Data bank information needs careful interpretation
■ How transparent should the National Practitioner Data Bank be?
Scenario: The National Practitioner Data Bank recently has been through a shutdown and restart. Some of the physician data can lead to mistaken conclusions. Some physicians worry about government reach into their practice matters and the media’s instinct to use any available data for a physician “gotcha” story.
Reply: In September 2011, the Dept. of Health and Human Services blocked access to the public use data file of the National Practitioner Data Bank, which contains information on malpractice payments and disciplinary actions of health care professionals. HHS’ action raised debate about whether such information should be kept from the public. Some medical organizations, including the American Medical Association, argue against release of data bank information on grounds of known inaccuracies and likely misinterpretation by the public. Patient advocacy groups, among others, stress that NPDB transparency is essential for advancing safety and quality.
Concerns regarding the misuse and misinterpretation of data bank files are indeed valid. However, in removing this data from the public domain, HHS seems to contradict its website claim to believe that:
“Transparency and data sharing are of fundamental importance to our ability to achieve HHS’ strategic goals of advancing the health and well-being of the United States & [and] help the public hold the private sector accountable & generate insights into how to improve health and well-being, mobilize public and private sector action and innovation to improve performance.”
Instead of pushing for less disclosure, the medical profession should work to contextualize these files and accept accountability when legitimate analysis of this data reveals strategies to improve health care.
Federal law mandates that state licensing boards and professional peer review entities report to the data bank any medical malpractice payments or disciplinary actions based on professional competence or misconduct. This addressed the problem of physicians with a questionable record being able to move from state to state without disclosing their disciplinary history.
By federal law, the public is not allowed to have direct access to the NPDB. Instead, the law provides that the public should have access to anonymized information, which HHS releases as special “public use files.”
What prompted HHS to block access to the public use files was an investigative report by The Kansas City Star that linked de-identified data bank public files to a specific doctor by cross-referencing other public information. “We have a responsibility to make sure under federal law that [data bank data] remain confidential,” an HHS representative said, referring to the news article. HHS further explained that the public-use files would remain closed until the agency completed a thorough review of the data fields contained in the files.
The AMA responded positively to HHS’ action on the grounds that “Duplicate entries, inaccurate data and inappropriate information in the NPDB provide, at best, an incomplete and haphazard indicator of a physician’s competence or quality.” The above objections are valid. For example, physician-specific NPDB information could be misleading to a typical patient who is not aware of the frequency of malpractice claims, especially in high-risk specialties such as neurosurgery and obstetrics. The longer a physician practices, for example, the more likely he or she is to have a malpractice claim. Thus, a patient wanting to select a physician with fewer claims might simply select a less-experienced (and perhaps less-competent) physician.
I have defended hospitals and physicians, so it is clear to me that malpractice payments, rather than signifying negligence, often represent a cost-benefit calculus designed to avoid litigation costs. But this critical nuance is not always captured in news articles. Citing NPDB data, The Kansas City Star found “about 200 doctors who have practiced in Kansas or Missouri since 1990 and have had five or more malpractice case payments made on their behalf, without ever being disciplined by the state’s board.” This statement implies inadequate professional oversight in Missouri and Kansas. But if a malpractice payment is based on a business decision and not evidence of incompetence or misconduct, there is good reason why it did not give rise to a disciplinary action. Further, as the newspaper acknowledges, using payments as proxy for individual claims can lead to overcounting, “because individual cases may involve payments from more than one source.”
Acknowledging that NPDB public-use files could be subject to misinterpretation by the public is not a compelling reason to block access to it. At least 17 states do not consider the malpractice and disciplinary histories of individual physicians to be confidential and have laws making such information accessible online. For instance, a patient in Georgia can easily look up an individual physician’s malpractice and discipline history, the same information “triangulated” by The Kansas City Star.
The liability field in the Georgia record contains the following language: “Disclaimer: Settlement of a claim may occur for a variety of reasons which do not necessarily reflect negatively on the professional competence or conduct of the physician. A payment in settlement of a medical malpractice action or claim should not be construed as creating a presumption that medical malpractice has occurred.”
In other words, the data are accessible to the public, but the information is contextualized to mitigate facile misinterpretations.
What, then, might be lost by blocking public access to NPDB data? HHS’ online statement on transparency summarizes it nicely. More concretely, consider the groundbreaking work of John Wennberg, MD, MPH, the founder of the Dartmouth Atlas Project, who analyzed Medicare utilization patterns and shed light on a shocking epidemic of overtreatment that endangered patients and escalated costs. This prompted policymakers to think seriously about evidence-based medicine and the incentives created by fee-for-service payments. Or consider the Institute of Medicine’s landmark 1999 report, "To Err is Human", which increased awareness of the number of medical errors in the U.S. health care system and spurred more robust patient safety and quality-control measures. Efforts like these would not be possible without sufficient data transparency.
Of course, one could distinguish the Dartmouth Atlas and the IOM report as aggregating systemic data and not singling out any health professional, as The Kansas City Star article did. The argument then becomes one of protecting individual physicians from unfair opprobrium rather than shielding the entire medical profession from scrutiny. The Internet was not a factor when the data bank public-use file was initiated, and HHS could not have anticipated the ease with which de-indentified data could be triangulated to a particular doctor.
Does the possibility of identifying a particular physician represent a harmful breach of confidentiality? In the case of morbidity and mortality conferences or other peer review sessions, yes, such breaches of confidentiality would be extremely harmful. Participants in these quality review processes would not be as honest and forthcoming, and that would impact health care quality and safety negatively.
But as states such as California and New York have discovered, providing public access to individual physician disciplinary and malpractice payment records has not undermined institutional quality-control measures like peer review of adverse outcomes, which weakens claims for keeping such information confidential.
Fazal Khan, MD, JD, associate professor, University of Georgia School of Law, Athens