Assessing craving and its relationship to subsequent prescription opioid use among treatment-seeking prescription opioid dependent patients

Highlights

• Drug craving predicts future drug use, especially in the short term.

• A 3-item measure predicted prescription opioid use in the next week.

• For each 1 unit increase on the scale, the odds of using opioid was 17% higher.

• Brief assessment of prescription opioid craving may provide a marker of risk.

Abstract

Background

Craving is viewed as a core feature of substance use disorders and has been shown to predict future drug use, particularly over the short term. Accordingly, craving is often assessed in treatment settings as a marker of risk for subsequent drug use. The identification of the briefest measure that maintains predictive validity is of particular value for both clinical and research settings to minimize assessment burden while maintaining utility for the prediction of use.

Methods

Data from a multi-site clinical trial of treatment for prescription opioid dependence were examined to evaluate whether a brief, 3-item craving scale administered each week predicted urine-confirmed self report of prescription opioid use in the subsequent week. Logistic regression models examining the association between craving and presence or absence of opioid use in the following week were conducted, controlling for opioid use in the previous week, treatment condition, and lifetime history of heroin use.

Results

Greater craving was associated with a higher odds of prescription opioid use in the following week. For each one-unit increase on this 10-point scale, the odds of using opioids in the subsequent week was 17% higher. In addition to an item assessing urges, items assessing cue-induced craving and perceived likelihood of relapse in an environment where drugs were previously used contributed uniquely to this association.

Conclusions

A brief measure of prescription opioid craving predicted prescription opioid use among individuals in treatment. This measure offers an efficient strategy to inform the assessment of risk for use in this population.

Introduction

Craving is a core feature of substance use disorders, as evidenced by its recent addition to the diagnostic criteria for these disorders in the Diagnostic and Statistical Manual of Mental Disorders (DSM-5; American Psychiatric Association, 2013). Much of the research on craving has focused on whether it may precipitate substance use and thus serve as a risk marker during treatment. A number of studies have examined whether craving at treatment entry predicts outcomes following treatment, with some studies supporting this association (Anton et al., 1996, Sinha et al., 2011), some finding no association (Ahmadi et al., 2009, Dreifuss et al., 2013), and others finding associations only for particular types of craving (e.g., stress-induced craving; Sinha et al., 2006).

Although self-reported craving may have inconsistent predictive validity for long-term substance use disorder treatment outcomes, several studies have demonstrated that drug craving is associated with proximal (e.g., within the next week) substance use (Hartz et al., 2001, Moore et al., 2013, Weiss et al., 2003). Because craving is influenced by contextual factors such as substance-related cues, stress, and withdrawal symptoms, it may serve as a marker for immediate use, and associations with substance use may weaken over longer periods of time (Shiffman et al., 1997).

The evidence supporting the predictive validity of craving for proximal substance use highlights the value of regular assessment of craving during treatment. Once heightened risk has been identified, providers may choose to implement specific strategies to attempt to mitigate the risk of use in the near future. However, demands on time and resources in clinical settings necessitate that assessments be as brief as possible while still yielding meaningful information. In research settings, efficient assessment strategies have the dual benefit of minimizing participant burden and providing additional time for the assessment of other pertinent constructs. Self-report craving measures vary dramatically in length, ranging from a single item (Connor et al., 2005, West and Ussher, 2010) to more than 40 items (Heishman et al., 2001, Tiffany et al., 1993). Although lengthy, multi-dimensional assessments may be of particular value in certain research settings, the briefest possible measure that still indicates level of risk is preferable clinically and in research settings in which assessment burden is a concern, or in which repeated measures are collected. A previous multi-site study of treatment for cocaine dependence demonstrated that a brief, 3-item craving measure developed by our group predicted cocaine use in the following week, when controlling for current cocaine use (Weiss et al., 2003). It is unclear whether the predictive validity of this measure extends to other substances of abuse.

As the prevalence of prescription opioid dependence has rapidly escalated in the past 15 years (Johnston et al., 2012, Substance Abuse and Mental Health Services Administration, 2012), studies have begun to examine craving in this population. However, the association between opioid craving and proximal opioid use has yet to be examined. A study in chronic pain patients receiving opioids for pain found that self-reported craving for medication was associated with aberrant drug behaviors (e.g., positive urine toxicology) assessed 6 months later (Wasan et al., 2009). Although this study included participants both with and without prescription opioid use disorders, these results suggest that craving for prescription opioids is a risk marker for later opioid misuse, at least among those with chronic pain.

Another reason to examine craving specifically in patients dependent on prescription opioids is the possibility that craving in this population differs in some ways from craving for illicit drugs. A recent report from our group found that craving in response to opioid cues was less robust in this population than in those dependent upon heroin (McHugh et al., 2014). One potential explanation for this difference is that the associative conditioning of prescription drug cues may differ from illicit drug cues because the pairing of drug use and contextual cues is less unique for prescription drug cues, which are encountered in numerous settings (e.g., pharmacies, advertisements). Thus, research examining the role of cue-induced craving in this population is of particular interest.

The aim of the current analysis of data from a multi-site study of prescription opioid dependence treatment was to examine whether craving for prescription opioids predicted use of these drugs in the following week. A brief, 3-item craving measure was used to assess three different domains of craving: general craving, cue-induced craving, and likelihood of use if exposed to an environment in which drugs were previously used. In this study, we examined the concurrent and predictive validity of this measure in a prescription opioid dependent population. We hypothesized that greater craving would be associated with a higher likelihood of opioid use in the following week, when controlling for current use. In an exploratory analysis, the associations between the individual craving scale items and subsequent use were examined to evaluate the relative importance of these three domains of craving. This study is novel in its examination of the relationship between craving and prescription opioid use over brief time intervals, its extension of a validated measure to a new population, and its consideration of the incremental importance of various domains of craving.