In covering health care technology, it’s generally thought that commercial market plans follow Medicare’s lead. In particular, I’ve heard it stated that private plans tend to deny coverage to technologies only after Medicare does so, and Medicare rarely does so. I never interpreted this statement in absolute terms—that literally no private plans would deny a coverage of a technology until Medicare does. I just thought it was a pretty firm tendency, a stylized fact.
An analysis by James Chambers and colleagues, just published in Heath Affairs, aims to debunk it, or maybe the absolute version of it (which probably nobody should have believed anyway). They compared Medicare’s national coverage determinations (NCDs) for medical devices (N=47) issued between February 1999 and August 2013 with coverage decisions made and published online by 16 of the 18 largest private insurers as of January 2014.
According to its policy, Medicare covers in some way 74% of the 47 devices in the analysis. Across all NCD-private payer combinations considered, 22% had more restrictive coverage than Medicare; 23% were less restrictive; about half were equivalently restrictive; and in 4% coverage varied by indication. These breakdowns vary by insurer. Some, like Blue Cross track Medicare even more closely; others, like Wellpoint, much less so.
The authors conclude, that their “findings cast doubt on the prevailing belief that private payers follow Medicare’s lead when covering medical technology.”
Unless you believed that literally no private insurer deviated from Medicare coverage policy, I think very little updating of priors is warranted by this study for two reasons. First, consider its limitations: It was an analysis of 47 national coverage decisions, for devices only (not drugs or procedures). Also, as the authors acknowledge, the vast majority of Medicare coverage decisions are made at the regional, not national level, but regional decisions were not considered in the analysis. It’s also not clear how much coverage decisions affect clinical practice. On TIE Nicholas Bagley wrote that there’s solid evidence that Medicare contractors don’t appear to do much to enforce Medicare’s regional coverage decisions—and suggestive evidence that even NCDs often get overlooked. A distinction between de facto and de jure coverage seems important here. Finally, only private payer policies that were available online were analyzed. (I don’t consider it a severe limitation that a subset of insurers were examined because they were among the largest insurers.)
Second, the analysis found that about three-quarters of plans’ coverage policies were at least as generous as Medicare’s. Only a minority (22%) were more restrictive than Medicare, though the extent to which less/greater restrictiveness is clinically meaningful is unclear. Appendix Exhibit 2 shows that plans tend not to cover the 12 of 47 devices not covered by Medicare. Nine plans don’t cover 80% or more of them; five plans don’t cover two-thirds or more of them; only two plans cover more than half, but still deny well over one-third of them. In other words, the analysis supports the idea that private plans tend not to cover technologies that Medicare doesn’t cover, which is precisely what I had thought.
The authors’ analysis does suggest that private insurers don’t invariably defer to Medicare’s coverage decisions. Given the dearth of evidence about coverage decisions, that’s a valuable contribution. But I don’t think the data presented warrant reevaluating the softer claim that private insurers typically follow Medicare’s lead.
A webinar about the paper with some of the authors and a former Director of the Coverage & Analysis Group at CMS is scheduled for Wednesday, September 9, 2015 at noon Eastern. Register here. More about Medicare’s NCDs by me here.