"METPOL" An automated system for the analysis of blood alcohol levels. D.A. Rudram

Transcription

1 "METPOL" An automated system for the analysis of blood alcohol levels D.A. Rudram Metropolitan Police Forensic Science Laboratory 109 Lambeth Road, London, SE1 7LP, England 1) Introduction Blood-alcohol analysis is a routine repetitive task in which large numbers of similar specimens have to be rapidly and accurately analysed. The Road Safety Act 1967 introduced a statutory blood alcohol level into the United Kingdom making it an offence to drive with more than 80 mgm. of ethanol in 100 ml. of blood and the number of samples submitted to the Metropolitan Police Laboratory climbed rapidly. (Fig.l) By the early 70's the original manual gas chromatographic method was becoming swamped and consideration was given to automation. Initially an analytical system based on one of the many automatic analytical instruments then available using enzymatic techniques was considered. However the advent of computer based gas chromatography data systems meant that the continued use of a familiar and well tried technique was possible. Injection of liquid blood -albeit diluted - onto the gas chromatograph presents problems with blockage of syringes and injection ports. However this could be overcome by head space analysis since both ethanol and the internal standard - n-propanol - are volatile. The Perkin - Elmer F40 automatic head space gas chromatograph (Perkin Elmer Ltd., Beaconsfield, Bucks.) solved the technical problems inherent in this approach. The sample is sealed in a bottle (Fig.2) with a rubber cap and the bottle inserted in a therm ostated water bath. The needle in the dosing head punctures the cap and the bottle is pressurised to full column pressure. The gas supply to the column is then shut off causing a pressure drop on the column. As a result an aliquot of the vapour is extracted from the 528

2 sample and is passed onto the column. Column pressure is restored and chromatography takes place as usual with extremely satisfactory results. A decision was made to purchase two Perkin Elmer F40's along with their PEP-1 data processor and these were delivered in the Spring of The equipment was operational by the late summer of the same year and grew eventually to six automatic gas chromatographs and two teletypes. The PEP-1 data processor was a purpose designed dedicated computer for gas chromatography. This "black-box" approach, along with its immediate availability as a working device and the reduced interfacing problems to be expected by using a single manufacturer's equipment, were among the principle reasons leading to the purchase of the Perkin Elmer PEP-1. The head space chromatographs were a success from the start and similar instruments are still in use today. However the limitations of the processor rapidly became apparent - although it provided a useful intermediate step to full automation - and a general purpose computer system was sought as a replacement. As a result the "METPOL" system based on Instem's Datachrome 2 was installed in 1977 (Kratos-Instem, Stone, Staffs.) The system is based around a Digital Equipment Corp. PDP mini-computer (Digital Equipment Corp., Reading, Berks.) linked via Datachrome to seven automatic gas chromatographs (a mixture of Perkin Elmer F42's and F45's) and it can also handle four manual instruments. The system has grown and now consists of (Fig.3) a central processor with 28K words of core store and two RK M word disc cartridges. The operator communicates with the system via an LA 36 Decwriter and there is an LA 180 high speed character printer for bulk output. A high speed papertape reader and punch are also available. Analogue signals are passed from the gas chromatograph via a multiplexer and analogue to digital converter both designed and built by Instem. Parallel digital links carry the control signals for the gas chromatographs. Modifications to provide positive sample identification using equipment from Vickers Ltd., (Eastleigh, Hants.) are now well advanced. 529

3 2) Softw are The "METPOL" software is made up of a number program modules which are combined under a Digital Equipment Corporation Operating System, RT-11. RT-11 operates on two levels simultaneously which enables the system to perform two tasks concurrently. The priority level - the foreground - is used to operate the gas chromatographs in real time. The programs for this are taken from Instem's Datachrome and they control the chromatographs, and monitor the output from the detectors. Datachrome automatically provides digital smoothing of the signal and compares the signal with an arbitrary baseline. It detects the beginning and end of peaks measuring their area and retention time. Provision is made for signal averaging and a reduction in scan rate as retention time increases and the peaks become broader. Theoretically the system can handle 22 gas chromatographs, although so far we only have seven automatic and one manual instrument connected. The subsid ary level - the background - is used for operator interaction with the system and the manipulation of the raw chromatography data to produce an output in the format required. These programs were specifically produced for "METPOL" and are all written in BASIC. The programmes are modular in design and are characterised by the high degree of operator assistance provided (Fig.4). These background programmes fall into four principle groups: METHOD BATCH REPORT to establish analytical parameters to set up and initiate the analysis of groups of samples to obtain analytical results from the system STATEMENT - to produce statem ents for common situations not catered for on the standard certificate

4 In addition there are several subsidiary groups of programs which enable the system supervisor to modify certain parameters - for example allowance for analytical error - which are the same for all analysts and only changed as a result of a deliberate policy decision. In addition to the operating flexibility the two level operating system has had other advantages. In particular it enabled a tried and tested G.C. Data system to be used for the sensitive real-time element of "METPOL" whilst the programs which adapted this system to our needs could be written in high level language which was easy to modify. Almost all of the background software has been modified or re-written as experience of the system was gained and its strengths and weaknesses became apparent. One notable addition is the automatic collection of statistical data. 3) Operation The system is designed to be as simple to operate as possible. Once "METPOL" has been loaded into the computer it prompts the operator by providing a list of the possible replies at each stage of the operation. The first step in the analytical process is the generation of analytical methods for the gas chromatographs. The analyst calls up "METHOD" and this prompts him to provide, all the information required to prepare an analytical method (Fig. 5). The method is divided into three sections. Data Aquisition Parameters which along with the Timing Functions determine how raw peak data is derived from the gas chromatograph signal when it is received by the processor. The Scan Rate is initially 10 samples a second and this is halved at regular intervals (every half a minute in the example) to give an approximately constant number of scans per peak. The Lowest Significant Peak defines (in microvolts) the smallest peak the system will recognise. Unlike an integrator a computer based system can decide whether overlapping peaks should be separated by a "tangent skim" or a "perpendicular drop" (fig. 6). The Peak Symmetry Factor and Relative Peak Height factors are ratios of peak width and peak height which must be exceeded before the system will apportion peak areas by "perpendicular drop" rather than by "tangent skim". The Timing 531

5 Functions used instruct the system on when - in relation to sample injection - to start looking for peaks. These parameters determine what raw peak data the system accepts and stores, thus they must be defined before analysis commences. On the other hand the third group - the Data Analysis Parameters - are used to process the raw data into its final form as required by the analyst. These may be defined or redefined after chromatography is complete. In practice the methods used in routine work are permanently stored and called up automatically when required. The daily standardisation routine is used to update the Data Analysis Parameters as necessary. The analyst having generated his methods is now ready to analyse a batch of samples. The samples are prepared physically - the administrative details are checked,the samples are diluted using a Hook & Tucker Auto-diluter, (Hook & Tucker Ltd. New Addington, Surrey) and the sealed bottles placed in the water bath of the chromatograph to equilibrate. Meanwhile details of the batch are entered onto the computer using the "BATCH" program. The batch name tells the system what kind of samples are to be analysed - standards, driver's specimens or some other type - and the system responds accordingly. The batch conversation (Fig.7) records the identity of the analyst and of his assistants, the methods and gas chromatographs to be used, the number of samples and their identity and source. These last items have to be entered manually, the Divisional Code refers to the Police Unit submitting the samples and is used to prepare statistical returns. The samples with identifiers commencing "99" are check standards, the second three digits giving the ethanol concentration in milligrammes per hundred millilitres. After creation of the batch and equilibration of the samples the analysis routine can be initiated and the system left to look after itself. After completion of analysis the system is asked to produce a report. The "REPORT" program is loaded and the type of report required selected. This module carries out base line correction, calculates concentrations and if required collates duplicate results on the same sample. If a batch of standards has been analysed the system 532

6 automatically updates the Data Analysis Parameters for the appropriate gas chromatograph. The report on driver's specimens can take two forms. A short summary of a complete batch (Fig. 8) giving the bare results for each sample along with an appropriate remark as to their acceptability - or otherwise. Alternatively a full report (Fig.9) on an individual case may be produced giving full analytical details for each of the duplicate determinations, the mean result, the allowance for analytical error and the concentration of ethanol to be reported. Samplejdentification At the moment the system provides for a limited degree of sample identification. All computer print-out contains the case number for the sample concerned. However, this number is dependent on accurate transcription of case numbers onto the computer console and correct positioning of samples in the carousel of the gas chromatograph. Samples are analysed in duplicate by separate operators and sample numbers are typed in along with a check digit to obviate typing errors. Careful house-keeping of this type has prevented mistakes so far. However, from the beginning some form of positive sample identification has been envisaged as an integral part of an automatic analytical system, Vickers Engineering Ltd have adapted the Plessey Bar-code and Light- Pen system to our needs (Plessey Communications and Data Systems, Poole, Dorset). The hardware has been developed successfully and the modifications to the software - which the Metropolitan Police Laboratory are doing in-house - are nearly complete. The heart of the system is a label bearing a number encoded as a series of thick and thin bars attached to the sample bottle. When the bottles are placed in the carousel the labels lie across the top. The light-pen is attached to a gantry and oscillates back and forth across the label attached to the sample currently being analysed. The design is such that the label is only in focus when the turntable is in the raised position. The number is read as a train of pulses which are decoded locally and stored in a buffer. At an appropriate point in the analytical cycle the computer 533

