Behavioral interventions to reduce inappropriate antibiotic prescribing: A randomized pilot trial

Stephen D. Persell*, Jason N. Doctor, Mark W. Friedberg, Daniella Meeker, Elisha Friesema, Andrew Cooper, Ajay Haryani, Dyanna L. Gregory, Craig R. Fox, Noah J. Goldstein, Jeffrey A. Linder

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

55 Scopus citations


Background: Clinicians frequently prescribe antibiotics inappropriately for acute respiratory infections (ARIs). Our objective was to test information technology-enabled behavioral interventions to reduce inappropriate antibiotic prescribing for ARIs in a randomized controlled pilot test trial. Methods: Primary care clinicians were randomized in a 2×2×2 factorial experiment with 3 interventions: 1) Accountable Justifications; 2) Suggested Alternatives; and 3) Peer Comparison. Beforehand, participants completed an educational module. Measures included: rates of antibiotic prescribing for: non-antibiotic-appropriate ARI diagnoses, acute sinusitis/pharyngitis, all other diagnoses/symptoms of respiratory infection, and all three ARI categories combined. Results: We examined 3,276 visits in the pre-intervention year and 3,099 in the intervention year. The antibiotic prescribing rate fell for non-antibiotic-appropriate ARIs (24.7% in the pre-intervention year to 5.2% in the intervention year); sinusitis/pharyngitis (50.3 to 44.7%); all other diagnoses/symptoms of respiratory infection (40.2 to 25.3%); and all categories combined (38.7 to 24.2%; all p<0.001). There were no significant relationships between any intervention and antibiotic prescribing for non-antibiotic-appropriate ARI diagnoses or sinusitis/pharyngitis. Suggested Alternatives was associated with reduced antibiotic prescribing for other diagnoses or symptoms of respiratory infection (odds ratio [OR], 0.62; 95% confidence interval [CI], 0.44-0.89) and for all ARI categories combined (OR, 0.72; 95% CI, 0.54-0.96). Peer Comparison was associated with reduced prescribing for all ARI categories combined (OR, 0.73; 95% CI, 0.53-0.995). Conclusions: We observed large reductions in antibiotic prescribing regardless of whether or not study participants received an intervention, suggesting an overriding Hawthorne effect or possibly clinician-to-clinician contamination. Low baseline inappropriate prescribing may have led to floor effects. Trial Registration: NCT01454960.

Original languageEnglish (US)
Article number373
JournalBMC Infectious Diseases
Issue number1
StatePublished - Aug 5 2016


  • Acute respiratory infections
  • Antibiotics
  • Behavioral economics
  • Clinical decision support
  • Social psychology

ASJC Scopus subject areas

  • Infectious Diseases


Dive into the research topics of 'Behavioral interventions to reduce inappropriate antibiotic prescribing: A randomized pilot trial'. Together they form a unique fingerprint.

Cite this