Cannabis Use for Chronic Pain: Advancing Knowledge and Cannabis Use Disorder Assessment with a Mixed Methods Approach

National Institute on Drug Abuse (NIDA), R01DA050032

3/15/26 – 1/31/31

Jacob T. Borodovsky (Geisel School of Medicine at Dartmouth); Deborah S. Hasin (Columbia University / New York State Psychiatric Institute)

Co-Investigators: Efrat Aharonovich (Columbia University / New York State Psychiatric Institute); Alan J. Budney (Geisel School of Medicine at Dartmouth); Ofir Livne (Columbia University / New York State Psychiatric Institute); Cara A. Struble (University of Maine); Melanie M. Wall (Columbia University / New York State Psychiatric Institute)

Consultants:Tammy Chung (Rutgers University); Jodi Gilman (Massachusetts General Hospital / Harvard Medical School)

U.S. adult rates of cannabis use and DSM-5 Cannabis Use Disorder (CUD) have now reached historic peaks, medical cannabis use for pain is legal in 40 states, and the FDA has recommended re-classifying cannabis from Schedule 1 (no medical use, high abuse potential) to Schedule 3 (use for pain, moderate-low dependence potential). These changes have prompted calls to re-examine and improve the DSM-5-TR CUD criteria, which have not been updated in over 10 years.

Opioid use disorder (OUD) is a useful reference for re-examining CUD because pain is the most common medical reason for using opioids or cannabis. Earlier, we showed excellent reliability and validity of 3 conceptual models of prescription opioid use disorder (OUD-P), with strongest validity evidence for the ‘fully-adjusted’ model that did not count DSM-5 criteria as positive if they occurred only when opioids were used as prescribed (under medical supervision; i.e., therapeutic use). This framework is our starting point to study potential improvements in CUD diagnostic criteria. However, unlike opioids, cannabis is ‘authorized’, not prescribed; medical supervision is rare; and dosing seldom specified. Knowledge gaps about key medical cannabis behaviors (e.g., efforts to get medical supervision, how therapeutic doses are determined) impede creating guidelines to assess therapeutic cannabis use.

In R01DA050032, we used cognitive interviews and online surveys to produce the Cannabis Exposure Index (CEI), a new measure of mg/THC (i.e., cannabis dose). We propose a 5-year renewal to use these methods to compare reliability and validity of 3 CUD models. 1) Unadjusted (DSM-5-TR); 2) Physiologically-adjusted (tolerance, withdrawal criteria not counted towards a CUD diagnosis if they arose solely from therapeutic cannabis use) and 3) Therapeutically-adjusted (novel, fully- adjusted model: any CUD criteria occurring solely during therapeutic use not counted towards a diagnosis). To test the models, we must first fill knowledge gaps on medical cannabis supervision/dosing in order to create guidelines to assess therapeutic use. We will then adapt our diagnostic measure, PRISM-5-OP, into a cannabis version (PRISM-CUD) to test if physiologically- or therapeutically-adjusted CUD models are improvements over DSM-5-TR CUD. Given FDA recommendations, we focus on cannabis for pain, but also explore sleep and anxiety. All participants will have chronic pain and use cannabis daily/near-daily.

AIM 1. Fill knowledge gaps about medical cannabis (e.g., efforts to obtain medical supervision or dose information; self-efforts to determine right dose) via online survey (n=1,000). AIM 2. Use findings to create guidelines to differentiate therapeutic from other use for PRISM-CUD measures of the CUD models; pilot and conduct cognitive interviews; refine items. AIM 3. Using PRISM-CUD, compare convergent and discriminant validity of the 3 CUD diagnostic models in online sample (n=3,000), re-testing 400 to determine reliability.

Hypothesis: Validity in capturing true cases of CUD will be ordered as: Therapeutically-adjusted>Physiologically-adjusted>Unadjusted. We aim to update and improve validity of CUD diagnoses, thereby improving public health in many areas that require such diagnoses.

Valid, accurate diagnoses are critical for many purposes: indicating referrals to appropriate treatment; determining eligibility for clinical trials; making accurate prevalence estimates, training clinicians; and for improving DSM-5-TR and potentially DSM-6. Despite the rapidly changing cannabis landscape and increasing use of cannabis for pain, the DSM-5-TR CUD criteria have not been updated since 2013, and the changed legal status of cannabis has left gaps in knowledge about how it is used medically and how existing DSM-5 guidelines for diagnosing substance use disorders for prescribed substances should be adapted for cannabis, which cannot be prescribed but only ‘authorized’. This project will fill major gaps in knowledge about medical use; adapt DSM- 5 CUD criteria and diagnosis using this knowledge; and test the adapted criteria and diagnoses to determine if they improve reliability and validity over the existing DSM-5-TR CUD.