Transcript DownloadSlides

The adventures of an epidemiologist giving expert
evidence in asbestos litigation
compensation claims for asbestos-related diseases in the
Netherlands
Prof Dr Alex Burdorf
Department of Public Health, Erasmus MC Rotterdam
Asbestos use and mesothelioma in the world
Mesothelioma deaths per million people
per year
Lin et al. Lancet
2007;369:844-49.
0
1
2
3
4
5
Asbestos consumption (kg per head per year)
Asbestos and occupation
Asbestos products are everywhere
Legal turning point: Asbestos Legislation 1977
Legislation:
-
first ban on specific products (asbestos spraying, crocidolite)
Key developments:
- rapidly emerging scientific knowledge within Dutch HSE and 1 university
- successful interference of industry, avoidance of a complete ban
(asbestos cement)
- rise of the political debate
Stakeholders:
-
Government: knowledge of secondary importance in policy-making process
Industry: active player in science, politics, and media
Unions: initiative for complete ban on asbestos, limited perseverance
Legal turning point: Asbestos Legislation 1977
The battlefield in science:
- no differences in opinion between (the few) academic and HSE scientists
- industry actively participated in advisory bodies, eg they drafted the first
brochure of the Labour inspectorate in 1971, agreed upon in the most
important advisory board to government
- industry actively participated in scientific meetings and research:
* creating uncertainty about fibre type, friability asbestos / solid matrix,
existence of safe threshold
* presented evidence for adequacy of control measures (safe use)
* distributed scientific articles and reports to scientists
- industry actively defended companies in emerging law suits
First asbestos litigation cases in the Netherlands
Critical arguments:
-
the health risk of asbestos were unknown at the time of exposure
(response: historical development of knowledge)
Development of knowledge about occcupational diseases
Burdorf et al.
Am J Ind Med 1991
First asbestos litigation cases in the Netherlands
Critical arguments:
-
the health risk of asbestos were unknown at the time of exposure
(response: historical development of knowledge)
-
working conditions were within the legal occupational limits at the time
(response: development of knowledge on control measures)
Development of knowledge on control measures
Swuste et al. IJOEH
2004
Legal turning point: Supreme Court decision 1990
Legislation:
Supreme Court judgement Janssen - Nefabas 1990: shift of burden of proof from
victim to employer.
* control measures are required from early onwards (< 1945)
* asbestos companies have an international network and, thus, have access
to state-of-the-art knowledge
* companies have the burden of proof, not the victim
The battlefield in science:
-
industry hired specialized lawyers, which defence was based on:
* claiming there was no or uncertain scientific knowledge in the past
* critical scientists are biased
* lengthy scientific exchange with lawyer of victim (> 100 pages)
Asbestos and mesothelioma - legal debate
Asbestos is only well-demonstrated cause of mesothelioma (almost 1:1 !)
Thus, debates about fibre types, routes of exposure and risks
Arguments:
-
fibre type: “low exposure to chrysotile present no detectable risk”
*
Bernstein et al Crit Rev Toxicol 2013;43:154-83
-
time at first exposure drives the individual risk, thus, everyone exposed
already at early age through environment and household products
*
Vecchia and Boffetta in Eur J Can Prev 2012;21:227-30
“Continued exposure at working age does not create additional risk”
Asbestos and mesothelioma
Potential attribution: apportionment of exposure history
1. Create a complete exposure history of an exposed worker with relevant sources
of exposure (different jobs & tasks, at home, in environment)
2. Use risk models to estimate risk for each source of asbestos exposure
(based on epidemiological research and risk assessment models)
3. Calculate relative contribution of each source (to the overall 100% !)
Price et al. Mesothelioma: risk apportionment among asbestos exposure sources.
Risk Analysis 2005;25:937-43.
Asbestos and mesothelioma
Apportionment of exposure history: (mis)use in litigation
1. Total cumulative exposure in the job (few) years is much lower than
1. total cumulative exposure in general environment over lifetime
2. Company is most likely not responsible for the critical exposure that has caused
the mesothelioma
Asbestos and disease
From population to individual person (with a disease !)
Levels of evidence for causality of the association (ILO-list Occupational Diseases):
1. Occupational disease
Disease is caused by specific agent, e.g. mesothelioma and asbestos
2. Work-related disease
Disease is multi-factorial, but strong evidence that a particular agent contributes
to the disease, e.g. lung cancer and asbestos
3. Diseases affecting working populations
Increased occurrence of disease in some populations, but association at
individual level is very uncertain, e.g. breast cancer and shift work
Asbestos and lung cancer
Classical epidemiological evidence
Classical study
Relative Risk for lung cancer
No smoking and no asbestos
1
Smoking
11
Asbestos
5
Smoking and asbestos
53
Cuyler Hammond et al. Asbestos exposure, cigarette smoking and death rates. Ann N Y Acad Sc
1979;330:473-90.
Asbestos and lung cancer - legal debate
Asbestos is less important than smoking !
Thus, debates about relative contribution of asbestos exposure and of fibre types
(again)
Arguments:
-
fibre type: “low exposure to chrysotile present no detectable risk”
*
Bernstein et al Crit Rev Toxicol 2013;43:154-83
-
smoking is much more important than asbestos, hence, no compensable
disease
Asbestos and lung cancer
Epidemiological measures of attribution
Population attributable =
proportion of lung cancer cases in the population
fraction
that is attributable to the exposure of interest
p ( RR  1)

p ( RR  1)  1
1
1
Smoking
PAF = 0.89
Asbestos
PAF = 0.55
Smoking and asbestos PAF = 0.94
1
1
Asbestos and lung cancer
Is population attributable fraction useful ?
YES, for public health purposes the best measure (population level!)
NO, not for assessment at individual level
Please note that  PAFs easily exceed 1 (mutual adjustment required):
Smoking
PAF = 0.89
Asbestos
PAF = 0.55
Smoking and asbestos PAF = 0.94
Asbestos and lung cancer
Epidemiological measures of attribution (etiological approach)
Attributable fraction = proportion of lung cancer cases among exposed that is
attributable to the exposure
RR1  1

RR1
Smoking
AF = 0.91
Asbestos
AF = 0.80
Smoking and asbestos AF = 0.98
R1  R0

R1
RR = relative risk, R = disease rate
Asbestos and lung cancer
Is attributable fraction useful ?
YES, for assessment of probability that disease among exposed persons is
caused by that exposure
Please note:
-
 AFs can easily exceed 1
-
AF often depends on the level of (cumulative) exposure, thus,
AF increases with higher exposure
-
how to compare AF of different causes?
Asbestos and lung cancer
Incorrect use of measure of attribution
Probability to die of lung cancer during lifetime
Not exposed to asbestos
8%
Exposed to asbestos (RR=1.35)
10.8%
Risk difference is 10.8% - 8% = 2.8%. Thus, likelihood that lung cancer is caused by
asbestos is 2.8% (= negligible)
Classical mistake: estimation of prior probability to contract a particular disease,
versus estimation of posterior probability whether the disease is caused by
exposure
Asbestos and lung cancer
Legal application of measures of attribution
Casus:
A 65 year old construction worker has been diagnosed with lung cancer. He has
worked with asbestos-cement products. He files a claim for compensation.
Questions:
Which measure of attribution is justified ?
What is interpretation of attribution ?
How do different jurisdictions deal with attribution ?
Asbestos and lung cancer
Use of measure of attribution (yes - no decision)
Cumulative exposure
Relative Risk of lung cancer
25 - 100 fibres/ml-years
2.00
(1997 Helsinki criteria)
By use of lower estimate of 25 fibres/ml-years:
Attributable fraction = (2 - 1) / 2 = 0.50 = 50%
Thus, if cumulative exposure > 25 fibres/ml-years then attribution > 50%, thus,
disease is caused primarily by asbestos
(compensation system in Germany, Belgium, and other countries)
Asbestos and lung cancer
Use of measure of attribution (proportional liability)
Cumulative exposure
Relative Risk of lung cancer
25 - 100 fibres/ml-years
2.00
1. risk = 1 fibre/ml-year will increase risk by 1% (upper estimate)
2. estimate cumulative exposure, eg. 35 fibre/ml-years (shipyard worker)
3. calculate relative risk; 35 * 1% = 35%, thus RR = 1.35
4. Calculate attributable fraction = 1.35 - 1
1.35
Schaier-De Schelde ruling 1999
= 26%
Asbestos and lung cancer
Complications: the role of smoking
1. Multiplicative effect of asbestos and smoking (Riskasbestos * Risksmoking) ?
2. Quitting smoking will reduce risk of lung cancer considerably
(crude estimation is elimination of risk after 20-30 years)
Quitting asbestos exposure will probably also reduce risk of lung cancer
3. Epidemiological evidence is largely based on results among smokers
(up to 1960s, smoking prevalence among men > 80%)
Reid et al. The risk of lung cancer with increasing time since ceasing exposure to asbestos and quitting
smoking. Occup Environ Med 2006;63:509-12.
Sandén et al. The risk of lung cancer and mesothelioma after cessation of asbestos exposure: a
prospective cohort study of shipyard workers. Eur Respir J 1992;5:281-5
Epidemiology and litigation
1. Use of appropriate measure of attribution:
-
attributable fraction
-
relevant for work-related disease
2. Understanding the correct use of the appropriate measure of attribution:
-
risk on disease  disease due to exposure
-
conditional probability (on disease status!)
3. Interpretation of attribution:
-
all or nothing (> 50%)
-
proportional liability
Risk management in the Netherlands
[email protected]