• Friday wonk dump: What I’ve been reading

    Below is a dump of thoughts and stuff I’ve read recently. I might circle back to some of it in the new year. Or I might not. I figured this is better than nothing.

    Market Death Spirals

    • As distinct from death spirals of plans and the prevailing issue with respect to exchanges.
    • Contrary to conventional wisdom, Buchmueller and DiNardo found that New York’s market did not experience a death spiral when it  implemented guaranteed issue and community rating without a mandate in the 1990s. (Adrianna touched on this paper recently.) Their results are consistent with the separating equilibrium theory of Rothschild and Stiglitz. Low risk consumers don’t exit the market. They just buy less complete coverage. (H/t Nick Bagley.)
    • Consistent with conventional wisdom, Lo Sasso and Lurie found large adverse selection into markets that implemented guaranteed issue/community rating reforms in the 1990s. (Ungated working paper version here.) See also Lo Sasso’s NIHCM Expert Voices brief.


    • Powell et al. provide a conceptual framework for overuse by primary care providers. It’s based on the theory of planned behavior that posits that behavior is determined by “an intention or behavioral plan that enables the attainment of a goal.” In brief, PCPs overuse care because they are trying to accommodate beliefs and expectations pertaining to outcomes for patients, themselves, and their organizations, as well as normative beliefs and beliefs about control. PCPs also may filter their intentions through a process of assessment that is constrained by aspects of care delivery such as time spent with patients.
    • Mathias and Baker discuss threats to development and potential unintended consequences of overuse measures.
    • Chan et al. describe and critique the current state of overuse measurement. A key point, among many, is that defining “low-risk” populations is difficult but necessary to turn back or prevent the spread of overuse of many therapies and services worthwhile for “high-risk” patients.


    • Obviously some overlap with overuse.
    • Bentley et al. provide a conceptual framework for US health system waste, decomposing it into administrative, operational, and clinical components.
    • Bohmer describes four habits of high-value health care organizations: (1) specify and plan what you do, (2) related, deliberately design all sub-systems, processes, and infrastructure to match delivery with patient need, (3) measure what you do with external oversight, (4) feedback measurement to refine (1) and (2). Isn’t roughly this the only way to get better at anything? Be clear about what you do and intend to do. Assess how well you do it. Feedback and refine. Seems obvious. And to the extent health providers and systems don’t already do this in serious ways is shameful, bordering on malfeasance in light of the vast public expenditures that support them.

    Threats to Reference Pricing


    Comments closed
    • The value of being able to identify precisely low-risk populations is huge. The creep of extensive margins in health care is a major driver of overuse. What’s good for some patients is not necessarily good for more/all patients.

    • I’m really not convinced by that NY Death Spiral paper. It may be that initially (2002 the paper was published) a death spiral hadn’t unfolded. But the state of NY’s individual market (something like 34,000 private commercial buyers last year with premiums over $1K, excluding the Healthy New York market) makes a pretty good case on its own. People certainly did leave the market (rather than just buying “less” insurance) over time. Casey Mulligan had a great NYT piece explaining why death spirals tend to be rare. But NY’s individual market (maybe not SG) certainly looks and smells like one.

    • Agree that counterfactual is important. But unsure if CT and PA over a short time-period offer that counterfactual. Would be nice to see this study extended to post-healthy NY years (HNY probably siphoned off some potential healthy purchasers). Haven’t seen any other papers on NY individual market, actually. If you’ve got links, would appreciate them.

      • YF, this paper of mine takes examines coverage changes in community-rated markets over a longer time horizon : http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2033424. It includes the introduction of “Healthy NY.” New York’s Medicaid expansions actually turned out to be quite effective at enrolling relatively unhealthy adults (see Figure 2 on Page 30).

        The analysis lumps the individual and small group markets together, much like the Buchmueller and Dinardo paper. In the Current Population Survey the samples of individuals purchasing coverage in the individual market are quite small, making these markets difficult to analyze in isolation. The paper’s Figures 1 and 2, for example, lump the similarly regulated individual and small group markets of Maine, New York, and Vermont together.