Importance of Quality Assurance for Forensic - Lab Manual

Document Sample
Importance of Quality Assurance for Forensic  - Lab Manual Powered By Docstoc
					   Importance of Quality
  Assurance for Forensic
Toxicology in North America
           Graham R. Jones, Ph.D.
              Chief Toxicologist
    Office of the Chief Medical Examiner
         Edmonton, Alberta, Canada

This presentation is based on an invited
lecture prepared for and given at the
annual meeting of the Japanese
Association of Forensic Toxicologists in
Fukuoka, Japan, August 27, 2005.
(Reproduced here with only minor changes)

Graham Jones

Importance of Forensic Toxicology
 Forensic testing is performed for a legal
  purpose and may affect:
  –   Criminal trial (freedom/jail or death of accused)
  –   Civil trial (denial of compensation/money)
  –   Loss of a job (if drug use proven)
  –   Loss of career (e.g. sports/athlete)
  –   Custody of children (if drug abuse proven)
 Toxicology testing must be ACCURATE!

Types of Forensic Toxicology
Laboratories in North America
 Federal (FBI – U.S.; RCMP – Canada)
 State (U.S.) / Province (Canada)
   – Police / Crime lab; Medical Examiner/Coroner
 County / City / Local (U.S.)
  – Police / Crime lab; Medical Examiner/Coroner
 Private (mostly in U.S.)
   – some have national coverage, some mainly

North American Courts - Adversarial
 In the U.S. and Canada, the court system is
   – Prosecution and defense lawyers may each hire their
     own toxicologists who may give different opinions
   – Prosecution toxicologists may perform testing and give
     an interpretation (opinion)
   – Usually defense toxicologists offer only an opinion and
     do not usually perform analytical work on cases
 The adversarial system should not affect the
  science, but allows more than one scientific
  opinion to be expressed in court

Why does an adversarial court
system make a difference?
 Prosecution laboratory results are more likely to
  be challenged by defense
   – Prosecution toxicologist may be required to provide all
     details of their analysis (chromatograms, methods,
     chain of custody etc)
   – Many (good) forensic toxicologists are available to
     review analytical data for the defense and provide a
     second opinion
   – Prosecution toxicologists are more likely to have to
     defend their analytical work in court

Everyone makes mistakes!!

Quality Assurance Program
 Every toxicologist in a laboratory will make a
  mistake. The problem is making sure that the
  mistake is detected before a result is reported
 A good QUALITY ASSURANCE program will
  help to minimize errors and detect errors when
  they occur
 “Quality assurance” covers every aspect of
  laboratory work that can affect the quality
  of the results reported

Example of a Preventable Error
 A person died in a psychiatric hospital
 Her medications were dispensed by hospital staff
 The pathologist determined death due to natural causes, but
  toxicology testing indicated an imipramine overdose!
 A public inquiry was held and data of the toxicologist was
  reviewed – finding calculation error of 10-fold
 Public disclosure of the error was very embarrassing for
  the toxicologist
 The error could have been discovered before public
  embarrassment if the data had been properly peer reviewed
  (i.e. by a colleague in the same laboratory).

Quality Assurance Includes:
 Education and training
 Standard Operating Procedures (SOPs)
 Validation of methods; Criteria for evaluation of
  calibrations, controls and case data
 Maintenance of equipment
 Review and reporting of results
   – Verifying results of a colleague
   – Someone else should check your transcriptions and calculations
 Proficiency testing
 Laboratory accreditation

Education and Training
 Education in chemistry (and pharmacology)
 Training in the laboratory
   – Develop good analytical skills
   – Learn acceptance criteria the laboratory uses for
     deciding which results are OK and which have to be
     repeated! (e.g. what is an acceptable calibration?)
   – Learn about the quality assurance program of the
 Continuing education
   – Additional training, reading, courses, conferences
   – May be minimum requirements for personal

