Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2015 May 1.
Published in final edited form as: Int J Drug Policy. 2014 Oct 23;26(5):484–491. doi: 10.1016/j.drugpo.2014.10.001

Urban vs. rural differences in prescription opioid misuse among adults in the United States: Informing region specific drug policies and interventions

Khary K Rigg 1, Shannon M Monnat 2
PMCID: PMC4397122  NIHMSID: NIHMS646888  PMID: 25458403

Abstract

Background

In the United States, prescription opioid misuse (POM) has increased dramatically over the past two decades. However, there are still questions regarding whether rural/urban differences in adult POM exist, and more important, which factors might be driving these differences.

Methods

Using data from the 2011 and 2012 National Survey on Drug Use and Health, we conducted unadjusted and adjusted binary logistic regression analyses to determine the association between metropolitan status and POM.

Results

We found that urban adults were more likely to engage in POM compared to rural adults because of their higher use of other substances, including alcohol, marijuana, and other illicit and prescription drugs, and because of their greater use of these substances as children.

Conclusions

This study fills an important gap in the literature by not only identifying urban/rural differences in POM, but by also pointing out factors that mediate those differences. Because patterns and predictors of POM can be unique to geographic region, this research is critical to informing tailored interventions and drug policy decisions. Specifically, these data suggest that interventions should be aimed at urban illicit drug users and adults in manual labor occupations.

Keywords: prescription opioid misuse, prevention, rural and urban drug use, poly-drug use, interventions, drug policy

INTRODUCTION

Over the last two decades, prescription opioid misuse (POM) has increased dramatically to become a serious behavioral health concern in the United States (Substance Abuse Mental Health Services Administration [SAMHSA], 2012). The morbidity, mortality, and economic burden caused by POM are substantial. A recent report estimated that in 2007, POM cost the U.S. over $55 billion in healthcare, lost workplace productivity, and criminal justice costs (Birnbaum et al., 2011). Additionally, the rate of treatment admissions for POM has quadrupled from 1999-2009 (SAMHSA, 2010), and prescription opioids are now implicated in more deaths annually than cocaine and heroin combined (Centers for Disease Control and Promotion [CDC], 2011).

POM is particularly problematic in some rural areas (Wunsch et al., 2009). Places like rural Maine and areas of Appalachia, including eastern Kentucky, southern Ohio, and western Virginia all have longstanding struggles with POM (Inciardi & Goode, 2007). In fact, the origins of the POM epidemic can be traced back to rural America, where media reports of OxyContin® abuse first began to surface (Hays, 2004). Recent data from rural communities across the country show that POM is still active in these regions (Havens et al., 2011).

POM is also endemic to many large metropolitan centers. For example, New York City and South Florida have had well-documented problems with POM, such as high rates of overdose deaths, excessive opioid diversion, and surges in treatment admissions (Florida Office of Attorney General, 2012; NYC Mayor's Task Force on Prescription Painkiller Abuse, 2013). These and other urban locations have been at the center of the POM epidemic for many years now (Lankenau et al., 2012). Over the past decade, prescription opioids have become a prominent fixture in the urban black market for drugs (Rigg, Kurtz, & Surratt, 2012) because they are used in conjunction with other street and club drugs, and are a common substitute for heroin (Rigg & Ibañez, 2010; Davis & Johnson, 2008; Kurtz et al., 2006).

A number of studies have examined rural/urban differences in POM, with conflicting findings. On one hand, researchers using various restricted samples have found that rural adults have greater odds of POM than their urban peers. For example, in a sample of felony probationers, Havens et al. (2007) found that rural probationers were almost five times more likely than their urban counterparts to have misused prescription opioids. In a recent study of prescription drug users in Kentucky, Young et al. (2012) found that rural Kentuckians were more likely than urban Kentuckians to begin using oxycodone and hydrocodone at earlier ages and have significantly greater odds of lifetime and recent use of various prescription opioids. Similarly, in a sample of methadone maintenance patients, Rosenblum et al. (2007) found that the odds of POM increased as urbanicity decreased. On the other hand, using nationally representative data, Wang et al. (2013) found similar prevalence of POM in urban and rural counties, and using U.S. poison center surveillance data, Spiller et al. (2009) similarly found that population density was not associated with POM.

In addition to lack of agreement about whether there are indeed rural/urban differences in adult POM, these studies often present only fully adjusted models with no attempt to identify the characteristics that might explain the rural/urban POM differences that exist. Understanding the demographic and psychosocial characteristics of individuals that might influence rural/urban differences in POM is important for tailoring interventions to the unique needs and differences of residents in these areas. There are a number of important compositional differences in human capital (embodied knowledge and skills that impact one's ability to work; Becker, 1962) and resource characteristics, perceptions of risk or prior experience with sanctions, health care experiences, and prior and recent substance use that may explain why residents in one type of area are more or less likely to engage in POM than residents in another type of area.

On the one hand, compared to urban residents, rural residents have been found to have lower human capital characteristics (e.g., income, educational attainment; Byun et al., 2012; Roscigno & Crowle, 2001), a higher proportion of manual labor occupations (McGranahan, 2003), less access to illicit street drugs (James et al., 2002), worse self-rated health (Monnat & Pickett, 2011), less psychological distress (Dhingra et al., 2009), greater frequency of emergency department (ED) use (Haggerty et al., 2014; Haggerty et al., 2007), and higher arrest rates for crimes like domestic violence and drunk driving (Logan, Walker, & Leukefeld; Olson, Weisheit, & Ellsworth). Each of these factors have been found to be associated with POM in various adult populations (Rigg & DeCamp, 2014; Dowling, Stor & Chilcoat, 2006; Becker et al., 2008; Harrell & Broman, 2009; Cicero et al., 2008). On the other hand, rural residents are more likely to attend religious services than their urban counterparts (Chalfant & Heller, 1991; Ellison & Sherkat, 1995) and are less likely to engage in illicit drug use (Gfroerer et al., 2008), both of which might reduce the likelihood that rural residents will engage in POM.

Given the ambiguity surrounding whether there are indeed rural/urban differences in adult POM, and given our lack of understanding about the characteristics of individuals that might influence these differences, this study builds on previous research in this area by using a nationally representative data set of US adults to identify rural vs. urban differences in POM and to identify the characteristics that mediate those differences.

