scorecardresearch Skip to main content

Researchers say Mass. effort to influence opioid prescribing failed


It seemed like a good idea.

Last March, the state Department of Public Health sent a confidential letter to every health care provider who prescribes opioids and other controlled substances, showing how each practitioner’s prescribing practices compared with those of his or her peers.

Presumably, those who were writing too many prescriptions would see the error of their ways and scale back.

But an article in Thursday’s New England Journal of Medicine concludes that the letter probably didn’t work. In the 12 weeks after it was sent, an analysis of prescribing by 284 primary care physicians found no change from the previous 12 weeks.


This points to a costly failure to evaluate government responses to health issues, said the lead author, Dr. Michael Barnett, who researches health care delivery and quality at the Harvard T.H. Chan School of Public Health. The letters were required by the sweeping state law intended to control opioid use, signed amid fanfare by Governor Charlie Baker in 2016.

“As a researcher, I am continually amazed at the pace of policy-making that just kind of races ahead of any kind of evaluation or monitoring, even when it’s not hard to do so,” Barnett said. The analysis was included in an opinion piece calling for collaboration between researchers and public officials.

But Dr. Monica Bharel, the state public health commissioner, said Barnett and colleagues were taking an overly narrow view of the letter’s purpose and results. The letter was merely intended to give prescribers objective information that they can incorporate into their clinical judgment, she said.

“This letter was a small component of our overall plan of how to work with prescribers to address this current opioid epidemic,” Bharel said. The state’s programs cannot be judged in isolation, she said. An array of efforts — including a requirement that prescribers consult a database that can alert them to doctor-shopping by patients — has led to a 29 percent decline in opioid prescribing since 2015, she said.


Bharel said she wouldn’t necessarily expect to see a drop-off in prescribing during the 12-week period that the researchers examined. She also questioned whether any conclusions can be drawn from 284 doctors, when the letters went out to 29,000 physicians, advanced practice nurses, physician assistants, and dentists.

Prescribing practices have been a focus in efforts to fight the addiction epidemic that claimed more than 2,000 lives in Massachusetts last year.

Barnett, who works part-time as an internist, was inspired to do the analysis after he received one of the health department’s letters.

The data showed that his volume of prescriptions was slightly below the median for other internists in the state. But the letter didn’t tell him why the state was providing the information or what health officials wanted him to do with it. It offered no advice on how much he should prescribe. He didn’t know what to make of it.

Talking to other doctors, including one whose prescribing was well above the median, Barnett found the reaction to the letter was “a collective shrug.”

A better approach, he said, would be to offer positive examples to emulate, rather than comparing doctors with an average or median.

“You really want to leverage the physician’s internal motivation to be the A student,” he said. But in the case of opioid prescribing, the optimal rate is unclear.


To conduct the analysis, Barnett teamed up with athenahealth, a Watertown-based company that provides electronic health records to many physicians. As a result, athenahealth can see what happens at each doctor visit, and the company is eager to aggregate this information to learn about trends in medical care.

Barnett and researchers from athenahealth decided to look at primary care doctors because, as a group, they prescribe the biggest volume of opioids, and they also represent a large proportion of athenahealth customers.

The athenahealth researchers compiled the number of doctor visits that resulted in opioid prescriptions during the 12 weeks before and after the letters were sent. The data included 284 doctors in Massachusetts and 864 in eight other states in the Northeast. They found no changes over time, and no differences between the two groups. They also found no reductions in opioid prescribing by the highest-volume prescribers.

Dr. Andrew Kolodny, codirector of Opioid Policy Research at Brandeis University’s Heller School for Social Policy and Management, was not involved in the research but agreed with the article’s message. “These urgent public health efforts need to be evaluated. They’re basically shooting in the dark,” Kolodny said. “I don’t blame anyone in the midst of an emergency for trying things . . . but you should simultaneously be evaluating what you’re doing.”

Felice J. Freyer can be reached at