Standard Operating Procedures
 Important that routine analytical methods
  and procedures are written and should be
  approved by the laboratory director
  – Helps to ensure that all toxicologists perform
    the assay the same way each time
  – Enables the courts to know exactly which
    method was used to generate a result in a
    particular case
  – Makes independent review of the analytical
    method easier

Calibrations – What criteria?
 Need to have consistent criteria for when a calibration
  is “acceptable” and whether controls have “passed”
 Several different criteria can be used to ensure that a
  calibration is acceptable
   – Correlation coefficient – (e.g. >0.98)
   – Multipoint calibration (e.g. 3 – 8 points)
   – Each calibrator can be read against calibration line (e.g. 80 – 120%
     of target)
   – Can use linear, exponential or other curve fit
       • Not all assays are inherently linear
 But you must use criteria of some kind…

Quality Control
 Used to verify the precision and accuracy of the
   – Preferably prepared as a “pool” that is verified prior to
     first use
   – Pre-prepared and frozen for use in many assays
 Analyte spiked into a similar matrix (e.g. blood)
 Should be independently prepared
   – Can be a external supplier (purchased material)
   – Can be another person in the same laboratory, but the
     control should ideally be made from another source

Controls – What criteria?
 Controls should be prepared by a person other
  than the toxicologist conducting the assay
 The source of the standard solution should ideally
  be different than that used for the calibration (e.g.
  different manufacturer; different lot number)
 Controls should usually read within +/-20% of
   – Exceptions – alcohol ±5-10%
   – Some difficult drugs or controls at very low
     concentrations may be ±25%

Proficiency Test (PT) Programs
 All toxicology laboratories should subscribe to
  one or more PT programs appropriate for their
  work (e.g. for alcohol and drugs)
 Corrective actions must be performed when results
  of the lab significantly different from that
  expected (e.g. for drugs >±20% or ±2 SD)
 It is NOT acceptable to review controls and ignore
  the results if they are outside the acceptable target
 Some laboratories participate in 3–5 PT programs
 PT programs can be internal, but external better

Example of Acceptable Proficiency Test Results

Mean value 413 mg/100ml- 2 S.D. range ±10.5% (Acceptable)
Example of Poor Proficiency Test Results

 Mean value 1021 mg/l - 2 S.D. range ±42.6% (POOR)
Corrective Action
 Bad control or PT results MUST be remediated
  – i.e. investigate why the results were out-of-range
 Corrective action could be:
  – Review of calibration and data calculation
  – Review of chromatograms and integration
  – Re-analysis of the sample
  – Troubleshooting the assay (precision, accuracy)
  – Inter-laboratory comparison of results
 Main reason for bad results: poor precision

Forensically unacceptable practice
 Once the calibration, controls and case
  samples are run:
  – Calibrators should not be dropped from the
    middle of the curve in order to improve the
    correlation coefficient
  – Failed controls should not be ignored
  – Failed controls should not be re-setup in
    isolation of the case sample(s)

Method Validation
 Analytical methods must be appropriately
   – Linearity (upper and lower)
      • Is the calibration accurate over the entire reporting range? The
        calibration parameters may change at the upper end
        (saturation) and lower end (adsorption)
   – Accuracy
   – Precision
   – Specificity
      • Is only the target compound being measured?
      • Mass spectrometry is increasingly preferred, but must be
        properly run

 Especially important to determine if range of
  calibrators limited (e.g. if 1 or 3 point calibration)
 Linearity may “fall off” at high end due to
  detector saturation
 Linearity and sensitivity may decrease at low end
  due to adsorption to glassware and GC injection
  port phenomena
 Important that controls run near upper AND lower
  end of reporting ranges if not covered by the

 Precision is important
   – Poor precision means that the “right answer”
     may only be obtained some of the time
 Replicate analyses?
  – Running samples in duplicate can help to
    determine if precision is poor

Professional Standards
 Professional standards
   – Professional Standards should be used as a guide to
     “good laboratory practice”
   – e.g. SOFT/AAFS Guidelines
 Can help a laboratory get the necessary resources
  to meet a nationally or internationally recognized
  professional standard
   – Useful to convince administrators to provide better
     funding (money!)

Case Study – Inadequate Validation
 The wife of a well-known pathologist died suddenly. He
    was accused or poisoning her with an unknown substance.
   Prosecution thought that neuromuscular blocking agents
    might be a good possibility
   A forensic laboratory developed an LC/MS/MS method for
    neuromuscular blocking agents
   The laboratory detected succinylmonocholine (SMC) in
    the postmortem tissue of the wife of the pathologist
   The pathologist was the charged with the murder of his
    wife, convicted and sentenced to life in prison

Succinylcholine (SCC)
 Succinylcholine (SCC) is an unstable ester of two
  choline molecules joined to succinic acid
 Because SCC is unstable, a lab can look for the
  mono-ester (succinylmonocholine - SMC), but
  that is also unstable - but more stable than SCC

Prosecution Position
 Prosecution toxicologists argued that
  hydrolysis of SCC to SMC and eventually
  to succinic acid and choline was in one
  direction only (not reversible)
 Therefore, they argued that detection of
  SMC proved beyond any doubt that the
  victim must have been administered SCC

Defense position
 Defense toxicologists argued that hydrolysis of SCC to
  SMC and to succinic acid and choline was an equilibrium,
  even though mostly towards complete hydrolysis
 Therefore, since all postmortem tissues contain high
  concentrations of succinic acid and choline (NOT from
  SCC), therefore combination of these compounds to form
  small amounts of SMC was a theoretical possibility
 It was up to prosecution to run adequate negative controls
  (postmortem tissues) to prove that trace amounts of SMC
  do not exist naturally. They did not do that prior to trial!!

What was the outcome?
 The jury trial did not take the defense argument
  seriously, and convicted the pathologist of
  murdering his wife
 Several months later, the laboratory performed the
  specificity studies and found that SMC was
  formed in postmortem tissues (in cases where
  SCC was known NOT to have been given)
 Therefore, there was no scientific evidence that
  the pathologist murdered his wife with SCC
 The legal outcome was complex…

Forensic Toxicology Accreditation
 Several different programs…
 Urine drug testing (workplace testing)
   – SAMHSA (U.S. Government - regulatory)
   – College of American Pathologists (CAP - voluntary)
 Forensic Toxicology
   – American Board of Forensic Toxicology (ABFT -
 Forensic Science / Police Labs
   – American Board of Crime Laboratory Directors – Lab
     Accreditation program (ASCLD/LAB)

What Does Accreditation Cover?
 Administration of the laboratory
 Personnel / staffing / training
 Security and chain of custody
 Procedure manuals / written documentation
   – Standard Operating Procedures (SOP)
 Analytical methods
 Quality control / quality assurance
 Reporting of results

Is Accreditation Mandatory?
 Urine drug testing
   – Only for less than 5 – 10% of testing
   – Mainly for transportation related jobs
 Forensic Toxicology / Science
   – Only in 3 states (New York State, Texas and Oklahoma
   – Not at all in Canada
 Ideally, every laboratory should be accredited and
  inspected by an accreditation program that is
  based on accepted international or national
  standards (e.g. ISO 17025; SOFT/AAFS

Are All Toxicology Labs Accredited?
 NO!!
 Many labs in Canada and the U.S. have no
  accreditation or independent assessment
 The only pressure for labs to become accredited is
  “peer pressure”, or a serious mistake that leads to
  public or political pressure for accreditation
 There is very little incentive for a lab manager to
  seek accreditation of a toxicology laboratory
   – It creates more work, with no promise of extra
     resources to address staffing and training needs!

The Future?
 The only way all forensic toxicology laboratories
  will adopt adequate and consistent QA standards -
  and become accredited - is if it is mandated by
  government legislation (personal opinion!)
 Mandated accreditation can have the benefit of
  additional resources in order to meet the
  professional standards that have been set
 Standards for forensic toxicology will continue to
  become more strict, perhaps following a trend that
  already exists for the environmental and
  pharmaceutical industries

Thank you for your attention
 and your kind hospitality!


Shared By: