What’s in your wallet? ‘Big data’ wants to tell doctors and health plans
Does privacy still matter? According to Facebook founder Mark Zuckerberg, privacy is no longer a “social norm.” But a recent report from the Pew Research Center found that Americans are indeed concerned about their privacy.
More than half (55%) of survey respondents said that the “state of your health and the medications you take” was “very sensitive” information, second only to Social Security numbers. Fewer people considered information about their relationship history (40%), religious or spiritual views (22%), or political views (20%) as very sensitive. At the bottom of the list, only 8% considered “basic purchasing habits” very sensitive.
What the respondents might not have realized is that “big data” can link their purchasing habits to their health information. So far, discussions of big data in healthcare have focused primarily on using the vast archive of clinical information collected in electronic health records and financial data from insurance claims to track and analyze utilization, costs and other factors—data of importance to providers and payers.
But consider big data from a different perspective—the enormous amount of personal information that can be drawn from public sources such as census records and credit reports, and from customer loyalty programs at local and national retailers, credit card purchases and social media. Should this nonmedical information be readily available to doctors and health plans?
Carolinas HealthCare System, which covers 2 million people, is already working with data brokers to use such information to develop an algorithm to identify patients whose unhealthy choices put them at high risk of hospitalization and readmission. Currently, doctors will get only a risk score, but other programs may soon take their own deep dives into individual items.
Privacy is a foundational principle of medical ethics and healthcare prac- tice. What happens in the doctor’s office stays there—unless, of course, there are valid reasons to disclose the information to other agencies and individuals. Many sensitive topics come up in medical encounters, but they are typically initiated by a provider’s question or a patient’s disclosure. Using nonmedical data from financial or public records to initiate discussions about smoking, drinking, unhealthy eating, unsafe sexual activity or other behaviors turns this view of privacy on its head.
Patients may withhold information—yes, even lie—to their doctors for many reasons. Some unhealthy behaviors are related to cultural, psychological and economic conditions that are difficult to change. It’s not that patients don’t know they should stop smoking or lose weight; they don’t want to be shamed and blamed for something that gives them pleasure and seems out of their control— or both.
Those who claim that “privacy is passé” point out that people are more likely to reveal information about their habits on Facebook or other social media than they are to tell their healthcare providers. And some patients, often those with chronic illnesses, set up their own information exchanges on sites such as MyHealthTeams.
These examples are very different from a scenario in which a doctor challenges a patient’s denial about following a diet, for example, by citing the charges on a restaurant loyalty card, or a health plan raises premiums for those who fail the “wallet test.” If the new spending information is likely to change a doctor’s or health plan’s behavior, the doctor-patient relationship is almost certain to be affected, and quite possibly for the worse.
We have come to expect marketers to use spending information available through big data to sell products, but consumers can always ignore a marketer’s ads. The blanket consent patients sign for use of their medical information does not typically include consent to mine their financial records for potentially unhealthy purchases.
Working with patients who cannot or will not change their unhealthy behaviors can be frustrating to practitioners, insurers and policymakers. It is difficult as well for the patients themselves who may want to change, but who lack the financial resources and social support to succeed.
Reducing risky behavior is certainly an important public health goal. Doctors and health plans can best serve this goal by respecting patient autonomy, building trust in healthcare and offering practical ways to meet realistic personal benchmarks. Patients are unique individuals, not collections of data points, no matter how big the source.
Carol Levine directs the Families and Health Care Project at the United Hospital Fund in New York City. She is a fellow of the Hastings Center, a bioethics research institute in Garrison, N.Y.