September 10, 2015
ARIZONA BLOOD ALCOHOL TESTING
Let’s make a list. A list of the things used to create the “number” printed by a gas chromatograph. That is, the “number” that the prosecution will use as their best evidence in any given DUI case.
In a laboratory we would call this list our methodology. Because the number printed out by a gas chromatograph in a DUI case is a measurement of a person’s blood alcohol concentration, this will be a measurement methodology.
A measurement by definition is not intended to be a true value. Metrology, which is the science of measurement, informs us that the number printed by the gas chromatograph is merely an estimation. In order to rely upon such an estimation for any important decision (i.e. innocence or guilt in a DUI case), the printed number must be:
- Reliable, and
Only after its proponent has proven all three of these qualities can the “number” be “fit for the purpose” of deterring a person alcohol concentration.
Creating a methodology that complies with minimum requirements of the scientific community can accomplish this task. These minimum requirements are found in what are known as consensus standards. The seminal consensus standard for creating a trustworthy measurement is ISO 17025. Accordingly, all the steps and procedures on our list must substantially comply with this consensus standard.
Demonstrating that these essential steps are followed falls under the categories of quality assurance and quality control. Quality assurance is utilizing a methodology to prove accuracy, reliability, and validity to the people in the lab. Quality control is to prove it to people outside the lab.
With the above concepts in mind, let’s look at our list and some of the associated issues. Our list is divided into three categories: Pre-Analytical, Analytical and Post-Analytical. These categories are taken from the concept of the “brain-to-brain loop.” In this manner of thinking errors are identified and classified by the stage they occur. In a DUI case, the loop starts when the brain of a person (a member of the police department) “decides” to order a blood alcohol test. The loop is “closed” when law enforcement receives the end results of the test.
Using this analytical framework clarifies the who is accountable for, and the requirement of, each stage in the testing processing.
Pre-analytical refers to steps in the measurement process before the sample is put into the instrument. Errors in this phase are just as damaging, if not more, than errors in the analytical phase. For example, if the person drawing the blood mislabels a blood tube, or the sample is contaminated, the gas chromatograph will not be able to identify the error. The pre-analytical stages commonly found in when using a gas chromatograph in an Arizona crime labs are listed below.
In Arizona DUI cases, the officer is given the choice of what type of test to use for attempting to measure the person’s alcohol concentration. There are two primary options: (1) Blood and (2) Breath. When the officer selects a blood test, the pre-anlystical steps include:
Sample Collection (Blood Draw)
The blood draw begins the process. Some of the issues commonly encountered that may affect the number the machine prints out at the end of the process are:
- Volume – is there an adequate volume in the blood tube?
- Tube Selection & Integrity – Did the “vacutainer” have the proper amount of anticoagulant and preservative?
- Inversion – where the correct amount of inversions done at the correct number of degrees?
- Labeling – is there evidence that right person’s label was placed on the right blood sample?
The goal of sample collection is to ensure that a representative blood sample is loaded into gas chromatograph for analysis. Put another way, the sample should be in substantially the same condition when analyzed, as it was when drawn from the person’s arm.
Impound & Identification
After blood is drawn an officer is required to impound the sample(s) with property and evidence section of the police department. The officer’s request for a blood alcohol measurement is entered into law enforcement’s information system. Eventually a computer processes the data and generates a type of pick-up list and work list.
Some of the issues to be examined here are:
- How long was the sample without refrigeration?
- Is the chain of custody accounted for?
- How was the sample assigned it identification for lab analysis
Transportation To The Lab
When it is ready to be tested, the sample is taken from the refrigerator (in the property room) and brought to the lab. Because testing is done in a batch multiple samples must be transported together. A batch means that multiple samples will be tested in in the same run. Accordingly, the about 30 to 40 blood samples will be taken to the lab. The concern in transportation to making sure the sample maintains its integrity.
Inspection In The Lab
Once the blood samples are transported to lab, the lab tech will do a visual inspection. These observations of the lab tech are documented and result in paperwork that is discoverable. For example, during this inspection is when the volume of blood sample in the tube is measured and documented.
Prior to being analyzed by the instrument, the blood sample must be brought to room temperature and mixed to make the homogenous. Many crime labs such a the City of Phoenix will place the blood tube on a rocker for approximately an hour. Failure to prepare the sample in this manner may result in an artificially high blood alcohol measurement.
A blood sample is collected in a “grey top” blood tube. In order for it to be analyzed in a gas chromatograph a small portion of the sample must be transported into a headspace vial. The headspace vial, containing the blood sample, is what is loaded into the gas chromatograph. Pipetting is the process used to transfer the blood sample into the headspace vial.
Some lab techs pipette directly from the blood tube into a headspace vial. However, many techs use a cuvette to aid them in this transfer.
A cuvette is a small container sealed at one end designed to hold a portion of the blood sample. If a cuvette is used it sanitation must be insured. Reusing a cuvette without properly sanitizing it could lead to cross-contaminstion.
The device used to transfer the blood sample is called a pipettor dilutor. After the lab tech uses the pipettor dilutor to draw a small amount of blood (approximately 100-250 ml) from the grey top tube used for collection, it is then used to deposit the sample into a headspace vial. While depositing the blood sample, it also deposits an internal standard into the headspace vial.
An internal standard, if properly used, attempts to minimize error. An internal standard (sometimes called a surrogate) is another chemical substance whose analytical behavior is well known and similar to the substance you are quantifying. A specific and consistent concentration of it is added to the subject samples, blank, calibrators and controls. The response from the analyte peak is compared to the internal standard. In blood alcohol testing, N-propanol is commonly used by Arizona crime labs as an internal standard. This because N-propanol has a similar structure and reacts similarly to ethanol in the testing process, at the same time it is not expected to be found in a blood ethanol sample.
Loading The Vials Into The Machine
Arizona crime labs manually load the headspace vials into the gas chromatography’s autosampler. Loading the wrong vials into the wring slot can cause the machine to print out a measurement belong to a different subject. Accordingly, the great care is required in the loading process. Some labs, such as the City of Phoenix’s Crime Lab has a second person double if the correct vials were laded into the correct slot. However, others like the Scottsdale Crime Lab have no such safeguard.
A good practice is to keep the headspace vials in a tray and only handle one subject at a time. As you can see from the illustration below, Arizona crime lab label the headspace vials – not with a bar code label – but with a sharpie marker. While this reduces costs, this practice is not recommended to prevent misidentification of a sample.
Eventually the the headspace vials are placed into the autosampler for processing. As you see below the autosampler can hold many vials. A full run will hold approximately 110 for most of the units in Arizona crime labs. The slots are not labeled. The identification process is controlled by software and the data that has been manually entered lab personnel.
All crime labs in Arizona use headspace gas chromatography for blood alcohol analysis. There process is designed to be largely automated. Samples are tested in batches. This is similar to production-line chemistry. A full batch of 110 vials can take over 10 hours to process. An analyst will commonly observe only about 10 to 20 minutes of the process. Thus, it is critical to ensure that the instrument and its software are working reliably if the test results are to be trusted.
The Result Is An Estimation
Like most State’s, Arizona has a legal definition of an alcohol concentration. A.R.S. 28-101(2)(a), provides an alcohol concentration is defined as “[t]he number of grams of alcohol per one hundred milliliters of blood.” However, the State does not use a 100 mili-liter blood sample. They merely use a blood sample between 100-250 micro-liters. A sample of this size will appear to be just smaller than a piece of M&M candy.
Moreover, the lab does not actually test the 100-250 micro-liters liquid. The sample is heated and a portion of the gas above it is captured to be analyzed by the machine. As such, the process estimates what the result would be if 100 mili-liters were analyzed. It is an indirect measurement relying on presumptions and extrapolations.
The Machine’s Essential Functions
During the Analytical phase, the instrument must perform two essential functions: (1) separating different substances; and then (2) quantifying them. Ultimately, it is during the analytical phase that these things occur. However, the analytical phase begins well before the machine is tuned on. It starts with teaching the machine to know specific alcohol concentrations.
A gas chromatograph does not come out of the box inherently knowing any alcohol concentration. While the device can separate different substances, it does not have the ability to identify them. It must be taught. This process is known as calibration.
For blood alcohol measurements, a gas chromatography is calibrated by using known alcohol concentration. That is, they purchase commercially made ethanol concentrations (i.e .40, .20., .10 and .02) build a calibration curve. Think of it like building a ruler. The calibrator are like specific points of measurement on the ruler and the instrument.s software connects the them. Subject samples are compared to the virtual ruler and a measurement is derived from a comparison.
The Blood Testing Machine
As you can see below the there are actually two devices. The gas chromatograph and autosampler. These devices and their corresponding software are commonly used to measure blood alcohol concentrations. The process is automated and the lab tech is not present for the majority of machine’s work.
Sampling The Air In The Headspace Vial
The machine does not analyze your blood. Recall that the sample is contained in a headspace vial which has been loaded into the machine. On the top of that headspace vial is a rubber septa.
After it heats the vial, the gas chromatograph injects a syringe through the rubber septa. It it then obtains a small sample of the air above the liquid in the vial. Thus, no blood is directly sampled, only the air above the blood sample. The capillary column is contained within an oven.
A Trip Though the Columns
Much like a human intestine, the instrument contains long thin tubes for substances to pass through. The mixture of the blood sample and the internal stand vapor in the headspace vial is transported by a carrier gas (usually hydrogen) through a column.
Separation In The Columns
Thin “capillary” columns have a coating in their internal wall. Based upon a variety of factors (temperature, flow rate, each substances affinity for parts of the column’s internal coating), different types of substances travel through the column at a different rate. Accordingly, each substance will start to separate from each other, and then come out the other end of the column at a different time.
The unique time each substance exists the column is called the time when a substance “elutes.” To illustrate, if ethanol elutes at 1.5 minutes (the time it takes to travel through the column) anything that comes out the column at that time will be identified as ethanol by the machine – even if it not ethanol.
The Two Columns
The end result of the measurement process is a graphical representation of the analysis in the form of a chromatogram. As the name suggests a chromatogram puts the information in a graph with vertical and horizontal lines. A chromatogram from an Arizona crime lab typically has two graphs on one page. This is because there are two columns in the gas chromatograph’s oven each performing an analysis at the same time. Thus, the processed data illustrates the results of both the “A” and “B” columns.
The “A” at the top of the chromatogram is the reported measurement. The “B” column is confirmation column.
The Flame After The Columns
The flame ionization detector (FID) is connected to the gas chromatograph by the columns discussed above. After a sample’s trip through the column, where it separates into it different substances, it exits the column and goes into the FID. This is where the substance is quantified.
The sample stream (the gas) from the column passes through a flame in the FID. The compound is burned by the flame. Burning a compound containing carbon creates ions by removing electrons. If you recall science class, an ion is a molecule with a net electric charge. As ethanol contains carbon it is capable of being ionized.
The newly freed electrons are then measured. The device works on the premise that the electrical current will be in direct proportion to the concentration of the hydrocarbons burned.
Current will pass between electrodes (an electrometer) located in the FID close to flame. The ions are attracted to the detector and eventually become a recordable digital signal. In the FID, the signal is amplified and converts to parts-per-million.
The Data Is Graphically Represented
After detection, the signal is processed by software. The end result of this process is called a chromatogram. A chromatogram is a visible representation of the data in a graph. It illustrates the results of separating the substances and their quantification.
The software does its analysis by comparison. It compares the known concentrations (calibrators and internal standard) to the the unknowns (subject samples).
A subject sample chromatogram will always have at least two peaks. Recall the headspace vial, shown above, contains both a blood sample and an internal standard. If ethanol is detected that will be represented by a peak, as will the internal standard.
Post-Analytical generally refers to the phase that starts after the instrument has processed the blood sample. In the Post-Analytical Phase the results of the blood alcohol analysis are reviewed, interpreted and conclusions are made. Post-analytical errors occur after the gas chromatograph has processed the sample. However, an error may have occurred in the pre-analytical or analytical phase but not detected until post-Analytical Phase.
Review of Printed Data
As stated above, the analyst is not present for the majority of the machine’s processing of the sample. While the instrument is processing the blood sample the machine generates an “original” chromatogram. When the analyst returns to work the next day, there is a pile of chromatograms waiting to be reviewed.
This paperwork printed from the instrument is called the “batch data.” The analyst reviews this paperwork to see if there are any visible abnormalities within chromatograms. Then batch data is then given to a second analyst to do a “technical review” of the paperwork. If the laboratory believes that their internal criteria for a measurement has be satisfied, then the results are released.
Uncertainty of Measurement
No measurement is intended to represent a true value. The most science can do is to provide a value believed to be within a range of values. The more important the purpose of the measurement, the more rigor is needed in determining the parameters of that range and confidence that the true value falls within it.
The measurement, which the machine prints at the end of its process, is known as a reported result. The reported result is the purported measurement. Simply providing a reported result, with nothing else, is often misleading. A reported result is only complete when accompanied by a “statement of its uncertainty.”
A Complete Result
A “statement of uncertainty” is the range of doubt that exists regarding a measurement. A test result is only complete when accompanied with a statement of its uncertainty. At a minimum, a complete test result must include the:
- Reported Result
- Range of Uncertainty; and
- Confidence Interval.
For example, take a reported result of a blood test of .100. This number alone is an incomplete result. To complete the result we need to look at the machine’s historical performance.
Lets assume, based on a review of the machine’s prior performance, a “range of uncertainty” was determined to be ±5 percent, with a “confidence interval” of 99.9999 percent. This means that the reported result could be as low as a .095 and as high as a .105.
This also means, if the same blood sample were repeatedly tested on this equipment, the result would only be outside of the ±5 percent range approximately one out of a million times. On the other hand, what if for the same reported result of .100 the range was ± 30 percent, with a confidence interval of 50 percent? Here, this means the reported result could be as low as .070 or as high as .130. Furthermore, if you continued to test this sample on the same equipment, it would fall outside the range dramatically more than the first example. In both examples, the reported result was the same, but the second seems a lot less trustworthy (especially without knowing what caused the results outside the specifications.)
When comparing the two complete test results, you can see that providing a mere reported result does not tell us the whole story. Merely telling us the reported result can actually tell us a very misleading story.