In the primary care practice of an urban safety-net hospital, 18.4% of patients with chronic pain met criteria for a PDUD and an additional 24.5% had other SUDs. The vast majority of participants with PDUD had at least one co-occurring other SUD, and those with PDUD were virtually indistinguishable from those with other SUDs. Seven patient characteristics were independently associated with PDUD compared to those without any lifetime SUD: time spent in jail, high degree of pain-related limitations, current smoking, family history of SUD, white race, male gender, and PTSD. The only difference between SUD and PDUD was the higher prevalence of white race in PDUD.
The finding that PDUD is highly associated with other SUDs corroborates prior clinical studies which relied on proxy measures for the determination of PDUD. 15; 43; 38; 51; 26; 17
Ives et al. found that prior cocaine or alcohol abuse predicted prescription opioid misuse among patients at a pain management clinic.26
Studies which included smoking status as a correlate for PDUD found a striking association, as did our study. This reflects the exceptionally high co-morbidity of smoking with alcohol and illicit drug dependence found in other studies.46; 38
For physicians who prescribe opioid and sedative medications, a critical concern is differentiating individuals who take medications for the intended therapeutic purposes from those with current, or potentially future, misuse and addiction. This cross-sectional study does not permit a longitudinal analysis of persons prescribed opioid analgesics. However, it does suggest that a propensity toward addiction (other SUD, family history of SUD, cigarette smoking) is a strong correlate of prescription drug use disorder. Fleming and colleagues found that positive urine toxicology screening for cocaine or marijuana and aberrant drug behaviors were among the significant predictors of SUD in primary care patients receiving opioids.19
Data in our study suggest that those with PDUD have similar characteristics to those addicted to illicit drugs and alcohol. The 2005 report from the Center on Addiction and Substance Abuse at Columbia University corroborates this association in a community sample: 75% of persons who misuse prescription drugs have at least one co-occurring other SUD.11
Of note, a physician may consider screening patients with pain and other risk factors for prescription drug use disorder even in the absence of prescribing controlled substances as most participants with PDUD in this study obtained the medication from sources other than a treating clinician.
A strong association of both PDUD and other SUDs was having spent time in jail. Almost two-thirds of those with PDUD reported having spent time in jail, compared to 15.3% of those with no SUD. The relationship between criminal activity and prescription drug abuse has been suggested in prior studies.11; 26
Akbik and colleagues reported prior legal problems predict subsequent opioid misuse among patients starting opioids for chronic pain.3
It is not known whether the jail history was due to crimes related to drug use, possession, manufacture or sale, which would suggest a history of SUD. It also may be a proxy for anti-social behavior, which is associated with PDUD.24
The associations of white race and male gender with PDUD in this sample reflect findings in other clinical and population samples.13; 24; 8
Whites are prescribed more opioid medications in Emergency Departments and primary care practices, perhaps reflecting a cultural bias by patients and physicians toward use of prescription opioids.12; 41
Male gender predominance reflects epidemiology of SUDs in general, although some data suggest more gender balance in PDUD. These trends suggest that future research should explore the social context, including gender and racial differences, of these associations.
Patients with PDUD reported a greater degree of pain-related limitation. Others have found low pain intolerance among those with active14
and past addictions32
compared to non-addicted controls. It may be that lower pain threshold is an increased risk for developing addiction, or that addiction itself lowers the pain tolerance. This may complicate pain management among those with PDUD.
Among this sample of urban primary care patients with chronic pain, PTSD was associated with PDUD. PTSD is known to be associated with SUD in clinical and community samples,29; 9; 27; 33
but the relationship with PDUD has not been described. The scientific evidence for neurological and physiological changes in PTSD,25; 48
pain 5; 31; 6
and substance use disorders22
is growing. Exploring these overlapping phenomena may allow development of tailored interventions.
Intimate partner violence and depression appeared to be individually associated with SUD but not PDUD, when accounting for the interaction between these phenomena. These associations may function differently in gender specific analyses which should be explored in future studies. The fact that Adverse Childhood Experiences were not independently associated with either PDUD or SUD after controlling for other variables suggests that its effect is mediated through other variables such as PTSD which remained significant in the final model.
These data strongly suggest that physicians treating patients with pain should assess for SUD. This can help direct care, including treatment for pain and substance use disorders. Specialty pain practices who commonly prescribe opioid analgesics are likely to screen for this, but primary care settings may not be as aware of the overlap between pain and addictions. Furthermore, patients do not always admit to SUD, particularly if they are intent on deceiving the treating physician to obtain prescription medication. Potential screening questions for patients with chronic pain could include assessment of smoking and specific questions used in this study: “Do you have a family history of alcohol or drug problems?”, “Have you ever spent time in jail?” Evaluations for pain disability and PTSD may be additional clinical tools to help identify those at highest risk for PDUD.
After identifying a patient with risk factors, should clinicians prescribe opioid medications? In an observational study, Wiedemer and colleagues examined the impact of a structured opioid clinic for patients with risk factors, including psychiatric and substance use problems.57
All patients with SUD but no aberrant behaviors were safely maintained on opioids.57
Clinical trials testing methods of opioid medication monitoring could inform clinicians about how to safely prescribe them to high risk patients.
This study adds to the literature in three ways. First, it examines an urban, largely poor and minority sample, which is underrepresented in the literature. Secondly, validated measures of PDUD used in this study improve upon the use of proxy measures that most other studies have employed. Finally, subjects were primary care patients with chronic pain, not limited to those being prescribed chronic opioid medication. This can illuminate issues about patients who may require opioid analgesics in the future, common clinical concerns of urban primary care pain patients.
The study limitations include possible misclassification of PDUD in individuals with pseudo-addiction, i.e. behaviors that resemble addiction but result from inadequate treatment of pain.15
We believe that this is not a significant limitation as the diagnostic criteria demand social or physical problems and compulsive use, which are not characteristic of pseudo-addiction. Furthermore, 71% had addictions to medications that were not prescribed for them, which lowers the probability of misclassification. Another limitation was that lifetime alcohol use disorders were not measured which should attenuate any associations found because of misclassification bias. Since numerous independent predictors of PDUD and SUD were found, it is not clear how information on lifetime alcohol disorders would change the associations. The cross-sectional design limits conclusions regarding cause and effect; the findings would be strengthened by studying a longitudinal cohort. The recruitment strategy may limit the generalizability of the findings. However, as this was a study of PDUD risk factors and not prevalence, the associations should remain stable in a similar sample. Another limitation is that no corroborating evidence of prescription drug misuse was obtained, such as urine toxicology screening. Such testing can be helpful as a supplement to self-report.
In an urban cohort of primary care patients with high levels of pain disability, unemployment and psychosocial stressors, PDUD was concentrated among those with a: family history of SUDs, having spent time in jail, current cigarette smoking, male gender, white race, pain-related functional limitations and PTSD. The vast majority had co-occurring other SUDs. This suggests that clinicians could gain clinical insight by carefully evaluating such patients for these risk factors when developing a comprehensive pain management strategy. It may also suggest which patients would benefit from a structured program for use of opioid medications. Refining the knowledge base on co-occurrence of addiction and pain could maximize safe and effective pain relief strategies.