ABSTRACT
BACKGROUND: Cannabis laboratory testing reliability is a scientific and policy challenge in US states with legal cannabis. Greater reported THC concentration yields higher prices, and media reports describe a well-known consumer and dispensary preference for flower products containing a minimum 20% THC content-an economically meaningful but biologically arbitrary threshold. This paper examines the frequency distribution of reported THC concentration in legal cannabis flower products in Nevada and Washington state for unusual shifts around the 20% threshold suggestive of potential manipulation of reported THC results. METHODS: THC concentration test results for 142,000 Chemotype 1 flower products from Washington state between June 2014 and May 2017 and 55,000 flower products from Nevada between December 2017 and January 2020 were analyzed for changes in the frequency distribution around the 20% threshold using the McCrary density test. Analyses were performed among all labs in each state, the highest volume lab in Washington, and two labs in Washington which had their licenses suspended for testing irregularities during the study period. RESULTS: Comparing just above the 20% THC threshold with just below it, the frequency of test results increased by about 43% in Nevada (z = 15.6, p < 0.001) and by about 17% in Washington (z = 11.0, p < 0.001). In Washington's highest volume testing lab, frequency increased by only about 1% (z = 0.39, p = 0.70), while it increased by about 47% (z = 12.7, p < 0.001) among the two suspended labs. Subset to those growers which sent products to both sets of labs, frequency of flower products just above the 20% threshold increased by 2% in Washington's largest lab (z = 0.50, p = 0.62) and by 52% among the two suspended labs (z = 12.8, p < 0.001). DISCUSSION: There is a statistically unusual spike in the frequency of products reporting just higher than 20% THC in both states consistent with economic incentives for products to contain at least 20% THC. This "bunching" of reported THC levels exists among some, but not all, cannabis testing labs, suggesting that laboratory differences (rather than precise manipulation by growers) drive this potential manipulation in reported THC content. These findings elaborate on prior research highlighting unexplained interlaboratory variation in cannabis testing results and highlight ongoing irregularities with legal cannabis testing. CONCLUSION: These findings highlight the need for industry oversight and cautions researchers working with reported cannabis THC concentration data, which may be biased by economic incentives to report higher THC.
ABSTRACT
BACKGROUND: Drug overdose deaths have risen precipitously over the last fifteen years. Substantial geographic variation, beyond a simple rural-urban dichotomy, exists in the concentration of overdose deaths, suggesting the existence of as-yet unidentified environmental variables that predict resilience (or vulnerability) to drug overdoses. Motivated by reports highlighting the role of community fragility in the opioid epidemic, we explore whether social capital attenuates overdose death rates. METHODS: We conducted an ecologic temporal trends study from 1999 to 2014 to investigate the association between mortality due to drug overdose and social capital. Data from multiple sources were compiled at the county-level to produce an analytic dataset comprising overdose mortality, social capital, and a host of potentially confounding variables indicated by the literature (N=49,664 county-years). Multinomial logistic regression was used to estimate the likelihood that a county falls in low (<4 deaths per 100,000), moderate, or high (>16 deaths per 100,000) categories of annual overdose morality. RESULTS: We observed a strong and statistically significant inverse association between county-level social capital and age-adjusted mortality due to drug overdose (p<0.01). Compared to the lowest quintile of social capital, counties at the highest quintile were 83% less likely to fall in the "high-overdose" category and 75% less likely to fall in the "moderate-overdose" category. CONCLUSION: This study finds large-sample evidence that social capital protects communities against drug overdose. This finding could help guide policymakers in identifying where overdose epidemics are likely to occur and how to ameliorate them.