7 interrogates the buffer and stores the number for the current analysis. This number will be used to identify these results from then on. 5) Conclusion The "METPOL" system has been operational for two and a half years. The system has been simple to operate and extremely flexible in use. As with all such systems operational experience has shown that modifications to the original specification were desirable. These have been simple to implement and the system has undergone considerable modification since installation. For example output of results has been considerably speeded up - partially due to improved hardware, statistics are now compiled automatically and the call-up of appropriate methods for different types of batch is now done automatically on entering the batch name and the gas chromatographs to be used. The advent of automatic sample identification will not only improve security but will also simplify operation of the system. One of the changes being implemented with automatic sample identification is conversion of all background programs to BASIC 11 a partially compiled version of standard BASIC which should cut run times - particularly at the report stage - even further. Not only has automation greatly simplified the analysis of drivers' blood samples, it has also reduced the number of qualified staff involved in this type of work. At the height of manual analysis, four qualified analysts would work in the section each week, and each analyst would require the services of a fully trained assistant. Staff assigned to blood-alcohol analysis would be occupied on this duty full time. Now whilst four analysts are still required only about half of their time is spent on blood-alcohol analysis, for the rest of their time they can continue with other duties. "METPOL" has been successful in reducing the amount of skilled staff time given up to a routine repetitive task. At the same time it has made the task both simpler and pleasanter to perform. 534

8 6) Acknowledgements I would like to thank the staff of the Alcohol Section at the Metropolitan Police Laboratory for much of the work which led up to this paper. 7) Reference Curry, A.S., Walker, G.W. and Simpson, G.S. The Analyst (1966). 535

9 30 ANNUAL TOTAL OF DRINK DRIVING CASES NUMBERS 15 - XI J i I I I I I I I I I 1 I I YEAR Fig. 1) Annual totals of Drink/Driving case received by the Metropolitan Police Laboratory,

10 Fig. 2. The F40 Automatic Gas Chromatograph sampling system. Key 1. Carrier gas inlet. 8. Silicone rubber membrane for closing 2. Fine pressure regulator space 7 3. Solenoid dosing valve 9. Flushing line 4. Inlet 10. Variable valve 5. Column 11. Solenoid venting 6. Dosing capillary 12. Sample bottle 7. Movable closing cylinder 13. Rubber cap 537

11 Fig. 3) 'METPOL' System - Block Diagram (Only one gas chromatograph shown) METPOL 09-MAY-80 BASIC V01B-02A *** M E T P O L S Y S T E M *** OPTIONS AVAILABLE ARE: A BATCH B METHOD C REPORT D STATEMENT E TERMINATE (A, B, C, D OR E)?C Fig. 4) Typical list of prompts produced by ^METPOL1 to remind the operator of the options available. 538

12 METHOl' EIGHT LISTE.Ii ON 09-MAY-B0 AT 13:5V LAST UFIiATE ON 29-FEfc-SO AT 16:46 DATA ACQUISITION PARAMETERS SCAN RATE HALVEIi AT 0.5 MINS LOWEST SIGNIFICANT FEAK : 5 FEAK SYMMETRY FACTOR : 2.0 RELATIVE FEAK HEIGHT FACTOR : 10X TIMING FUNCTIONS PEAK STORING : STARTS AT 0.7 MINS ENIiS AT 3.0 MINS THERE ARE NO TIME il,efeni«ent FUNCTIONS IiATA ANALYSIS PARAMETERS ZONE START TIME ENl' TIME ZONE NAME MINS 0.9 MINS ETHANOL MINS 2.25 MINS F'ROF'ANOL RESPONSE FACTOR ZONE 2 (F'ROF'ANOL) IS REFERENCE ZONE* CONCENTRATION 100 Fig. 5) Contents of a typical METHOD file. PERPENDICULAR DROP Fig. 6) Methods of apportioning the area under a pair of overlapping peaks. 539

13 ich MAKE! MAY 99 ON &' - ANALYST'S DETAILS ASSISTANT'S DETAILS : D.KUDRAM.* M-KORRISON ANALYSIS METHOD EIGHT ON GC 6 C NALYSIS METHOD NINE ON GC 9 NO. OF SAMFLES 3 SAMPLES RUN NO* IDENTIFIER B HIV CODE J Fig. 7) Information contained in BATCH file. A batch can contain up to thirty cases. The 'Div. Code' is used to compile a statistics file. CHECK LIST OF RESULTS FOR DATCH MAY 99 CONCENTRATION RUN NO. SAMFLE NO. GC 8 GC 9 REMARKS OK STD N/L REPEAT CASE STATISTICS FILES HAVE PEEN UPDATED Fig. 8) Print-out giving a summary of all results on a single batch. 540

14 NUN 1 SAMFLE BATCH MAY 99 ANALYST : D.A.RUDRAM ASSISTANT ANALYST J M.MORRISON DATE ANALYSED J 09-MAY-80 KEF'ORT FRINTED ON 09-MAY-e0 GC 8 USING METHOD EIGHT TIME RRT F'EAK AREA RF CGNC NAME ETHANOL F'ROF'ANOL (MARKER) GC 9 USING METHOD NINE TIME RRT F'EAK AREA RF CONC NAME »s ETHANOL F'F'OF'ANOL (MARKER) r-iean CONCENTRAT ION OF ETHANOL IS 1so m g : LESS ALLOtJANCE or 9 hcx REF'OFvTED CONCENTRATION OF ETHANOL IS 141 MGZ (O.K.) Fig. 9) Full print out on the analysis, in duplicate, of a single case sample. 541