Physician Organizations’ Use of Behavioral Nudges to Influence Physician Behavior

ABSTRACT

Objectives: Because physicians’ decisions drive health care costs and quality, there is growing interest in applying behavioral economics approaches, including behavioral nudges, to influence physicians’ decisions. This paper investigates adoption of behavioral nudges by health system–affiliated physician organizations (POs), types of nudges being used, PO leader perceptions of nudge effectiveness, and implementation challenges.

Study Design: Mixed-methods study design (PO leader survey followed by in-depth qualitative interviews). Purposive sample of 30 health system–affiliated POs in 4 states; POs varied in size and quality performance.

Methods: We collected data between October 2017 and June 2019. The survey asked PO leaders to report their organization’s use of 5 categories of nudges to influence primary and specialty physicians’ actions. We conducted semistructured phone interviews to confirm survey responses, elicit examples of the nudges that POs reported using, understand how nudges were structured, and identify implementation challenges. We present descriptive tabulations of nudge use and effectiveness ratings. We applied thematic analysis to the interview data.

Results: Almost all POs in this study reported nudge use. Clinical templates, patient action lists, and altered order entry were most commonly used. However, PO leaders reported that nudge use was limited to a narrow range of clinical applications, not widespread across the organization, and mostly structured as suggestions rather than default actions or hard stops.

Conclusions: Nudge use remains limited in practice. Opportunities exist to expand use of nudges to influence physician behavior; however, expanding use of behavioral nudges will require PO investment of resources to support their construction and maintenance.

Am J Manag Care. 2022;28(9):473-476. https://doi.org/10.37765/ajmc.2022.89223

Takeaway Points

Physician organizations (POs) are applying behavioral nudges in the physician practice environment to influence physician behavior.

• Types of nudges most commonly used were clinical templates, patient action lists, and altered order entry for prescriptions, tests, procedures, and referrals.
• Nudge use was limited to a few clinical applications and not widespread across POs.
• POs cited challenges in building nudges and maintaining their currency given the evolving clinical evidence.
• Opportunities exist to expand the use of nudges to improve care and reduce wasteful spending. Research is needed to evaluate the effectiveness of nudges, which could promote greater use of nudges in clinical practice.

Physicians make many decisions that drive the cost and quality of care that patients receive.^1,2 Numerous studies detail widespread variation in quality of care,3 missed opportunities for evidence-based care,⁴ and substantial use of low-value care⁵ resulting in wasteful spending. To spur quality improvements and reduce health spending, policy makers and payers have applied various strategies, including performance-based financial incentives, to influence physician and health care organization behavior. Financial incentives have had modest and mixed effects,^6,7 and concern has been expressed that financial incentives may reduce providers’ intrinsic motivation to deliver the best care.⁸

With US health spending totaling $3.8 trillion and 17.7% of the gross domestic product in 2019,⁹ identifying mechanisms to control spending is a policy priority. Because physicians are central actors in efforts to improve health care costs and quality, interest in applying behavioral economics approaches, such as behavioral nudges, to influence physician behavior is growing.¹⁰ Such approaches seek to structure the decision-making environment in ways that support achievement of desired goals and reduce systematic errors that occur when individuals rely on mental shortcuts or heuristics to quickly make decisions.¹¹ Behavioral nudges are described as any aspect of the way in which choices are presented that influences or alters decision-making in a predictable way but does not forbid or penalize any choice.¹² Nudges can be designed in various ways: “opt in” vs “opt out,” defaults, checklists, social norms, commitment or attestations, or reframed options.

Physicians commonly are presented with information in ways unrelated to their clinical or cost effectiveness (eg, alphabetically), rather than strategically to influence choice (eg, first-line therapy or generic medication at the top of the list).¹³ Although behavioral economic interventions in health care targeting patient behaviors¹⁴ have been studied, assessments of the effectiveness of behavioral economic interventions to influence physician behavior are just emerging.¹⁵

Health system–affiliated physician organizations (POs) may be among the best equipped physician practices to implement behavioral nudges given their size, infrastructure, and organizational resources; as such, they may serve as a bellwether of use of nudges in the broader PO community. To understand the current landscape of nudge use to influence frontline physician behavior, we conducted a study of health system–affiliated POs to measure the extent of nudge use in practice, the types of nudges being deployed, how they were structured to influence behavior, PO leader assessments of nudge effectiveness, and implementation challenges that might shed light on issues related to uptake of nudges and expanded use.

