Citation

Romano EO, Moore C, Kelley-Baker T, Torres-Saavedra PA. Traffic Injury Prev. 2019; 20(7): 667-672.

Affiliation

Department of Mathematical Sciences, University of Puerto Rico at Mayagüez , Mayagüez , Puerto Rico.

Copyright

(Copyright © 2019, Informa - Taylor and Francis Group)

DOI

10.1080/15389588.2019.1635690

PMID

31356118

Abstract

Objective: Blood and/or urine are typical drug detection matrices used by law enforcement. There are some concerns about using oral fluid (OF) in the identification of drivers potentially impaired by cannabis, particularly regarding their accuracy when compared to blood. The study objectives were to (1) examine the accuracy of predicting delta 9-tetrahydrocannabinol (THC) in blood from THC measured in OF and (2) examine factors influencing prediction accuracy. Methods: Using data from the 2007 and 2013-2014 National Roadside Survey (NRS) of Alcohol and Drug Use, 7,517 drivers with known laboratory results in both OF and blood were included in this study. OF samples were collected using the Quantisal® device and analyzed at the same private laboratory in both the 2007 and 2013-2014 NRS. The Quantisal device has consistently shown to collect 1 mL ±10%. Descriptive statistical analyses were used to examine and compare the distribution of THC concentrations in OF and blood. A hurdle model was applied to examine factors influencing the accuracy of the THC_blood predictions based on THC_OF while accounting for the decisions of cannabis consumption. We estimated the number of true positives (TPs), false positives (FPs), true negatives (TNs), false negatives (FNs), sensitivity, specificity, and positive predicted value (PPV). Results: This study found that THC measured in OF (THC_OF) is a good predictor of THC measured in blood (THC_blood), in particular when THC_OF > 0 ng/mL is used to predict being positive for THC_blood (THC_blood > 0 ng/mL). However, as blood and OF concentrations depart from 0 ng/mL, the proportion of TPs (sensitivity) decreases, which might be a concern for law enforcement. The likelihood of accurately predicting THC_blood from THC_OF is lower for drivers who were simultaneously using cannabis and other drugs. Conclusions: The findings of this study are based on THC measures obtained in a laboratory, which may not be the same as those conducted by police using point-of-care devices. However, this study is unique due to its large sample of drivers obtained in similar roadside locations and times to actual law enforcement activities. Though a positive THC_OF may assist law enforcement in probable cause for a blood draw, efforts to develop reliable methods to detect drug impairment based on OF should continue.


Keyword: Cannabis impaired driving

Language: en

Keywords

Cannabis; THC; blood; oral fluid