Decision Fatigue, Running Late, and Population Health Management—Screening Out of Time

Oh dear! Oh dear! I shall be too late!

The White Rabbit, Alice’s Adventures in Wonderland, Lewis Carroll

In clinic sessions, the amount of work necessary to achieve quality or performance metrics for patients is overwhelming. In primary care, it was previously estimated that the typical physician, providing preventive and chronic care for a typical patient panel—never mind acute care—required 18 hours per day.¹ Physicians also spend 1 to 2 hours interacting with the electronic health record for every hour they spend with patients.^2,3 The volume of work may be one reason the quality of ambulatory care in the United States has not improved appreciably in the past decade.⁴ To fit it all in, many of us feel like Lewis Carroll’s White Rabbit, constantly late, checking our watches, running from task to task, and from patient to patient.

Beyond the volume of work, quality of care may depend on the time of day. Hsiang and colleagues⁵ describe a retrospective cohort study of patients due for breast or colorectal cancer screening who were seeing their primary care physician. For both forms of cancer screening, unadjusted rates of screening orders and screening completion within 1 year were highest for patients seen first thing in the morning. Rates of orders and completed screening progressively decreased during the morning, moderately increased around noon, and again decreased during the afternoon. The absolute difference between morning and late afternoon screening and screening completion rate ranged across outcomes from approximately 10% to 16%. In adjusted analyses, the investigators observed inverse associations of later hours of the day with screening orders and screening completion for both breast and colorectal cancer screening. For each hour over the course of the day, there were 3% to 6% lower odds of cancer screening being ordered or completed.

Despite the study’s many strengths, such as its large size and exclusion of acute visits, it has limitations. The exposure was clinic appointment time, not actual visit time. There were no adjustments for how busy clinicians were at the time of an individual visit. In addition, the investigators adjusted for many potential confounders but not prior screening completion, a behavior associated with subsequent screening.⁶ There was a lack of contextual details associated with observed rates of screening completion, like patient preferences and refusals or the length of patient-physician relationship.

Even for this universally insured cohort, absolute screening rates were unimpressive. After 1 year, only 33% of those with 8 am clinic appointments completed breast cancer screening and only 28% completed colorectal cancer screening. From data in the Supplement, one can calculate that only about two-thirds of all eligible patients had completed colorectal cancer screening 1 year after the initial visit.

Two possible explanations for the results found by Hsiang and colleagues⁵ are decision fatigue and running late. Decision fatigue is a progressive erosion of ability, self-control, or willpower that occurs as people make choices. Decision fatigue was described most famously in an analysis of Israeli judges making serial parole decisions. Prisoners were much more likely to be granted parole at the beginning of court sessions; the rate of parole decisions dropped to almost 0 before breaks. Presumably, as judges became more cognitively fatigued, they defaulted to doing the easy thing, denying parole. Outside of medicine, businesses increase revenue by taking advantage of decision fatigue. Car dealerships offer more expensive—and probably unnecessary—options toward the end of the purchasing process. Supermarkets offer sugary, unhealthy foods at the checkout counter. Within medicine, decision fatigue has been associated with increasing antibiotic prescribing for respiratory infections, increasing opioid prescribing for back pain, decreasing influenza vaccination, and decreasing handwashing.⁷ Unlike these more time-sensitive services, a major addition by Hsiang and colleagues⁵ is to show that appointment time was associated with patient care 1 year into the future.

Another, simpler explanation for the association described by Hsiang and colleagues⁵ could be that clinicians have a tendency to get further behind as shifts go on. In becoming further behind and perhaps trying to catch up, clinicians may naturally omit certain aspects of care that are not immediately urgent, like ordering cancer screening. However, if getting further behind were the sole explanation, one might expect screening rates in the morning and afternoon to mirror each other more closely rather than erode over the entire day.

Hsiang and colleagues⁵ focused on clinician ordering behavior, but it is important to consider the role of staff. As presented, completion tracked with rates of ordering, a necessary but insufficient step in cancer screening. Clinic staff may have also been more fatigued at the end of sessions and therefore less likely to take next, necessary actions for patients to arrange breast and colorectal cancer screening.

What about patients? Perhaps depleted, end-of-the-day patients were less likely to make immediate arrangements for follow-up or remember that they needed follow-up. To begin to isolate the relative associations of physician, staff, and patient behavior, Hsiang and colleagues⁵ could have reported whether screening completion rates were lower, conditional on ordering, depending on time of day. Based on the present and other analyses showing decision fatigue, patients should be concerned that their physician is relatively impaired later in the day. Should physicians be concerned that staff and patients are similarly impaired?

Regardless of the exact cause, it is tempting to conclude that decreased cancer screening ordering over the course of the clinic day indicates the need for interventions to address decision fatigue, such as modified schedules or mandatory breaks. However, cancer screening is operationally different from the other, previously described examples of decision fatigue in medicine. Visits for respiratory tract infections or back pain, by definition, require the presence of clinician and patient. Preventive services generally do not.

Many high-performing cancer screening interventions are delivered by nonclinician members of the care team, including proactive efforts to address screening gaps before office visits as well as mail- or phone-based outreach for patients due for screening.⁶ In addition to being an effective preventive approach, population-based screening efforts should decrease the amount of face-to-face work, freeing clinicians to devote greater attention to patients’ short- and long-term needs.

In demonstrating decreasing screening rates over the day, Hsiang and colleagues⁵ have provided evidence and argument for population health management. Efforts devoted to clinical preventive services should not depend on one’s appointment time, nor should they be confined to face-to-face visits. Payers should recognize that supporting a fee-for-service, face-to-face dominant system encourages overstuffed visits and probably impairs enrollees’ quality of care. Payers who support population health management could improve the quality of preventive care and improve the quality of within-visit acute and chronic care.

Practices and health systems should be organized so clinic time only contains high-value activities requiring the presence of physician and patient. To break the association of clinic time with cancer screening, preventive services should be delivered out of time. Physicians, to fit it all in when together with patients, should be able to focus on necessary care in the moment and not need to rush about like latter-day White Rabbits.

Accepted for Publication: March 26, 2019.

Published: May 10, 2019. doi:10.1001/jamanetworkopen.2019.3402

Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2019 Liss DT et al. JAMA Network Open.

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

Conflict of Interest Disclosures: Dr Liss reported receiving grants from the Agency for Healthcare Research and Quality, the Health Resources and Services Administration, and the National Institute of Diabetes and Digestive and Kidney Diseases; and a contract from United Healthcare Services, Inc. Dr Linder reported receiving grants from the National Institute on Aging, the National Institute on Drug Abuse, the Agency for Healthcare Research and Quality, the Gordon and Betty Moore Foundation, and the Peterson Center on Healthcare; and a contract from the Agency for Healthcare Research and Quality.