December 2016 Data BRIEF TRENDS IN PHYSICIAN NETWORKS IN THE MARKETPLACE IN 2016 Daniel Polsky, Yuehan Zhang, Laura Yasaitis, and Janet Weiner In this brief, we describe the breadth of physician provider networks offered on the health insurance marketplaces in 2016, and present differences by plan type, physician specialty, and state. We also compare networks in 2016 to those in 2014. We find little change in overall prevalence of narrow networks, but we find important geographic shifts and a trend towards x-small networks among plans with narrow networks. We discuss the policy implications of our findings for consumers, regulators, and health plans. INTRODUCTION In this brief, we describe the physician provider for plans of otherwise equivalent design, a plan networks offered in the marketplace in 2016, with an extra-small network had a monthly With the fourth open enrollment period and compare how networks have changed premium that was 6.7% less expensive than underway and with election results behind from 2014 to 2016. We describe the steps we that of a plan with a large network. For a typical us, the health insurance Marketplace is took to develop comparable data across years, plan, consumers were saving between $212 and experiencing uncertainty, instability, and turmoil. and present summaries of network size overall, $339 a year. In 2017, premiums increased by 22% and many by plan type, specialty, and by state. issuers declined to participate in the public To date, consumers have had little indication marketplace. The designs of the plans offered, of network size when choosing a plan. Many however, have remained fairly stable, reflecting marketplaces have a feature that allows a regulatory environment that has community BACKGROUND consumers to search for a specific provider or rating, essential health benefits, standardized Because the breadth of a provider network can to see all participating providers by specialty, actuarial levels characterized by metal level, and dictate how consumers access care covered but the overall breadth of the network remains no dollar limits on benefits. Insurers still have by their health plan, they should be aware of opaque. To address the issue of transparency, flexibility in plan design through the provider the breadth of network of the plan they are for Plan Year 2017 the Centers for Medicare networks of qualified plans. choosing. While network breadth is not the and Medicaid Services (CMS) is piloting a only characteristic of a provider network, we display of network breadth information on It is possible that the breadth of plan networks showed in our first brief on this topic that this the marketplaces in four states: Maine, Ohio, has changed, either because insurers have measure is easily calculated and can quickly Tennessee, and Texas. During open enrollment, increased their offerings of narrow networks, or capture the relative differences in provider consumers in these states see information because insurers with broader networks have networks across plans. classifying the breadth of the plans’ provider disproportionately exited the marketplaces. networks, as compared to other plans in the The only longitudinal data we have comes Transparency of provider networks is county. Consumers can compare networks for from McKinsey & Co., who categorized particularly important given the price sensitivity three provider types, including adult primary network size by the proportion of participating of Marketplace consumers, who tend to care providers, pediatricians, and hospitals. The hospitals in a rating area. They found that prefer lower-premium plans, and lower price new labels categorize a network as ‘Standard’ the proportion of plans with narrow hospital plans tend to be narrow network plans. In a if within a standard deviation of a baseline networks (defined as a network with fewer Health Affairs article, we quantified how much Provider Participation Rate, or ‘Broad’ or ‘Basic’ than 70% of hospitals in a rating area) stayed consumers were saving by choosing a narrow if above or below it respectively. relatively flat from 2014-2016 (42% in 2014, 39% network on the marketplace. Within a market, in 2015, 43% in 2016). COLONIAL PENN CENTER | 3641 LOCUST WALK | PHILADELPHIA, PA 19104-6218 | LDI.UPENN.EDU | P: 215-898-5611 | F: 215-898-0229 | @PENNLDI DataBRIEF LDI Another concern raised by narrow networks 544 unique provider networks offered by process of collecting the 2014 data is described is one of adequacy. In a study of 2015 federal 292 different issuers. We obtained the list in our previous Data Brief. Because methods Marketplace plans, nearly 15% had no in- of providers participating in each of these of data collection and cleaning have improved network physicians within 50 miles for at least networks from Vericred, a healthcare data since that time, we returned to the 2014 file to one specialty. Endocrinology, rheumatology, services company. The provider network reconcile differences. This primarily required and