stethoscope pills and first aid kit

Personalized Nudges Aren’t Just for Patients; They’re for Clinicians Too

read time - icon

0 min read

Jan 26, 2022

In my previous article, I discussed personalized nudging in healthcare specifically in the context of improving patient health behavior. Although a recent trend, it’s a rather widely known concept that has gained a lot of traction. Here, I explore how the same concept of personalization could possibly be applied to improving clinician behavior. 

Over the last decade, several studies have demonstrated the effectiveness of nudges for improving clinician decision-making. However, very few have focused on how different physician characteristics may be associated with varying levels of nudge effectiveness. Furthermore, electronic health record (EHR) usage and activity patterns can provide valuable data that can be used to identify suboptimal decision-making by clinicians. The combination of these insights could aid in personalizing nudges, but why might we not simply deliver nudges to all clinicians?

Before I dive into why personalizing nudges might be important for clinicians, I want to discuss two specific groups of studies to help contextualize the discussion. ​​I focus on these two particular examples for a few reasons:

  1. They are both examples of nudges targeted at improving clinical decision-making by aligning physician behavior with established treatment guidelines;
  2. The interventions in the studies work by disrupting clinical workflow, unlike other forms of nudges such as setting default choices; and
  3. Concurrent research has been conducted on the clinician characteristics associated with suboptimal decision-making patterns of interest. 

Behavioral Science, Democratized

We make 35,000 decisions each day, often in environments that aren’t conducive to making sound choices. 

At TDL, we work with organizations in the public and private sectors—from new startups, to governments, to established players like the Gates Foundation—to debias decision-making and create better outcomes for everyone.

More about our services

Therapeutic inertia among neurologists

Therapeutic inertia (TI) is defined as the failure to escalate treatment when warranted. In a recent study where neurologists were exposed to simulated case scenarios of patients with multiple sclerosis (MS), it was found that over 70% of neurologists demonstrated TI during treatment decisions. This led to suboptimal decisions in approximately 20% of cases.1

Interestingly, the research group was able to conduct a deeper analysis and identify certain physician characteristics that were associated with an increased likelihood of TI. These factors include aversion to ambiguity, level of expertise in MS care, and even country of practice.

The same group also developed a short educational intervention that leveraged what is known as a “traffic light system.” The traffic light system uses color associations to help clinicians identify high-risk patients (i.e. red signifies high-risk and escalates treatment while yellow encourages re-evaluation in 6 months). More details about the intervention can be found in their study. A randomized controlled trial proved its effectiveness in reducing the prevalence of TI during MS care. This is promising for improving clinical outcomes for MS patients.2

It is important to note that the traffic-light intervention was not delivered as a nudge in this study. Rather, it was delivered to study participants that were given case scenarios that represented real-world clinical situations. However, in a real-world clinical setting, it could be presented as a nudge by being delivered through electronic health records or another format during patient care.

Prescribing statins for cardiovascular disease patients

Statins have been proven to significantly lower the risk of cardiovascular events and mortality. Several studies, however, have demonstrated their underutilization: about 50% of patients who would benefit from statins (according to national guidelines) are not prescribed them by physicians.3 On top of this, those who are prescribed statins more often than not receive a lower-than-optimal dose. Researchers at the Penn Nudge Unit, a behavioral design team embedded within the University of Pennsylvania Health System, sought to evaluate whether nudging could improve guideline-concordant statin prescribing.4

Their first study compared a control with two intervention arms: an active choice intervention with and without peer comparison performance feedback. Physicians in the first arm (just the active choice intervention) received an email that included a list of their patients who would benefit from statins but were not yet prescribed them. Other data in the email included relevant patient clinical characteristics and the national guidelines for statin prescribing. Here, the active choice nudge was asking the physician to review the list and select whether or not to prescribe a statin within 1 week.

In the peer comparison arm, physicians also received information about how their statin prescribing rates compared with those of other physicians at their institution. This nudge was based on what is known about social norms and their effects on behavior change. The active choice + peer comparison intervention arm demonstrated a significant increase in appropriate statin prescribing. 

In another study, the same research group identified clinician characteristics that were associated with statin prescribing. Specifically, being female, being a physician assistant, and having more years of experience were associated with a lower likelihood of prescribing statins.5

The future of clinician-directed nudges

Why does personalization hold so much promise for the future of clinician nudging? The first reason is relatively straightforward, and parallels that of the reasoning for patients: certain clinicians may be more responsive to a particular type of nudge than another. For example, clinicians who are more competitive may respond better to peer-comparison nudges.

It would be interesting if, in the future, EHRs were able to test an array of different nudges for particular clinical contexts and monitor clinician activity and potential behavior change to then deliver more appropriate and personalized nudges. Knowing certain clinician characteristics (i.e. their level of expertise) can help fuel the machine learning algorithms in identifying the appropriate nudge. 

Other value additions for nudge personalization may be more discrete. Flaws in clinician decision-making extend beyond a particular treatment context. Clinicians are dealing with a variety of screening techniques, different health conditions, prescription drugs, treatment options, etc. As the body of research in this field continues to grow, it will reveal the need for many other nudges in the clinical workflow.

How personalization can curb burnout

However, in real-world practice, implementing several nudges at a time is simply not feasible. Explicit nudges like the ones described in the examples above often require a clinician to redirect their attention. Too many such interruptions can contribute to “alert fatigue,” leading clinicians to become desensitized and ignore these nudges altogether. Tangentially, EHR alerts are predictive of burnout among the healthcare workforce, so curbing the number of alerts remains a priority.6

Personalization can be a potential solution to minimizing alert fatigue from nudges. By identifying which nudges are most urgent for particular clinicians, future technology can prioritize different interventions based on EHR data and clinician characteristics. For example, the active choice nudge described above may only be delivered to clinicians who consistently under-prescribe statins, or the traffic light intervention to the ~70% of neurologists that exhibit therapeutic inertia.

