Random and systematic errors in case-control studies calculating the injury risk of driving under the influence of psychoactive substances

Alcohol and driving-How bad is the combination? A meta-analysis

OBJECTIVE

The objective of this study was to summarize existing knowledge on the association between driving under the influence of alcohol and road safety outcomes and to assess factors that may explain why these estimates vary.

METHODS

Based on studies of the association between blood alcohol concentration (BAC) levels and crashes, we used multilevel metaregression to calculate summary effects of BAC level and analyze potential moderators of the association.

RESULTS

Based on 60 studies and 393 effect estimates, we found that BAC level, outcome severity, use of hospital data, and geographical region contributed to heterogeneity in results.

CONCLUSION

The effect of BAC on crash and injury risk and culpability is stronger at higher BAC levels and for more serious outcomes. The relationship between BAC level and outcome is approximately exponential. The relationship is stronger in studies from Nordic countries than in studies from other countries, possibly because of the relatively low level of drunk driving in these countries. Studies based on hospital data and studies using non-crash-involved control groups find smaller effects on average.

The detection of cannabinoids in veterinary feeds by microNIR/chemometrics: a new analytical platform

In this work, the capabilities of a novel miniaturized and portable microNIR spectrometer were investigated in order to propose a practical and intelligible test allowing the rapid and easy screening of cannabinoids in veterinary feeds. In order to develop a predictive model that could identify and simultaneously quantify the residual amounts of cannabinoids, specimens from popular veterinary feeds were considered and spiked with increasing amounts of cannabidiol (CBD), Δ9-tetrahydrocannabinol (THC), and cannabigerol (CBG). Partial least squares discriminant analysis (PLS-DA) and partial least squares regression (PLSr) were applied for the simultaneous detection and quantification of cannabinoids. The results demonstrated that the microNIR/chemometric platform could statistically identify the presence of CBD, THC and CBG in the simulated samples containing cannabinoids from 0.001 to 0.01%w/w, with the accuracy and sensitivity of the official reference methods actually proposed. The method was checked against false positive and true positive responses, and the results proved to be those required for confirmatory analyses, permitting to provide a fast and accurate method for monitoring cannabinoids in veterinary feeds.

Real time detection of amphetamine in oral fluids by MicroNIR/Chemometrics

In this work, a novel coupled approach MicroNIR/Chemometrics based on a miniaturized and portable spectrometer is proposed for the on site detection of amphetamines (AMP) in non pretreated oral fluids. In particular, the coupling of MicroNIR with chemometrics was investigated with the aim of developing a fast and accurate approach able to perform the on-site prediction of AMP abuse. A predictive model to be used in real cases was developed by collecting specimens from volunteers and spiked samples with increasing amounts of AMP were prepared to optimize calibration. Partial Least Square-Discriminant Analysis (PLS-DA) and Partial Least Square regression (PLS) were involved to detect and quantify AMP. Results demonstrated that MicroNIR/Chemometric platform is statistically able to identify AMP abuse in simulated oral fluid samples containing, with the accuracy and sensitivity of the actual proposed official reference methods. The method was checked against false positive and true positive response and results proved to be those required for confirmatory analyses. This method would permit to simplify AMP abuse monitoring for roadside drug testing or workplace surveillance and may be of help at first aid points.

MicroNIR/Chemometrics: A new analytical platform for fast and accurate detection of Δ9-Tetrahydrocannabinol (THC) in oral fluids

BACKGROUND

Δ⁹-Tetrahydrocannabinol (THC) is already considered one of the most addictive substances since an increasing number of consumers/abusers of THC and THC based products are observed worldwide. In this work, the capabilities of a novel miniaturized and portable MicroNIR spectrometer were investigated in order to propose a practical and intelligible test allowing the rapid and easy screening of Δ⁹-Tetrahydrocannabinol (THC) oral fluids without any pretreatment.

METHODS

Specimens from volunteers were collected in order to consider any sources of variability in the spectral response and spiked with increasing amount of THC in order to realize predictive models to be used in real cases. Partial Least Square-Discriminant Analysis (PLS-DA) and Partial Least Square regression (PLSr) for the simultaneously detection and quantification of THC, were applied to baseline corrected spectra pre-treated by first derivative transform.

RESULTS

Results demonstrated that MicroNIR/Chemometric platform is statistically able to identify THC abuse in simulated oral fluid samples containing THC from 10 to 100 ng/ml, with a precision and a sensitivity of about 1.51% and 0.1% respectively.

CONCLUSIONS

The coupling MicroNIR/Chemometrics permits to simplify THC abuse monitoring for roadside drug testing or workplace surveillance and provides the rapid interpretation of results, as once the model is assessed, it can be used to process real samples in a "click-on" device.

