The Journal of the American Medical Association reports on research aimed at finding methods to deter inappropriate antibiotic prescribing by primary care doctors. (JAMA Article) Excessive use of antibiotics can lead to adverse events, higher spending and a greater prevalence of antibiotic-resistant bacteria. Respiratory diseases have high rates of antibiotic misuse so three interventions were tested among 248 primary care doctors in Boston and Los Angeles. The primary outcome was antibiotic prescribing rate for upper respiratory tract infection, flu, and acute bronchitis. The interventions were modeled on behavioral science principals, designed to “nudge” physicians in the right direction.
The first intervention was “suggested alternatives”, which upon certain diagnoses being entered popped-up a screen in the EHR saying antibiotics were not recommended for the diagnosis and giving alternative treatments. The second intervention was referred to as “accountable justification” and also was EHR-based, asking the physician to write a free-text justification for antibiotic use and saying it would be available in the record and if no justification was given, the EHR would note “no justification”. The third intervention was an email-based peer comparison, which told the top 10% of appropriate prescribers that they were top performers, and told the rest they weren’t and were given data on the number and type of their inappropriate prescriptions and shown comparisons with the high-performers.
The doctors in the study received 0, 1, 2 or all 3 of the interventions. Over 18 months antibiotic prescribing rates declined from 24.1% to 13.1% for control practices, from 22.1% to 6.1% for the suggested alternative intervention, from 23.2% to 5.2% for the accountable justification intervention and from 19.9% to 3.7% for the peer comparison one. The interventions did not appear to have any interaction with each other when multiple ones were used. There was a slight but generally insignificant increase in return visits in the intervention groups. It is interesting to note that even among the control group there was a substantial decline in prescribing, potentially due to either broader awareness generally of the issue or since some of the physicians were in the same practice, there may have more intra-practice discussion of the issue. But overall, the results suggest that fairly simple and inexpensive interventions may be able to improve not only prescribing behavior but other treatment behaviors.