Hospital networks: Updated national view of configurations on the exchanges

Articles

This updated view of the network configurations being offered on public exchanges across the country suggests that consumer choice of health plan design is expanding.

In December 2013, we released an Intelligence Brief on exchange hospital network trends based on analyses of the silver-tier plans offered in 20 geographically dispersed urban areas.1 We have since enhanced our hospital network database to include all products in all tiers in all 501 rating areas in the U.S., which has given us a comprehensive view of the exchange network landscape. By leveraging the new database, we were able to generate additional insights into local market differences and patterns of network formation. We augmented these insights with findings from our April national open enrollment period (OEP) consumer survey, which explored (among other things) consumers’ purchasing patterns during the 2014 OEP.2

Our database includes all 282 payors filing on the 2014 exchanges and all 4,773 acute care hospitals in the U.S.3 The payors offered a total of 20,818 on-exchange products across the five metal tiers; these products included 2,366 unique individual exchange networks. These networks had to meet adequacy requirements that were in place before, and then expanded by, the ACA; these requirements include the minimum number and types of providers, and the maximum driving distance and wait time, to ensure that patients have adequate access to care.4

For a few of the more detailed analyses discussed in this Intelligence Brief, we focused on silver-tier networks specifically. The vast majority of exchange networks (93 percent) are offered on the silver tier, and this tier is the only one in which income-eligible consumers can receive both federal premium and cost- sharing subsidies. More than 60 percent of all consumers who enrolled in an exchange product chose one in the silver tier.

For comparison with the 2014 hospital networks offered on individual exchanges, we used network data from the 2013 individual market products of all incumbent payors. In our analyses, we categorize each network based on the extent of hospital participation, as follows: broad networks have more than 70 percent of all hospitals in the rating area participating, narrow networks have 31 to 70 percent of all hospitals in the rating area participating, and ultra-narrow networks have 30 percent or less of all hospitals in the rating area participating. We classified a network as tiered if the payor put different hospitals into different tiers with different co-payment requirements. In the remainder of this Intelligence Brief, we use the phrase narrowed network to refer to narrow, ultra-narrow, and tiered networks in the aggregate.5

Seven key observations emerged from our analyses:

  • Consumers now have an expanded choice of network offerings at the point of health plan purchase on exchanges. Broad networks are available to close to 90 percent of the addressable population.6 In addition, narrowed networks are available to 92 percent of that population; they make up about half (48 percent) of all exchange networks across the U.S. and 60 percent of the networks in the largest city in each state. The increased prevalence of narrowed networks gives consumers a wider range of value propositions and prices among health insurance plans. But, if a consumer purchases a narrowed network product, then at the point of access, the choice of providers is reduced.
  • Compared to plans with narrowed networks, products with broad networks have a median increase in premiums of 13 to 17 percent (when the analysis is controlled for payor, product type, rating area, and metal tier); the maximum increase is 53 percent. Across the country, close to 70 percent of the lowest- price products are built around narrow, ultra-narrow, or tiered networks.7
  • There is no meaningful performance difference between broad and narrowed exchange networks based on Centers for Medicare and Medicaid Services (CMS) hospital metrics such as the composite value-based purchase score as well as its three sub-components (outcome, patient experience, and clinical process scores).8 However, broad networks have higher rates of academic medical center participation.
  • Certain market conditions are associated with a greater prevalence of narrowed networks—specifically, higher excess bed capacity, greater provider or payor fragmentation, and more significant potential for growth from the uninsured than from people who previously had coverage. Each of these market conditions is associated with 1.4 to 1.9 times as many ultra- narrow networks, and the combination of factors is associated with an even higher prevalence of ultra-narrow networks (up to 4.7 times as many).
  • In those rating areas in which at least two different payors offer ultra-narrow, silver-tier networks,9 67 percent of the ultra-narrow networks share the majority of their hospitals (on average, over 80 percent) with at least one other ultra-narrow network. Fourteen percent of all acute-care hospitals participate in ultra-narrow networks; of them, 23 percent are in more than one such network.10
  • Seventy-five percent of all ultra-narrow, silver-tier products include only some of the hospitals within participating health systems. Forty-four percent of these products exclude at least one hospital from every single participating health system. Ultra-narrow networks excluding hospitals from every participating health system are priced an average of 13 percent lower than ultra-narrow networks containing entire health systems.
  • In our April consumer survey, 42 percent of the respondents who indicated they had enrolled in an ACA plan and were aware of the network type reported purchasing a product with a narrowed network. However, 26 percent of those who indicated they had enrolled in an ACA plan were unaware of the network type they had selected.
  1. “Hospital networks: Configurations on the exchanges and their impact on premiums,” McKinsey Center for U.S. Health Reform, December 2013.
  2. Our national consumer survey included 2,874 consumers eligible for qualified health plans. More details about the survey and the methodology it used can be found in our May 2014 Intelligence Brief, “Individual market: Insights into consumer behavior at the end of open enrollment.”
  3. Includes general, medical, and surgical hospitals, orthopedic hospitals, heart and cancer hospitals, ear, nose and throat hospitals, and children’s general hospitals, as defined by the American Hospital Association. See appendix for further detail.
  4. Before passage of the ACA, network adequacy requirements existed for HMOs in almost all states and for PPOs in about half of the states. The ACA set network adequacy requirements for all QHPs but left it to the states to define and regulate adequacy. Both national agencies and the states may continue to add regulations over time.
  5. Narrowing of provider networks can occur across hospitals or physicians. For the purposes of this Intelligence Brief, we have focused on hospital networks.
  6. Addressable population is defined as people who are eligible to purchase qualified health plans on the exchanges (i.e., non-elderly adults with incomes above 100 percent FPL in non-Medicaid expansion states and above 138 percent FPL in Medicaid expansion states).
  7. Out of all rating areas where ultra-narrow or narrow networks are present (329 of the 501 rating areas). For each rating area, when the same payor offered multiple products based on the same network, the lowest-price product was used to determine the price of the network. Payor count represents unique payors at a state level. See methodology in the appendix for further details.
  8. Specifically, we used the following metrics: the outcomes score of 30-day mortality rate for acute myocardial infarction, the patient-reported score of hospital rating, and the clinical process scores for surgery patient antibiotics delivery.
  9. 211 of the 501 rating areas in the U.S.
  10. Of all acute-care hospitals in the U.S., 96 percent participate in an exchange network—84 percent in broad, 38 percent in narrow, and 14 percent in ultra-narrow networks. (Numbers add up to more than 100 percent because some hospitals are in multiple networks).