This is a randomized controlled trial that uses an audit and feedback intervention to alert primary care physicians who are outliers in one or more metrics related to high risk prescribing of participants' outlier status. Primary care physicians will be randomized to the intervention or control arm, except in California, where all outliers will be notified. The investigators will evaluate the impact of the intervention on prescribing patterns.

The investigators have developed five different metrics related to low-value (and/or high risk) prescribing habits based on guidelines. The five metrics are listed below. The investigators are using 100% capture Medicare claims data to evaluate the prescribing habits of all US primary care physicians who prescribed to ten or more Medicare patients. A patient is attributed to the primary care physician who prescribed the patient the most medications that year.

The investigators have calculated the mean, median, and standard deviation for each of the five metrics based on data from 2016. All primary care physicians who are two or more standard deviation above the mean of a given metric are considered outliers. In each state (excluding California), half of the outliers while be randomly assigned to the intervention group and half will be assigned to the control group, using a random number generator. Those outliers who are assigned to the intervention group will receive a cover letter signed by a members from the Physician Engagement Council (PEC), which is composed of physicians from the Society of General Internal Medicine (SGIM) which explains the study. Participants will also receive a report of participants' status as an outlier which shows in both text and graphic representation, how participants compare to participants' peers. These communications will be sent by mail. The control group will not receive any communication. All outliers in California will receive the intervention.

Brief summary of metrics:

• Metric 1: Average number of concurrent medications per older patient, by physician
• Metric 2: Proportion of older patients who received at least one medication that is a designated high-risk medication in older patients, by physician
• Metric 3: Proportion of older patients with concurrent use of multiple central nervous system-active medications, by physician
• Metric 4: Proportion of older patients with chronic use of benzodiazepine sedative hypnotic medications, by physician
• Metric 5: Proportion of older patients with concurrent use of opioids and benzodiazepines, by physician

Inclusion Criteria:

• Primary care physician
• Must be the primary care physician for at least ten Medicare patients

Exclusion Criteria:

• None