Use of Behavioral Economics to Improve Treatment of Acute Respiratory Infections (Main Study) (BEARI)

Bacteria resistant to antibiotic therapy are a major public health problem. The evolution of multi-drug resistant pathogens may be encouraged by provider prescribing behavior. Inappropriate use of antibiotics for nonbacterial infections and overuse of broad spectrum antibiotics can lead to the development of resistant strains. Though providers are adequately trained to know when antibiotics are and are not comparatively effective, this has not been sufficient to affect critical provider practices. The intent of this study is to apply behavioral economic theory to reduce the rate of antibiotic prescriptions for acute respiratory diagnoses for which guidelines do not call for antibiotics. Specifically targeted are infections that are likely to be viral. The objective of this study is to improve provider decisions around treatment of acute respiratory infections. The participants are practicing attending physicians or advanced practice nurses (i.e. providers) at participating clinics who see acute respiratory infection patients. A maximum of 550 participants will be recruited for this study. Providers consenting to participate will fill out a baseline questionnaire online. Subsequent to baseline data collection and enrollment, participating clinic sites will be randomized to the study arms, as described below. There will be a control arm, with clinic sites randomized in a multifactorial design to up to three interventions that leverage the electronic medical record: Order Sets that are triggered by electronic health record (EHR) workflow containing exclusively guideline concordant choices (SA, for Suggested Alternatives); Accountable Justifications triggered by discordant prescriptions that populate the note with provider's rationale for guideline exceptions (AJ); and performance feedback that benchmarks providers' own performance to that of their peers (PC, for Peer Comparisons). The outcomes of interest are antibiotic prescribing patterns, including prescribing rates and changes in prescribing rates over time. The intervention period will be over one year, with a one-year follow up period to measure persistence of the effect after EHR features are returned to the original state and providers no longer receive email alerts.

Each consented provider will be randomized to 1 of 8 cells in a factorial design with equal probability. If results of retrospective data analysis imply that design will be improved by stratification, randomization will be stratified by factors that could influence outcomes. Data will be collected from the clinics' Enterprise Data Warehouses which store copies of data recorded in the electronic health record. Data elements from qualifying office visits will be collected from coded portions of the electronic health record. An encounter is eligible for intervention if the patient's diagnosis is in the selected group of acute respiratory infections. The intervention EHR functions will be triggered when clinicians initiate an antibiotic prescription or enter a diagnosis for an acute respiratory infection that has a defined Order Set. If an antibiotic from a list of frequently misprescribed antibiotics is ordered and a diagnosis has not yet been entered, providers will be prompted to enter a diagnosis. If the diagnosis entered is acute nasopharyngitis; acute laryngopharyngitis/acute upper respiratory infection; acute bronchitis; bronchitis not specified as acute or chronic; or flu; the interventions will be triggered. The diagnosis-appropriate order set will pop-up for providers in the SA arm, while clinicians randomized to the AJ arm will receive an alert and be required to enter a brief statement justifying their antibiotic prescription if antibiotics are not indicated for the diagnosis entered. This note will then be added to the patient's medical record. Clinicians randomized to the Peer Comparison condition will receive email updates about their antibiotic prescribing practices relative to other clinicians in their practice.

Participants receive the Suggested Alternatives and Accountable Justification interventions, but not the Peer Comparison intervention.

Participants receive the Suggested Alternative and Peer Comparison interventions, but not the Accountable Justification intervention.

Participants receive the Accountable Justification and Peer Comparison interventions, but not the Suggested Alternative intervention.

Participants receive the Peer Comparison intervention, but do not receive the Suggested Alternatives or Accountable Justification interventions.

Participants receive the Suggested Alternatives intervention, but not the Accountable Justification or Peer Comparison interventions.

Participants receive the Accountable Justification intervention, but do not receive the Suggested Alternatives or Peer Comparison interventions.

Accountable Justifications triggered by discordant prescriptions that populate the electronic health record (EHR) note with provider's rationale for guideline exceptions (AJ).

Performance feedback that benchmarks providers' own performance to that of their peers (PC, for Peer Comparison).

Order Sets that are triggered by EHR workflow containing exclusively guideline concordant choices (SA, for Suggested Alternatives).

Assess inappropriate antibiotic prescribing rates (relative to all practices that did not receive the intervention) for antibiotic-inappropriate acute respiratory tract infection visits and no concomitant reason for antibiotic prescribing. based on the following non-antibiotic-appropriate International Statistical Classification of Diseases, version 9 (ICD-9) diagnoses: 460 Acute nasopharyngitis (common cold) 465 Acute laryngeopharyngitis/acute upper respiratory infection 466 Acute bronchitis 490 Bronchitis not specified as acute or chronic 487 Flu

We will monitor overall prescribing for the specified diagnoses and other acute respiratory infection diagnoses, including cough/fever and pneumonia.

Inclusion Criteria: A practicing attending physician or advanced practice nurse ("provider") at a participating clinic in 2011-2013 who sees acute respiratory infection patients. Exclusion Criteria: None.

Meeker D, Linder JA, Fox CR, Friedberg MW, Persell SD, Goldstein NJ, Knight TK, Hay JW, Doctor JN. Effect of Behavioral Interventions on Inappropriate Antibiotic Prescribing Among Primary Care Practices: A Randomized Clinical Trial. JAMA. 2016 Feb 9;315(6):562-70. doi: 10.1001/jama.2016.0275.

Gong CL, Zangwill KM, Hay JW, Meeker D, Doctor JN. Behavioral Economics Interventions to Improve Outpatient Antibiotic Prescribing for Acute Respiratory Infections: a Cost-Effectiveness Analysis. J Gen Intern Med. 2019 Jun;34(6):846-854. doi: 10.1007/s11606-018-4467-x. Epub 2018 May 8.

Gong CL, Hay JW, Meeker D, Doctor JN. Prescriber preferences for behavioural economics interventions to improve treatment of acute respiratory infections: a discrete choice experiment. BMJ Open. 2016 Sep 22;6(9):e012739. doi: 10.1136/bmjopen-2016-012739.

Persell SD, Friedberg MW, Meeker D, Linder JA, Fox CR, Goldstein NJ, Shah PD, Knight TK, Doctor JN. Use of behavioral economics and social psychology to improve treatment of acute respiratory infections (BEARI): rationale and design of a cluster randomized controlled trial [1RC4AG039115-01]--study protocol and baseline practice and provider characteristics. BMC Infect Dis. 2013 Jun 27;13:290. doi: 10.1186/1471-2334-13-290.