More than250millionoutpatient antibiotic prescriptions are written in the US each year,1 most of which are for acute respiratory tract infections.A largeproportionof theseprescriptions are unnecessary.2 Antibiotic overuse leads to unnecessary adverse drug effects and antimicrobial-resistant infections, which harm patients. To address this threat, professional organizations and regulatory agencies have called for acute care hospitals to establish antimicrobial stewardshipprograms,which are quality improvement and patientsafety initiatives tooptimize antimicrobial use.3 The vast majority of antibiotic use, however, occurs in the outpatient setting, mostly in internal medicine, familypractice, pediatric, anddental offices,1where there is virtually no formal antimicrobial stewardship. In addition, although themost resistantorganismsusually involve infections among hospitalized patients, difficult-to-treat infections such as methicillin-resistant Staphylococcus aureus and extended-spectrumbeta lactamase–producingEnterobacteriaciae arenowcommonly isolated fromcommunity-onset infections. Thus, adapting the principles of antimicrobial stewardship to the ambulatory setting should be prioritized. The appropriate use of antibiotics for the most common acute respiratory tract infections isknownbutnot followed.For example, numerous randomized trials have shown that otherwisehealthyadultsshouldnotreceiveantibiotics foracutebronchitis.However, roughly75%ofadults receiveantibiotics for this condition, a proportion that has not changed in more than 20 years.4 Even when an antibiotic is indicated, often the wrong one is chosen: roughlyhalf of antibiotics for childrenarebroadspectrum, second-line agents,5 and the most commonly prescribed antibiotic for adults is azithromycin, despite this drug being recommended as the first-line choice for relatively few conditions. So how can these issues be addressed? Changing clinician behavior is challenging. Public health campaigns help, but they are time consuming and expensive. Education is important, butwhenasked, internists andpediatricians agree that antibiotics are overused and that antibiotic resistance is a problem, but not in their practice.6 Factors other than specific medical need drive prescribing behavior. These include perceived patient (or parent) pressure,7,8 the presence of trainees,9 and even the time of day10 or the race of the patient.11 Different levers that shape clinician behavior need to be considered, especially at the point of care, when thedecision toprescribe ismade. In the acute care setting, including the ambulatory setting, behavioral determinants and social norms clearly influence antibiotic prescribing.12 For example, a simple “behavioral nudge” in the form of a signed commitment letter posted in the patient room was associatedwith a 20%reduction in inappropriate prescribing for viral acute respiratory tract infections.13 In this issue of JAMA, Meeker and colleagues14 report the results of a cluster randomized trial to reduce inappropriate antibiotic prescribing in 47 primary care practices in Boston orLosAngeles staffedby248clinicians.The investigators compared the effect of 3 different behavioral interventions that were triggered when an antibiotic was ordered for a targeted acute respiratory tract infection: suggestedalternatives,which triggeredanelectronichealth record–basedpop-upstating that “antibiotics are generally not indicated for this” and provided suggestions for nonantibiotic alternatives (eg, decongestants); accountable justification, which required the clinician to document the justification for antibiotic treatment (using free text) or, if nothing was entered, “no justification given”wouldpopulate thechart; orpeer comparison,whereby clinicianswereemailedamessagestatingthat theywerea“high performer” if their inappropriateantibioticprescribing ratewas in the topdecile and“nota topperformer” if their ratewasoutside this range. Practiceswere randomly assigned to0, 1, 2, or all 3 of the interventions and followed up for 18 months. The rate of antibiotic prescribing for upper respiratory tract infection, acute bronchitis, or influenza—all conditions for which antibiotics are not indicated—was then compared across each of the 3 intervention groups. The resultsof these simple interventionswerepromising. Over the18-month intervention, inappropriateprescribingdecreased ineachgroupsubstantially: from22%to6%withsuggestedalternatives, 23%to5%with accountable justification, and 20% to 4% with peer comparison. However, because inappropriate prescribing declined substantially in the control group (24% to 13%), the absolute between-group reductions inprescribingwereonlyapproximately5%foreachgroupand wereonly statistically significant in the accountable justificationgroup(differenceofdifferences inprescribingrates,−7.0% [95%CI,−9.1%to−2.9%];P < .001)andpeercomparisongroups (difference of differences, −5.2% [95% CI, −6.9% to −1.6%]; P < .001) but not the suggested alternative groups (difference of differences, −5.0% [95%CI, −7.8% to0.1%];P = .66). Investigatorsobservednocodeshifting fromviraldiagnoses tobacterialdiagnoses,whichcouldsimulatea reduction in inappropriate use based on the outcome definition used, and did not find a clinically significant increase in infectious complications in patients who did not receive antibiotics. Related article page 562 Opinion