Networks, Networks, Networks
This is the second post covering the Third Annual Health Insurance Exchanges Conference, hosted by LDI’s HIX working group, which took place in April 2015. The full agenda is here, and the first post, which focused on choice architecture, is here.
With the growing popularity of narrow network health insurance plans on the federal and state-based exchanges, and recent research showing that narrow networks play a role in controlling health care costs, there’s no way to avoid vexing questions about adequate access, quality and consumer preferences.
Narrow networks give patients access to a limited number of providers. They vary in their ‘narrowness’ across different types of providers and hospitals. Premiums for narrow-network plans are typically lower than for broad networks, and analysis by McKinsey found that these premium differences are widening. About 90% of consumers on the health insurance exchanges had at least one narrow network in their market in 2015.
No matter where the discussion started at the HIX LDI conference, it always came back to networks.
How do consumers evaluate a network and how much do they care about its breadth?
Although narrow networks are popular with consumers, it is not clear whether consumers are consciously making the trade-off between cost and network breadth and how they judge a network’s suitability.
LDI’s Amanda Starc presented new research exploring how consumers make these decisions. She has found that consumers are willing to pay higher prices for broader networks. However, many consumers distinguish networks by using a proxy for network status. Specifically, consumers may use insurer brand as a proxy for network quality. In addition, consumers use a flagship, ‘big brand’, hospital as a sign-post for a suitable network, and are willing to pay substantially more to have a key hospital in-network.
Is it possible to measure the ‘quality’ of a network?
Although the ACA established network adequacy regulations, their implementation varies state-by-state. States that set up their own network adequacy standards tend to be those that have a state-based or partnership health insurance exchange, indicating a more proactive engagement with the ACA.
But these minimum standards do not measure the quality of a network nor its value to consumers. Presenting at the conference, LDI’s Ashley Swanson made the case that these measures, though harder to define than numerical ones, need to play a role in network adequacy regulations.
In ongoing work, she is seeking to fill this information gap by collecting data to define the ‘relative value’ of a network to consumers. She uses inpatient claims and hospital data for Colorado and New York to estimate consumers’ preferences over hospitals as a function of driving distance, illness characteristics, and hospital quality and service offerings. As expected, she finds that patients prefer lower driving distances and higher quality service offerings, particularly in the service area relevant to their medical condition. She then uses the results to estimate the value of networks offered by the most popular issuers in those states, which include both broad and narrow network plans.
Initial findings have defied some expectations. Very restrictive or narrow networks don’t necessarily mean less relative value. For instance, some of the narrow networks in Swanson’s study include the best hospitals while competing networks can be broader but include lower-quality providers. New research by Simon Haeder (University of Wisconsin – Madison) and colleagues appears to support this initial finding. Haeder and colleagues compared the quality of hospitals included in California’s state Marketplace plans with those in non-Marketplace plans. They found that depending on the measure of hospital quality employed, the Marketplace plans have networks with comparable or even higher average quality than the networks of their commercial counterparts.
When complete, Swanson’s work will also allow for comparisons of network value across geographic markets, accounting for the total number of available providers and differences in average network inclusiveness. Comparisons across markets can help inform the debate about network adequacy at the state level.
How do consumers know what their provider network looks like?
In a previous LDI HIX conference, the idea of ‘t-shirt sizing’ emerged: that is, simple, universal measures for provider network breadth that would be easy for consumers to understand. The task is complicated by the difficulty in knowing who is actually participating in local networks. Physician directories are notoriously inaccurate; physicians often don't know exactly which plans include them and insurers often don't know if a practice's location changes or if it has closed. There is no central information source updated in real time, so every time networks need updating, someone has to cull the relevant information, likely from multiple databases. This painstaking work does not scale well.
Taking on this challenge, LDI Executive Director Dan Polsky is assembling a database of all physicians in all networks on all plans offered on the ACA marketplaces. This allows for a standard way of ‘t-shirt sizing’ network physician provider networks for each rating area. He previewed this data resource, but also offered a caution, as the relative size of physician networks can be very different between communities within rating areas. He proposed a consumer interface which would allow for any consumer to estimate network size within a specified radius from any geographic location. This could complement current practices which mostly involve checking the “find a doctor” website to make sure a given provider is in network.
These findings are garnering attention among regulators seeking new ways of measuring network adequacy and among exchange officials seeking to help consumers make an informed choice.
Further resources from the LDI blog: