Access to data

Austin Frakt @afrakt tweeted this important idea today about data access and health reform (it was across two tweets so I put it in text box below):

…require all insurers and providers that accept any federal subsidies for their products to make claims/util data available to researchers (with appropriate, but not research-crippling, safeguards for privacy and data security).

The current Medicare Advantage program (where beneficiaries are covered by private insurance, paid for by Medicare) does not make available their claims data for researchers. I have published many studies using Medicare claims data, but the only data available in such databases is Fee for Service or traditional Medicare. Thus, there is a selection effect introduced into all of these studies because you are systematically missing the subset (~ one fourth) of beneficiaries who are in Medicare Advantage plans. These patients tend to be healthier than average. The degree to which this is a problem really depends on the question being asked.

There is another source of selection in Medicare claims data that occurs when beneficiaries who are part of surveys such as HRS are asked whether they wish their Medicare claims records to be linked to their survey responses (in a de-identified, protected manner). This adds a second source of potential selection bias to such a study. A recent paper* I did with colleagues investigating the question of how good are Medicare claims as an epidemioloigcal tool to track the prevalence and impacts of Alzheimer’s Disease makes note of both sources of potential bias:

Seven hundred ninety (92%) of the individuals in the ADAMS sample gave consent to link their Medicare claims records with their HRS survey and dementia assessment (449 female, 56.8%, and 341 male, 43.2%). Persons with no Medicare claims records (N = 32) were excluded because they had no chance of being identified as having dementia from the claims. Most of these individuals were in Medicare HMOs for the entire period, and therefore did not have claims records.

The point is that we need as much data as possible to best understand the use, cost and outcomes of health care services that we provide. It is ridiculous that current Medicare Advantage claims are censored, or not made available, again in a way that protects individuals. Austin’s suggestion of one requirement for any more private options in Medicare should be full data access is a good one, and it should apply to Medicare Advantage going forward.

Finally, I believe that the initial Medicare law (1965) included a requirement that if you accepted Medicare’s payment for your health care that you agreed to allow data related to your care to be used for research and improvement, so long as you were not harmed in doing so. I am heading out the door, and don’t have the reference to the law….but I recall being in a meeting about 10 years ago when the potential reauthorization of the National Long Term Care Survey (NLTCS) was being considered and it was stated that the implied consent of the above provision, which had to that point meant 100% linkage of FFS Medicare claims to NLTCS (but not Medicare Advantage) would no longer be possible and that active consent would be necessary going forward. This would be a hindrance to the database. The HRS (from which the ADAMS sub study came on which the paper linked is based) required consent and 8% did not grant it in the paper described above. So, there are at least two sources of selection bias that can cause trouble when using Medicare claims. The degree of the problem introduced depends on the question being addressed.

Moving toward all data generated from any health care payer that got any public subsidy (meaning all of them) being available for research would be a good thing.

*Donald H. Taylor, Jr., Truls Østbye, Kenneth M. Langa, David Weir, Brenda L. Plassman,   The accuracy of Medicare claims as an epidemiological tool: the case of dementia revisited.  Journal of Alzheimer’s Disease 2009;17(4):807-812.

Related papers on the use of Medicare claims to study dementia:

Donald H. Taylor, Jr., Gerda G. Fillenbaum, Michael E. Ezell. The Accuracy of Medicare Claims Data in Identifying Alzheimer’s Disease. The Journal of Clinical Epidemiology 2002;55:929-937.

Truls Østbye, Donald H. Taylor, Jr., Elizabeth C. Clipp, Lynn Van Scoyoc, Brenda L. Plassman  Identification of Dementia—Agreement Among National Survey Data, Medicare Claims and Death Certificates.  Health Services Research 2008;43:313-26.

About Don Taylor
Professor of Public Policy at Duke University (with appointments in Business, Nursing, Community and Family Medicine, and the Duke Clinical Research Institute). I am one of the founding faculty of the Margolis Center for Health Policy, and currently serve as Chair of Duke's University Priorities Committee (UPC). My research focuses on improving care for persons who are dying, and I am co-PI of a CMMI award in Community Based Palliative Care. I teach both undergrads and grad students at Duke. On twitter @donaldhtaylorjr

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: