Reducing High-Risk Geriatric Polypharmacy Via EHR Nudges: Pilot Phase

Polypharmacy is common among older adults in the United States and is associated with harms such as adverse drug reactions and higher costs of care. This pilot-phase project is designed to test two electronic health record (EHR)-based behavioral economic nudges to help primary care clinicians reduce the rate of high-risk polypharmacy among their older adult patients.

Study Overview

• Status: Withdrawn
• Conditions: Fall; Congestive Heart Failure; Adverse Drug Event; Chronic Kidney Failure
• Intervention / Treatment: Behavioral: Commitment nudge; Behavioral: Justification nudge

Detailed Description

Polypharmacy increases the likelihood of being prescribed and harmed by high-risk medications. As noted in the 2014 National Action Plan for Adverse Drug Event (ADE) Prevention, polypharmacy both increases the likelihood of being prescribed high-risk medications and increases the likelihood that these high-risk medications will lead to adverse drug events. This pilot-phase study is intended to test clinicians' perceptions of EHR-based nudges designed to reduce high-risk polypharmacy among patients aged 65 years or more, thereby enabling investigators to refine the nudges, and to generate outcomes data that will inform power calculations for a subsequent larger study (the main study) of the nudges' effectiveness.

In this pilot-phase study, the investigators will deploy 2 EHR-based behavioral nudges (a commitment nudge and a justification nudge) among 18 or more primary care clinicians in 3 primary care practices (6 clinicians or more per practice) affiliated with Northwestern University for approximately 4 months. The 3 practices participating in the pilot will be a convenience sample of Northwestern-affiliated practices known to study investigators.

The investigators will randomly assign each of the 3 participating pilot practices to 1 of 3 arms: (1) commitment nudge, (2) justification nudge, or (3) both commitment and justification nudges. Randomization will be at the practice level, without replacement, thus assigning exactly 1 practice to each arm. All participating clinicians within a given practice will receive the same nudges.

Northwestern-affiliated practices that do not participate in the pilot will constitute a fourth arm of this pilot study.

The investigators will ask leaders of participating practices for their qualitative observations on how clinicians and patients experience the nudges (e.g., how the nudges affect workflows). The investigators also will collect data on the outcome measures before and during the approximately 4-month pilot period and compare these data to contemporaneous outcomes measures generated by Northwestern-affiliated practices that do not participate in the pilot.

• Study Type: Interventional
• Phase: Not Applicable

Contacts and Locations

Study Locations

• United States
  • Illinois
    • Chicago, Illinois, United States, 60611
      • Northwestern Medicine

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: 18 years and older (Adult, Older Adult)
• Accepts Healthy Volunteers: No

Description

Inclusion Criteria:

• Primary care clinicians practicing in one of the participating Northwestern-affiliated practices

Exclusion Criteria:

• none

Study Plan

How is the study designed?

Design Details

• Primary Purpose: Health Services Research
• Allocation: Randomized
• Interventional Model: Factorial Assignment
• Masking: None (Open Label)

Number of Arms

4

Arms and Interventions

Participant Group / Arm	Intervention / Treatment
Experimental: Commitment nudge; Primary care clinicians in the practice assigned to this arm will receive only the commitment nudge.	Behavioral: Commitment nudge; The commitment nudge will ask clinicians to commit to discussing high-risk polypharmacy at the next office visit with patients who have high-risk polypharmacy. Clinicians who commit will receive a reminder just before the next office visit begins. The patient will also receive notification of the commitment via EHR patient portal. This nudge will be operationalized via two sequential clinician-facing EHR best practice alerts (BPAs): the first at time of opening any encounter (including encounters other than a face-to-face office visit, e.g., medication refills), and the second (contingent on making a commitment to discuss high-risk polypharmacy) at time of opening the subsequent encounter for a face-to-face office visit. Each of these BPAs will describe the specific high-risk polypharmacy criterion (or criteria) the patient meets, the specific harms associated with the medication(s) triggering the criterion/criteria, and lower-risk alternative treatment strategies.
Experimental: Justification nudge; Primary care clinicians in the practice assigned to this arm will receive only the justification nudge.	Behavioral: Justification nudge; The justification nudge will ask clinicians who prescribe or renew a drug that meets high-risk polypharmacy criteria (in the context of the patient's other medications) to write a brief justification for prescribing this high-risk medication. This written justification will be recorded in the patient's medical record. The best practice alert requesting the justification will also describe the specific high-risk polypharmacy criterion (or criteria) the patient meets, the specific harms associated with the medication(s) triggering the criterion/criteria, and lower-risk alternative treatment strategies.
Experimental: Commitment + Justification nudges; Primary care clinicians in the practice assigned to this arm will receive both the commitment nudge and the justification nudge.	Behavioral: Commitment nudge; The commitment nudge will ask clinicians to commit to discussing high-risk polypharmacy at the next office visit with patients who have high-risk polypharmacy. Clinicians who commit will receive a reminder just before the next office visit begins. The patient will also receive notification of the commitment via EHR patient portal. This nudge will be operationalized via two sequential clinician-facing EHR best practice alerts (BPAs): the first at time of opening any encounter (including encounters other than a face-to-face office visit, e.g., medication refills), and the second (contingent on making a commitment to discuss high-risk polypharmacy) at time of opening the subsequent encounter for a face-to-face office visit. Each of these BPAs will describe the specific high-risk polypharmacy criterion (or criteria) the patient meets, the specific harms associated with the medication(s) triggering the criterion/criteria, and lower-risk alternative treatment strategies.; Behavioral: Justification nudge; The justification nudge will ask clinicians who prescribe or renew a drug that meets high-risk polypharmacy criteria (in the context of the patient's other medications) to write a brief justification for prescribing this high-risk medication. This written justification will be recorded in the patient's medical record. The best practice alert requesting the justification will also describe the specific high-risk polypharmacy criterion (or criteria) the patient meets, the specific harms associated with the medication(s) triggering the criterion/criteria, and lower-risk alternative treatment strategies.
No Intervention: Non-participating; Primary care clinicians in Northwestern-affiliated practices other than the 3 pilot-participating practices will not receive any study interventions.

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
High-risk polypharmacy criterion: Fall condition-drug interaction	Denominator: count of patients aged >=65 years AND who have >=1 fall marker within the prior 2 years (ICD-10 codes for falls or hip fractures; OR who are recorded in the EHR as being high-risk for falls). Numerator: count of patients in the denominator who have 5+ medications in their EHR med list AND who have >=1 medication in their EHR med list that is a tricyclic antidepressant, Z-drug (e.g., zolpidem), benzodiazepine, antipsychotic, anticonvulsant, SSRI/SNRI, or opioid.	28 months
High-risk polypharmacy criterion: Fall drug-drug interaction	Denominator: count of patients aged >=65 years. Numerator: count of patients in the denominator who have 5+ meds AND who have >=3 medications in their EHR med list that are tricyclic antidepressants, Z-drugs (e.g., zolpidem), benzodiazepines, antipsychotics, anticonvulsants, SSRIs/SNRIs, or opioids.	28 month
High-risk polypharmacy criterion: CKD-glyburide/glimepiride interaction	Denominator: count of patients aged >=65 years AND who have most-recent estimated glomerular filtration rate (eGFR) < 60 as estimated by the Cockcroft-Gault equation. Numerator: count of patients in the denominator who have 5+ meds AND glyburide or glimepiride in their EHR med list.	28 months
High-risk polypharmacy criterion: CKD-NSAID interaction	Denominator: count of patients aged >=65 years AND who have most-recent eGFR < 30 as estimated by the Cockcroft-Gault equation. Numerator: count of patients in the denominator who have 5+ meds AND a systemically-absorbed non-steroidal anti-inflammatory drug (NSAID) in their EHR med list.	28 months
High-risk polypharmacy criterion: CHF-NSAID interaction	Denominator: count of patients aged >=65 years AND who have 5+ meds AND congestive heart failure (identified via ICD-10 code within prior 2 years). Numerator: count of patients in the denominator who have a systemically-absorbed non-steroidal anti-inflammatory drug (NSAID) in their EHR med list.	28 months
High-risk polypharmacy criterion: CHF-non-dihydropyridine calcium channel blocker interaction	Denominator: count of patients aged >=65 years AND who have low-ejection-fraction congestive heart failure (identified via ICD-10 code within prior 2 years). Numerator: count of patients in the denominator who have 5+ meds AND a non-dihydropyridine calcium channel blocker in their EHR med list.	28 months
High-risk polypharmacy criterion: CHF-thiazolidinedione interaction	Denominator: count of patients aged >=65 years AND who have low-ejection-fraction congestive heart failure (identified via ICD-10 code within prior 2 years). Numerator: count of patients in the denominator who have 5+ meds AND a thiazolidinedione in their EHR med list.	28 months

Collaborators and Investigators

• Sponsor: RAND
• Collaborators: University of California, Los Angeles; Northwestern University; University of Pittsburgh; National Institute on Aging (NIA); University of Southern California
• Investigators: Principal Investigator: Mark W Friedberg, MD, MPP, RAND

Study record dates

Study Major Dates

• Study Start (Actual): February 1, 2019
• Primary Completion (Actual): August 1, 2019
• Study Completion (Actual): September 1, 2019

Study Registration Dates

• First Submitted: December 29, 2018
• First Submitted That Met QC Criteria: December 29, 2018
• First Posted (Actual): January 2, 2019

Study Record Updates

• Last Update Posted (Estimated): December 19, 2025
• Last Update Submitted That Met QC Criteria: December 17, 2025
• Last Verified: December 1, 2025

More Information

Terms related to this study

• Keywords: Primary Health Care; Polypharmacy; Geriatrics; Decision Support Systems, Clinical; Electronic Health Records; Economics, Behavioral; Deprescriptions
• Additional Relevant MeSH Terms: Urogenital Diseases; Cardiovascular Diseases; Pathologic Processes; Male Urogenital Diseases; Kidney Diseases; Urologic Diseases; Female Urogenital Diseases; Female Urogenital Diseases and Pregnancy Complications; Heart Diseases; Chronic Disease; Disease Attributes; Chemically-Induced Disorders; Renal Insufficiency; Renal Insufficiency, Chronic; Pathological Conditions, Signs and Symptoms; Heart Failure; Drug-Related Side Effects and Adverse Reactions; Kidney Failure, Chronic; Behavior
• Other Study ID Numbers: STU00207210; R21AG057396 (U.S. NIH Grant/Contract)

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: NO

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No