Luck may play a role in determining who ends up taking prescription opioids long-term, according to a study released Wednesday that found a link between the prescribing habits of hospital emergency room doctors and extended use of narcotic painkillers.

People seen by an emergency room physician who tends to prescribe opioids frequently are about 30 percent more likely to take them for at least six more months over the ensuing year, according to the study by researchers at the Harvard T.H. Chan School of Public Health.

The study also found a remarkable difference in opioid prescribing habits by doctors in the very same emergency rooms: “High-intensity prescribers” doled out narcotics during 24.1 percent of patient visits, on average, while “low-intensity prescribers” called for them only 7.3 percent of the time. In addition, patients who received a large dose of opioids at their initial visit were more likely to end up as long-term users.

“We just don’t have a lot of evidence on when and how to prescribe opioids effectively in a large number of clinical situations,” said Michael J. Barnett, an assistant professor in the public health school and lead author of the study. “And when we lack evidence . . . doctors tend to do what we feel like.”

Overprescribing by physicians has been widely blamed for helping to start an epidemic of prescription opioid abuse, which since 2000 has killed about 180,000 people through overdoses. But it has been difficult to apportion doctors’ responsibility for the crisis or for policymakers to agree on measures to rein in their habits.

The new study, published in the New England Journal of Medicine, looked at nearly 380,000 Medicare patients who visited thousands of emergency rooms across the United States between 2008 and 2011, complaining of pain. As a retrospective analysis, the research was not a controlled experiment and does not prove cause and effect. But the large sample size and the clear results allowed the researchers to confidently conclude that the initial prescriber’s decision is associated with the likelihood of long-term use.

With opioid prescribing, “there isn’t a really unified principal around the way we do things. That’s in contrast to, say, heart disease,” Barnett noted.

Long-term use, in this context, does not mean addiction or physical dependence on the drug. Some people use opioids for years without becoming addicted or dependent. But there is no solid evidence that long-term use of narcotics is effective for chronic pain, and the Centers for Disease Control and Prevention has recommended that doctors sharply limit opioid prescribing to the lowest possible dose for the shortest possible time.

In a December poll conducted by The Washington Post and the Kaiser Family Foundation, one-third of Americans who had taken prescription opioids for at least two months said they became addicted to, or physically dependent on, the powerful painkillers.

One question raised by the study is why other doctors offering follow-up care after an ER visit would continue to prescribe opioids for so long. Colleen Barry, chair of the Department of Health Policy and Management at the Johns Hopkins Bloomberg School of Public Health, said physicians are often reluctant to change treatment regimens when patients are happy with what they have.

“We see this kind of natural human behavior play out in lots of settings in health care,” said Barry, who was not involved in the study. “A physician in an outpatient setting is . . . trying to see a lot of patients in the course of a day. An easy option will be to refill a script.”

Barnett described this phenomenon as “clinical inertia.” He added, “In the clinical reality, it can be hard to say no to a patient who you think is really suffering.”