DATA AND METHODS

This study is one component within a larger study examining actions that health systems are taking to deliver evidence-based care and improve cost and quality performance.¹⁶ The larger study selected a purposive sample of 24 systems (several with multiple POs) of varying size and performance for in-depth assessment¹⁷ in California, Washington, Minnesota, and Wisconsin (see eAppendix [available at ajmc.com]). We report findings from 30 POs (22 of the 24 health systems) that completed a survey and follow-up interview.

We collected data from PO leaders between October 2017 and June 2019. We first surveyed each PO on their use of 5 categories of nudges to influence primary and specialty physicians’ actions: (1) templates (eg, order sets or checklists); (2) altered order entry for prescribing (eg, arranging or grouping to highlight generic vs brand or preferred medication classes); (3) altered order entry for tests, procedures, or referrals (eg, prioritized listing of preferred options, approvals required for nonpreferred options); (4) patient action lists (ie, reminders to address care gaps); and (5) use of formal commitments to a preferred course of action (eg, physician pledge to select high-value care options). PO leaders rated the extent to which they agreed (5-point Likert scale¹⁸ ranging from “strongly disagree” to “strongly agree”) that various mechanisms, including nudges, were effective in improving physician performance. We then conducted 1-hour semistructured phone interviews with PO leaders to confirm survey responses, elicit examples of the nudges they reported using and how nudges were structured (default/required action or as informational/suggestive), and identify implementation challenges.

We present descriptive tabulations of nudge use and effectiveness ratings. Two analysts coded and summarized interview transcripts using Dedoose, with a pooled kappa interrater reliability score across all codes of 0.79 (details in eAppendix). From the qualitative interviews, we identified specific examples of nudge use, the structure of the different types of nudges that PO leaders reported their organizations were using, and challenges they cited with implementing nudges. The RAND Human Subjects Protection Committee approved this study.

RESULTS

All 30 system-affiliated POs were not-for-profit entities; 14 (47%) were academic medical center affiliated. By practice type, 25 (83%) were medical groups and 5 (17%) were independent practice associations (IPAs). Fifteen (50%) were in California, 7 (23%) were in Minnesota, 5 (17%) were in Wisconsin, and 3 (10%) were in Washington. The systems ranged in size from 45 primary care physicians (PCPs) to 1000, with a median of 173 PCPs.

Use of Nudges

Roughly two-thirds (19/30) of POs reported using 4 or 5 of the 5 types of nudges (eAppendix Figure); only 1 PO reported zero nudge use. Templates were the nudge type most frequently cited (27/30) (Figure 1), followed by patient action lists (26/30) and altered order entry for tests, procedures, or referrals (25/30). About two-thirds of POs reported using altered order entry for prescribing (21/30), whereas physician commitments were less commonly used (7/30). Regarding order sets, one PO leader commented, “If the order is fed, it pushes physicians to do the right thing for the right patient,” and another noted that order sets “help make it easy for providers to choose the pathway that’s most likely to result in the best outcome for patients.”

Design of Altered Order Entry Nudges

Among the 25 POs reporting use of altered order entry for tests, procedures, and/or referrals (Figure 2), most structured the nudge to make preferred options easier to access and select (22/25), followed by requiring exceptions or additional review (19/25), purposeful ordering of options (17/25), and prompts suggesting alternatives (17/25). A number of POs structured order entry to make it easier to order procedures or make referrals within their own system, with one PO stating, “We’re trying to keep care within our system. We think we give better care. It makes it easier to coordinate care.” Fewer POs reported making less-preferred options harder to access or select (12/25). Several POs indicated that they use advanced imaging decision support, requiring physicians to justify orders against indication criteria and, in some cases, suggesting alternatives, but never preventing ordering. One PO indicated that it was in the process of building nudges for laboratory orders that would show laboratory tests’ costs, whereas another had built Choosing Wisely low-value care guidelines into order entry. One PO leader commented, “We make it harder to do dumb things. [For instance,] there’s no evidence that vitamin D testing has helped anyone. We put in extra clicks to make people think about whether vitamin D testing is beneficial.”

Among those (21/30) using altered order entry for prescribing (Figure 2), nearly all made the preferred option easier to access or select (19/21), followed by purposeful ordering of prescription options (16/21). Several PO leaders stated that their nudges were designed to make it easy to do the right thing. For example, one PO presented physicians “with pain [medication] orders that are our preferred drugs and preferred quantities.” Defaults, when used, were used for generic prescribing. Less common approaches included prompts suggesting alternatives (13/21) and additional review or exception requirements (12/21).

