CH 6 - Foul Waters / Forest For Trees
Apr 15, 2009

CH 7 - "Hit and Run" Homicide / No Batch Experiments!
Apr 15, 2009

CH 8 - The dose makes the poison even with non-toxic paints!
Apr 15, 2009

CH 9 - Automotive Engine Evaluations from Oil Filters
Apr 15, 2009

CH 10 - Powder on the Peanuts / Contamination of Foodstuffs
Apr 15, 2009


CH 15 - Toxicology Opinions / Critique/Cross Examination of Expert Witness and Data
2009-04-15
Toxicology Opinions

We have been consulted for toxicology opinions in both civil and criminal actions. In the majority of cases encountered the problem presented involves calculating the reported blood alcohol level based upon reported consumption vs the blood alcohol levels measured. We have also been asked to render opinions as to the expected results of interactions between prescribed and over the counter medications and the effects of alcohol on both prescription and OTC products.



Critique/Cross Examination of Expert Witness and Data

A wise and very experienced analytical professor taught in our undergraduate courses that "First the analysis is made reproducible, then it can be made accurate." The reproducible-accurate statement is a very contrite summary of the basis of statistical process control, which applies to ,all, the measurement procedures known and used today.

Scientific data presented in both courtrooms and used in forensic investigations is often, mistakenly accepted, as "irrefutable fact". All measurement data is subject to the laws of statistical variation and when the data generated is shown to be conforming to those laws then a definable numerical level of confidence exists for the measurement made. The level of confidence when expressed as a percentage can vary from virtually nothing to values in the high nineties.

When measurement numbers are received from a laboratory or analysis facility the written reports should have the requested measurement results and at least several statistical parameters, which should have been recorded for the measurement process used. A quantity called the minimum detection limit or MDL for the measurement made, indicates the least amount or smallest quantity of the requested measurement which can be seen with an acceptable degree of confidence using the appropriate measurement process. Generally speaking when the values of the MDL and the measured data are far apart the measurement may be very reliable. The reliability of the data degrades rapidly as the values draw closer together.

For measurement processes which use chemical analysis a number called the reagent blank should also be reported. The reagent blank is a measure of the contamination within the process and should be very small in comparison with the value of the requested quantity. If the value of the reagent blank is not small in comparison to the value being reported then the reliability of the reported data is lowered.

Chemical analysis processes should also report a quantity called the Matrix Spike Recovery. The MSR is a percentage that should be close to one hundred. The MSR number is generated by adding a known amount of the material being measured to the sample and measuring the new total. Deviations from one hundred percent recovery indicate the severity of interferences in the analytical process. Large deviations from a value of one hundred indicate that the reliability of the data measured is lowered in proportion to the deviation.