Customize your JAMA Network experience by selecting one or more topics from the list below.
3 tables omitted
During 1999-2006, the number of poisoning deaths in the United States nearly doubled, from approximately 20 000 to 37 000, largely because of overdose deaths involving prescription opioid painkillers.1 This increase coincided with a nearly 4-fold increase in the use of prescription opioids nationally.2 In Washington, in 2006, the rate of poisoning involving opioid painkillers was significantly higher than the national rate.1 To better characterize the prescription opioids associated with these deaths and to reexamine previously published results indicating higher drug overdose rates in lower-income populations,3 health and human services agencies in Washington analyzed overdose deaths involving prescription opioids during 2004-2007. This report describes the results of that analysis, which found that 1668 persons died from prescription opioid-related overdoses during the period (6.4 deaths per 100 000 per year); 58.9% of decedents were male, the highest percentage of deaths (34.4%) was among persons aged 45-54 years, and 45.4% of deaths were among persons enrolled in Medicaid. The age-adjusted rate of death was 30.8 per 100 000 in the Medicaid-enrolled population, compared with 4.0 per 100 000 in the non-Medicaid population, an age-adjusted relative risk of 5.7. Methadone, oxycodone, and hydrocodone were involved in 64.0%, 22.9%, and 13.9% of deaths, respectively. These findings highlight the prominence of methadone in prescription opioid–related overdose deaths and indicate that the Medicaid population is at high risk. Efforts to minimize this risk should focus on assessing the patterns of opioid prescribing to Medicaid enrollees and intervening with Medicaid enrollees who appear to be misusing these drugs.
For this analysis, the Washington State Department of Health defined an overdose death involving prescription opioids as a death in Washington during 2004-2007 of a state resident whose death certificate had (1) a manner of death of “accidental” or “natural”; (2) one or more contributing causes coded to “poisoning by narcotics” or a “mental and behavioral disorder due to use of opioids” (based on International Classification of Diseases, 10th Revision codes T40.0–T40.6 and F11*); (3) specific words compatible with an acute drug intoxication recorded in any of the cause of death fields (eg, “overdose”); and (4) a prescription opioid term in any of the cause of death fields. Examples of prescription opioid terms sought on manual review of the certificates were “oxycodone,” “methadone,” and “hydrocodone.” Although morphine is a prescription opioid painkiller, it is also a metabolite of heroin. Therefore, mention of morphine on a death certificate was only accepted as evidence that a death was prescription opioid–related when the certificate specified that the morphine was a prescription drug. As a result, 82 deaths involving morphine and no other opioids (36.6% of all deaths in which morphine was mentioned) were excluded from this analysis.
The Washington State Health and Recovery Services Administration (WSHRSA), which operates Medicaid and several associated medical-assistance programs, determined which deaths occurred among persons who were enrolled in Medicaid at some time during the year of their death. During 2004-2007, the Medicaid-enrolled population (5 109 363 person-years) represented 20.2% of the Washington population (25 287 800 person-years). WSHRSA also linked the deaths from prescription opioids to records of clients in the Medicaid Patient Review and Coordination (PRC) program, a special state program for clients who overuse or inappropriately use medical services.† PRC program members (5858 person-years) represented 0.1% of the Medicaid population during 2004-2007. Rates were age adjusted because the Medicaid population was younger than the non-Medicaid population.
During 2004-2007, a total of 2282 deaths in Washington met the manner and cause of death case definition criteria. Of these, 2194 (96.1%) had a death certificate that included a term indicating acute drug intoxication. Of these 2194, a total of 1668 (76.0%) had a death certificate that included a prescription opioid term and were included in this analysis. The age-adjusted prescription opioid overdose rate was 6.4 per 100 000 per year. The male mortality rate was 1.4 times the female rate. Rates increased with age to a peak of 15.0 per 100 000 in the 45-54 years age group and then declined.
Among all decedents, 758 (45.4%) were enrolled in Medicaid at some point during the year of their death. Medicaid-enrolled decedents had an age distribution comparable with that of decedents statewide. However, the percentage of females was greater among Medicaid-enrolled decedents (52.2%) than among decedents statewide (41.1%). A total of 34 Medicaid-enrolled decedents were in the PRC program, representing 4.5% of all Medicaid-enrolled decedents.
The risk for prescription opioid overdose death varied substantially by Medicaid status. The crude annual risk for prescription opioid overdose death was approximately one in 6757 in the Medicaid-enrolled population and one in 172 in the Medicaid-enrolled PRC program population.
Medical examiners and coroners recorded methadone on death certificates nearly three times more often than the next most common opioid, oxycodone. At least one nonopioid prescription drug was reported in 54.6% of the deaths. A benzodiazepine was listed on the death certificate in 20.9% of the deaths, and an antidepressant in 31.7%. An illegal drug was reported in 21.8% of the deaths. Cocaine was involved in 15.7%, methamphetamine in 5.5%, heroin in 2.4%, and alcohol in 17.1% of the deaths. More than one drug was listed for 72.3% of decedents. The mean and median numbers of drugs per death were 2.7 and 2.0, respectively.