METHODS

Data sources and respondents

We use data from the National Survey of Drug Use and Health (NSDUH), which is an annual survey of the U.S. population aged 12 and older that has been collected for over 30 years. The main focus of the NSDUH is prevalence and correlates of substance use in the U.S. The NSDUH uses a multi-stage area probability sample of all 50 states and the District of Columbia. The survey is administered using computer assisted in-person and audio self-interviewing (U.S. Department of Health and Human Services [DHHS], 2012). We aggregated data from the 2011 and 2012 survey years to enable a large enough sample for robust analysis of respondents aged 18 and older. Our analyses are restricted to non-Hispanic Whites, non-Hispanic Blacks, and Hispanics. Due to the importance of controlling for race/ethnicity in this analysis and the small NSDUH sample sizes for Asians, Native Americans, Native Hawaiians and Pacific Islanders, and multi-race groups living in rural counties, we excluded those groups from our analysis. In 2011 and 2012, the weighted screening response rates were 87.0% and 86.1%, respectively, and the weighted interview response rates were 74.4% and 73.0%, respectively (DHHS, 2011; DHHS, 2012).

Variables selected for analysis

Our outcome of interest was POM in the past year. Respondents were asked: “How long has it been since you last used any prescription pain reliever that was not prescribed for you or that you took only for the experience or feeling it caused?” We dichotomized respondents into those who reported POM in the past year vs. those who did not (including those who never engaged in POM). The survey asked respondents about several specific opioids, including codeine, hydrocodone, oxycodone, Demerol, Dilaudid, methadone, and morphine, among others.

Our main independent variable of interest was metropolitan status. The NSDUH categorizes respondents’ counties of residence into large metropolitan, small metropolitan, and nonmetropolitan. Nonmetropolitan counties are those that fall outside of a metropolitan statistical area (MSA) as defined by the Office of Management and Budget (OMB). We collapsed respondents living in large metropolitan and small metropolitan counties into one metropolitan category (Wang et al., 2013). Throughout the paper, we use the term “urban” in lieu of metropolitan and the term “rural” in lieu of nonmetropolitan. These terms have been used interchangeably in prior studies on drug use in rural and urban areas (Havens et al. 2011). As a sensitivity check, we also ran models using the all three metropolitan status categories instead of the dichotomous operationalization. The results from those models were the same, and there were no significant differences in POM between respondents in small metropolitan versus large metropolitan counties.

We examined a number of covariates as potential mediators in the association between metropolitan status and POM. These are characteristics that have been found to vary between urban and rural residents and are also likely to be associated with POM. First, human capital characteristics included family income, receipt of any public assistance in the past year, educational attainment, employment status, having health insurance, and working in a manual labor occupation (farming, fishing and forestry, installation, maintenance and repair work, and construction trades and extraction work). Perceived risk and sanctions variables included respondent's perception that heroin is easy to obtain, perception that heroin use is risky, whether the respondent had ever been arrested for breaking the law, and number of religious services attended in the past year. Clinical characteristics included self-rated health, visiting an ED for medical treatment in the past year, and a psychological distress variable measured by a series of questions that comprise the K6 scale – a validated scale that measures respondents symptoms of distress: feeling nervous, hopeless, restless or fidgety, sad or depressed, everything is an effort, and no good or worthless. Respondents rated the severity of each symptom on a scale of 0 to 4 during the past year when they felt the worst. For each respondent, the NSDUH provides a summed K6 score, and this score was used to create a variable for past-year psychological distress, where a score greater than 13 indicates the presence of serious psychological distress (Kessler et al., 2003; Wang et al., 2013). Substance use variables included past year use of nicotine, alcohol, marijuana, other illicit drugs (cocaine, crack, heroin, methamphetamine, and hallucinogens), and non-medical use of non-opioid prescription drugs (tranquilizers, sedatives, and stimulants). We also included age of first alcohol use, age of first cigarette use, and a composite variable for age of first illicit drug use (marijuana, cocaine, crack, heroin, hallucinogens, and methamphetamine).

Finally, we controlled for a number of demographic characteristics based upon results from previous studies that indicate higher odds of POM among certain demographic groups (Becker et al., 2008; Cicero, Inciardi & Muñoz, 2005). Control variables included race/ethnicity, age, gender, marital status, and presence of children in the household. To standardize the family income variable, we also control for total number of people living in the household.

Statistical analysis

We first used difference of means and difference of proportion t-tests to compare characteristics between respondents who indicated past-year POM vs. those who indicated no past-year POM as well as between urban and rural respondents. We then conducted unadjusted and adjusted binary logistic regression analysis to determine the association between metropolitan status and POM as well as to determine whether human capital and resource characteristics, perceptions of risk and sanctions, clinical characteristics, or substance abuse characteristics explain the rural/urban differences that exist. Each group of variables is introduced in a sequential manner (Keith, 2006). To determine whether the changes in the magnitude of the coefficients are a function of true mediation vs. the addition of covariates, we use the KHB method in Stata – a method that allows researchers to statistically assess the influence of confounding relative to the influence of rescaling (Karlson, Holm & Breen, 2012). To assess for risk of multicollinearity, we examined correlations between all covariates and found no strong associations (i.e., greater than 0.6) that would suggest a problem with multicollinearity in our regression models. All analyses were performed using StataMP version 13. We used study-calculated weights and robust standard errors via the ‘cluster’ option in Stata to adjust for the complex sampling design and non-response bias.

RESULTS

Our sample was composed of 47,440 respondents, with 5.7% of respondents reporting POM in the past year. Characteristics of respondents are presented in Table 1 for the sample as a whole, as well as by POM and by county type. There were a number of significant differences between respondents who reported POM vs. those who did not. Probability of POM was significantly higher among urban respondents, Whites, younger respondents, individuals who were never married, those with children living in the household, low income and low education respondents, those receiving public assistance, respondents who were either employed or unemployed, those working in manual labor occupations, respondents who perceive that heroin is easy to obtain, those who have been arrested, those who attended no religious services in the past year, individuals with serious psychological distress, individuals with at least one ED visit in the past year, respondents who used nicotine, alcohol, marijuana, other illicit drugs, or other prescription drugs (non-opioids) non-medically in the past year, and those who first used alcohol, cigarettes, and illicit drugs under the age of 18. Probability of POM was significantly lower among females, married respondents, respondents with health insurance, retired individuals or those keeping house, individuals who perceive heroin use as risky, and individuals who attended religious services more than 24 times in the past year.

Table 1.

