Default Nudges in Medicine—Designing the Right Choice

Defaults work. With defaults, a designer or “choice architect” alters the environment and provides a preselected option.¹ An individual must take action to avoid the default occurring. Defaults might be described as making the right choice the easy choice. Although universal agreement on what constitutes “the right choice” is not necessary, defaults should target actions or events about which broad consensus exists. Defaults should preserve individual choice for the target of the action, enable those who do not want to take the action to easily opt out, and occur at a time and setting that make sense for the recipient and the action. Defaults should be transparent or risk being perceived as sneaky and eroding trust.

Outside medicine, defaults have been used to increase organ donation, retirement plan enrollment, use of green energy, and healthy food selection. In medicine, defaults—notably those embedded in the electronic health record—have been used to increase vaccinations, increase generic drug prescribing, decrease opioid prescribing amounts, decrease unnecessary imaging, and increase cancer screening.²

What about a default to increase hepatitis C virus (HCV) screening? Hepatitis C virus infection is the leading cause of cirrhosis in the US. Owing to HCV’s associated morbidity and mortality, the Centers for Disease Control and Prevention has made recommendations on HCV screening since 1998. The most recent expansion of screening guidelines in 2020 include 1-time HCV screening for all nonpregnant adults. Despite these recommendations, most eligible US residents have not been screened, and effective interventions to improve screening and treatment are needed.³

To increase screening, some states have implemented laws on HCV screening. Since 2016, a Pennsylvania law has mandated offering HCV screening to adults born between 1945 and 1965—the target of earlier HCV screening recommendations—in primary care and inpatient settings. To better adhere to Pennsylvania law and increase screening, Mehta et al⁴ conducted a stepped-wedge randomized clinical trial in 2 academic Pennsylvania hospitals; results are published elsewhere in JAMA Network Open.

The intervention embedded a default HCV laboratory order within admission order sets. The order included text reminding clinicians of the state mandate and that positive results would be followed up by a dedicated hepatitis linkage team. The intervention was compared with a pop-up alert that included information about the state mandate and that positive test results would be analogously followed up by the hepatitis linkage team; clinicians were required, in 1 click, to either accept the laboratory order or dismiss the alert and order.

With the pop-up alert comparison, HCV tests were ordered for 38.7% of admitted patients, and screening was completed for 34.9%. The default intervention increased laboratory ordering by 38.1 percentage points to 2599 of 3229 patients (80.5%); screening completion increased by 31.8 percentage points to 2257 of 3220 patients (69.9%). The intervention was associated with a 0.4% increase in detection of HCV viremia, and, of the 63 patients with positive HCV viral loads, 39 were connected with a treating clinician.

Clearly, as with many defaults, this one worked. The default order was associated with increased ordering, testing, and HCV detection. However, the Pennsylvania law and the default intervention by Mehta et al⁴ raise questions about legislating medical practice and the application of defaults.

The Pennsylvania law is well-intentioned—increasing screening will ensure more patients receive effective treatment—but the law, enacted in 2016, has not kept up with expanded screening recommendations. In addition, mandating screening for inpatients is unconventional: the marginal population increase of screening inpatients within a state or nationwide is likely to be small. Each year in the United States, 85% of Americans have a physician visit, but only 8% of adults are hospitalized. However, hospitalized patients may be less able to engage in follow-up care and treatment. Legislating specific medical practices is potentially fraught with problems and should be undertaken with care.

The default order is also well-intentioned but may not meet the goals of HCV screening recommendations or the intent of the Pennsylvania law. According to the US Preventive Services Task Force, HCV screening should be “voluntary…undertaken only with the patient’s knowledge and understanding that HCV screening is planned…providing patients the opportunity to ask questions and to decline screening.”⁵ Similarly, the Pennsylvania law requires hospitals and primary care practices to “offer” HCV screening to patients, requires that patients “consent to the test,” and requires that “if an individual accepts the offer of a hepatitis C screening test and the screening test is reactive, the health care provider shall either offer the individual follow-up health care or refer the individual to a health care provider who can provide follow-up health care.”⁶ Voluntary informed consent is essential, especially for illnesses such as HCV that carry social stigma.

Primary care has several conditions to do this well. We can efficiently inform our patients about the recommendation, recommend screening, ask whether patients have questions, and verify that the patient wants to proceed with screening. We usually have a preexisting relationship on which to base follow-up conversations, especially in the event of a positive test finding.

In the hospital, clinicians order many tests for patients. Initial order sets often include dozens or hundreds of tests. In a perfect world, we would be able to discuss the rationale, risks, benefits, alternatives, and implications of every test with every patient. In the real world, hospital care would grind to a halt. Also, orders are seldom entered in the presence of the patient.

In addition, as the authors acknowledge, the default order probably led to unnecessary repeated testing. In the fragmented US system, patients may have received care outside the system in which the study was conducted. Health information exchange has improved in recent years, but for HCV screening—which only needs to be done once in a lifetime for nonpregnant patients and during every pregnancy—the potential for overscreening seems high.

Perfect is the enemy of good, of course. Legislation will never work precisely as intended. Not all US residents have access to primary care, where HCV screening would ideally be offered. With any screening program, some overscreening and loss-to-follow-up is inevitable. We should work to increase offering HCV screening to US residents who have not had it done.

However, legislating mandated HCV screening for inpatients does not recognize the reality of how hospital medicine is practiced. Responding to that law, the investigators implemented a default order that facilitates ordering HCV tests but may not facilitate offering HCV screening to an informed patient.

Yes, defaulting to “the right choice” works.⁷ Defaults can be great. However, defaults need to be carefully designed, implemented transparently, occur at the appropriate time and setting, and target the intended action.

Published: March 17, 2022. doi:10.1001/jamanetworkopen.2022.2437

Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2022 Jia J et al. JAMA Network Open.

Corresponding Author: Jeffrey A. Linder, MD, MPH, Division of General Internal Medicine and Geriatrics, Northwestern University Feinberg School of Medicine, 750 N Lake Shore Dr, 10th Floor, Chicago, IL 60611 (jlinder@northwestern.edu).

Conflict of Interest Disclosures: Dr Linder is supported by contract HHSP233201500020I from the Agency for Healthcare Research and Quality, grants R33AG057383, R33AG057395, P30AG059988, and R01AG069762 from the National Institute on Aging, grants R01HS026506 and R01HS028127 from the Agency for Healthcare Research and Quality, and the Peterson Center on Healthcare, pertaining to behavioral economics and use of electronic health records.