"Big data" in health care raises some ethical concerns
The Institute of Medicine has advanced "learning health care systems" which involve collecting large amounts of data electronically during every health care encounter.1 However, some bioethicists are voicing concerns about how patients’ privacy will be protected.
These concerns are significant because there are no limits on how long data will be stored, and long-term storage presents additional possibilities for security breaks, argues Ana S. Iltis, PhD, associate professor in the Department of Philosophy and director of the Center for Bioethics, Health and Society at Wake Forest University in Winston-Salem, NC.
"Moreover, health care data can affect not only individual patients but also their families, who might be presumed to be at higher risk for various conditions based on family history," says Iltis. "Patients also have reasons to be concerned that data could be used against them in various ways in the future."
Privacy concerns are sometimes dismissed by suggesting data will be de-identified or heavily protected using other mechanisms. "We have seen enough evidence from other industries that collect and store large amounts of data that protections can fail," says Iltis. "People have good reasons to be suspicious of claims that their data will be well-guarded."
Others point to the fact that so much information about individuals already is collected, stored, and used for other purposes, such as retailers that analyze and sell information about their customers’ buying habits. "Many people have argued that this is inappropriate as well. And if they are permissible, they do not give health care a free pass," says Iltis. "In many ways, we treat health care as special."
Health care professionals and organizations are held to higher standards with respect to protecting privacy and confidentiality, and they expect from patients a higher level of honesty and transparency than retailers do, for instance.
"I might be justified in lying to a retailer about my birthday, but most would agree that I should not lie to my physician about my age," says Iltis. "There are good grounds for treating health care differently. Practices employed by other industries are not necessarily justified in health care."
Great promise, major concerns
The application of big data to health care holds great promise at an individual patient care level, as a population heath management tool, a quality improvement system, and in facilitation of comparative effectiveness research and longitudinal observational studies, according to Blair Henry, BSc, MTS, an ethicist at Sunnybrook Health Sciences Centre in Toronto, Ontario, Canada.
However, privacy remains a major concern. "Many fear truth hidden in Paul Ohm’s warning: Data can be either useful or perfectly anonymous — but never both!’" says Henry.2 "How will the inherent tension between the protection of privacy and a reality that effective analysis requires sufficiently detailed data be managed, and by whom?"Concepts such as consent and data sharing agreements will need re-assessment in the face of big data’s far-reaching effects, adds Henry. "Data ownership and downstream access also needs to be considered," he says. Here are some ethical concerns involving "big data" in health care:
The data could potentially be applied to consumer preferences.
"Big data is particularly alarming for its ability to tailor advertising specifically at patients for demand-driven health services," says John Henning Schumann, MD, associate professor of medicine and director of the internal medicine residency program at the University of Oklahoma School of Community Medicine in Tulsa.
The "good" of big data includes providing patients with necessary reminders for screenings, vaccinations, and medication refills, acknowledges Schumann, but he says direct-to-consumer medical marketing becomes even more worrisome if big data is applied to consumer preferences.
"We already see this with Google and Amazon, as well as brick-and-mortar retailers who track our buying habits via their frequent buyer programs," he notes. "Will patients with certain diagnostic codes continue to be able to exert free will? Or does the future hold a Big Brother-like medical nanny state nagging our patients to conform, comply with, and adhere to treatment, even when their values conflict with the best available evidence?’"
Patients’ data could be used for purposes that they do not support.
"Parents of children with special needs might not want their children’s medical records used to support research that suggests that these children are too expensive’ or have lives that involve extensive suffering, or to justify recommending abortion to women whose fetuses are diagnosed with particular conditions," says Iltis.
Data may be biased.
Clinicians may believe falsely that by collecting, storing, and utilizing large data sets, they are able to make more objective, evidence-based decisions.
"Work in the philosophy of science and the philosophy of medicine teaches us that data collection, interpretation, and use are fraught with biases," says Iltis. "There are numerous biases in how we collect, store, and use data that may go unnoticed."
Data collection efforts could negatively affect the clinical encounter.
Requiring clinicians to collect certain data could be very helpful to patients and improve the clinical encounter. For example, physicians might be prompted to ask certain questions that they would otherwise skip. "However, efforts to collect all required data can adversely affect the clinical encounter," says Iltis.
Clinicians can become engrossed in computer screens, in asking rote questions without paying much attention to whether the question truly was relevant to the patient or whether the patient understood it, and in fulfilling their data-driven obligations, rather than on the encounter with the patient.
"Even something as simple as an office that is not set up for a clinician to use a computer and also look at and engage a patient can downgrade the clinical encounter," says Iltis. "Moreover, the desire to have one’s data look good’ and be complete can become more important than other aspects of the clinical encounter."
1. IOM (Institute of Medicine). Best care at lower cost: The path to continuously learning health care in America. Washington, DC: The National Academies Press, 2013.
2. Paul Ohm, Broken promises of privacy: Responding to the surprising failure of anonymization. UCLA Law Review 2010;57:1710.
- Blair Henry, BSc, MTS, Ethicist, Sunnybrook Health Sciences Centre, Toronto, Ontario, Canada. Phone: (416) 480-6100 ext. 7178. E-mail: firstname.lastname@example.org.
- Ana S. Iltis, PhD, Associate Professor, Department of Philosophy, Director, Center for Bioethics, Health and Society, Wake Forest University, Winston-Salem, NC. Phone: (336) 758-4254. E-mail: Iltisas@wfu.edu.
- John Henning Schumann, MD, Associate Professor of Medicine/Program Director, Internal Medicine Residency, University of Oklahoma School of Community Medicine, Tulsa. Phone: (918) 579-1217. E-mail: email@example.com.