"Lab-on-Click" Detection of Illicit Drugs in Oral Fluids by MicroNIR-Chemometrics

A novel, entirely automated MicroNIR-chemometric platform was developed for the "lab-on-click" detection of illicit drugs in nonpretreated oral fluids, and a novel tool for the first-level test is proposed. Calibration of the method was achieved by collecting oral-fluid specimens from volunteers, and chemometric analysis was considered for the development of models for prediction for cocaine, amphetamine, and Δ9-tetrahydrocannabinol. In addition, a comprehensive model was optimized for the simultaneous prediction of positive-negative samples and the specific illicit drug used by abusers in a single "click". The detection ability of the method was checked for true-positive and false-positive outcomes, and results were validated by a GC-MS reference official method. The MicroNIR-chemometric platform provided the simultaneous prediction of the three most frequently abused addictive drugs with the sensitivity and accuracy of the confirmatory analyses, offering the advantages of rapidity and simplicity and demonstrating that it is a promising tool for supporting public-health surveillance.

Challenges and common weaknesses in case-control studies on drug use and road traffic injury based on drug testing of biological samples

PURPOSE

To determine and discuss common weaknesses and errors in case-control studies on the association between drug use and road traffic crash injury among drivers and recommend improvements for future studies.

METHODS

A search for case-control studies published between 2000 and 2016 was performed using PubMed and other databases in addition to manual search. The used methodologies were compared with requirements and recommendations for case-control studies as well as current knowledge on the interpretation of drug concentrations in biological samples.

RESULTS

Seventeen studies were identified. The major difficulties in the studies were related to likely selection bias, information bias, and confounding. In some studies, the definition of drug exposure was different for controls than for cases, generating potentially serious errors in the odds ratio estimations. Other weaknesses include lacking explanation of the assessment of drug exposure, missing covariates, lacking description of statistical methods, and lack of discussion of bias and confounding.

CONCLUSIONS

Many of the observed challenges and weaknesses can be overcome or reduced. Recommendations for future studies are presented.

The effects of cannabis intoxication on motor vehicle collision revisited and revised

AIMS

To determine whether and to what extent acute cannabis intoxication increases motor vehicle crash risk.

DESIGN

Study 1 replicates two published meta-analyses, correcting for methodological shortcomings. Study 2 is an updated meta-analysis using 28 estimates from 21 observational studies. These included studies from three earlier reviews, supplemented by results from a structured search in Web of Science and Google Scholar, and by the personal libraries of the research team. Risk estimates were combined using random-effects models and meta-regression techniques.

SETTING

Study 1 replicates the analysis of Asbridge et al., based on nine studies from five countries, published 1982-2007; and Li et al., based on nine studies from six countries, published 2001-10. Study 2 involves studies from 13 countries published in the period 1982-2015.

PARTICIPANTS

In study 1, total counts extracted totalled 50 877 (27 967 cases, 22 910 controls) for Asbridge et al. and 93 229 (4236 cases and 88 993 controls) for Li et al. Study 2 used confounder-adjusted estimates where available (combined sample size of 222 511) and crude counts from the remainder (17 228 total counts), giving a combined sample count of 239 739.

MEASUREMENTS

Odds ratios (OR) were used from case-control studies and adjusted OR analogues from culpability studies. The impact of the substantial variation in confounder adjustment was explored in subsample analyses.

FINDINGS

Study 1 substantially revises previous risk estimates downwards, with both the originally reported point estimates lying outside the revised confidence interval. Revised estimates were similar to those of study 2, which found cannabis-impaired driving associated with a statistically significant risk increase of low-to-moderate magnitude [random-effects model OR 1.36 (1.15-1.61), meta-regression OR 1.22 (1.1-1.36)]. Subsample analyses found higher OR estimates for case-control studies, low study quality, limited control of confounders, medium-quality use data and not controlling for alcohol intoxication.

CONCLUSIONS

Acute cannabis intoxication is associated with a statistically significant increase in motor vehicle crash risk. The increase is of low to medium magnitude. Remaining selection effects in the studies used may limit causal interpretation of the pooled estimates.

