On April 27, 2016, CMS held a MEDCAC meeting on "Treatment Resistant Depression," covered a bit in the CMS press. The CMS webpage for the meeting, including agenda, prepared documents, and (eventually) a transcript, is here.
I noticed that two different trade journal articles, MedPageToday (here) and Modern Healthcare (here), noted that CMS had not stated "a reason for holding the meeting."
The rationale have its roots in a 2014 appeal of the one NCD that addresses treatment resistant depression, an NCD for vagal nerve [electrical] stimulation, which covers therapy for epilepsy but does not allow coverage for treatment resistant depression.
In 2014, there was an administrative court action at CMS where a party moved to have the NCD overturned. The home page for this process is here, and the listing of cases over the years is here.
The judicial finding, issued on December 30, 2014, makes clear in a 96-page opinion that CMS staff went to great lengths to find problems with existing data for VNS for TRD (here), which ensured that the NCD was not overturned by the judge. It's quite a read.
I noted separately that CMS rigor in data evaluation swings widely, depending on context. Compared to the case just cited, CMS applied a far, far less rigorous analysis to its internal one-arm study of diabetes prevention programs as cost saving (see story here).