Homicide and the Opioid Epidemic: A Longitudinal Analysis

• Published date: 01 August 2023
• DOI: http://doi.org/10.1177/10887679211054418
• Author: Richard Rosenfeld,Randolph Roth,Joel Wallman
• Date: 01 August 2023
• Subject Matter: Articles

https://doi.org/10.1177/10887679211054418

Homicide Studies

2023, Vol. 27(3) 321 –337

© 2021 SAGE Publications

Article reuse guidelines:

sagepub.com/journals-permissions

DOI: 10.1177/10887679211054418

journals.sagepub.com/home/hsx

Article

Homicide and the Opioid

Epidemic: A Longitudinal

Analysis

Richard Rosenfeld1, Randolph Roth2,

and Joel Wallman3

Abstract

Recent cross-sectional research has disclosed a positive relationship between

opioid-related death rates and homicide rates. The current study adds a longitudinal

dimension to this research. We estimate fixed effects panel models of the temporal

relationship between race-specific homicide rates and opioid-related death rates

within U.S. counties and county clusters between 1999 and 2015. The results reveal a

positive association between change over time in homicide and opioid-related deaths,

net of multiple socioeconomic and demographic controls, in both the Non-Hispanic

White and Black population. The association is stronger in the Appalachian counties,

where the opioid epidemic has been particularly severe.

Keywords

homicide, drugs, opioids, illegal markets, race, longitudinal analysis, trends

Prior research has revealed a significant association between the opioid epidemic and

homicide rates in the United States (Rosenfeld, Wallman et al., 2021; see also Gaston

et al., 2019). Rosenfeld, Wallman et al. (2021) argued that the demand for opioids,

particularly in street drug markets, increased homicide rates. They did not draw strong

causal conclusions, however, because their analysis was based on cross-sectional data.

They concluded simply that, controlling for other conditions, places with higher levels

of opioid demand, indexed by opioid death rates, had higher rates of homicide. They

1University of Missouri, St. Louis, USA

2Ohio State University, Columbus, USA

3The Harry Frank Guggenheim Foundation, New York, NY, USA

Corresponding Author:

Richard Rosenfeld, University of Missouri, One University Blvd., 8411 Roanoke Dr., Street, Louis,

MO 63121, USA.

Email: richard_rosenfeld@umsl.edu

1054418HSXXXX10.1177/10887679211054418Homicide StudiesRosenfeld et al.

research-article2021

322 Homicide Studies 27(3)

recommended that future research use longitudinal data to strengthen causal infer-

ences about the relationship between homicide and the opioid crisis.

The current study follows this recommendation. We hypothesize that a positive

relationship exists between changes over time in the prevalence of opioid use, which

we term “opioid demand,” and homicide rates in U.S. counties between 1999 and

2015. The results of the current study support our hypothesis. Growing demand for

opioids between 1999 and 2015 contributed to increases in homicide rates, in both the

Black and White populations. The paper proceeds by describing the empirical and

theoretical background for our hypothesis and the data, methods, results, and conclu-

sions of the current study.

Background

Opioid overdose death rates have risen sharply since 1999, generally considered to be

the beginning of the current opioid crisis in the United States. The age-adjusted over-

dose death rate from natural and semisynthetic opioids (e.g., hydrocodone and oxyco-

done) increased 340% from 1.0 per 100,000 population in 1999 to 4.4 per 100,000 in

2016. The age-adjusted rate of drug overdose deaths involving heroin changed little

from 1999 to 2005 and then increased from 0.7 per 100,000 in 2005 to 4.9 per 100,000

in 2016, a 600% increase. Overdose deaths involving fentanyl and its analogs rose

from 1.0 per 100,000 in 2013 to approximately 6.0 per 100,000 in 2016, a 500%

increase (Hedegaard et al., 2020). It is not an exaggeration to refer to the rise in opioid

overdose deaths as an epidemic.

We use the race-specific drug death rate, from the National Center for Health

Statistics (NCHS), to measure opioid demand. The NCHS contains a category com-

prising overdose deaths due to “narcotics and hallucinogens.” This category is domi-

nated by opioid overdoses, as deaths due to hallucinogen overdose are extremely rare.1

There is reason to believe that this classification substantially undercounts the number

of opioid overdose deaths, however, as many death certificates for drug overdose do

not specify a drug (Rudd et al., 2016; Ruhm, 2017). To increase the number of cases

available for analysis, therefore, we combined the overdose death rates for all drugs

(excluding alcohol) as our measure of opioid demand. That this polydrug measure

consists mainly of opioid overdoses is evident in the high correlation between it and

the NCHS narcotics and hallucinogen measure: r = .976 and 0.971 in the Non-Hispanic

White and Non-Hispanic Black populations, respectively.

Nonmedical use of opioids increased between 1999 and 2015, yet just 6% of users

reported in 2016 that their most recent source was a “drug dealer or other stranger”

(Substance Abuse and Mental Health Services Administration, 2017). An estimated

11.5 million users, however, means that nearly 700,000 of them utilized a street source

at least once to obtain opioids. As Rosenfeld, Wallman et al. (2021:552) point out,

“Even if the likelihood of violent conflict in a single illicit exchange is low, therefore,

the volume of activity in the illicit markets for opioids provides hundreds of thousands

of encounters with a potential for violence.”

As has been amply documented in prior research, illicit drug markets are risky

social spaces (e.g., Blumstein, 1995; Donohue & Levitt, 1998; Fagan & Chin, 1990;