P Coolen, MN, S Best, Patient Review and Coordination Program, Washington State Health and Recovery Svcs Admin; A Lima, Center for Health Statistics, J Sabel, PhD, Injury and Violence Prevention Program, Washington State Dept of Health. L Paulozzi, MD, Div of Unintentional Injury Prevention, National Center for Injury Prevention and Control, CDC.
CDC Editorial Note:
The number of deaths attributed to poisoning, more than 90% of which involve drugs, has risen steadily in the United States for the past decade.1 Poisoning became second only to motor-vehicle crashes among leading causes of injury death in the United States in 2004.4 By 2006, poisoning had become the leading cause of unintentional injury death in Washington, five other states,‡ and the District of Columbia. Overdoses associated with prescription opioid painkillers are driving increases in poisoning death rates nationally,1 which parallel increases in opioid prescribing in the United States.2 Opioids are subject to abuse and are frequently used recreationally in combination with other drugs, including alcohol. In 2006, Washington's opioid overdose death rate was 8.2 per 100 000 population, compared with a national rate of 4.6 per 100 000.1 Some of this might be attributable to Washington's high rate of self-reported nonmedical use of prescription opioid painkillers, which was the fourth highest in the United States during 2006-2007.5 The findings of this analysis indicate that deaths from prescription opioid drug overdose in Washington occurred disproportionately among males and persons aged 45-54 years. This analysis also is the first to show an increased risk among persons enrolled in Medicaid. The age-adjusted risk of such a death for a Medicaid enrollee was 5.7 times the risk for a person not enrolled in Medicaid. These findings are similar to previous research showing a higher risk for such deaths in lower-income populations3 and can be used to better focus preventive interventions.
The cause of the higher death rate in Washington's Medicaid enrolled population might be related, in part, to differences in opioid prescribing in the Medicaid population. Although comparable prescribing data for Medicaid and non-Medicaid populations are not available for Washington, studies indicate that opioid prescribing rates among Medicaid enrollees are at least 2-fold higher than rates for persons with private insurance.6,7 In one of these studies, both the percentage of patients with pain being treated with opioids and the opioid dose per prescription were higher in Medicaid patients than in non-Medicaid patients.6 The higher death rate among Medicaid enrollees in Washington also might be related to a higher prevalence of substance abuse and other mental health problems, which has been found in other Medicaid populations.8 In this analysis, medical examiners and coroners reported the presence of an illegal drug (eg, cocaine, methamphetamine, and heroin) in nearly a fifth of deaths, and psychotherapeutic drugs such as benzodiazepines and antidepressants were reported in a high proportion of deaths.
Methadone, a drug used both for treatment of heroin addiction and as a long-acting, inexpensive painkiller, has become increasingly prominent among drugs involved in overdoses, both nationally and in state studies.1,9,10 Methadone's use as a painkiller increased more than 12-fold in the United States and Washington during 1997-2006,2 driven in part by its low cost. Washington ranked fourth among states in the per-capita consumption of methadone in 2005 and 2006.2
The findings in this report are subject to at least two limitations. First, the number of overdoses involving prescription opioids might be underestimated because (1) such drugs might not have been specified on the death certificates even though they contributed to death and (2) some deaths involving morphine and no other opioids were not included because the morphine detected might have been a metabolite of heroin. Second, some deaths labeled unintentional might have been suicides by poison or vice-versa, but the net effect of such errors likely is minimal.
Surveillance for prescription drug overdose deaths should be improved. Drugs listed on death certificates for overdoses are coded into broad categories, making identification of specific drugs difficult. Use of uncoded text in the cause-of-death fields on death certificates, as was done in this study, might be a promising strategy at the state or national level. Health authorities (eg, state and local health departments, coroner and medical examiner offices, and substance abuse programs) in other states should examine trends in and risks for prescription opioid-related overdose death in their jurisdictions, especially among Medicaid clients.
What is already known on this topic?
Since 1999, deaths from overdoses of prescription opioid painkillers have been increasing in the United States, but no study has determined whether the rate of such deaths is higher in the Medicaid population.
What is added by this report?
The rate of prescription opioid–related overdose death during 2004–2007 in Washington state was 30.8 in the Medicaid population and 4.0 per 100 000 in the non-Medicaid population (a relative risk of 5.7), and methadone was involved more frequently than any other prescription opioid.
What are the implications for public health practice?
Health authorities (eg, state and local health departments, coroner and medical examiner offices, and substance abuse programs) in other states should examine trends in and risks for prescription opioid–related overdose death in their jurisdictions, especially among Medicaid client.
*Available at http://apps.who.int/classifications/apps/icd/icd10online.
†During 2004-2007, approximately 90% of clients in the Washington PRC program misused prescription opioids by doctor shopping, frequent cycling through emergency departments, and prescription forgery. WSHRSA attempted to limit such misuse by restricting PRC clients to one primary-care provider, one narcotics prescriber, one pharmacy, and one hospital for nonemergency care. In addition, WSHRSA could require prior authorization for all opioid prescriptions.
‡Connecticut, Massachusetts, New Jersey, Ohio, and Rhode Island.
Overdose Deaths Involving Prescription Opioids Among Medicaid Enrollees—Washington, 2004-2007. JAMA. 2010;303(1):21–28. doi:
Create a personal account or sign in to: