Saturday, November 5, 2011

Are Opioid Pain Reliever Deaths an Epidemic?

CDC Logo The U.S. Centers for Disease Control and Prevention (CDC) recently launched a new campaign to alert the public about deadly risks of opioid analgesics. While the government’s systematic collection and skillful presentation of data are impressive and overwhelming, this may be tilting away from a balanced perspective in order to incite fears of a so-called “epidemic” of opioid misuse, overdoses, and deaths.

There seems no doubt that America has serious substance misuse problems. And, by all accounts, coming from groups like the World Health Organization, such problems are endemic globally. Overdoses and deaths associated with strong analgesics are increasing, but whether these trends represent an epidemic and the proposed solutions have valid evidence for their effectiveness remain open questions.

Government agencies seem to have adopted a viewpoint of prescription analgesics as threats to public health. And, despite frequent assurances by these agencies that they want to stem the tide of opioid overdoses and deaths while ensuring that patients with pain continue to have access to these vital medications, we have wondered if this might reflect a “myth of beneficence”: that is, it appears to have the best interests of patients in mind while, at the same time, inciting fears and encouraging actions and regulations that are questionable in terms of their actual benefit to patients [discussed in UPDATE here].

Are the Data Portraying Fact or Fiction?

On November 1, 2011, the CDC released several reports to demonstrate that overdose deaths involving opioid analgesics are a present and worsening epidemic. The 3 papers, each drawing upon the same data for different presentations, are as follows…

  1. Paulozzi LJ, Jones CM, Mack KA, Rudd RA. Vital Signs: Overdoses of Prescription Opioid Pain Relievers — United States, 1999-2008. MMWR. 2011(Nov 1);60 [PDF here].

  2. Policy Impact: Prescription Painkiller Overdoses. CDC. 2011(Nov) [PDF here].

  3. Vital Signs: Prescription Painkiller Overdoses in the US. CDC. 2011 (Nov) [PDF here].

While the report titles seem straightforward, use of the term “painkiller” in the titles and throughout papers #2 and #3 for the general public discussing opioid analgesics, immediately projects an adversarial bias that these drugs kill rather than help people. In paper #1, a more technical version for professionals, the less sensational phrase “opioid pain relievers (OPR)” is used.

News media reported widely on the CDC reports — using the “painkiller” label — without questioning their accuracy, veracity, or biases. Here are highlights of some of those data along with our commentary…

  • In 2008, drug overdoses in the U.S. were associated with 36,450 deaths of which 20,044 involved one of more prescription drugs. Opioid pain relievers were implicated in 14,800 overdose deaths, 41% of the total, or, as one of the CDC reports stated it: “Prescription painkiller overdoses killed nearly 15,000 people in the US in 2008. This is more than 3 times the 4,000 persons killed by these drugs in 1999.”

    Apart from the inflammatory CDC rhetoric suggesting that opioids are “killer drugs,” it must be understood that mortality data are gathered from medical examiner reports attributing cause of death and the role that opioids might have played. In many cases, opioids are found present at autopsy, and reported as such, but their causative role might be challenged. Also, it is not uncommon for drug poisoning deaths to be miscoded, and the CDC concedes that some deaths involving both illicit and prescription opioids are not counted correctly — so the accuracy of mortality data and the sole role of opioids should be questioned.

    Furthermore, whether the 3-fold increase in number of deaths during the 10-years is also a significant relative-rate increase can only be determined if the numbers of persons being prescribed, using, or misusing opioids during those respective time periods is known, but this was not reported. In other words, if 3 times more persons were exposed to opioid analgesics in 2008 than in 1999, the relative incidence rate of deaths for each of those years could be identical; still a problem, but not one that is mushrooming out of control.

  • Sales of OPR — to pharmacies, hospitals, clinics — quadrupled between 1999 and 2010. Or, as the CDC described this: “Enough OPR were prescribed last year to medicate every American adult with a standard pain treatment dose of 5 mg of hydrocodone taken every 4 hours for a month.”

    While this may seem like an extraordinarily huge amount of opioid medication, as the CDC report writers no doubt intended it to sound, the latest data from the Institute of Medicine indicate that 116 million adult Americans suffer chronic pain [see UPDATE here]. So, if just a significant fraction of those persons need opioid medication daily, year round (12 months), for pain relief, the CDC numbers actually do not amount to very much OPR.

