Back to the future: Retrograde alcohol calculations an uncertain science

Abstract

To respond to Labay and Logan's call for a scientific consensus regarding retrograde extrapolations in alcohol DUI cases by describing the existing practices in England and Wales and outlining changes that need to be made. Retrograde extrapolations, known as back calculations in England and Wales, are widely used in toxicology to estimate the blood alcohol concentration of an individual at some prior time. In 2018, Labay and Logan (Journal of Forensic Sciences, 2018, Vol 63. No 5, 1602–1603) published a call for a scientific consensus regarding back calculations. In the UK guidelines have been issued by the United Kingdom and Ireland Association of Forensic Toxicologists (UKIAFT, 2021, Guidelines for Alcohol Calculations, Version 4.3), while in the US, the Organization of Scientific Area Committees (OSAC) for Forensic Science have issued draft guidelines (2020, Guidelines for Performing Alcohol Calculations in Forensic Toxicology). However, these guidelines are not fully agreed and open for interpretation. Alcohol elimination rates have been discussed in many publications since Widmark's original data were published. The current guidance from UKIAFT, is to report the most likely back calculated result together with a range of results based on the 95% confidence limit elimination rates from Jones's 1996 study of drinking drivers (Journal of Forensic Sciences, 1996, Vol 41, No 5, 922–926). That range of elimination rates being from 9 to 29 mg/100 mL/hour (0.009 to 0.029 g/dL/hour). There is an acceptance of this 95% confidence interval for the reported final result by practitioners. The Divisional Court, upheld by the House of Lords, however ruled that in order to convict someone for being over the prescribed limit on the basis of any back calculation, the case must be proven beyond reasonable doubt. While the 95% confidence interval covers a large portion of the population, 1 in 20 individuals may fall outside the reported values. It is clearly not sufficient to prove a defendant's guilt so clearly that it must be accepted as fact, a common requirement of UK law. A confidence interval increased to 3 standard deviations or 99.73% would provide a greater factual basis for the court and cover a much larger proportion of the population. There is a reluctance within the forensic community to use this larger range as it often renders the calculation ineffective. However, that is simply a consequence of the considerable population variation. Retrograde extrapolations also rely on the subject being post absorptive at the prior time. In the UK, it is considered safe to back calculate if the subject has not eaten or consumed alcohol in the hour prior to the back calculation time. Where the subject has eaten recently prior to the back calculation time, experts are instructed to consider whether the back calculation is valid. In Germany back calculations are not permitted to a time within 2 hours after last drink consumed. The 2 hour limit would better meet the highest standard of ‘beyond reasonable doubt’ burden of proof, and should be used. There is a lot of debate around certain elements of the back calculations and experts have the freedom to interpret variables differently to an extent. Based on the review of many publications the have proposed modifications to the current practice in forensic casework to provide a more robust uncertainty and confidence interval. This will ensure consistency in the testimony given by different experts to the court and provide greater clarity for judges and juries in the assessment of scientific concepts and evidence.