Erin Krebs started medical school in 1996, just months after OxyContin was approved for sale in the U.S. Over the next seven years, as she earned her M.D. and trained in internal medicine, she watched in astonishment as “oxy” and other potent opioids became the reflexive prescription for all manner of pain while worries about addiction and prolonged use were brushed aside. “As a natural skeptic, I went looking for a good reason why we changed our practice,” she recalls, “and it wasn't there.”

It was then that Krebs conceived of a “dream study” that, amazingly, had never been done: a long-term randomized controlled trial comparing opioids with nonopioids for treating serious chronic pain. It took a while, but with funding from the U.S. Department of Veterans Affairs, Krebs, now at the Minneapolis VA Health Care System, began such a study in 2012, after enrolling 240 veterans who suffered from persistent moderate-to-severe back pain or a similar level of arthritic pain in their hips or knees.

Patients were randomly assigned to either an opioid group or a nonopioid group, both starting with low-intensity drugs but able to move to stronger stuff as needed. The results, published in March, were eye-opening. Patients given alternative drugs did just as well as those taking opioids in terms of how much pain interfered with their everyday life. In fact, they reported slightly less pain and had fewer side effects.

Why hadn't such a study been done before? “At some level, physicians, as well as the general public, were willing to believe we don't need studies to show us that these drugs work,” Krebs suggests. Powerful drug-marketing efforts had somehow swamped science. Dentists were also caught up in the rush to opioids. Apart from Tylenol with codeine, they had been reluctant to offer such drugs before the 1990s, says Harold Tu, director of oral and maxillofacial surgery at the University of Minnesota School of Dentistry. Now about 95 percent of dentists and oral surgeons prescribe hydrocodone or oxycodone for patients undergoing painful procedures, such as wisdom tooth extraction.

Like Krebs, Tu was not convinced that opioids were a superior choice. Dental research said otherwise. He was also horrified by the possibility that his profession had helped open the gateway to today's opioid addiction crisis, which caused more than 42,000 overdose deaths in the U.S. in 2016. “The evidence shows that dentists—and in particular, oral surgeons—are one of the largest prescribers to people between ages 10 and 19,” he notes. Research also shows that high school students who are prescribed opioids have a 33 percent increased risk of later misusing the drugs.

In early 2016, under Tu's leadership, the Minnesota dental school introduced a mandatory protocol stipulating that the first-line treatment for pain would be nonsteroidal anti-inflammatory drugs such as high-dose ibuprofen. Opioids were permissible after a difficult surgery, but providers had to use the lowest adequate dose and register the prescription in a digital tracking system.

The result: in 15 months the school's 30 practitioners cut opioid prescriptions nearly in half, according to a report to be published later this year. Nevertheless, they saw no increase in after-hours calls or return visits related to pain. Since the study was completed, opioid use has further plummeted, Tu says: in 2015 95 percent of painkillers prescribed after a procedure were opioids; in 2017 it was just 21 percent. The average number of opioid pills per patient also dropped—an important change because unused pills often go astray. Surveys show that dentists typically prescribe 16 to 24 opioid pills, and yet patients use about eight.

Changing practice isn't easy, even in the face of a crisis. The amount of opioids prescribed in the U.S. fell by 18 percent between 2010 and 2015—not nearly enough—and the number of pills per script actually rose! Krebs is a firm supporter of using opioids for acute surgical pain or for easing the agony of dying, but there has never been good evidence for deploying them against chronic pain. And yet colleagues told her it would be “unethical” to withhold the drugs from patients in her study. Patients may also balk. Some told Tu, “The only thing that works is Vicodin.”

I know how that is. Last year I was stunned when my periodontist presented me with a script for 800-milligram tablets of ibuprofen following surgery in which she sliced tissue from my palate and stitched it to my gums. “Is this all I need?” I mumbled incredulously through the novocaine. “You'll be fine,” she told me. And indeed, I was.