Since late 2013, CMS has been scrubbing substance use disorder claims from key Medicare and Medicaid data files used for a wide range of health services and health economics research. As Austin and I have explained many, many times—including in the New York Times and the New England Journal of Medicine—this is extremely shortsighted, especially given the epidemic of opioid abuse in this country.
In August, the Substance Abuse and Mental Health Services Administration asked the White House’s Office of Management and Budget to review and approve a proposed amendment to the regulations that direct suppression of these data—42 CFR Part 2. It’s not entirely clear that SAMHSA’s proposal contains a research fix—the agency has been considering other changes to the regulations—but we’re cautiously optimistic.
Under Executive Order 12,866, however, OMB was supposed to complete its review by November 20. We’re still waiting—and time is running out to complete the notice-and-comment process before the end of the Obama administration.
For those of you new to the controversy, here’s a recap of where we stand.
- In 1974, Congress enacted a statute protecting the privacy of those seeking treatment for alcohol abuse. The statute was later expanded to cover other substance use disorders. Identifiable patient records were to be kept confidential unless the patient consented to their release—except when they were disclosed for research purposes.
- Two years later, the agency that became the Substance Abuse and Mental Health Services Administration (SAMHSA) issued rules, now codified at 42 CFR Part 2, to implement the statute. Those rules contained a research exception. As the agency explained, “it would be wholly inappropriate to use the rulemaking process to impose an absolute requirement of patient consent for [research] activities which by statute may be conducted without it.”
- Forty years ago, however, researchers didn’t use insurance claims to do much research. Claims data wasn’t electronic and computers weren’t yet common. The agency’s rules thus prohibited third-party payers—including private insurers, Medicare, and Medicaid—from sharing patient records, even for research purposes.
- Nevertheless, CMS has for decades shared identifiable records on substance use disorders with researchers. The prohibition in 42 CFR Part 2 went unnoticed or unheeded. Its benign neglect facilitated a ton of important research, including the path-breaking Dartmouth studies documenting massive variations in practice patterns across the country.
- Recently, SAMHSA noticed (or decided to act upon the fact) that 42 CFR Part 2 prohibited CMS from disclosing data on substance use disorders to researchers. Without informing researchers of the change in agency policy, CMS started scrubbing substance use claims from Medicare and Medicaid data.
- The data suppression affects about 8% of inpatient Medicaid claims and almost 7% of inpatient Medicare claims. Further details on which claims are redacted and from what files are found on this ResDAC page. Only Research Identifiable Files are affected, not Limited Data Sets. But the former are widely used by researchers to conduct studies that are infeasible with the latter.
- Without those claims in the Medicare and Medicaid data files, it is impossible to conduct the same kind of research on substance use disorders that researchers routinely conduct on other conditions. Once private insurers follow suit—they’re also bound by 42 CFR Part 2—such research will be impossible with private data. Because the data suppression is non-random, any research using health claims—even research having nothing to do with substance use—will also be skewed.
- After Austin and I started raising a ruckus, we heard last April that SAMHSA was planning on proposing a rule to restore research access to Medicare and Medicaid data. It’s that proposal that’s now languishing at OMB. In the meantime, our efforts to get Congress to fix the problem haven’t gone anywhere.