    From the CDC data, we do not know the numbers of persons with chronic pain in 1999, so there is no way to determine if there is a correspondence between that and sales of OPR at that time. However, there has been increased recognition during the past decade of chronic pain as a serious disease, so a corresponding increase in OPR prescribing (even a 4-fold increase in sales) for better pain management could and should be naturally expected.

  • The CDC reports that the rate of admissions to “substance abuse treatment” programs in 2009 was almost 6 times the rate in 1999.

    Apparently, this figure is supposed to depict the growing burden on society of increased OPR distribution; although, the CDC does not provide any information about whether “treatment” also involves therapy for alcohol and illicit opioid misuse, as is commonly the case. So, perhaps, this is merely presented in their reports for some sort of shock value. At any rate, this could actually be a favorably positive trend if it means the public is becoming more sensitive to substance misuse problems and more willing to seek appropriate treatment for those conditions.

  • Opioid Deaths The CDC reports incorporate a version of the figure at right to visualize the data described in the above 3 points. While the trends may look impressive at first glance, they also are somewhat misleading.

    The rate calculations are population based and have been adjusted to make OPR death rates seem of a magnitude directly comparable to OPR sales and treatment admissions. In actuality, OPR death rates (per 100,000 persons) are 10 times smaller than the other rate measures (based on per 10,000 persons). If all curves were drawn to the same scale, OPR deaths would be mostly off the bottom of the graph. At its peak in 2008, the OPR overdose death incidence rate was only 0.48 per 10,000 persons, and the 14,800 deaths represented roughly 0.005% of the total population.

    The figure also portrays a common fallacy in medical research:
    cum hoc, ergo propter hoc
    , or "with this, therefore because of this." This wrongly reasons that when two or more event rates or trends seem to change in parallel there is a cause-effect relationship between them. For example, to the naïve reader, the figure implies that increased opioid prescribing (sales) through the years has directly resulted in more opioid overdose deaths and more addiction requiring treatment. Although this may seem reasonable at face value, the CDC has not built a scientifically credible argument or statistical basis to support such a cause-effect relationship.

    Perhaps of greatest interest, looking at the trend lines, the OPR deaths trend flattens from 2006 to 2008. The CDC authors do not comment on this, but it could suggest that overdose fatalities actually might have started moderating as of nearly 6 years ago, even as the opioid distribution rate continued to largely increase. The CDC could be falsely saying that there is a burgeoning epidemic today, while their own data on overdose mortality from nearly 4 to 6 years ago suggest it could be otherwise.

  • One of the reports notes that in 2010, 2 million people used OPRs nonmedically for the first time, nearly 5,500 persons each day. Another of the reports states that for every person who dies of an OPR overdose there are 825 others who are nonmedical users of opioid analgesics. “Nonmedical” is defined by the CDC as “use of a prescription pain reliever without a prescription belonging to the respondent or use for the experience or feeling the drug causes.”

    This definition, as commonly used by the government, is vague and imprecise. Most clearly, the recreational use of opioids (eg, to get “high”) would be nonmedical use. However, the category also includes persons with pain who use opioids without a prescription of their own for analgesic purposes, which is a form of “medical misuse” that is more common than imagined these days due to a continuing undertreatment of pain.

    For example, one research study found that 12% of high school students surveyed had misused opioids for so-called “nonmedical” reasons; however, among those teenagers, 45% said that they used the medications solely for a medical purpose — to relieve physical pain [discussed in UPDATE
    here]. This is still a bad situation, but the motivations behind what might best be called “medical misuse” make an important difference in defining, measuring, and solving the problems.

    Another term frequently used in government data is opioid “abuse.” This blurred and overly broad category usually reflects both medical misuse (also including medication noncompliance by persons who take more or less of an opioid than prescribed) and nonmedical use. Thereby, data regarding any deviations from opioid use exactly as prescribed are easily conflated and inflated to make the problems look of greater magnitude than they actually are.

