Providing Insights that Contribute to Better Health Policy
Insurance Coverage & Costs Access to Care Quality & Care Delivery Health Care Markets Employers/Consumers Health Plans Hospitals Physicians Issue Briefs Data Bulletins Research Briefs Policy Analyses Community Reports Journal Articles Other Publications Surveys Site Visits Design and Methods Data Files
Tiered-Provider Networks: Patients Face Cost-Choice Trade-Offs
Issue Brief No. 71
Tiered Networks in Concept and Practice
aced with rapidly rising health care costs, employers and health plans are seeking new cost-containment strategies, including ways to boost plans negotiating leverage over hospitals and medical groups. One emerging approach is to develop health insurance products that group providers into tiers based on the cost or efficiency of care they deliver and then steer patients to choose these providers through lower premiums or cost sharing.
Conceptually similar to tieredpharmacy benefits now common in health plans, tiered-provider networks allow patients to make trade-offs between provider choice and costs rather than having health plans make these decisions. Tiered networks are expected to constrain overall health plan costs by steering patients to lower-cost providers, while encouraging hospitals and physicians to improve their efficiency or accept discounted payment rates in exchange for preferred-tier placement.
Sorting providers into tiers, however, is much more difficult than tiering pharmaceuticals with the same or similar therapeutic effects but different prices.1 Measuring the cost and efficiency of hospital and physician services is difficult. Providers in a given community can vary widely based on costs, quality and accessibility of care. Consequently, some fear tiered-provider networks could limit lower-income patients access to high-quality but costly providers. Moreover, tiered-network designs may prove difficult to maintain over time in the face of market pressure for broad and unrestricted provider choice.
More Health Plans Experiment
ine health plans in six of the 12 communities were experimenting with tiered networks by 2003 (see Data Source and Table 1). While some employers are interested in tieredprovider networks, uptake of existing products has been slow.Many plans also have encountered operational difficulties and provider resistance, raising questions about the ability of these new designs to reduce health care costs significantly.
Health plans in the 12 markets first started developing tiered-network designs in 2000-01, largely in response to the growing negotiating leverage of prominent hospitals and large medical groups.2 In Seattle, health plans faced considerable network instability, as hospitals and medical groups terminated contracts and negotiated aggressively for more lucrative payment arrangements. One of the markets largest insurers, Premera Blue Cross, tested a three-tier network design in 2001 and rolled out the design statewide in 2002. One goal was to improve provider relations by allowing hospitals and medical groups more flexibility in establishing their payment rates, thereby avoiding tense and protracted contract talks. Rather than negotiating, the insurer sorted hospitals and medical groups into tiers based on the payment rates they required and the resulting average cost per episode of treatment. High-cost providers could seek reassignment to preferred tiers by agreeing to lower their payment rates.
Health plans in Boston and Orange County also began to develop tiered-network designs in 2000-01. In Boston, plans developed tiered networks to differentiate more costly teaching hospitals from community hospitals. During the 1990s, the proportion of patients admitted to teaching hospitals increased steadily, and Tufts Health Plan and Blue Cross and Blue Shield of Massachusetts launched two-tier products in 2001 with higher patient copayments at teaching hospitals.
In Orange County, Blue Shield of California introduced a two-tier hospital network product in early 2002 that distinguished high-cost hospitals from their lower-cost counterparts. Blue Shield of Californias design uses lower copayments to encourage consumers to choose lower-cost hospitals. Two other health plans launched similar products in Orange County in 2002.
Health plan interest in tiered networks continued to grow during 2002-03, with health plans in three additional markets experimenting with the concept. In Miami, Blue Cross and Blue Shield of Florida tested a three-tier hospital network in south Florida, while health plans in Syracuse and northern New Jersey began testing tiered products with the employees and dependents of several large hospital systems. Under these designs, members pay lower copayments for care at the sponsoring hospital and affiliated physicians.
Design Variation Abounds
ealth plans vary considerably in the methodologies used to develop network tiers and in the benefit designs used to steer employers and consumers to preferred providers. To date, most health plans have based their tiering methodologies primarily on prices or costs. One Orange County plan, for example, established tiers based on negotiated hospital payment rates, noting that methodological simplicity was a key advantage of this approach.
Other plans, such as Blue Cross and Blue Shield of Florida and Seattles Premera Blue Cross, established tiers by using hospital and physician claims data to estimate the average cost of an entire episode of care, controlling for differences in the severity of patients conditions. Providers with significantly above-average costs are assigned to the nonpreferred tier. Although more complex, this approach potentially allows providers with higher unit prices to be placed in preferred tiers if they perform well in limiting overall costs by reducing unnecessary services and avoidable complications.
Providers and employers in several markets have criticized initial tiered-network designs because they were based solely on measures of cost, prompting health plans to explore ways of incorporating quality-ofcare measures. Blue Shield of California, for example, revised its cost-based tiering methodology to allow higher-cost hospitals into the preferred tier if the hospitals participated in quality projects involving public reporting of patient satisfaction data. Although this approach falls short of creating tiers based on hospitals actual quality scores, the insurer viewed it as a first step toward that goal. In Seattle, Premera Blue Cross recently began a pilot project that involves generating comparative reports on quality measures for a subset of medical groups participating in its tiered network, with a goal of eventually using the quality measures to tier providers.
Health plans also vary approaches to steer consumers to providers in preferred tiers. The most common approach is to require consumers to pay higher copayments or coinsurance at the point of service when receiving care from nonpreferred providers. For example, Tufts Health Plan in Boston requires a $600 copayment for teaching hospital admissions, compared with a $350 copayment for community hospital admissions. In Orange County, Blue Shield of California charges health maintenance organization members an additional $100 copayment and preferred provider organization members an additional 10 percent coinsurance for care at hospitals in the nonpreferred tier. In contrast, Premera Blue Cross in Seattle offers a more flexible product design that allows employers and consumers to choose a network tier at the point of enrollment and/or at the point of service, with premiums that vary depending on the tier chosen as the base network and on the amount of coverage desired for care provided outside the tier. Employers can obtain lower premiums by choosing a lower-cost network tier as the base network and by requiring higher copayments and deductibles for access to hospitals and medical groups outside the base network.
Hospitals and Physicians Resist
any health plans have encountered significant challenges to establishing tiered-network products. Provider opposition to network tiering has surfaced in markets where these products dont even exist, with some hospital systems preemptively negotiating contract language prohibiting the practice. Health plans in Indianapolis, Cleveland, Greenville and Little Rock took initial steps to develop tiered networks in 2002-03 but abandoned the efforts after large hospital systems refused to participate and threatened to drop out of the network altogether. Moreover, in several markets where tiered networks were launched, large hospital systems reportedly used their negotiating clout and political influence to be included in preferred tiers despite their higher costs.
In some smaller markets, there are too few providers to make tiered networks workable. Health plans in Lansing, for example, questioned the viability of a tiered network based on the communitys two hospital systems, noting that purchasers and consumers want products that include both systems.
Health plans also have encountered technical difficulties in differentiating among providers based on cost and efficiency measures. Health plans and benefit consultants noted that data limitations and methodological problems often make it difficult to detect significant differences in costs across hospitals and medical groups within a given market. In some cases, the only providers with significantly higher costs are hospitals that offer unique and vital specialty services, such as trauma care, transplant services and burn units—services that would be difficult to exclude from preferred tiers.
Moreover, health plans that use tiers based on measures of provider efficiency often must use information that is several years old to capture all claims associated with an episode of care, thereby eliciting criticism from providers that the tiers do not reflect current practice patterns. Other technical challenges include developing stable and reliable cost measures for medical groups and individual physicians who serve relatively small numbers of plan members and risk adjusting the measures to account for differences in patient mix and health status across providers.
Health plans face the additional difficulty of handling cases where low-cost physicians admit to high-cost hospitals. At least one plan has responded to this problem by grouping physicians into tiers based on the costs of the hospitals to which they admit patients. This strategy, however, may undermine patient incentives to choose low-cost physicians while penalizing physicians who admit to high-cost hospitals. The difficulties associated with creating tiers for physicians and medical groups have led most health plans to develop tiers for hospitals only.
Provider Choice and Costs
inding the right balance between provider choice and costs remains an important challenge for tiered-network products. If a plan is forced to water down tiering criteria to secure participation of the communitys most popular but costliest hospitals and medical groups, then tiered networks are unlikely to produce significant cost savings for employers. However, if such providers end up in nonpreferred tiers or refuse to participate in the product because of more stringent tiering criteria, then employer and consumer interest in the product—an important prerequisite for success—is likely to remain low.
Most tiered-network products to date exclude relatively few providers from their preferred tiers, raising questions about the ability of these products to save money. The three-tier network product offered by Premera Blue Cross, for example, excludes only two of Seattles hospitals from the preferred tier. In Miami, Blue Cross and Blue Shield of Floridas three-tier network includes all but two hospitals in its preferred tier, and Blue Shield of California includes all Orange County hospitals in its preferred tier. These results suggest that tiered-network designs may have limited ability to constrain costs by steering consumers to lower-cost providers.
Some employers and other observers interpret the relatively inclusive tierednetwork products developed so far as evidence that provider resistance and technical difficulties have largely thwarted health plan efforts to differentiate providers based on cost and efficiency. Indeed, several health plans acknowledged that, contrary to employer expectations, tiered-network products offered only modest premium savings over single-network products because of the large number of providers included in preferred tiers.
However, several health plans indicated that the inclusive nature of their preferred tiers resulted at least in part from some highcost hospitals and medical groups accepting lower payment rates in exchange for preferred- tier placement. Together, these experiences suggest that tiered-network designs so far have proved more useful in increasing plans negotiating leverage with providers than in channeling patients to more efficient providers.
Employers adoption of tiered-network products has been slow in most markets because of uncertainty about cost savings and reluctance to limit consumer choice of providers. Among the six communities where tiered-network products have been launched, enrollment appeared to be highest in Orange County and Seattle, where these products have been under development for several years. In Orange County, Blue Shield of California recently made its tiered hospital network mandatory for all small employer groups and nongroup subscribers to boost participation. In Seattle, Premera Blue Cross has moved approximately 35 percent of its 1.2 million members statewide into its tiered-network platform, although employers can choose to use a single network tier rather than giving consumers a choice of tiers. Other health plans offering these products still considered them experimental, with only a small number of employer groups enrolled to date.
n equally important concern involves how tiered networks may affect the quality of health care available to patients. Some observers fear that designs based primarily on cost will result in the most desirable providers—which could be more costly—being placed in nonpreferred tiers, making them accessible only to those who can pay extra. At the same time, these products could penalize hospitals and medical groups that engage in quality improvement efforts or that produce other important public goods, such as medical education and charity care, if these activities cause providers to be placed in high-cost tiers. To address these possibilities, some health plans have begun to explore methods of incorporating quality measures into their tiering criteria, but the absence of readily available data on quality of care for hospitals and physicians remains an important barrier. Continued progress in improving both quality and cost measurement is needed to ensure that the use of tiered networks does not limit consumer access to high-quality care and undermine incentives for providers to improve quality.
Because health plan experimentation with tiered networks is still in the early stages, the effects of these arrangements on the health care marketplace will depend on how they mature and evolve over time in relation to other health plan design elements. If the movement toward consumer-driven health plans continues to progress, tiered-network designs could become important mechanisms for helping consumers make informed choices among providers based on the cost and quality of care offered.
Tiered networks also can potentially stimulate price-based competition among health care providers, but only if sufficient numbers of employers and consumers show interest in these products. Although most employers remain reluctant to offer health plans that limit provider choice, rapidly rising health insurance premiums may lead employers to consider more aggressive costcontainment strategies, allowing tiered networks to become increasingly important features of health plan design. Because many tiered-network designs allow consumers to make trade-offs between provider choice and costs, these designs are likely to prove more attractive to employers than more restrictive cost-containment strategies. Employer interest and support will be critical for overcoming provider resistance to tiered networks, particularly in communities where providers have significant market power.
Every two years, HSC researchers visit 12 nationally representative metropolitan communities to track changes in local health care markets. The 12 communities are Boston; Cleveland; Greenville, S.C.; Indianapolis; Lansing, Mich.; Little Rock, Ark.; Miami; northern New Jersey; Orange County, Calif.; Phoenix; Seattle; and Syracuse, N.Y. HSC researchers interviewed key individuals in each community, including representatives of health plans, employers and other stakeholders. This Issue Brief is based on an analysis of these individuals assessments of tiered-network health insurance products and expectations for their impact on health care costs, provider choice and access to care.
ISSUE BRIEFS are published by the Center for Studying Health System Change.