Prices 2018): 1348-54; David Blumenthal, Lovisa Gustafsson, Paid to Hospitals by Private Health Plans Are High and Shanoor Seervai, “Price Transparency in Health Relative to Medicare and Vary Widely: Findings from Care Is Coming to the U.S. — But Will It Matter?” an Employer-Led Transparency Initiative (RAND Harvard Business Review (July 3, 2019). Corporation, 2019). commonwealthfund.org Report, December 2020 State All-Payer Claims Databases: Tools for Improving Health Care Value PART 218 25. Center for Improving Value in Health Care, Regional Claims Databases,” American Journal of Managed Price Information as a Percent of Medicare Now Care 25, no. 5 (May 1, 2019): e138-e144. Available (CIVHC, Sept. 9, 2019). 32. For example: University of Colorado researchers 26. Colorado Health Institute, Peak Health Alliance: sponsor bimonthly APCD user group meetings, in Summit in Sight (CHI, Nov. 11, 2019). Colorado collaboration with the Center for Improving Value in instituted a state-based reinsurance program for the Health Care, at which stakeholders can get updates individual insurance market in 2020. Taking into about the APCD and share their experiences on how account the effect of reinsurance, the total premium to make effective use of the data. savings for Summit County residents purchasing Peak 33. David Schleifer, Rebecca Silliman, and Chloe Rinehart, Health Alliance plans amounted to 39 percent to 47 Still Searching: How People Use Health Care Price percent; see: Colorado Department of Regulatory Information in the United States (Public Agenda, April Agencies, Summit County Residents’ Individual Health 6, 2017). Insurance Plans to See up to 47% Decrease in 2020 over 2019 Plans. 34. New Hampshire officials report that most of the new traffic to the HealthCost website is from organic 27. For example, the New Hampshire Health Web Google searches on health insurance questions, for Reporting and Query System included a Healthcare which the website offers frequently asked questions. Claims Indicator Report Module that drew data The website often appears as the top search response from the APCD for officials to use in community and visitors often then look to compare price health needs assessments; see: Association of State estimates for medical procedures. and Territorial Health Officers, Case Study: New 35. Roslyn Murray, Suzanne Delbanco, and Jaime S. Hampshire’s All-Payer Claims Database (ASTHO, n.d.). King, Promoting Health Care Transparency via State 28. Tanya Bernstein and Mary Jo Condon, Fighting Legislative Efforts (JAMA Health Forum, March 9, COVID — Using Claims Data for Tracking, Insights and 2020). Directing State Responses (Freedman Healthcare, July 36. Gerald F. Riley, “Administrative and Claims Records 22, 2020). as Sources of Health Care Cost Data,” Medical Care 47, 29. For example, see: Center for Improving Value no. 7 Supplement 1 (2009): S51-S55. in Health Care, Potential Impact of COVID-19 37. A summary of state approaches to assure APCD Temporary Cessation of Elective Procedures (CIVHC, credibility can be found in: California Office of July 22, 2020); Utah Office of Health Care Statistics, Statewide Health Planning and Development, Health Preliminary COVID-19 Healthcare Trends: A Snapshot Care Payments Data Program: Report to the Legislature from Utah’s All Payer Claims Database (OHCS, Aug. 25, (OSHPD, March 9, 2020): 139. 2020). 38. National Association of Health Data Organizations 30. For example, see: State Health Access Data Assistance and the APCD Council, Confidentiality of Substance Center, Evaluation of the Minnesota Accountable Use Disorder (SUD) Patient Records: Final Rule Health Model: Final Report (Minnesota Department of Guidance for State Health Data Organizations Human Services, Sept. 2017). (NAHDO, 2017). 31. Maria de Jesus Diaz-Perez et al., “Producing 39. United States Supreme Court, Gobeille v. Liberty Comparable Cost and Quality Results from All-Payer Mutual Insurance Co., Inc. commonwealthfund.org Report, December 2020 State All-Payer Claims Databases: Tools for Improving Health Care Value PART 219 40. States can require data submission from the third- privacy concerns. Purchasers say they appreciate the party administrators (TPAs) of self-insured public expertise the state has developed and hope to see employers not subject to ERISA including cities, broader use to support value-based purchasing. counties, schools, and state employee benefit plans. 47. For example, the Virginia Center for Health Several APCDs collaborate with employer groups to Innovation, a collaborative of public and private encourage voluntary submission of claims data by the stakeholders that seeks to accelerate the adoption TPAs of ERISA plans. of value-driven care, partners with Virginia Health 41. Massachusetts projects statewide spending using Information to fund analyses of low-value care aggregate trend data reported by insurers, which and apply the results for improvement. Similarly, is available sooner than claims data. This approach the Maine Health Data Organization collaborates is workable because a small number of dominant with the Maine Health Quality Forum to develop insurers also serve as third-party administrators for and operate its CompareMaine consumer shopping many self-insured plans in the state; see: Lisa Waugh website. These symbiotic relationships allows each and Douglas McCarthy, How the Massachusetts organization to focus and excel in its area of expertise. Health Policy Commission Is Fostering a Statewide Commitment to Contain Health Care Spending Growth 48. Several states offer convenings with data submitters (Commonwealth Fund, March 2020). to discuss and identify opportunities to improve data submission. Based on feedback from providers, 42. One such proposal was included in the Lower Health Colorado’s Center for Improving Value in Health Care Care Costs Act of 2019 (S. 1895) approved by the Senate established a process for previewing transparency Health, Education, and Labor Committee on June data with stakeholders before releasing it to the 26, 2019; it has not received further consideration in public. Congress. For an in-depth analysis of the potential value of a national APCD, see: Matthew Fiedler 49. Mathew B. Miles and A. Michael Huberman, “Cross- and Christen Linke Young, Federal Policy Options Case Displays: Exploring and Describing,” Chapter to Realize the Potential of APCDs (USC-Brookings 7 in: Qualitative Data Analysis, 2nd Ed. (Sage Schaeffer Initiative for Health Policy, Oct. 2020). Publications, 1994). 43. Minnesota Department of Health, Minnesota Health 50. Matthew Panhans, Ted Rosenbaum, and Nathan E. Care Spending: 2015 and 2016 Estimates and Ten-Year Wilson, “Prices for Medical Services Vary Within Projections (Report to the Minnesota Legislature, Feb. Hospitals, but Vary More Across Them,” Medical Care 2019). Research and Review, June 19, 2019. 44. Katherine L. Gudiksen, Samuel M. Chang, and 51. Sarah H. Gordon et al., “The Impact of Medicaid Jaime S. King, The Secret of Health Care Prices: Why Expansion on Continuous Enrollment: A Two-State Transparency Is in the Public Interest (California Analysis,” Journal of General Internal Medicine 34, no. Health Care Foundation, July 2019). 9 (Sept. 2019): 1919–24. 45. For example, officials in New Hampshire, the 52. Sarah H. Gordon et al., “Effects of Medicaid Expansion first state to develop a medical shopping website, on Postpartum Coverage and Outpatient Utilization,” shared its methodology and computer code with Health Affairs 39, no. 1 (Jan. 2020): 77–84. counterparts in Maine. 53. Annie Haakenstad et al., “Rural-Urban Disparities 46. For example, the Minnesota legislature created the in Colonoscopies After the Elimination of Patient state’s APCD to report on provider performance Cost-Sharing by the Affordable Care Act,” Preventive but repurposed it for research due to industry and Medicine 129 (Dec. 2019): 105877. commonwealthfund.org Report, December 2020 State All-Payer Claims Databases: Tools for Improving Health Care Value PART 220 54. Noah J. Orfield, “New Long-Term Opioid Prescription- Filling Behavior Arising in the 15 Months After Orthopaedic Surgery,” Journal of Bone and Joint Surgery 102, no. 4 (Feb. 19, 2020): 332–9. 55. James M. Whedon et al., “Association Between Utilization of Chiropractic Services for Treatment of Low-Back Pain and Use of Prescription Opioids,” Journal of Alternative and Complementary Medicine 24, no. 6 (June 1, 2018): 552-6. 56. Jonathan P. Goldstein et al., “The Cost and Outcome Effectiveness of Total Hip Replacement: Technique Choice and Volume-Output Effects Matter,” Applied Health Economics and Health Policy 14, no. 6 (Dec. 2016): 703–18. 57. Project Angel Heart, Small Intervention, Big Impact: Health Care Cost Reductions Related to Medically Tailored Nutrition (PAH, June 2018). 58. Yu Jiani et al., “Population-Level Estimates of Telemedicine Service Provision Using an All-Payer Claims Database,” Health Affairs 37, no. 12 (Dec. 2018): 1931-9. commonwealthfund.org Report, December 2020 State All-Payer Claims Databases: Tools for Improving Health Care Value PART 221 ABOUT THE AUTHOR Maine Health Data Organization: Karynlee Harrington, Douglas McCarthy, M.B.A., is senior research advisor for the executive director Commonwealth Fund and president of Issues Research, Inc., Minnesota Department of Health: Stefan Gildemeister, Ph.D., in Durango, Colorado. He has supported the Commonwealth director, Health Economics Program; and Karl Fernstrom, Fund’s work on a high-performance health system since manager, Health Data Services Center 2002 through the development of chartbooks and scorecards on health system performance and case study research on Minnesota Health Action Group: Deb Krause, vice president promising practices and innovations in health care delivery. Mr. McCarthy’s 30-year career has spanned roles in government, Minnesota Legislature: Scott Jensen, state senator corporate, and nonprofit organizations, including the Institute National Association of Health Data Organizations: Norman for Healthcare Improvement and UnitedHealth Group’s Center Thurston, Ph.D., executive director; and Denise Love, consultant for Health Care Policy and Evaluation. He was a public policy fellow at the University of Minnesota’s Humphrey School of New Hampshire Department of Health and Human Services: Public Affairs and a leadership fellow of the Denver-based Mary Fields, CHIS project manager and business systems analyst Regional Institute for Health and Environmental Leadership. Mr. McCarthy serves on the boards of Colorado’s Center for New Hampshire Insurance Department: Tyler Brannen, Ph.D., Improving Value in Health Care and the Peak Health Alliance. director of Health Economics; and Maureen Mustard, director of Healthcare Analytics University of New Hampshire, Institute for Health Policy and Editorial support was provided by Paul Berk. Practice: Jo Porter, director and cochair, APCD Council; and Ashley Wilder, project director Utah Department of Health, Office of Health Care Statistics: ACKNOWLEDGMENTS Carl Letamendi, Ph.D., bureau director; and Sterling Petersen, analytics lead The author thanks Lovisa Gustafsson, vice president of the Controlling Health Care Costs program at the Commonwealth Virginia Health Information: Michael Lundberg, CEO; and Kyle Fund, for guidance on the project, and the Communications Russell, director of Strategy and Analytics staff for editing and production. The author is grateful to the following organizations and individuals for sharing information Wisconsin Health Information Organization: Dana Richardson, and insights for the reports: CEO; and Jim Auron, director of Customer Solutions Arkansas Center for Health Improvement: Joseph Thompson, M.D., president & CEO; and Craig Wilson, J.D., director of Health Policy For more information about this case study, Arkansas Insurance Department: Allen Kerr, former please contact: commissioner; and Lesia Carter, assistant director and APCD Douglas McCarthy liaison Senior Research Advisor The Commonwealth Fund Center for Improving Value in Health Care (Colorado): dmcmwf.org Ana English, CEO & president; Cari Frank, vice president of Communication and Marketing; and Kristin Paulson, J.D., vice president of Innovation and Compliance See the first brief in this series for insights and Colorado Department of Health Care Policy & Financing: John lessons learned about how states establish APCDs Bartholomew, chief financial officer and equip them to meet their intended uses. Freedman Healthcare: Linda Green, executive vice president; and Jonathan Mathieu, Ph.D., senior consultant See the companion state profiles for more information on each state’s APCD. commonwealthfund.org Report, December 2020 About the Commonwealth Fund The mission of the Commonwealth Fund is to promote a high-performing health care system that achieves better access, improved quality, and greater efficiency, particularly for society’s most vulnerable, including low-income people, the uninsured, and people of color. Support for this research was provided by the Commonwealth Fund. The views presented here are those of the author and not necessarily those of the Commonwealth Fund or its directors, officers, or staff.