Intended for healthcare professionals

Feature Opioid Prescription

Containing the opioid overdose epidemic

BMJ 2012; 345 doi: (Published 14 December 2012) Cite this as: BMJ 2012;345:e8340
  1. Michael McCarthy, editor
  1., Seattle, Washington, USA
  1. mxmc{at}

Michael McCarthy looks at moves in Washington State to stem the rise in prescription opioid misuse

In the late 1990s, Washington State began to relax its rules regulating the prescription of opioids. Shortly thereafter, overdose deaths began to climb.

“We saw the deaths increase within a year,” says Gary Franklin, medical director for the Washington Department of Labor and Industries, which administers compensation for job related injuries and illnesses for more than 3.2 million workers in the state.

“These were productive people who were working the day they came into the system with a back sprain or whatever, and three years later they were dead from an accidental overdose of opioids,” Franklin says. “I had never seen anything so sad.”

The deaths triggered an investigation which found that as opioid regulations had been relaxed in the state there followed a modest increase in overall opioid prescriptions, but a marked shift towards prescriptions for more potent, long acting opioids, such as transdermal fentanyl, oral methadone, and oral oxycodone, the proportion for these formulations rising from 19.3% in 1996 to 37.2% in 2002.1

During the same time period, the average daily morphine equivalent doses (MED) for long acting opioids had risen 50% to 132 mg/day. “Doses,” says Franklin, “were going through the roof.”


US rates of sale of prescription painkillers, deaths, and admissions for substance misuse (1999-2010)2

The investigation also identified 32 deaths that were “definitely or probably” related to unintentional overdoses among patients covered by Franklin’s agency and 12 more that were possibly related. Most deaths were in men; the average age was 40 years.

Only two of the 32 cases were in people who had what would be considered catastrophic injuries, Franklin says. Most were being treated for conditions like low back pain (20 cases) and carpal tunnel syndrome (three).

The Washington report was the first to document a trend that the US Centers for Disease Control and Prevention now calls a national “epidemic” of prescription drug overdoses, which were involved in 14 800 deaths in the US in 2008, more than cocaine and heroin combined.2


Drug overdose death per 100 000 people by state, 20082

While the epidemic continues to accelerate across the country, Washington has been one of the first to reform its opioid prescription regulations to try to stem the tide.

New guidelines

Washington, like many states in the US, began relaxing rules on opioid prescribing in the late 1990s in response to heavy lobbying by drug companies and advocacy groups, many of them industry funded, which argued that opioids were a safe, effective, and underused treatment for chronic, non-cancer pain. It lifted most restrictions on the use of opioids for chronic pain, stating that “No disciplinary action will be taken against a practitioner based solely on the quantity and/or frequency of opioids prescribed.”3

That language told doctors, “It doesn’t matter if you hand out bags of opioids to people, you will not be not be hammered by our disciplinary board,” says Franklin.

Alarmed by the overdose deaths detected by the 2005 Labor and Industries’ study, the state’s Agency Medical Directors’ Group, a collaboration of all state agencies that provide or oversee healthcare coverage, including the state employee plan, the Department of Health and Medicaid, held a series of meetings to revise their guidelines.

The new guidelines, issued in 2007, emphasised the risks of opioid treatment, advised their more cautious use, and designated an opioid dose threshold of 120 mg/day MED as a “yellow flag” warning. At that dose, the guidelines advised, if the patient’s pain and functioning had not substantially improved, the physician should consult with a pain specialist before increasing the dose.4

“We weren’t saying don’t use opioids,” says Franklin, “but if you choose to use them for chronic pain and the dose is creeping up and you get to 120 MED and the patient’s not better in terms of pain control and function, do not keep moving the dose up—ask for some help.”


Physicians at Group Health, a Seattle based, consumer run, non-profit health plan with more than 600 000 members in the Pacific Northwest, had also noticed an uptick in opioid prescribing among its providers.

What’s more there were wide variations in care. Use and dosing of opioids varied from clinic to clinic and from provider to provider, with some prescribing high doses and others refusing to prescribe opioids at all.

Both approaches are “lazy medicine,” says Claire Trescott, a family physician and medical director of primary care at Group Health. Doctors who are “overly easy” or “overly strict” with opioids are often trying to avoid having to address the patient’s problem, “whether it’s addiction or pain,” says Trescott.

“I wanted our clinicians to understand what is good medicine in treating this population of patients and how we should do it,” says Trescott.

At a recent conference in Seattle, Trescott and her coworkers described how they changed opioid prescribing practices throughout their system using “rapid process improvement workshops”—the “lean” quality improvement technique developed by the Japanese carmaker Toyota and adopted by several US medical systems to improve healthcare delivery.5

Drawing on the new state guidance as well as its own in-house expertise, Group Health spent a year creating its own guidelines. The implementation was then handed over to the rapid process improvement workshop team, made up of frontline workers, including physicians, nurses, and pharmacists. The team drew up an efficient, “mistake proofed” process to put guidelines into practice.

The elements of the protocol devised by the workshop group include the requirement that only one physician manage a patient’s opioid prescriptions, that the physician and the patient develop an individualised care plan using a standard template, that refills are scheduled every 28 days so the refill date never falls on a weekend, and that patients at high risk of abuse have urine drug testing to detect the abuse of other drugs.

The response, says Trescott, has been overwhelmingly positive. “This is the easiest thing that we have rolled out at Group Health. People were hungry for it; they pulled it to the clinics. I didn’t have to push it or talk anyone into it. People wanted it.”

The response from patients has been positive as well, says Trescott, largely because, with a clearly laid out standardized refill schedule, patients’ anxiety about whether they will get their medications has been greatly reduced.

Group Health is still collecting data on the impact of the new protocol, says Trescott, but “we do know our doctors prescribe much lower doses than they did. And they do more urine drug screening.”

Early data on the state’s reforms have also been promising, says Franklin. A recent study of opioid dosing for patients covered by the Labor and Industries program found that since the state implemented its guidelines in 2007 the percentage of workers receiving long acting opioids had fallen by 27% and the proportion that were prescribed doses of 120 mg/day MED or more had fallen by 35%. At the same time, deaths linked to prescription opioids, which had been rising through 2009, dropped 50% in 2010.6

This drop in deaths is promising, says Franklin, lead author of the report, but he acknowledges the 2010 number represents only one data point. An analysis of more recent data is under way.

Meanwhile, the national epidemic, which is has now taken more American lives than the Vietnam war, continues to take an estimated 100 lives a day in the US, Franklin says. “There is no reason for people not to move quickly on this. It’s a national emergency. Our hair should be on fire.”


Cite this as: BMJ 2012;345:e8340


  • Competing interests: The author has completed the ICMJE unified disclosure form at (available on request from the corresponding author) and declares no support from any organisation for the submitted work; no financial relationships with any organisation that might have an interest in the submitted work in the previous three years; and no other relationships or activities that could appear to have influenced the submitted work.

  • Provenance and peer review: Commissioned; not externally peer reviewed.


View Abstract