Descriptive Statistics and Difference of Proportions tests for Variables Included in Sample, 2011-2012

Past Year Nonmedical Opioid Use County Type
Full Sample (N=47,440) No (N=43,169) Yes (N=4,271) p-value Urban (N=37,586) Rural (N=9,854) p-value
Non-medical use of prescription opioids 5.7 ------ ------ ------ 5.9 4.9 0.0002
County Type
    Urban 84.0 83.9 86.3 <.0001 ------ ------ ------
    Rural 16.0 16.1 13.7 <.0001 ------ ------ ------
DEMOGRAPHIC CHARACTERISTICS
Race/Ethnicity
    non-Hispanic white 74.0 73.8 76.2 0.0005 71.5 87.0 <.0001
    non-Hispanic black 11.6 11.8 9.3 <.0001 12.5 7.1 <.0001
    Hispanic 14.4 14.4 14.5 0.812 16.0 5.9 <.0001
Age
    18-25 15.1 14.0 34.1 <.0001 15.4 13.6 <.0001
    26-34 16.8 16.1 28.2 <.0001 17.0 15.6 0.0007
    35-49 27.3 27.3 24.1 <.0001 27.4 26.0 0.006
    50 and older 40.9 42.6 13.5 <.0001 40.2 44.8 <.0001
Female 49.2 49.6 43.3 <.0001 49.2 49.3 0.873
Marital status
    Married 52.0 53.3 29.6 <.0001 51.2 55.7 <.0001
    Divorced, separated, widowed 20.0 20.2 16.2 <.0001 19.6 22.1 <.0001
    Never married 28.1 26.5 54.2 <.0001 29.2 22.2 <.0001
Number of people in household 3.00 (1.40) 2.99 (1.42) 3.28 (1.12) <.0001 3.03 (1.44) 2.87 (1.20) <.0001
Any children in household 39.0 38.8 42.6 <.0001 39.4 37.0 <.0001
HUMAN CAPITAL & RESOURCE CHARACTERISTICS
Family Income
    Less than $20,000 18.4 18.1 24.1 <.0001 17.6 22.7 <.0001
    $20,000-39,999 21.1 20.9 25.4 <.0001 20.1 26.7 <.0001
    $40,000-74,999 28.0 28.1 27.1 0.177 27.7 29.9 <.0001
    $75,000 or more 32.4 32.9 23.4 <.0001 34.6 20.6 <.0001
Any public assistance in past year 20.1 19.4 31.3 <.0001 19.3 24.1 <.0001
Educational attainment
    Less than high school 13.0 12.8 17.4 <.0001 12.3 16.7 <.0001
    High school grad/some college 57.7 57.4 63.1 <.0001 56.4 64.5 <.0001
    College graduate 29.3 29.9 19.5 <.0001 31.3 18.8 <.0001
Employment status
    Employed 67.0 66.8 71.0 <.0001 68.0 62.0 <.0001
    Unemployed 8.1 7.7 14.9 <.0001 8.1 8.3 0.524
    Disabled 5.6 5.6 4.8 0.015 5.0 8.5 <.0001
    In School 2.6 2.5 4.3 <.0001 2.7 2.1 0.0003
    Keeping house or retired 16.8 17.5 5.0 <.0001 16.3 19.2 <.0001
Works in manual labor occupation 6.8 6.7 10.3 <.0001 6.6 8.4 <.0001
Has health insurance 83.2 83.9 71.7 <.0001 83.5 81.5 <.0001
RISK/SANCTIONS
Perceives that heroin is easy to obtain 17.3 16.8 25.7 <.0001 17.6 15.6 <.0001
Perceives heroin use as risky 84.3 85.0 73.3 <.0001 84.3 84.5 0.610
Ever been arrested and booked for breaking law 22.8 21.6 43.4 <.0001 22.5 24.6 <.0001
# of religious services attended in past year
    None 40.3 39.8 49.7 <.0001 40.5 39.5 0.061
    1-5 20.6 20.3 25.5 <.0001 20.9 19.0 <.0001
    6-24 13.4 13.6 11.0 <.0001 13.6 12.9 0.088
    more than 24 25.7 26.4 13.9 <.0001 25.1 28.6 <.0001
CLINICAL CHARACTERISTICS
Poor/fair self-rated health 13.2 13.1 14.1 0.071 12.4 17.1 <.0001
Psychological distress 11.9 10.9 29.4 <.0001 11.8 12.4 0.130
Had at least one emergency room visit in past year 29.0 28.3 41.7 <.0001 28.4 32.3 <.0001
SUBSTANCE USE
Nicotine use 40.9 38.9 73.2 <.0001 40.1 44.8 <.0001
Alcohol use 76.1 75.2 91.6 <.0001 77.5 68.8 <.0001
Marijuana use 16.4 13.9 57.7 <.0001 17.1 12.9 <.0001
Any illicit drug use (not incl. marijuana) 4.2 2.7 29.2 <.0001 4.6 2.6 <.0001
Non-medical use of other non-opioids 4.2 2.1 38.1 <.0001 4.3 3.6 0.003
Age of first cigarette use
    Never used 21.6 22.5 7.4 <.0001 21.8 20.2 0.0005
    under 18 57.3 56.2 75.1 <.0001 56.7 60.7 <.0001
    18-25 19.4 19.7 16.2 <.0001 19.9 17.3 <.0001
    older than 25 1.7 1.7 1.4 0.301 1.7 1.7 0.745
Age of first alcohol use
    Never used 7.9 8.3 1.9 <.0001 7.6 9.8 <.0001
    under 18 58.7 57.3 81.9 <.0001 59.1 56.4 <.0001
    18-25 31.7 32.7 15.8 <.0001 31.6 31.9 0.576
    older than 25 1.7 1.8 0.5 <.0001 1.7 1.9 0.062
Age of first illicit drug use
    Never used 36.4 38.1 8.5 <.0001 35.4 42.1 <.0001
    under 18 36.4 34.3 70.0 <.0001 37.0 32.8 <.0001
    18-25 23.3 23.6 20.0 0.019 23.8 20.7 <.0001
    older than 25 3.9 4.0 1.5 <.0001 3.8 4.3 0.016
Open in a new tab

Notes: Two-tailed difference of proportions t-tests; weighted

There are also a number of important differences between respondents living in urban vs. rural areas. Those living in rural areas are more likely to be White, older, married, have lower incomes and lower educational attainment, receive public assistance, be disabled, keeping house, or retired, and work in a manual labor occupation. Those in rural areas are also more likely than those in urban areas to have: been arrested, attended church more than 24 times in the past year, rate their health as fair or poor, visited the ED for medical care at least once in the past year, used nicotine in the past year, and started smoking under the age of 18. Those in rural areas are less likely than those in urban areas to: have children living in the household, be employed, have health insurance, perceive that heroin is easy to obtain, have used alcohol, marijuana, other illicit drugs, and non-opioid prescriptions within the past year, and have started using alcohol or illicit drugs as a child.

Regression results

In an unadjusted binary logistic regression model, we found that rural respondents had about 18% lower odds of POM compared with urban respondents (OR 0.82, 95% CI 0.72-0.95, p<.008). Results of adjusted binary logistic regression models predicting POM are presented in Table 2. Model 1 shows that, adjusting for demographic characteristics, respondents residing in rural areas have about 15% lower odds of POM compared with urban respondents (OR 0.85, CI 0.73-0.92, p<.019). The addition of human capital and resource characteristics in Model 3 increased the magnitude of the urban/rural difference to 25% (OR 0.75, CI 0.64-0.86, p<.001); not accounting for human capital and contextual characteristics in Model 1 masked the true magnitude of the association between metropolitan status and POM. Supplemental analysis (not shown but available upon request) indicated that the addition of the educational attainment variable led to the decrease in the odds ratio for ‘rural’. This is because rural respondents have significantly lower educational attainment than urban respondents, and low educational attainment is associated with increased POM. Our results suggest that rural residents have lower POM than would be expected based on their educational attainment alone, and if rural residents had the same higher level of education as urban residents, their odds of POM would be 25% lower than their urban peers.

Table 2.

Odds Ratios and 95% Confidence Intervals from Logistic Regression Models Predicting POM

Model 1 Model 2 Model 3
OR 95% CI P OR 95% CI p OR 95% CI p
County Type (urban=ref)
    Rural 0.845 0.734-0.972 0.019 0.745 0.644-0.861 <.001 0.770 0.664-0.892 <.001
DEMOGRAPHIC CHARACTERISTICS
Race/Ethnicity (non-Hispanic white=ref)
    non-Hispanic black 0.575 0.475-0.696 <.001 0.479 0.392-0.584 <.001 0.494 0.404-0.603 <.001
    Hispanic 0.661 0.557-0.786 <.001 0.573 0.480-0.684 <.001 0.631 0.528-0.754 <.001
Age (18-25=ref)
    26-34 0.920 0.803-1.054 0.229 0.908 0.790-1.043 0.173 0.795 0.691-0.914 0.001
    35-49 0.518 0.437-0.613 <.001 0.512 0.433-0.606 <.001 0.458 0.387-0.542 <.001
    50 and older 0.191 0.115-0.251 <.001 0.219 0.165-0.290 <.001 0.220 0.166-0.291 <.001
Female 0.802 0.716-0.897 <.001 0.882 0.782-0.995 0.041 1.081 0.958-1.220 0.206
Marital status (married=ref)
    Divorced, separated, widowed 1.892 1.547-2.315 <.001 1.515 1.227-1.870 <.001 1.372 1.108-1.697 0.004
    Never married 1.999 1.681-2.378 <.001 1.682 1.413-2.002 <.001 1.477 1.243-1.756 <.001
Number of people in household 1.087 1.033-1.144 0.001 1.046 0.992-1.104 0.098 1.049 0.994-1.107 0.081
Any children in household 0.895 0.764-1.047 0.165 0.844 0.720-0.989 0.036 0.859 0.732-1.007 0.060
HUMAN CAPITAL & RESOURCE CHARACTERISTICS
Family Income ($75,000 or more=ref)
    Less than $20,000 1.011 0.833-1.227 0.913 1.012 0.832-1.231 0.904
    $20,000-39,999 1.142 0.944-1.381 0.172 1.131 0.934-1.369 0.207
    $40,000-74,999 1.116 0.939-1.327 0.214 1.115 0.936-1.328 0.221
Any public assistance in past year 1.595 1.374-1.852 <.001 1.456 1.248-1.699 <.001
Educational attainment (college=ref)
    Less than high school 1.499 1.202-1.870 <.001 1.288 1.032-1.609 0.026
    High school graduate 1.318 1.109-1.566 0.002 1.200 1.008-1.428 0.040
Employment status (employed=ref)
    Unemployed 1.277 1.091-1.496 0.002 1.182 1.009-1.385 0.039
    Disabled 0.904 0.665-1.228 0.517 0.835 0.613-1.138 0.254
    In School 0.838 0.682-1.029 0.091 0.894 0.728-1.098 0.286
    Keeping house or retired 0.514 0.386-0.684 <.001 0.526 0.396-0.698 <.001
Manual labor occupation 1.268 1.028-1.565 0.027 1.225 0.990-1.515 0.062
Has health insurance 0.812 0.712-0.926 0.002 0.861 0.755-0.982 0.026
RISK/SANCTIONS
Perceives that heroin is easy to obtain 1.536 1.347-1.752 <.001
Perceives heroin use as risky 0.554 0.486-0.631 <.001
Ever been arrested and booked for breaking law 2.118 1.880-2.387 <.001
# of religious services attended in past year (none=ref)
    1-5 1.125 0.981-1.290 0.092
    6-24 0.822 0.680-0.993 0.042
    more than 24 0.735 0.613-0.881 0.001
Pseudo R-square 0.083 0.099 0.126
Model 4 Model 5
OR 95% CI p OR 95% CI p
County Type (urban=ref)
    Rural 0.768 0.662-0.892 0.001 0.886 0.746-1.052 0.167
DEMOGRAPHIC CHARACTERISTICS
Race/Ethnicity (non-Hispanic white=ref)
    non-Hispanic black 0.514 0.421-0.628 <.001 0.799 0.641-0.998 0.048
    Hispanic 0.667 0.558-0.798 <.001 0.941 0.778-1.139 0.534
Age (18-25=ref)
    26-34 0.814 0.707-0.938 0.004 1.017 0.859-1.203 0.846
    35-49 0.490 0.413-0.582 <.001 0.777 0.636-0.948 0.013
    50 and older 0.250 0.187-0.333 <.001 0.502 0.370-0.682 <.001
Female 0.983 0.869-1.112 0.789 1.058 0.917-1.220 0.442
Marital status (married=ref)
    Divorced, separated, widowed 1.294 1.041-1.607 0.020 1.150 0.909-1.454 0.245
    Never married 1.471 1.235-1.751 <.001 1.103 0.901-1.349 0.342
Number of people in household 1.044 0.989-1.101 0.121 1.029 0.965-1.097 0.388
Any children in household 0.875 0.747-1.025 0.098 0.989 0.826-1.185 0.907
HUMAN CAPITAL & RESOURCE CHARACTERISTICS
Family Income ($75,000 or more=ref)
    Less than $20,000 0.943 0.774-1.150 0.563 0.954 0.759-1.199 0.688
    $20,000-39,999 1.087 0.897-1.318 0.395 1.231 0.994-1.524 0.057
    $40,000-74,999 1.089 0.913-1.298 0.343 1.225 1.005-1.493 0.044
Any public assistance in past year 1.355 1.160-1.583 <.001 1.386 1.164-1.651 <.001
Educational attainment (college=ref)
    Less than high school 1.213 0.971-1.517 0.090 1.336 1.034-1.726 0.027
    High school graduate 1.165 0.978-1.517 0.087 1.148 0.944-1.397 0.166
Employment status (employed=ref)
    Unemployed 1.113 0.949-1.306 0.118 0.962 0.801-1.155 0.679
    Disabled 0.610 0.448-0.832 0.002 0.596 0.417-0.852 0.005
    In School 0.895 0.726-1.103 0.298 0.820 0.638-1.054 0.121
    Keeping house or retired 0.516 0.388-0.685 <.001 0.658 0.481-0.899 0.009
Manual labor occupation 1.248 1.006-1.548 0.044 1.387 1.104-1.743 0.005
Has health insurance 0.859 0.753-0.981 0.025 0.931 0.802-1.079 0.341
RISK/SANCTIONS
Perceives that heroin is easy to obtain 1.436 1.255-1.642 <.001 1.213 1.034-1.422 0.018
Perceives heroin use as risky 0.550 0.482-0.628 <.001 0.582 0.499-0.678 <.001
Ever been arrested and booked for breaking law 2.003 1.776-2.258 <.001 1.206 1.051-1.384 0.008
# of religious services attended in past year (none=ref)
    1-5 1.116 0.972-1.281 0.119 1.236 1.057-1.446 0.008
    6-24 0.810 0.668-0.983 0.033 0.976 0.782-1.217 0.826
    more than 24 0.751 0.626-0.902 0.002 1.191 0.975-1.454 0.087
CLINICAL CHARACTERISTICS
Poor/fair self-rated health 1.136 0.946-1.364 0.172 1.148 0.931-1.417 0.197
Psychological distress 2.107 1.848-2.402 <.001 1.544 1.328-1.794 <.001
At least one visit to emergency room in past year 1.465 1.303-1.647 <.001 1.347 1.183-1.535 <.001
SUBSTANCE USE
Nicotine use 1.321 1.116-1.564 0.001
Alcohol use 1.485 1.124-1.963 0.005
Marijuana use 2.137 1.835-1.963 <.001
Any illicit drug use (not incl. marijuana) 2.128 1.779-2.545 <.001
Non-medical use of other non-opioids 10.378 8.672-12.420 <.001
Age of first cigarette use (never used=ref)
    under 18 1.015 0.754-1.365 0.924
    18-25 0.937 0.689-1.274 0.676
    older than 25 1.435 0.650-3.167 0.372
Age of first alcohol use (never used=ref)
    under 18 0.951 0.533-1.697 0.865
    18-25 0.802 0.459-1.402 0.439
    older than 25 0.883 0.345-2.256 0.795
Age of first illicit drug use (never used=ref)
    under 18 2.009 1.514-2.666 <.001
    18-25 1.647 1.230-2.204 0.001
    older than 25 1.207 0.665-2.191 0.536
Pseudo R-square 0.142 0.298
Open in a new tab

