Using our interactive chartsOn this page, we focus on drug-caused deaths from the King County Medical Examiner with an identified place of death (not necessarily the same as the place of residence) that involve drugs that can be abused and which are known to be potentially lethal; these deaths would commonly be known as "drug overdoses". (See our statewide opioid deaths page and major drug deaths page for more information.) The overall rate of death is based upon deaths of individuals--the number of drug-caused deaths, inclusive of all drug types. The rates for specific drugs are based upon the number of deaths with a particular drug present--these are largely duplicated data across drug types because most deaths involve multiple drugs, often more than two.
Several drug categories warrant additional definition. "Other opioids" includes those available pharmaceutically as well as illegal opioids other than heroin, such as acetyl fentanyl. Probable heroin deaths are categorized based upon both toxicological and death investigation information, with the methodology described in this report (see sidebar). Alcohol reporting has changed over time such that early data are not comparable with more recent data and so the early data are not presented.
Turn off the higher rate categories (All drug-involved deaths, Any opioid, etc.) by clicking the series name in the legend to rescale the plot and better see rates for less common drugs.
In the chart below, we trace the relative prominence of a given substance among all drug-caused deaths (as defined above). The sum of the proportions across all drugs exceeds 100% because most deaths involve multiple drugs. Opiates were involved in roughly 75% of drug-caused deaths over this period, but heroin and other opioids displayed strikingly mirror-imaged trends (turn off the other series to see heroin and other opioids): As heroin deaths declined, prescription-type opioids became more prominent in the 2000s, reaching a peak in terms of percentage of drug-caused deaths in 2009. Then the trends reversed, and heroin-caused deaths surpassed those involving other opioids in 2014. Methamphetamine accounted for 0 deaths in 1997 through 1999, but in 2015 meth was involved in over one quarter of drug-involved deaths.
Here we examine the rate of deaths associated with the most prominent drugs among those younger than 30. Rates are expressed as deaths (among those aged less than 30) per 100,000 county residents less than 30 years of age. From 2012-2015 heroin was the most common drug detected in deaths among young adults.
In the video below, we show where drug-involved deaths occur. Each dot represents a death in the two-year period displayed. In all, there were 4,584 drug-involved deaths in King County over the years 1997 through 2015.
In the next video, we concentrate on deaths from the most common drugs involved in accidental overdoses and poisonings: cocaine, methamphetamine, and opioids (including heroin and prescription-type). Deaths involving at least one of these substances comprise 3,988 out of the 4,584 drug-involved deaths, or 87%, from 1997 through 2015.
Above, we showed that there have been many drug deaths in central Seattle, where there are also more residents, but that many deaths have occurred outside the urban core. In this section, we examine more rigorously whether deaths have indeed been concentrated geographically, given population size and changes, and whether rates have changed over time anywhere in the county.
Spatial-temporal analyses are used to examine patterns over space and time statistically to determine if there are significant patterns in where and when something is happening. A software program, SaTScan, searches for patterns and can determine if there are high or low clusters of an event, such as a drug overdose, versus how many are expected in the area if the average death rate applied. These clusters can be in a certain place (geography) and the software determines where and how big the cluster is, e.g., one Census block group or several contiguous block groups. Clusters can also come and go over time or persist, so the years during which a cluster exists can also be determined.
In the analysis presented here, major drug overdose deaths occurring in King County are examined by year for the period from 1997-2015. Deaths involving methamphetamine, cocaine, heroin (probable), and/or pharmaceutical opioids are included. Note that most deaths involve multiple drugs and/or alcohol. Place data are based upon where a death occurred, with deaths occurring at or on the way to the hospital (e.g., in the ambulance) excluded to not create artificial clusters around hospitals. The analysis is based on the number of deaths in a Census block group or cluster of block groups divided by the population of the block group(s) for that year. This creates a local death rate that is comparable between places with different size populations and over time as population increases or decreases in an area.
The results of the analysis indicate two clusters that lasted during the entire timeframe. A large low cluster beginning on the west shores of Lake Washington and including much of Eastern King County represents an area in which the major drug death rate was consistently below the county-wide average for 1997-2015. While the number of deaths in Eastern King County did increase over time, as shown above, the rate of deaths (i.e. adjusting for changes in population size) did not. A high cluster was present in 48 block groups in downtown Seattle throughout 1997-2015, meaning that the rate in this region was higher than the county average and that the geography of the cluster remained consistent over time. The major drug death rate in the high cluster is over 10 times the rate in the low cluster.
Within King County, central Seattle--including parts of SoDo, Rainier Valley, First and Capitol Hills, Cascade/South Lake Union, and Belltown--demonstrated high rates of major drug deaths statistically beyond what was expected given population size. East Seattle and East King County had consistently low rates of major drug deaths. The other areas not in either cluster had rates throughout the period consistent with the county average. The increase in the number of deaths for all areas shown above are largely consistent with population growth.