psychiatry were the most common data used in this research was obtained by identifying one unique geographic location per excluded specialties. The ACA set a national Vericred in September 2016 either directly from provider based on SK&A data when matched, standard for network adequacy requiring “a insurers or through machine readable provider and based on NPI data when not matched. network that is sufficient in number and types directories released by the insurers. Because the 2014 data were collected for silver of providers,” and that “all services will be plans only, all comparisons are restricted to accessible without unreasonable delay.” But the Providers were matched to the National silver plans. interpretation of “sufficient” and “reasonable” Provider Identifier (NPI) registry, available from was left to the states. To guide state adequacy CMS, to obtain a consistently coded specialty, standards, in November 2015 the National provider type, and location. We restricted the list to matched providers and verified physicians QUANTIFYING PHYSICIAN Association of Insurance Commissioners updated its 1996 Managed Care Plan Network as active by matching listed physicians to the NETWORK SIZE Adequacy Model and renamed it the Health SK&A office-based physician dataset. The We estimate network size only for the parts Benefit Plan Network Access and Adequacy of a state where plans are sold using that Model Act. The Act specifies that state particular provider network. Network size insurance commissioners, not health plans, CONSUMERS SAVED is estimated by the ratio of the number of determine if provider networks are adequate, physicians participating in each network to sets standards for the accuracy of provider 6.7% ON PREMIUMS BY the total number of physicians eligible for that directories, and includes consumer protections network in each state. A physician’s eligibility against “surprise medical bills” when out-of- CHOOSING NARROW to be included in a network was determined by network providers deliver care in in-network NETWORK PLANS, BUT whether he or she was practicing in a county facilities. However, the NAIC model act did where a plan associated with the network was not recommend quantitative standards of CONCERNS ABOUT sold. Likewise, participating physicians were only adequacy, nor is it binding on states. counted in the numerator of this measure if their TRANSPARENCY AND practice location was within a county where a Amid the slow progress addressing plan associated with the network was sold. As transparency and adequacy of networks, two ADEQUACY REMAIN. in 2014, we categorized network size into five recent studies quantified the changes in health groups using arbitrary cutoffs that might provide care use and spending attributable to a narrow meaningful information to consumers: x-small network. In the context of a state employee SK&A dataset telephone verifies location and (< 10%), small (10%-25%), medium (25%-40%), plan, Gruber & McKnight found that consumers specialty information every six months and thus large (40%-60%), and x-large (≥ 60%). were very price sensitive when given the option provides validated, updated, and consistently of a narrow network plan, and that it was an collected specialty and location information for We assess the number of plans with networks effective strategy for controlling costs. Those 594,776 physicians. For these physicians we use of each size. Networks are typically attached to who switched to a narrow network plan (10% of SK&A specialty and location information. multiple plans, but we use the plan as the unit employees) spent almost 40% less on medical of analysis. (Our first brief used the network care, with savings primarily coming from We excluded 13 networks managed by 11 as the unit of analysis, but because consumers specialist and hospital care. In the context of a issuers where we had no valid data. Our purchase plans rather than networks, we small group market, Atwood & Lo Sasso found analysis dataset consisted of 523,503 physicians believe plans are the most appropriate level an overall spending reduction of 25% associated participating in plans issued by the 281 carriers of analysis.) To adjust for the fact that some with enrollment in a narrow network plan, with across 531 networks. We matched 392,856 plans are only offered regionally within a state reductions in primary care, specialist care, allied of these physicians to the SK&A file. For the while others are sold state-wide, we summarize health, and prescription drugs. 130,557 unmatched, we used the specialty and plans by weighting by the fraction of the state’s location information from the NPI data. We population living in counties where the plan also created a dataset of the 229,644 physicians was offered. We chose this approach as it that were found to not be participating in any reflects consumers’ experiences in choosing WHAT WE DID marketplace network and were verified as active between different plans, rather than networks. From the 2016 list of all 5,022 qualified health office-based physicians by the SK&A data. We examined network sizes associated with plans (and 108,448 unique plan/county different types and levels of plans, as well as combinations) sold in the marketplaces for In addition to describing the networks in among several different physician specialty all 50 states plus DC as provided by the the Marketplace in 2016, we compared how subsets. RWJF HIX Compare dataset, we identified networks have changed from 2014 to 2016. The 2 DataBRIEF LDI WHAT WE FOUND Figure 1. Network size categories, overall and by metal The distribution of physician networks, overall and by metal tier, in 2016 are shown in Figure 1. PhysicianNetworkSize:OverallandbyMetalin2016 By our measures, 31% of networks are small or 100% 10 x-small: 12% of networks are x-small, meaning 90% 15 17 16 14 17 they include less than 10% of office-based 80% 28 X-large:>=60% practicing physicians in the area and another 70% 31 30 32 30 31 19% are small, including between 10% and 25% 60% Large:40-59.9% 21 of physicians. At the other end of the spectrum, 50% 24 25 Medium:25-39.9% 40% 22 24 23 15% are x-large, which we define as networks 30% Small:10-24.9% that include at least 60% of physicians. There 19 19 17 20 32 17 20% is little difference in network breadth across 10% X-small:<10% metal tiers, except for the platinum plans, which 0% 12 12 12 11 9 12 feature more small networks. However, the Overall Gold Silver Bronze Platinum Catastrophic 100% 23% 35% 30% 5% 7% platinum tier has just 5% of the plans offered. Most networks offered on the marketplace are Preferred Provider Organizations (PPOs) or Figure 2. Network size categories, overall and by plan type Health Maintenance Organizations (HMOs): 28% are PPOs, 53% are HMOs, 9% are Exclusive Provider Organizations (EPOs), and PhysicianNetworkSize:OverallandbyPlanTypein 10% are Point of Service (POS) plans. Within 2016 each plan type we categorize network size, 100% 3 as shown in Figure 2. We found meaningful 90% 15 14 13 differences across type, with 44% of HMOs 80% 34 25 having x-small or small networks, compared to 70% 31 35 X-large:>=60% 50 35% of PPOs. HMO plans have twice as many 60% 28 Large:40-59.9% 50% 30 x-small networks as PPOs (14% vs. 7%). POS 40% 24 16 Medium:25-39.9% plans have a surprisingly large proportion of 30% 30 9 19 21 Small:10-24.9% x-small plans (27%), while on the other extreme, 20% 19 27 12 EPOs have a large proportion of large and 10% 12 7 7 14 6 X-small:<10% 0% x-large networks (63%). Overall PPO HMO POS EPO 100% 28% 53% 10% 9% We then characterize network size within broad specialty group categories and display results in Figure 3. The most common specialty groups among practicing physicians are primary care Figure 3. Network size categories, overall and by provider specialty group (29%), hospital-based (17%), and surgery- related (16%). We find a striking similarity in PhysicianNetworkSizeforSelectedSpecialtyGroupsin2016 network sizes across specialty groups with the exception of hospital-based specialties 100% 90% 15 21 7 6 11 18 34 33 32 31 (radiology, anesthesiology, emergency 80% 70% 31 37 36 24 12 28 30 medicine, and pathology). Network size for 60% 50% 20 24 25 24 24 24 24 23 X-large:>=60% primary care physicians is very similar to overall 40% 24 18 16 16 19 21 17 20 22 20 Large:40-59.9% 30% network size with 31% having x-small or small 20% 19 16 13 13 12 13 15 19 49 19 Medium:25-39.9% 23 networks. For pediatrics, only 23% of plans have 10% 0% 12 10 11 12 11 10 13 12 13 Small:10-24.9% small or x-small networks. With the notable X-small:<10% exception of psychiatrists, specialist groups are less likely to be narrow networks compared to primary care groups. The one striking outlier are hospital-based specialty groups, where 72% of plans have either x-small or small networks. This is notable given that this is the group of physicians most likely to lead to a surprise out- of-network bill. 3 DataBRIEF LDI Figure 4. State-level percentage of narrow networks (plans associated with network We found meaningful differences in the sizes < 25%) prevalence of narrow networks by state. In Figure 4, we summarize this information using states grouped by their propensity to offer PercentofNarrowNetworksbyStatein2016 narrow networks (x-small or small) in their plans. 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% As shown, plans in some states offer mostly narrow networks, (KY, VA), while in other states, KY VA narrow networks are quite rare or non-existent. CA NV CO Comparing Network Size in 2014 and 2016 AZ We found relative overall stability in network NY OH size when we compared 2014 and 2016 silver MA plans, except among the narrowest networks, OK HI where the number of plans with x-small NH IL networks doubled from 6% to 12%. As shown in AR Figure 5, the sum of x-small and small networks FL IA declined slightly: from 31% to 29% and the TX fraction of plans with larger network sizes also MN GA remained remarkably stable. Thus the doubling NE RI of x-small networks was matched by the KS reduced prevalence of small networks. CT WA NC Stateswith 60%+Narrow What might account for the increase in x-small ME Networks TN networks? We explored shifts in the distribution IN Stateswith 40- <60%Narrow of plan types given the noted higher prevalence OR Networks NM of narrow networks among HMO plans. We MI PA Stateswith 20- <40%Narrow find that the prevalence of PPOs and HMOs MD Networks changed from 2014 to 2016: 29% are PPOs WV LA Stateswith <20%Narrow (down from 36% in 2014) and 51% are HMOs MO Networks (up from 46% in 2014). As shown in Figure 6, AL ND the shifts in the prevalence of different plan WI types are not as meaningful as the shifts within UT SC plan type. We find the greatest jumps in x-small AK DC networks among HMO plans and POS plans. DE ID On a state level, the propensity to offer narrow MS MT networks changed considerably, with narrow NJ SD networks emerging in some states that had none VT (IA, AR, NH) while disappearing in others in WY which they had been prevalent (for example, NJ, AK). Figure 7 shows these differences by state. Figure 5. Comparison of network size for silver plans in Figure 6. Comparison of network size for silver plans overall and by plan 2014 and 2016 type in 2014 and 2016 PhysicianNetworkSizeforSilverPlansin PhysicianNetworkSizeforSilverPlans: 2014and2016 OverallandbyPlanTypein2014,2016 100% 7 3 3 100% 90% 15 16 12 14 12 26 24 90% 15 16 80% 35 24 30 80% 70% 29 32 39 34 X-large:>=60% 60% 52 X-large:>=60% 70% 29 32 27 30 32 50% 31 Large:40-59.9% 12 Large:40-59.9% 60% 40% 25 24 48 17 50% 30% 22 27 Medium:25-39.9% 25 Medium:25-39.9% 19 36 8 28 19 40% 24 20% 25 17 16 Small:10-24.9% 30% 10% 7 17 23 10 Small:10-24.9% 6 12 7 7 16 12 6 0% 5 2 X-small:<10% 20% 25 17 10% X-small:<10% 2014 2016 2014 2016 2014 2016 2014 2016 2014 2016 6 12 100% 100% 36% 29% 46% 51% 11% 10% 7% 11% 0% 2014 2016 Overall PPO HMO POS EPO 4 DataBRIEF LDI Figure 7. State-level percentage of narrow networks for silver plans in 2014 and 2016 POLICY IMPLICATIONS We found narrow physician provider networks in 31% of all qualified silver plans offered in PercentofNarrowNetworksforSilver 2016, compared to 29% in 2014. While the PlansbyStatein2014,2016 average prevalence of narrow network plans has remained stable over time, there are important 100% 80% 60% 40% 20% 0% 20% 40% 60% 80% 100% shifts in the plans with narrow networks. X-small KY networks have doubled from 6% of silver plans VA CA to 12% of silver plans. On a state level, the NV propensity to offer narrow networks changed OK AZ considerably, with narrow networks emerging in NY CO some states that had none (IA, AR, NH) while IA MA HI disappearing in others in which they had been FL prevalent (for example, NJ, AK). IL OH NE The trend of more x-small networks highlights AR NH key issues for narrow networks going forward. GA KS Enforcing adequacy standards and further TX WA developing enforceable adequacy standards MN RI in all states has been a challenge, but the NC increasing prevalence of x-small networks ME CT adds pressure and immediacy to the task. OR IN Providing greater transparency regarding TN MI networks is critical for consumers who shop NM PA based on price and possibly whether their MD WV primary care physician is in their desired plan. MO ND These consumers, when seeking care, may find LA 2016plans WI themselves with a plan they would not have AL 2014plans selected had they know more about the choices SC UT of physicians available to them. The CMS AK DC pilot to label network breadth and offer this DE ID information to consumers when selecting plans MS MT is an important development. NJ SD VT The high prevalence of narrow networks among WY hospital-based physicians, however, is stunning. Given that these physicians are the ones most likely to send surprise out-of-network bills, this remains a concern for those with narrow network plans and broad plans. ABOUT LDI The emergence of narrow networks is an important health plan innovation, as it offers Since 1967, the Leonard Davis Institute of Health Economics (LDI) has been the leading the opportunity for providing lower-priced university institute dedicated to data-driven, policy-focused research that improves our nation’s plans in the marketplace. For this innovation to health and health care. Originally founded to bridge the gap between scholars in business be executed fairly and safely for consumers, it (Wharton) and medicine at the University of Pennsylvania, LDI now connects all of Penn’s must be accompanied by continued innovation schools and the Children’s Hospital of Philadelphia through its more than 200 Senior Fellows. among regulators to ensure transparency, network adequacy, and elimination of surprise LDI Data Briefs are produced by LDI’s policy team. For more information please contact Janet out-of-network bills. Weiner at weinerja@mail.med.upenn.edu. COLONIAL PENN CENTER | 3641 LOCUST WALK | PHILADELPHIA, PA 19104-6218 LDI.UPENN.EDU | P: 215-898-5611 | F: 215-898-0229 | @PENNLDI 5