Notes: two tailed tests, weighted

Notes: two tailed tests, weighted

The addition of perceived risk and sanction variables in Model 3 and clinical characteristics in Model 4 did little to change the magnitude of the rural/urban difference in POM. Net of these controls, rural residents remained significantly less likely to report POM (OR 0.77, CI 0.66-0.89, p<.001). However, the introduction of substance use variables in Model 5 resulted in a decrease in and elimination of significance for the rural/urban difference. Results of our KHB test for statistical mediation (not shown due to space constraints, but available upon request) found a significant difference in the coefficient for ‘rural’ between Model 4 and Model 5. In addition, the KHB test indicated that, net of the rescaling that occurs due to the introduction of new variables, illicit drug use before the age of 18 accounted for 15.4% of the change in the coefficient for ‘rural’ and use of alcohol, marijuana, other illicit drugs, and other prescriptions (non-opioids) in the last year collectively accounted for 36.6% of the change in the coefficient for ‘rural.’

DISCUSSION

The current study is important for a number of reasons. Whereas prior work in this area tends to rely on restricted samples, we used a nationally representative data set of US adults which allows for broader generalization. Additionally, our study fills an important gap in the literature by not only identifying urban/rural differences in POM, but by also pointing out factors that mediate those differences. Because the drivers of POM are often area specific, this research is critical to informing tailored prevention and policy decisions.

Results showed that the profile of U.S. adult prescription opioid misusers is largely consistent with those found in other studies, particularly with regards to demographics and psychosocial characteristics (Sigmon, 2006; Cicero, Inciardi & Muñoz, 2005). Interestingly, however, we found that residents living in urban areas were more likely to engage in POM than those in rural regions. This differs with the results of other studies that found rural adults more at risk for POM (Havens et al., 2007; Rosenblum et al., 2008; Young et al., 2012), but those studies were restricted to small samples of very specific populations (probationers, methadone maintenance patients, and residents of Kentucky). Our results are more in line with those of Wang et al. (2013), who using nationally representative data, also found POM prevalence to be slightly higher among urban residents than rural residents. It is worth noting that our estimate of POM is slightly higher than those found in the Wang et al. study. This may be partly explained by our sample being limited to White, Black, and Hispanic respondents. Respondents who were excluded due to small nonmetropolitan samples, particularly Asians, have very low rates of POM (Wu et al., 2013; Watkins & Ford, 2011). Accordingly, interpretation of our results should be considered in light of the racial/ethnic context of our sample. In addition, Wang et al. used data from the 2008 and 2009 versions of the NSDUH while our analysis uses more recent data from 2011 and 2012.