Extent of Nudge Use

Only 2 of 30 POs reported widespread nudge use across their practice environment. PO leaders reported that application of nudges varied by specialty and were most commonly used in primary care, hospital medicine, larger practice sites, and in medical group practice models compared with IPA practice models. Most widely applied were smart sets, best practice alerts, and checklists for routine care.

Perceived Effectiveness of Nudges

Twenty-four of 30 PO leaders agreed or strongly agreed that nudges were effective mechanisms to improve physician performance, although system-level assistance (eg, care managers, chronic care clinics) (26/30), unblinded performance feedback and peer comparisons (25/30), and financial incentives (25/30) were perceived to be somewhat more effective (eAppendix Table).

Implementation Issues

Many PO leaders reported that more opportunities to apply nudges exist within their organization but would require time and resources to develop and be implemented. Leaders also cited challenges in customizing nudge tools, which requires resources, modifications to electronic health records (EHRs), and keeping up with evolving evidence. One leader commented, “Changing the opioid prescribing logic based on new evidence would require making 6000 changes in our EHR where opioid prescribing occurs.”

Some PO leaders expressed caution about applying too much pressure through nudges, remarking “it isn’t like we’re doing this to people” but rather “we engage with our physicians to determine the order set defaults based on the evidence.” Another leader stated that “…it sounds like we’re nagging people to do the right thing rather than recognizing we’ve engineered it wrong to begin with and we’re working with them to reengineer it. Rather than a nudge, it’s a support.”

PO leaders reported variability in physician use of nudge tools, noting that this reflects that some nudges are optional and can be ignored or bypassed. One leader commented, “Our challenge right now is how [to] start to drive behavioral change [so] that when you see those things, providers automatically think ‘This is a gap I need to close.’”

DISCUSSION

Among our sample of 30 health system–affiliated POs, most are using nudges, mainly order sets, checklists/patient action lists, and making preferred medications or care options easier to select. Among smaller POs and those not affiliated with health systems, we would expect less use of nudges because these POs may have less infrastructure and resources to support nudge development, implementation, and maintenance. Although nudges were being applied in practice, PO leaders’ descriptions of their nudges reveal that the applications tend to be limited in their scope and strength to drive behavior change.

We suspect that the frequent use of templates and action lists occurs because the EHR systems purchased by system-affiliated POs have similar, relatively standard features as part of the core EHR platform, prompted by national meaningful use policy standards. Beyond these standard EHR features, fewer custom nudge applications were occurring. Other sources of nudges were third-party add-ons and PO-developed tools. Some POs reported that clinician leaders were engaged in nudge development, particularly in designating which medications were preferred or ensuring that the clinical logic of the nudge was reflective of best practice within the PO.

The PO leaders’ experiences provide useful insights for other organizations that are considering the implementation of nudges. First, time and resources are required to develop and implement nudge tools, whether they are developed internally or with external entities. Although a nudge may target a single application, the targeted action frequently occurs in many different clinical scenarios and practice environments, adding complexity to building and maintaining nudges. Second, PO leaders grappled with how much pressure to apply, specifically whether prompted actions are optional or suggestive rather than defaults. Future research examining the effects on physician behavior of nudges that vary in strength could yield useful information for nudge design. Third, PO leaders highlighted the importance of engaging physicians in nudge design.

Limitations

We examined only system-affiliated POs that likely have greater resources and more advanced health information technology capabilities to build nudges and to ensure that they remain up to date with scientific evidence; therefore, our estimates of nudge use likely are biased upward relative to the broader set of POs in the United States. Self-reported data may suffer from social desirability bias (ie, reporting greater use than is true); however, participants understood that results would be blinded for reporting. Our study was not designed to represent a summative evaluation of the effectiveness of nudge efforts deployed by the POs. POs did not appear to be engaging in formal evaluations of their nudge applications. This represents an opportunity for future research.

CONCLUSIONS

Our study findings suggest that further opportunities exist to expand the use of nudges to influence physician behavior. Expanding the use of behavior nudges will require investment of resources to support nudge construction and maintenance. Research to evaluate the effectiveness of nudges across more clinical scenarios and applications could generate needed evidence to support greater use of nudges in clinical practice. 

Author Affiliations: RAND Corporation (CLD, AT, ROR), Santa Monica, CA.

Source of Funding: This work was supported through the RAND Center of Excellence on Health System Performance, which is funded through a cooperative agreement (1U19HS024067-01) between the RAND Corporation and the Agency for Healthcare Research and Quality (AHRQ). The content and opinions expressed in this publication are solely the responsibility of the authors and do not reflect the official position of AHRQ or HHS.

Author Disclosures: Dr Reid is employed as a physician by Brigham and Women’s Hospital and has received AHRQ and National Institutes of Health research grants and CMS research contracts. The remaining authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

Authorship Information: Concept and design (CLD, ROR); acquisition of data (CLD, AT, ROR); analysis and interpretation of data (CLD, AT, ROR); drafting of the manuscript (CLD, AT); critical revision of the manuscript for important intellectual content (CLD, AT, ROR); statistical analysis (CLD, AT); provision of patients or study materials (CLD, AT); obtaining funding (CLD); administrative, technical, or logistic support (CLD, AT); and supervision (CLD).

Address Correspondence to: Cheryl L. Damberg, PhD, RAND Corporation, 1776 Main St, Santa Monica, CA 90401. Email: damberg@rand.org.

REFERENCES

1. Emanuel EJ, Steinmetz A. Will physicians lead on controlling health care costs? JAMA. 2013;310(4):374-375. doi:10.1001/jama.2013.60073

2. Crosson FJ. Change the microenvironment: delivery system reform essential to control costs. Mod Healthc. 2009;39(17):20-21.

3. Brook RH, Lohr K, Chassin M, Kosecoff J, Fink A, Solomon D. Geographic variations in the use of services: do they have any clinical significance? Health Aff (Millwood). 1984;3(2):63-73. doi:10.1377/hlthaff.3.2.63

4. McGlynn EA, Asch SM, Adams J, et al. The quality of health care delivered to adults in the United States. N Engl J Med. 2003;348(26):2635-2645. doi:10.1056/NEJMsa022615

5. Reid RO, Mafi JN, Baseman LH, Fendrick AM, Damberg CL. Waste in the Medicare program: a national cross-sectional analysis of 2017 low-value service use and spending. J Gen Intern Med. 2021;36(8):2478-2482. doi:10.1007/s11606-020-06061-0

6. Damberg CL, Sorbero ME, Lovejoy SL, Martsolf GR, Raaen L, Mandel D. Measuring success in health care value-based purchasing programs: findings from an environmental scan, literature review, and expert panel discussions. Rand Health Q. 2014;4(3):9.

7. Eijkenaar F, Emmert M, Scheppach M, Schöffski O. Effects of pay for performance in health care: a systematic review of systematic reviews. Health Policy. 2013;110(2-3):115-130. doi:10.1016/j.healthpol.2013.01.008

8. Berwick DM. The toxicity of pay for performance. Qual Manag Health Care. 1995;4(1):27-33. doi:10.1097/00019514-199504010-00003

9. Martin AB, Hartman M, Lassman D, Catlin A; National Health Expenditure Accounts Team. National health care spending in 2019: steady growth for the fourth consecutive year. Health Aff (Millwood). 2021;40(1):14-24. doi:10.1377/hlthaff.2020.02022

10. Halpern SD, Ubel PA, Asch DA. Harnessing the power of default options to improve health care. N Engl J Med. 2007;357(13):1340-1344. doi:10.1056/NEJMsb071595

11. Tversky A, Kahneman D. Judgment under uncertainty: heuristics and biases. Science. 1974;185(4157):1124-1131. doi:10.1126/science.185.4157.1124

12. Thaler RH, Sunstein CR. Nudge: Improving Decisions About Health, Wealth, and Happiness. Yale University Press; 2008.

13. Patel MS, Volpp KG, Asch DA. Nudge units to improve the delivery of health care. N Engl J Med. 2018;378(3):214-216. doi:10.1056/NEJMp1712984

14. Volpp KG, Loewenstein G, Troxel AB, et al. A test of financial incentives to improve warfarin adherence. BMC Health Serv Res. 2008;8:272. doi:10.1186/1472-6963-8-272

15. Meeker D, Knight TK, Friedberg MW, et al. Nudging guideline-concordant antibiotic prescribing: a randomized clinical trial. JAMA Intern Med. 2014;174(3):425-431. doi:10.1001/jamainternmed.2013.14191

16. Ridgely MS, Buttorff C, Wolf LJ, et al. The importance of understanding and measuring health system structural, functional, and clinical integration. Health Serv Res. 2020;55(suppl 3):1049-1061. doi:10.1111/1475-6773.13582

17. Ridgely MS, Duffy E, Wolf L, et al. Understanding U.S. health systems: using mixed methods to unpack organizational complexity. EGEMS (Wash DC). 2019;7(1):39. doi:10.5334/egems.302

18. Likert R. A technique for the measurement of attitudes. Arch Psychol. 1932;22(140):5-55.