This approach shows promise. A 2020 study used EHR data to phenotype physician practice patterns and categorize physicians based on variables such as years of training, estimated workload, and number of patients seen.7 They found that an active choice EHR nudge was significantly effective in increasing influenza vaccination rates among clinicians with higher clinical workload, but had no effect among those with a lower clinical workload.

As clinician behavior changes, AI could also potentially identify when a nudge is no longer necessary to produce the desired behavior change, and progressively transition to a different nudge to target a different behavior. Even factors as simple as the time of day, or how many hours a clinician is into their shift, can pave the way for personalization. For example, as the day progresses, clinicians tend to order fewer cancer screening tests,8 but active choice nudging through the EHR was proven effective in increasing physician orders for colonoscopies and mammograms.9 Time-specific nudges can also help to decrease alert fatigue.

Future directions

Nudges can encourage more guideline-concordant prescribing behavior among clinicians, but before these nudges are delivered at scale, several questions must still be answered. Future research must focus on how different nudges will interact with each other, and how this could affect clinician behavior. We also need more research to look at the long-term effects of these nudges. 

From this future research, while the intuitive next step would be for health systems to prioritize certain decision tools and EHR changes over others, I argue that personalized nudging for clinicians offers a solution with fewer trade-offs while maximizing effectiveness.

While it’s true that in an ideal world, all nudges in practice will be discrete and embedded seamlessly into the design of EHRs (i.e. default nudges), this is rather impractical for a variety of reasons. For example, some individuals may respond to other forms of nudges (i.e. active choice), and default nudges may not be feasible in certain contexts. Personalized nudging can help to identify which form of nudge is most effective for an individual clinician, and also help to identify those who are most in need of a nudge.

While this sounds rather imaginative, the technology is already being used for improving patient behavior. Why not work towards applying it to clinicians too?

References

1. Almusalam N, Oh J, Terzaghi M, Maurino J, Bakdache F, Montoya A, et al. Comparison of Physician Therapeutic Inertia for Management of Patients With Multiple Sclerosis in Canada, Argentina, Chile, and Spain. JAMA Netw Open. 2019;2: e197093–e197093.

2. Saposnik G, Mamdani M, Montalban X, Terzaghi M, Silva B, Saladino ML, et al. Traffic Lights Intervention Reduces Therapeutic Inertia: A Randomized Controlled Trial in Multiple Sclerosis Care. MDM Policy Pract. 2019;4: 2381468319855642.

3. Ma J, Sehgal NL, Ayanian JZ, Stafford RS. National trends in statin use by coronary heart disease risk category. PLoS Med. 2005;2. doi:10.1371/journal.pmed.0020123

4. Patel MS, Kurtzman GW, Kannan S, Small DS, Morris A, Honeywell S, et al. Effect of an Automated Patient Dashboard Using Active Choice and Peer Comparison Performance Feedback to Physicians on Statin Prescribing: The PRESCRIBE Cluster Randomized Clinical Trial. JAMA Network Open. 2018;1. doi:10.1001/jamanetworkopen.2018.0818

5. Kannan S, Asch DA, Kurtzman GW, Honeywell S Jr, Day SC, Patel MS. Patient and physician predictors of hyperlipidemia screening and statin prescription. Am J Manag Care. 2018;24: e241–e248.

6. Gregory ME, Russo E, Singh H. Electronic Health Record Alert-Related Workload as a Predictor of Burnout in Primary Care Providers. Appl Clin Inform. 2017;8: 686.

7.Changolkar S, Rewley J, Balachandran M, Rareshide CAL, Snider CK, Day SC, et al. Phenotyping physician practice patterns and associations with response to a nudge in the electronic health record for influenza vaccination: A quasi-experimental study. PLoS One. 2020;15: e0232895.

8. Hsiang EY, Mehta SJ, Small DS, Rareshide CAL, Snider CK, Day SC, et al. Association of Primary Care Clinic Appointment Time With Clinician Ordering and Patient Completion of Breast and Colorectal Cancer Screening. JAMA Netw Open. 2019;2: e193403–e193403.

9. Patel MS, Volpp KG, Small DS, Wynn C, Zhu J, Yang L, et al. USING ACTIVE CHOICE WITHIN THE ELECTRONIC HEALTH RECORD TO INCREASE PHYSICIAN ORDERING AND PATIENT COMPLETION OF HIGH-VALUE CANCER SCREENING TESTS. Healthcare (Amsterdam, Netherlands). 2016;4: 340.

About the Author

Sanketh Andhavarapu portrait

Sanketh Andhavarapu

Staff Writer

Sanketh is an undergraduate student at the University of Maryland: College Park studying Health Decision Sciences (individual studies degree) and Biology. He is the co-Founder and co-CEO of Vitalize, a digital wellness platform for healthcare workers, and has published research on topics related to clinical decision-making, neurology, and emergency medicine and critical care. He is also currently leading business development for a new AI innovation at PediaMetrix, a pediatric health startup, and previously founded STEPS, an education nonprofit. Sanketh is interested in the applications of behavioral and decision sciences to improve medical decision-making, and how digital health and health policy serve as a scalable channel to do so.

Read Next

Notes illustration

Eager to learn about how behavioral science can help your organization?