In April, the Centers for Disease Control and Prevention announced that prescriptions for opioid pain relievers, like Percocet, Vicodin and OxyContin, declined by 10 percent last year alone, part of a downward trend that has been happening since 2011. This good news, however, does not necessarily indicate an end to the years-long increase in opioid-related deaths.

How to explain the paradox? Isn’t overprescribing behind the opioid epidemic?

[Explore America’s in-depth coverage on Mental Health.]

Not today. In the late 1990s and into the 2000s, prescription opioids were indeed the main cause of overdose deaths (at least when those pills were combined with alcohol and Valium-type drugs), but today the drivers are heroin and, even more, fentanyl, a synthetic opioid that is 50 times as potent as heroin. Fentanyl or heroin (which itself is usually tainted to some extent with fentanyl) was present in more than two-thirds of the roughly 42,000 opioid-related deaths in 2016.

Today the drivers of the opioid epidemic are heroin and fentanyl, a synthetic opioid that is 50 times as potent as heroin.

Pain relievers remain widely abused, and doctors and dentists need to prescribe them carefully. But caution will be effective only if we understand more precisely who is at risk for abusing these drugs and how they get access to them.

First, it is important to address the matter of addiction. Contrary to common belief, prescription opioids do not pose a significant risk of addiction to the average person who has had a tooth extraction or a surgical procedure.

According to a 2016 national survey conducted by the Substance Abuse and Mental Health Services Administration, 87.1 million U.S. adults had used a prescription opioid—whether prescribed directly by a physician or obtained illegally—sometime during the previous year. Only 1.6 million of them, or about 2 percent, developed a “pain reliever use disorder,” which includes behaviors ranging from overuse to overt addiction.

People who are given a few days’ worth of an opioid pain reliever like Percocet or Vicodin rarely develop an addiction. Yes, a subset is at risk; it includes people with histories of addiction and those struggling with psychiatric conditions, and caution is particularly important with young adults. It is wise to discuss if opioids are really needed (as they often are not), to prescribe as few pills as possible and to enlist a family member to manage them for a young adult.

People who are given a few days’ worth of an opioid pain reliever like Percocet or Vicodin rarely develop an addiction.

Addiction in chronic pain patients is a somewhat different story. Estimates for new addictions resulting from opioids prescribed to manage chronic pain cover a wide range, but the C.D.C. guideline on prescribing opioids cites a range of 0.7 percent at a low dose to 6.1 percent at higher doses from one study, though lower and higher estimates can be found.

Still, just because opioids do not produce high rates of addiction in patients does not mean doctors are not overprescribing. Few patients need (or use) more than a week of pain relievers for a temporary problem. The problem is that excess pills remain in medicine chests around the country and fall into the wrong hands. According to the Substance Abuse and Mental Health Services Administration, among people who misused prescription pain relievers in 2013 and 2014, about half said that they obtained them from a friend or relative, while only 22 percent said they received the drugs from their doctor.

While physicians must be cautious about prescribing, they must also be sensitive to the needs of patients with long-term and intense pain. Take the case of Anne Fuqua, a 37-year-old nurse from Birmingham, Ala. Previously, Ms. Fuqua worked in hospice care, but she has not been able to take care of patients for the past 14 years. In her late teens, she was diagnosed with primary generalized dystonia, a neurological disorder marked by involuntary movements and painful muscle spasms. She was unable to tolerate the usual medications prescribed for dystonia, but Ms. Fuqua and her doctor found that opioids dramatically reduced the rigidity.

Ms. Fuqua is among millions suffering from afflictions like rheumatoid arthritis, neurological illness and inflammatory conditions who receive opioids, sometimes at high doses. Without these drugs, many of them cannot get out of bed or hold a job.

While a strong case can be made for carefully reducing doses in patients with their consent, no data support nonconsensual or forcible dose reductions or curtailment in otherwise stable patients. There is also anecdotal evidence of harm (emotional trauma, medical or psychiatric deterioration and even suicide) from forced reductions or outright cessation of painkillers.

A compassionate approach to the many casualties of the opioid crisis is possible. Doctors need to be careful prescribers, and health systems and payers need to make more treatment available for those struggling with addiction. At the same time, we need to allow people crippled by pain to retain access to the medicine that enables them to live as full a life as possible.

This article appears in May 28 2018.

Sally Satel is a psychiatrist who practices in a methadone clinic. She is a resident scholar at the American Enterprise Institute.