Treating patients versus numbers

As a group of Canadian physicians interested in the management of patients with chronic pain and addiction, we feel compelled to respond to the recent paper by Dhalla et al regarding opioid-related deaths in Ontario.¹

Similar to a previous paper by some of the same authors,² Dhalla and colleagues once again confuse association with causation. They overinterpret statistics from administrative databases to make pronouncements on clinical pain management practice—an area in which none of the authors profess to have any expertise.

They have failed to discuss relevant confounders. It is like saying that cardiac surgeons in leading institutions have worse results than those in provincial hospitals, without taking into account the severity of the conditions treated. They have failed to consider alternative explanations for their results or discuss other important limitations of their study.

In 2006, there were approximately 12 million people in the province of Ontario. This would mean approximately 2.3 million people with moderate to severe chronic pain.³ If we assumed that somewhere from 30% to 50% might be taking regular opioid therapy (likely higher than this), 406 deaths would result in a crude death rate of 50 to 60 per 100 000 people with pain who were taking opioids (or about 3 per 100 000 total population). With such small numbers, any flaws in the methodology of this study that change the numbers would have a very big effect on the reported percentages.

The reported suicide rate in the Canadian population is about 15 per 100 000 people.⁴ In patients with persistent pain that number is at least doubled, to about 30 per 100 000 people.⁵ Higher doses of opioids often are consistent with a longer time in treatment, poorer efficacy of other treatments, and more opportunity for patients to realize that their pain will not go away. There is little organized support and a lack of other, nonpharmacologic treatment options compared with other chronic diseases. All of us can recount hearing patients with chronic severe pain say they feel like they have nothing to live for. If we subtract the number of people who might have committed suicide from Dhalla’s numbers, then the number of deaths “caused” by opioids shrinks substantially.

The authors used data from coroners to assign deaths “related to opioids.” With all due respect to our hardworking coroners, assigning a cause of death in the case of a patient taking therapeutic opioids can be an extremely difficult challenge.^6–8 There can be a very large overlap between the blood levels of someone stable on long-term opioid therapy and those of someone found dead with opioids in their blood, and there is a poor correlation between opioid blood levels and death.⁹ The definition of opioid-related death among coroners can be variable and can have a large influence (up to a 2-fold difference) on reported death rates.¹⁰

How did the authors account for the effects of other substances also found in the blood of decedents? Which substance actually caused the patients’ deaths? Were the deaths most likely due to substance abuse or addiction (more than 90% in a study by Hall et al¹¹) or was it therapeutic misadventure? The authors stated that they adjudicated questionable cases among themselves to come to a decision on cause of death, yet they did not report any expertise to allow them to do so.

The authors have suggested that there is an association between the number of deaths and the number of prescriptions written, particularly in the antemortem period. No information is provided regarding the drug or quantity prescribed. Therefore, the authors provide no evidence that the deaths among the patients of high prescribers are due to the drug they prescribed or the dose prescribed. What was the length of time between death and last prescription? If a physician writes a prescription for an opioid 12 months before the patient dies, is that doctor somehow responsible?

Most opioids prescribed for palliative care are prescribed by FPs who are often not identified as palliative care physicians, as they have had no extra training that would justify a formal label even though it might be the focus of their work. How was this population of physicians identified or accounted for in the high-prescribers group? What about those who do both palliative care and chronic pain management?

The authors have provided no denominators. Although more deaths occurred among the patients of the high prescribers, the death rate per prescription written, drug used, or dose taken is not calculated. If high prescribers were “responsible” for 63% of the deaths, but wrote more than 63% of the prescriptions, then it might be that they are actually safer prescribers compared with those who prescribe less frequently.

Dhalla et al conclude from their data that a small number of FPs are irresponsibly prescribing high-dose opioids, causing deaths among patients. They suggest targeting these doctors for further education or regulatory attention. What we know from our communities is that a large proportion of FPs are already reluctant to take on the care of patients with complex pain and will not prescribe any opioids even when appropriate. This leaves a small number of compassionate family doctors who have a special interest in pain to take on a disproportionate number of patients, which likely accounts for the small number of high-dose opioid prescribers.

The high prescribers were noted to be older and more experienced. Does this mean that, on average, they also have an older practice with a higher prevalence of chronic pain? The authors have provided no evidence to support any assertion that older, experienced practitioners are less likely to follow safe prescribing guidelines and hence are more in need of, or amenable to, “academic detailing” or regulatory scrutiny. On the contrary, those who write more opioid prescriptions might be more likely to seek further training, particularly given the regulatory environment for opioids that already exists in Canada. Family physicians are expected to manage complex patients experiencing pain, despite very little formal education and very few funded nonpharmacologic treatment resources available. Subjects used in this study were drawn from the Ontario Drug Benefit Program database. This group comprises a vulnerable population that has less access to the determinants of health and that is more at risk than the general population of Ontario. This group might also have even less access to appropriate nonpharmacologic care than most Ontarians. Dhalla et al offer a blanket criticism of, and a call for a reduction in, the use of opioids for chronic pain, yet do not suggest alternative solutions. Although the published scientific evidence for the use of opioids to treat chronic pain is still evolving, the balance of current evidence suggests that opioids can be an effective treatment in some people, with a low overall risk of adverse effects, including addiction.¹² All physicians recognize that no treatment is risk-free and the potential benefits must always be balanced with the potential harms. Even those clinicians with expertise in pain management know that there are no risk-free treatment options for severe pain. Acetaminophen, available for decades over the counter, has risks of organ toxicity when used chronically.¹³ The recent Scottish Health Survey found an increased risk of cardiovascular events and death in people taking tricyclics.¹⁴ The chronic use of nonsteroidal anti-inflammatory drugs is associated with an increased risk of stroke and cardiovascular deaths as well as the known risks of death due to upper gastrointestinal bleeds and perforations.^15,16 The best type of care for chronic pain is an interdisciplinary biopsychosocial-spiritual approach. Opioids can be an important pharmacologic component of such multimodal care. Unfortunately, prescribing medication is often the only type of treatment funded by our health care system. When prescribed carefully and monitored appropriately, opioid therapy can result in reduced pain and suffering and improved quality of life. The key is to educate all physicians on appropriate assessment, efficacious treatment, and careful prescribing rather than targeting a small number for “overprescribing” opioids. Studies, such as that of Dhalla et al, that report on numbers without considering the clinical context do nothing to advance solutions for the epidemic of poorly treated chronic pain in Canada. They only make family doctors even more reluctant to treat patients with pain. We would have appreciated a more thorough peer review of this paper before publication. We are doing our utmost to promote more dialogue and understanding to optimize patient care, and we hope for the same from all our colleagues.

• Copyright© the College of Family Physicians of Canada