Americans increasingly believe cannabis to be a harmless substance with therapeutic benefits; cannabis is increasingly legal in the U.S.; a burgeoning industry has introduced novel, high potency products; and the prevalence of use and cannabis use disorder (CUD) has increased. The scientific and public health communities have struggled to keep pace in determining the impact of this changing cannabis landscape. An important barrier to progress is the lack of an adequate measure of cannabis consumption. Unlike measures of alcohol use that can discriminate high and low risk drinking patterns, current cannabis measures are inadequate to investigate cannabis risks and benefits, the impact of policy changes, and the outcomes of clinical trials. Developing valid measures faces challenges, e.g., estimation of quantity; data collection from sufficiently large and diverse samples for validation. Leveraging our expertise in cannabis research, measure development, and social media survey methods, we propose to address these challenges via a rigorous mixed-methods study to develop and test a family of measures of cannabis exposure: The Cannabis Exposure Inventory (CEI; timeframe, past 30 days), a short form (CEI-S), and a daily form (CEI-D; timeframe, last 24 hours). Aim 1: Prepare initial CEI. Our expert team will assemble and program an initial version of the CEI using novel items and images to estimate use. Through cognitive interviewing and a preliminary test-retest study, we will examine how users understand the items and response categories, and iteratively adjust the CEI. Aim 2: Initial examination of CEI validity. We will administer an on-line survey with the CEI, validators (e.g., CUD severity) and covariates to 3,000 cannabis users using well-tested social media survey methods and research panels. Analyses will assess associations between the different exposure item domains (construct validity) and identify the combination of items most associated with external validators (convergent validity) to inform further refinement of the CEI. A definitive test- retest substudy (n=600) will indicate reliability. Aim 3: Confirm CEI validity in a large sample of current users; develop the CEI-S. We will administer the CEI to 12,000 users to confirm construct and convergent validity, overall and across major subgroups (e.g., gender, race/ethnicity). We will derive the CEI-S for use in studies where time does not permit the full CEI. We will conduct a biological validation substudy with n=150 participants, examining THCCOOH. Aim 4: Prepare and validate the CEI-D. We will create the CEI-D, adjusting the CEI-S timeframe to the prior 24 hours for use as a daily measure, and examine its validity in a subsample from Aim 3 (n=400). They will complete the CEI-S, then complete the CEI-D, functioning, and mood items for 30 days on a mobile device. Consistent with FDA guidelines on demonstrating validity of outcome measures, we will test how CEI-D scores, functioning and mood covary over time. Accomplishing these aims will provide the field with a set of greatly improved measures of cannabis use that will enhance clinical and epidemiologic research, and lead to more informed communication about cannabis to clinicians, health educators and policy makers.

Project Narrative The scientific community has struggled to keep pace with the impact of the changing landscape of cannabis use and policy in the U.S. An important barrier to progress is the lack of adequate measures of cannabis consumption that can be used to identify low vs. high risk use patterns. We propose to address this challenge via a rigorous cannabis measure development / validation project that will enhance the capacities of clinical, epidemiologic and policy studies and lead to more informed communication about cannabis and its risks to clinicians, health educators and policy makers.