Which drugs are associated with highest risk for being arrested for driving under the influence? A case-control study

The aim of this study was to determine the association between drug type and arrest for driving under the influence of drugs (DUID) by calculating odds ratios (ORs) using a case-control design. A DUID arrest is in most cases related to aberrant or risky driving and might therefore be regarded as a proxy for a drug related traffic crash. The 'cases' were 2738 drivers arrested on suspicion of drugged driving from April 2008 to March 2009 with blood alcohol concentrations below the legal limit of 0.2g/L; 794 were arrested due to involvement in road traffic crashes, whereas 1944 were arrested for other reasons, mainly dangerous driving, suspected impairment when stopped in traffic controls, or because of phone calls to the police from other road users or observers. The 'controls' were 9375 random drivers in normal traffic, also with alcohol concentrations below this limit. Blood samples from 'cases' and oral fluid samples from 'controls' were analyzed for 15 drugs that have legislative concentration limits in Norway, in addition to two other commonly detected psychoactive drugs. The most prevalent illicit drug in the control group was tetrahydrocannabinol (THC; 0.58%), which was also commonly found in samples from drivers arrested due to road crash (15.6%) or arrested for other reasons (21.8%). Amphetamine/methamphetamine was most prevalent among arrested drivers involved in crashes (30.6%) and drivers arrested for other reasons (56.9%), whereas only 0.18% of the control group was positive for amphetamine/methamphetamine. The single-use substances which gave highest OR for police arrest were amphetamine/methamphetamine, alprazolam, clonazepam and oxazepam. The majority of the alprazolam and clonazepam findings were probably due to non-therapeutic use of medicinal drugs purchased on the illegal market. Combinations of two or more drugs yielded higher ORs than the use of single substances; combinations of amphetamine/methamphetamine and benzodiazepines gave the highest risk.

Trends in alcohol and other drugs detected in fatally injured drivers in the United States, 1999-2010

Drugged driving is a safety issue of increasing public concern. Using data from the Fatality Analysis Reporting System for 1999-2010, we assessed trends in alcohol and other drugs detected in drivers who were killed within 1 hour of a motor vehicle crash in 6 US states (California, Hawaii, Illinois, New Hampshire, Rhode Island, and West Virginia) that routinely performed toxicological testing on drivers involved in such crashes. Of the 23,591 drivers studied, 39.7% tested positive for alcohol and 24.8% for other drugs. During the study period, the prevalence of positive results for nonalcohol drugs rose from 16.6% in 1999 to 28.3% in 2010 (Z = -10.19, P < 0.0001), whereas the prevalence of positive results for alcohol remained stable. The most commonly detected nonalcohol drug was cannabinol, the prevalence of which increased from 4.2% in 1999 to 12.2% in 2010 (Z = -13.63, P < 0.0001). The increase in the prevalence of nonalcohol drugs was observed in all age groups and both sexes. These results indicate that nonalcohol drugs, particularly marijuana, are increasingly detected in fatally injured drivers.

Commentary: why is the odds ratio for involvement in serious road traffic accident among drunk drivers in Norway and Finland higher than in other countries?

OBJECTIVE

Recent Norwegian and Finnish studies have found high odds ratios for serious or fatal injury in road traffic accidents among drivers after drinking alcohol. In this report we have compared the odds ratios with results from studies in other countries.

METHODS

A literature review was conducted.

RESULTS

The odds ratios were significantly higher than in countries where drunk driving is more common.

CONCLUSION

The calculated odds ratios are not only related to alcohol use or blood alcohol concentration per se but also related to the study design-for example, the inclusion of nonculpable drivers among cases-and confounding factors not included in statistical analysis; for example, risk-taking behavior. Those two issues may contribute to explaining why the reported odds ratios are higher for Norway and Finland.

Examining the impact of opioid analgesics on crash responsibility in truck drivers involved in fatal crashes

INTRODUCTION

Commercial motor vehicle (CMV) drivers, particularly drivers of large trucks continue to be a population of concern regarding traffic safety despite the reduction in large truck crash rates over the past decade. Medication and drug use while driving is one important risk factor for large truck crashes. Work-related exposures, such as vibration, manual handling and poor ergonomics contribute to an increased risk for injuries and chronic conditions and are common reasons for opioid analgesic (OA) use by CMV truck drivers. The objectives of this study were to examine the role of OA use in CMV truck drivers involved in fatal crashes by: (a) generating prevalence estimates of OA use; (b) documenting the relationship between OA use and crash responsibility.

METHODS

Case-control study using logistic regression to compare Fatality Analysis Reporting System (1993-2008) record of one or more crash-related unsafe driver actions (UDAs--a proxy measure of responsibility) between drivers with a positive drug test and drivers with a negative drug test for OA, controlling for age, other drug use, and driving history.

RESULTS

The annual prevalence of OA use among all CMV drivers of large trucks involved in fatal crashes did not exceed 0.46% for any year in the study period and mostly ranged between 0.1 and 0.2%. Male truck drivers using OA had greater odds of committing an UDA (OR: 2.80; 95% CI: 1.64; 4.81). Middle-aged users had greater odds than younger or older users.

CONCLUSION

The results of our study indicate that the presence of OAs is associated with greater odds of committing an UDA. This association may have implications for the commercial transport industry and traffic safety. However, the limited prevalence of OA use is encouraging and further research is needed to address the limitations of the study.