We also observed that religiosity was protective against POM. Few studies examine this association among adults. Prior work has shown that religiosity can protect against numerous types of drug abuse (Palamar, Kiang & Halkitis, 2014; Herman-Stahl et al., 2007); our study is among the first, however, to show that religiosity can specifically protect against adult POM. Additionally, results showed that adults in manual labor occupations were more likely to misuse opioids than adults who were not in these professions. This finding suggests that individuals in manual labor occupations are at increased risk for POM and should be monitored more closely for chronic pain and opioid use/misuse. Employers and labor unions in these professions would do well to develop POM prevention programs for their workers that aim to educate on the dangers of POM and also provide them with evidence-based alternatives to using opioids for pain management (e.g., cognitive behavioral therapy, massage treatment).

Our finding of lower odds of POM among rural adults is different from adolescent- focused studies that have found rural teenagers to be at greater risk for POM than their urban counterparts (McCauley et al., 2010; Dew, Elison & Dozier, 2007). A recent population-based study (Havens, Young & Havens, 2011), for example, found that adolescents living in rural areas were significantly more likely to report lifetime POM than adolescents residing in urban areas. These previous studies suggest that rural adolescents are more likely to engage in POM probably because drugs such as heroin are less available to adolescents in rural regions (Draus & Carlson, 2006).

Our results suggest though that urban adults have a higher likelihood of POM compared to rural adults because of their higher use of other substances, including alcohol, marijuana, and other illicit and prescription drugs and because of their greater use of these substances as children. This finding suggests that, in urban areas, prescription opioids are more often used in conjunction with other drugs than they are in rural areas, and that this type of poly-drug use is contributing to elevated levels of urban POM. Additionally, our findings suggest that a reduction in illicit drug use among adolescents will lower rates of POM in urban areas. Policymakers and practitioners who are interested in reducing urban POM should take note of the close connection between POM and illicit drug use in these communities and therefore aim interventions at illicit drug users, particularly those under age 18 (Rigg & Murphy, 2013).

Our results should be interpreted in light of some methodological considerations. First, the NSDUH is a household sample of the non-institutionalized population of the US and therefore may not necessarily represent persons not permanently attached to one particular household. For example, it is unlikely for persons who experience excessive housing instability or are serious long-term substance users to appear in the NSDUH. Second, this study utilized cross-sectional data which prohibited us from being able to establish temporal order and ascribe causality. Third, the data used are based on self-reports of substance use and may be subject to under-reporting due to recall and/or social desirability bias. These effects are believed to have been mitigated, however, through the use of a trained interviewer. Last, we acknowledge as a limitation our inability to distinguish residents in large nonmetropolitan counties that are adjacent to metro counties from those living in small remote rural counties. Further, counties are heterogeneous geographic units where many outlying residents of an urban county (e.g., exurban or urban fringe) may experience a life that is more rural in character than residents living in the center city of an urban county. Unfortunately, the NSDUH do not allow for a more fine grained geographic analysis. These more nuanced definitions of rural vs. urban may lead to different findings.

Despite these limitations, this research begins to provide answers to the question of whether rural/urban differences in adult POM exist, and more important, which factors might be driving these differences. This study was among the first to demonstrate a significantly higher prevalence of POM among urban adults using a nationally representative sample. Moreover, we were able to specifically show that these differences are mediated, in large part, by the fact that urban adults tended to initiate illicit drug use prior to age 18 and were more likely to have used other drugs in the past year. Future research should attempt to determine if rural/urban differences are more pronounced among certain subgroups of users (e.g., racial/ethnic minorities) or different classes of controlled medications (e.g., benzodiazepines). Additional data on the unique POM onset patterns in these areas would also help facilitate the development of more region specific drug policies and interventions.

Acknowledgements

Dr. Monnat would like to acknowledge support from the Population Research Institute at Penn State which receives core funding from the National Institute of Child Health and Human Development (Grant R24-HD041025).

Contributor Information

Khary K. Rigg, Department of Mental Health Law & Policy Louis de la Parte Florida Mental Health Institute University of South Florida 13301 Bruce B. Downs Blvd. Tampa, FL 33612 (USA)

Shannon M. Monnat, Department of Agricultural Economics, Sociology, and Education Population Research Institute The Pennsylvania State University 110B Armsby Bldg. University Park, PA 16801 (USA) smm67@psu.edu

References

  1. Becker WC, Sullivan LE, Tetrault JM, Desai RA, Fiellin DA. Non- medical use, abuse and dependence on prescription opioids among US adults: psychiatric, medical and substance use correlates. Drug and Alcohol Dependence. 2008;94(1):38–47. doi: 10.1016/j.drugalcdep.2007.09.018. [DOI] [PubMed] [Google Scholar]
  2. Becker GS. Investment in human capital: A theoretical analysis. The Journal of Political Economy. 1962:9–49. [Google Scholar]
  3. Birnbaum HG, White AG, Schiller M, Waldman T, Cleveland JM, Roland CL. Societal costs of prescription opioid abuse, dependence, and misuse in the United States. Pain Medicine. 2011;12:657–667. doi: 10.1111/j.1526-4637.2011.01075.x. [DOI] [PubMed] [Google Scholar]
  4. Byun SY, Meece JL, Irvin MJ. Rural-nonrural disparities in postsecondary educational attainment revisited. American Educational Research Journal. 2012;49(3):412–437. doi: 10.3102/0002831211416344. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Centers for Disease Control and Prevention [CDC] Vital signs: Overdoses of prescription opioid pain relievers—United States, 1999-2008. Morbidity and Mortality Weekly Report. 2011;60(43):1487. [PubMed] [Google Scholar]
  6. Chalfant HP, Heller PL. Rural/urban versus regional differences in religiosity. Review of Religious Research. 1991:76–86. [Google Scholar]
  7. Cicero TJ, Lynskey M, Todorov A, Inciardi JA, Surratt HL. Co-morbid pain and psychopathology in males and females admitted to treatment for opioid analgesic abuse. Pain. 2008;139(1):127–135. doi: 10.1016/j.pain.2008.03.021. [DOI] [PubMed] [Google Scholar]
  8. Cicero TJ, Inciardi JA, Muñoz A. Trends in abuse of OxyContin® and other opioid analgesics in the United States: 2002-2004. The Journal of Pain. 2005;6(10):662–672. doi: 10.1016/j.jpain.2005.05.004. [DOI] [PubMed] [Google Scholar]
  9. Dhingra SS, Strine TW, Holt JB, Berry JT, Mokdad AH. Rural-urban variations in psychological distress: findings from the Behavioral Risk Factor Surveillance System, 2007. International Journal of Public Health. 2009;54(1):16–22. doi: 10.1007/s00038-009-0002-5. [DOI] [PubMed] [Google Scholar]
  10. Dew B, Elifson K, Dozier M. Social and environmental factors and their influence on drug use vulnerability and resiliency in rural populations. The Journal of Rural Health. 2007;23(s1):16–21. doi: 10.1111/j.1748-0361.2007.00119.x. [DOI] [PubMed] [Google Scholar]
  11. Dowling K, Storr CL, Chilcoat HD. Potential influences on initiation and persistence of extramedical prescription pain reliever use in the US population. The Clinical Journal of Pain. 2006;22(9):776–783. doi: 10.1097/01.ajp.0000210926.41406.2c. [DOI] [PubMed] [Google Scholar]
  12. Draus PJ, Carlson RG. Needles in the haystacks: The social context of initiation to heroin injection in rural Ohio. Substance Use & Misuse. 2006;41(8):1111–1124. doi: 10.1080/10826080500411577. [DOI] [PubMed] [Google Scholar]
  13. Ellison CG, Sherkat DE. The “semi-involuntary institution” revisited: Regional variations in church participation among black Americans. Social Forces. 1995;73(4):1415–1437. [Google Scholar]
  14. [August 2014];Florida Office of Attorney General: Florida's Prescription Drug Diversion and Abuse Roadmap 2012-2015. 2012 from: http://myfloridalegal.com/ webfiles.nsf/wf/kgrg-8t8l5k/$file/Prescriptiondrugdiversionandabuseroadmap.pdf.
  15. Gfroerer JC, Larson SL, Colliver JD. Drug use patterns and trends in rural communities. The Journal of Rural Health. 2007;23(s1):10–15. doi: 10.1111/j.1748-0361.2007.00118.x. [DOI] [PubMed] [Google Scholar]
  16. Haggerty JL, Roberge D, Lévesque JF, Gauthier J, Loignon C. An exploration of rural–urban differences in healthcare-seeking trajectories: Implications for measures of accessibility. Health & Place. 2014;28:92–98. doi: 10.1016/j.healthplace.2014.03.005. [DOI] [PubMed] [Google Scholar]
  17. Haggerty JL, Roberge D, Pineault R, Larouche D, Touati N. Features of primary healthcare clinics associated with patients' utilization of emergency rooms: urban–rural differences. Healthcare Policy. 2007;3(2):72. [PMC free article] [PubMed] [Google Scholar]
  18. Harrell ZA, Broman CL. Racial/ethnic differences in correlates of prescription drug misuse among young adults. Drug and Alcohol Dependence. 2009;104(3):268–271. doi: 10.1016/j.drugalcdep.2009.05.017. [DOI] [PubMed] [Google Scholar]
  19. Havens JR, Young AM, Havens CE. Nonmedical prescription drug use in a nationally representative sample of adolescents: Evidence of greater use among rural adolescents. Archives of Pediatrics & Adolescent Medicine. 2011;165(3):250–255. doi: 10.1001/archpediatrics.2010.217. [DOI] [PubMed] [Google Scholar]
  20. Havens JR, Oser CB, Leukefeld CG, Webster JM, Martin SS, O'Connell DJ, Inciardi JA. Differences in prevalence of prescription opiate misuse among rural and urban probationers. The American Journal of Drug and Alcohol Abuse. 2007;33(2):309–317. doi: 10.1080/00952990601175078. [DOI] [PubMed] [Google Scholar]
  21. Hays LR. A profile of OxyContin addiction. Journal of Addictive Diseases. 2004;23(4):1–9. doi: 10.1300/J069v23n04_01. [DOI] [PubMed] [Google Scholar]
  22. Herman-Stahl MA, Krebs CP, Kroutil LA, Heller DC. Risk and protective factors for methamphetamine use and nonmedical use of prescription stimulants among young adults aged 18 to 25. Addictive Behaviors. 2007;32(5):1003–1015. doi: 10.1016/j.addbeh.2006.07.010. [DOI] [PubMed] [Google Scholar]
  23. Inciardi JA, Goode JL. OxyContin: Miracle medicine or problem drug? In: Inciardi JA, McElrath K, editors. The American drug scene: An anthology. 4th ed. Roxbury; Los Angeles: 2007. pp. 163–173. [Google Scholar]
  24. James KE, Wagner FA, Anthony JC. Regional varation in drug purchase opportunity among youths in the United States, 1996–1997. Journal of Urban Health. 2002;79(1):104–112. doi: 10.1093/jurban/79.1.104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Karlson KB, Holm A, Breen R. Comparing Regression Coefficients Between Same-sample Nested Models Using Logit and Probit A New Method. Sociological Methodology. 2012;42(1):286–313. [Google Scholar]
  26. Keith TZ. Multiple regression and beyond. Allyn and Bacon; Boston, MA: 2006. [Google Scholar]
  27. Kessler RC, Barker PR, Colpe LJ, Epstein JF, Gfroerer JC, Hiripi E, Zaslavsky AM. Screening for serious mental illness in the general population. Archives of General Psychiatry. 2003;60(2):184–189. doi: 10.1001/archpsyc.60.2.184. [DOI] [PubMed] [Google Scholar]
  28. Kurtz SP, Inciardi JA, Surratt HL, Cottler L. Prescription drug abuse among ecstasy users in Miami. Journal of Addictive Diseases. 2006;24(4):1–16. doi: 10.1300/j069v24n04_01. [DOI] [PubMed] [Google Scholar]
  29. Lankenau SE, Schrager SM, Silva K, Kecojevic A, Bloom JJ, Wong C, Iverson E. Misuse of prescription and illicit drugs among high-risk young adults in Los Angeles and New York. Journal of Public Health Research. 2012;1(1):22–30. doi: 10.4081/jphr.2012.e6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Logan TK, Walker R, Leukefeld Rural, urban influences, and urban differences among domestic violence arrestees. Journal of Interpersonal Violence. 2001;16(3):266–283. [Google Scholar]
  31. McCabe SE, West BT, Morales M, Cranford JA, Boyd CJ. Does early onset of non-medical use of prescription drugs predict subsequent prescription drug abuse and dependence? Results from a national study. Addiction. 2007;102(12):1920–1930. doi: 10.1111/j.1360-0443.2007.02015.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. McCauley JL, Danielson CK, Amstadter AB, Ruggiero KJ, Resnick HS, Hanson RF, Kilpatrick DG. The role of traumatic event history in non medical use of prescription drugs among a nationally representative sample of US adolescents. Journal of Child Psychology and Psychiatry. 2010;51(1):84–93. doi: 10.1111/j.1469-7610.2009.02134.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. McGranahan DA. How People Make A Living in Rural America. In: Brown DL, Swanson LE, editors. Challenges for Rural America in the Twenty-first Century. Penn State University Press; University Park, PA: 2003. pp. 135–51. [Google Scholar]
  34. Monnat SM, Beeler Pickett C. Rural/urban differences in self-rated health: Examining the roles of county size and metropolitan adjacency. Health & Place. 2011;17(1):311–319. doi: 10.1016/j.healthplace.2010.11.008. [DOI] [PubMed] [Google Scholar]
  35. NYC Mayor‘s Task Force on Prescription Painkiller Abuse. 2013. [August, 2014]. from: http://www.nyc.gov/html/cjc/downloads/pdf/rx_stat_september_2013_report.pdf. [Google Scholar]
  36. Olson DE, Weisheit RA, Ellsworth T. Getting down to business: A comparison of rural and urban probationers, probation sentences, and probation outcomes. Journal of Contemporary Criminal Justice. 2001;17(1):4–18. [Google Scholar]
  37. Palamar JJ, Kiang MV, Halkitis PN. Religiosity and exposure to users in explaining illicit drug use among emerging adults. Journal of Religion and Health. 2014;53(3):658–674. doi: 10.1007/s10943-012-9660-3. [DOI] [PubMed] [Google Scholar]
  38. Rigg KK, DeCamp W. Explaining prescription opioid misuse among veterans: A theory-based analysis using structural equation modeling. Military Behavioral Health. 2014;2(2):210–216. [Google Scholar]
  39. Rigg KK, Murphy JW. Understanding the etiology of prescription opioid abuse: Implications for prevention and treatment. Qualitative Health Research. 2013;23(7):963–975. doi: 10.1177/1049732313488837. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Rigg KK, Kurtz SP, Surratt HL. Patterns of prescription medication diversion among drug dealers. Drugs: Education, Prevention and Policy. 2012;19(2):145–155. doi: 10.3109/09687637.2011.631197. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Rigg KK, Ibañez GE. Motivations for non-medical prescription drug use: A mixed methods analysis. Journal of Substance Abuse Treatment. 2010;39:236–247. doi: 10.1016/j.jsat.2010.06.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Rigg KK, March SJ, Inciardi JA. Prescription drug abuse & diversion: role of the pain clinic. Journal of Drug Issues. 2010;40(3):681. doi: 10.1177/002204261004000307. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Roscigno VJ, Crowle ML. Rurality, Institutional Disadvantage, and Achievement/Attainment. Rural Sociology. 2001;66(2):268–292. [Google Scholar]
  44. Rosenblum A, Parrino M, Schnoll SH, Fong C, Maxwell C, Cleland CM, Haddox JD. Prescription opioid abuse among enrollees into methadone maintenance treatment. Drug and Alcohol Dependence. 2007;90(1):64–71. doi: 10.1016/j.drugalcdep.2007.02.012. [DOI] [PubMed] [Google Scholar]
  45. Sigmon SC. Characterizing the emerging population of prescription opioid abusers. The American Journal on Addictions. 2006;15(3):208–212. doi: 10.1080/10550490600625624. [DOI] [PubMed] [Google Scholar]
  46. Spiller H, Lorenz DJ, Bailey EJ, Dart RC. Epidemiological trends in abuse and misuse of prescription opioids. Journal of Addictive Diseases. 2009;28(2):130–136. doi: 10.1080/10550880902772431. [DOI] [PubMed] [Google Scholar]
  47. Substance Abuse and Mental Health Services Administration [August, 2014];Results from the 2012 National Survey on Drug Use and Health: Summary of National Findings. 2012 http://www.samhsa.gov/data/NSDUH/2012SummNatFindDetTables/NationalFindings/NSDUHresults2012.pdf.
  48. Substance Abuse and Mental Health Services Administration [August, 2014];Treatment episode data set report: Substance abuse treatment admissions involving abuse of pain relievers. 2010 :1998–2008. htttp://www.oas.samhsa.gov/2k10/230/230PainRelvr2k10.htm.
  49. United States Dept. of Health and Human Services . Substance Abuse and Mental Health Services Administration. Center for Behavioral Health Statistics and Quality. National Survey on Drug Use and Health. Codebook; Bethesda, MD.: 2011. [Google Scholar]
  50. Wang KH, Becker WC, Fiellin DA. Prevalence and correlates for nonmedical use of prescription opioids among urban and rural residents. Drug and Alcohol Dependence. 2013;127(1):156–162. doi: 10.1016/j.drugalcdep.2012.06.027. [DOI] [PubMed] [Google Scholar]
  51. Watkins WC, Ford JA. Prescription drug misuse among Asian-American adults: results from a National Survey. Substance Use & Misuse. 2011;46(13):1700–1708. doi: 10.3109/10826084.2011.605415. [DOI] [PubMed] [Google Scholar]
  52. Wu LT, Blazer DG, Swartz MS, Burchett B, Brady KT. Illicit and nonmedical drug use among Asian Americans, Native Hawaiians/Pacific Islanders, and mixed-race individuals. Drug and Alcohol Dependence. 2013;133(2):360–367. doi: 10.1016/j.drugalcdep.2013.06.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Wunsch MJ, Nakamoto K, Behonick G, Massello W. Opioid deaths in rural Virginia: a description of the high prevalence of accidental fatalities involving prescribed medications. The American Journal on Addictions. 2009;18(1):5–14. doi: 10.1080/10550490802544938. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Young AM, Havens JR, Leukefeld CG. A comparison of rural and urban Nonmedical prescription opioid users' lifetime and recent drug use. The American Journal of Drug and Alcohol Abuse. 2012;38(3):220–227. doi: 10.3109/00952990.2011.643971. [DOI] [PubMed] [Google Scholar]

RESOURCES