The reports go on to compare and contrast overdose and mortality rates from the various states, and the extreme burdens of this on society: eg, “Abuse of OPR costs health insurers approximately $72.5 billion annually in health-care costs.” Unfortunately, the data are not entirely current and the accuracy and validity of the database might be questioned, as above; so, these numbers might be worthy of further and more rigorous scrutiny by independent, unbiased researchers.

Data in Perspective: What is an Epidemic?

The CDC reports follow after another recent paper, from the White House Office of National Drug Control Policy (ONDCP), titled “Epidemic: Responding to America’s Prescription Drug Abuse Crisis” [PDF here]. So this epidemic agenda is not entirely new. And, while it makes for sensational news headlines that grab public attention, the use of this term and its accompanying panic-provoking rhetoric might be akin to falsely shouting “fire” in a crowded movie theater.

The typical dictionary definition of “epidemic” is a disease or condition “affecting or tending to affect a disproportionately large number of individuals within a population, community, or region at the same time” [Merriam-Webster online]. The dimensions of “disproportionately” and “large” are open-ended and vague; yet, the government’s liberal use of “epidemic” in the context of analgesic overdoses/deaths is seemingly intended to connote newly emerging and unique problems of monumental proportions needing urgent and sweeping action.

However, an important question is: How do the roughly 15,000 OPR-related deaths, noted above, compare with other causes of death in the U.S.?

The most recent government mortality data, from National Vital Statistics Reports [PDF here], lists the 15 leading causes of death in the U.S. for 2008-2009. These range from Diseases of the Heart (598,607) at the top to Assault/Homicide (16,591) at the bottom. Drug-induced deaths, in general, are not among the top causes, but the report observes that their rate decreased by 1.6% from 2008 to 2009, which might support the moderating trend noted above.

Curiously, none of those 15 leading causes is being touted by the government as an epidemic; not even Assault/Homicide, which, like OPR-related deaths, might be construed as a large, disproportionate, and preventable problem. In a separate analysis, the Vital Statistics Reports note that in 2008 there were 24,263 Alcohol-Induced deaths; another large and preventable problem. While it is unclear why this was not among the top 15, the rate of alcohol-related deaths is 62% greater than for OPR-related mortality — yet, it is not considered an epidemic.

Of particular interest and concern, Intentional Self-Harm/Suicide is ranked number 10 on the list — 36,547 deaths in 2009 — and the only cause of death in the top 15 to have increased from 2008. Suicide should be of great concern because, as we have previously noted [here] and [here], mortality risks of poorly managed chronic pain are high and suicide attempts and completed suicides among persons with unrelieved pain may be drastically increasing. Yet, the government has not declared suicide — at more than twice the number of OPR-related deaths — as an epidemic.

Some might argue that rates of homicides, suicides, and alcohol-induced deaths have not risen as dramatically during the past decade as OPR-releated deaths. However, even if that were so, it would imply that current incidences of those other tragedies, which are significantly higher than for OPR deaths, are acceptable and not worthy of being labeled epidemics.

Perhaps, there may be a government directive for determining what is or is not an epidemic that is driven by something other than an unbiased scientific viewpoint. Solving problems of substance misuse, overdoses, and deaths certainly deserve a high priority. However, the public has a right to expect better from tax-supported agencies than clever manipulations of data and flawed arguments to advance hidden agendas. In the long run — as greater controls and restrictions are placed on the prescribing of opioid analgesics — this may prove detrimental to a very large segment of the population: the 116 million persons with chronic pain.

ADDENDUM 11/8/2011 — It has just come to our attention that, last September, the CDC issued a request for a proposal [here] that appears to acknowledge current inaccuracies of medical examiner reports regarding the role opioids may play as a cause of death (as we suggested above). The document states, “The net result of these problems has been a wide variability in case definitions and practices used in autopsy and toxicologic evaluations for deaths involving heroin or opioid analgesics. Such variability has made it more difficult for public health agencies to make accurate comparisons of the rates of overdose in different jurisdictions.” The requested project is to develop a position paper and guidelines on determining deaths from opioid intoxication.

The document further states, “Such a process could lead to a uniform case definition and uniform standards for autopsy and toxicologic evaluations resulting in an opportunity for consistent death certification and better public health surveillance.” We believe that, until such standards are developed and implemented, mortality data on opioid-related deaths coming from government agencies should be cautiously considered as having any validity.