Leveraging Behavioral Economics to Reduce Low-Value Prostate Cancer Screening

EDITORIAL TEXT

Low-value care (LVC) is common in the United States health care system, contributing up to $28 billion in waste annually, and innovative solutions are vital if meaningful reduction is to occur.¹ Behavioral economic theory holds promise for such innovations, interweaving human psychology and the economics of incentives to describe decision-making as it occurs in the real world. In contrast to traditional economic theory, behavioral economics acknowledges that humans are not perfect and their decisions are not always rational. Currently, applications of behavioral economics to improve urologic oncology care are lacking.

Value can be defined as outcomes achieved relative to costs. Consequently, LVC represents a scenario when either the outcomes achieved are poor, or the cost to achieve them is high. Evidence suggests that low-cost items ordered in high volume are the largest contributors to overall waste created by LVC, especially when both direct and downstream costs are considered.² Low-value prostate specific antigen (PSA) testing sits firmly in this category.

Experts have often debated whether to offer PSA-based prostate cancer (PCa) screening amidst discordant guideline statements. A notable shift occurred in 2018, when new developments in two large randomized trials drove the USPSTF to update their Grade D recommendation (“recommend against”) to Grade C (“offer in select patients”) for men aged 55–69. Higher interest in PSA screening is likely in the wake of this shift, raising concern for increased utilization of LVC as providers adapt their practice without clear and definitive guidance. Furthermore, the field’s recent focus on debating whether to test at all has distracted from educating those who do test on how to preserve value.

Screening for PCa works best when targeted to the appropriate population. Current evidence suggests the greatest benefit using a blood test (PSA) is in men aged 55–69 who have a life expectancy of greater than 10 years.³ Still, substantial testing is done in men younger than 45, older than 70, and in those with limited life expectancy. In our large health system, 40%–50% of all PSA testing is low-value.⁴

Currently, the underlying drivers of LVC are inadequately understood with many factors likely contributing. The practice of medicine comprises art and science alike, and constant advancements make keeping abreast of current evidence-based best practices challenging. This combination produces substantial practice variation due to learned behavior or habit. Furthermore, incorporating new evidence into routine practice takes years, with such delays in dissemination and implementation certainly contributing.

Evidence is lacking on the efficacy of efforts to reduce LVC, and a 2017 systematic review identified only two studies focused on cancer.⁵ Interventions may target both demand for and supply of LVC. Currently, supply-side efforts are more thoroughly studied, with the greatest evidence behind clinical decision support and clinical education (Table 1). These domains provide an opportune arena for the injection of behavioral economic principles.

Table 1 –

Supply-side interventions with potential to reduce low-value care

Intervention Type	Description
Provider Education	Educating providers on the prevalence and specific types of LVC often leveraging national guidelines and coupled with reminders. (e.g. Choosing Wisely campaign)
Provider Feedback*	Supplying providers with feedback combined with benchmarks on provider-specific utilization of LVC.
Peer Comparisons*	Special type of provider feedback in which performance and utilization of LVC is compared against that of peers.
Clinical Decision Support*	Integrated support for decisions to reduce LVC through the electronic health record, administrative ordering restrictions, or promotion of clinical pathways/guidelines.
Public Commitments*	Providers make public pledges to avoid or reduce certain low-value services.
Risk Sharing	Contrasting payment model to fee-for-service in which providers are paid for the care of a defined population or episode of care (e.g. Accountable Care Organization, Bundled Payment).
Pay-for-Performance	Financial rewards for reducing low-value (or increasing high-value) care.
Insurer Restrictions	Using practices such as prior-authorization and network construction to discourage provision of LVC.

LVC = low-value care

^*May leverage principles of behavioral economics

In theory, provider supply and patient demand for LVC should always follow the strict rules of economics. In practice, however, they tend to align with decidedly human patterns of behavior.⁶ Unlike purely rational economic models, humans are fallible, imperfect, and susceptible to the fundamental pitfalls described by behavioral economics.

Providers experience decision fatigue when attempting to maintain high-quality decisions throughout a busy clinic or shift. Low-value antibiotic prescribing rises 26% by the fourth hour of clinic compared to the first, with similar patterns observed in breast and colorectal cancer screening.^7,8 PSA testing decisions, which are cognitively taxing due to ambiguous and conflicting guidelines, likely exhibit a similar pattern.

Indeed, in forthcoming work, we found that the likelihood of PSA testing was greatest at 8am and declined steadily towards noon, remaining depressed through the afternoon. Appropriate testing experienced a relatively larger decline than inappropriate testing. This may represent providers falling back on a default position, such as not testing, as decision fatigue accumulates.

Loss aversion may also drive aspects of low-value screening. A cancer diagnosis constitutes a loss from the patient perspective, while missing a diagnosis when still curable is the provider equivalent. Patients often overestimate the risks of PCa and the ability of screening tests to catch it early. Reflecting this loss aversion, providers tend to overemphasize the benefits of screening while minimizing the associated risks.

Patients and providers alike may also overly discount the future and indulge in care that seems satisfying in the present; the quintessential example of this being poor antibiotic stewardship. Either party may pursue low-value PCa screening to assuage present fears, despite potential downstream harm in the form of unnecessary diagnostic procedures and risky treatments.

Novel approaches for reducing low-value PCa screening emerge when behavioral economic principles are appropriately applied. Our team is tapping into these principles to develop a strategic clinical decision support (CDS) tool that leverages the electronic health record (EHR) to shape provider behavior. When a PSA test is ordered, our algorithm will analyze key patient-specific data and clearly identify if the test is low-value. Interventions aimed at reducing low-value PCa screening will directly employ this EHR-driven tool.

Behavioral economic principles exploited by this tool include relative social ranking (peer comparisons), nudges (best practice alerts), and defaults (workflow pre-selections). Peer comparisons have shown promise in reducing LVC in numerous medical and surgical domains,⁹ while nudges have been widely successful in realms such as repeat lab testing.¹⁰ Furthermore, thoughtfully crafted default selections will prioritize high-value practices while warding off low-value behaviors such as those produced by decision fatigue.

Conclusions

The US health care system is plagued by LVC which generates substantial financial waste. While the etiology of LVC in urologic oncology remains incompletely understood, both rational and irrational behaviors likely contribute. Novel interventions aimed at reducing low-value PCa screening should consider incorporating strategies from the behavioral economics playbook—including peer comparisons, nudges, and defaults—to combat mechanisms like decision fatigue and ultimately improve care. Implementing these strategies via carefully designed CDS, which has been effective at reducing LVC in a variety of settings, may prove to be an important strategy.

TAKE HOME MESSAGE.

Behavioral economic principles model decision-making behavior and offer promising and unexplored mechanisms for understanding the etiology of low-value care in urologic oncology. Clinical decision support built around these principles is poised to substantially reduce wasteful spending in prostate cancer screening.