Unanticipated Ethical Issues Arise When Data Are Collected and Analyzed

At Seattle Children’s Hospital, the information services department was becoming increasingly aware of the ethical implications of products and analyses they were asked to produce.

“They began to talk among themselves about these issues, and then decided to create a working group where these issues could be discussed among a broader group of information professionals,” says Douglas S. Diekema, MD, MPH, an attending physician and director of education for the Treuman Katz Center for Pediatric Bioethics. These are some examples of ethical issues that were arising:

• While working, data specialists could see private patient data. “This felt at times like a violation of privacy or confidentiality,” Diekema says.
• Information services might be asked to provide data or an analysis they thought could be biased in certain ways.
• Staff were asked to create algorithms they realized might pose equity issues or build in unrecognized biases.
• Information services were concerned about someone using patient data for marketing campaigns.
• Some staff believed they were asked to provide metrics that might lead to organizational changes that were not necessarily appropriate, without understanding the context of the data.

For instance, clinical performance can be judged based on metrics like how many patients clinicians see in an hour, or the number of lab tests ordered in one day. “Using metrics like that to determine incentives or promotion may not be fair, and may not really be a good indicator of quality care,” Diekema argues.

Seeing more patients per hour increases revenue, but might adversely affect the patient experience or increase the risk of safety events. It also might lead to equity issues.

Patients requiring an interpreter will take more time. Clinicians who are overly focused on seeing a certain number of patients per hour may be tempted to avoid using an interpreter, or ask fewer open-ended questions, to save time.

Discussions about these ethical concerns led to the development of a data ethics checklist to raise awareness of ethical issues that arise in their daily work.¹ Diekema says a person trained in ethics can give in-person training, either monthly or quarterly, to a group of information services professionals who bring up cases from their own work.

“An ethicist can help moderate the discussion, ask probing questions, and provide an ethical framework for thinking about the issues,” Diekema explains.

REFERENCE

1. Montague E, Day TE, Barry D, et al. The case for information fiduciaries: The implementation of a data ethics checklist at Seattle Children’s Hospital. J Am Med Inform Assoc 2021;28:650-652.