Transcript Causality_Seville_2008 wiki

EAPCCT 2008: Causality Assessment in Poisoning
Causality Assessment
in Poisoning
Essential for Data Quality
Hugo Kupferschmidt, M.D.
Director
Swiss Toxicological Information Centre
Zuerich
Seville, May 8, 2008
XXVIII EAPCCT Congress, Melia Sevilla
Swiss Toxicological Information Centre
1
EAPCCT 2008: Causality Assessment in Poisoning
Overview
 Definitions
 History
 Rationale
 Causality assessment in adverse drug reactions
 Limitations and weaknesses
 Causality in poisoning
 A proposal for standarized causality assessment

in poisoning and drug overdose
Discussion
Swiss Toxicological Information Centre
2
EAPCCT 2008: Causality Assessment in Poisoning
Definitions
Causality assessment
 is the evaluation of the likelyhood that a
particular event (exposure) is the cause of an
observed effect.
 investigates the relationship between the
exposure and the occurrence of an effect.


is an important component of pharmaco- and
toxicovigilance
contributes to better evaluation of risk-benefit
profiles
Auriche M et al. Drug Saf 1993; 9: 230-5
Edwards IR et al. Drug Saf 1994; 10: 93-102
Meyboom RHB et al. Drug Saf 1997; 17: 374-89
Agbabiaka TB et al. Drug Saf 2008; 31: 21-37
3
Swiss Toxicological Information Centre
EAPCCT 2008: Causality Assessment in Poisoning
Rationale
Source of data on human poisoning
 Prospective cohort studies and RCTs are still
lacking for most questions and aspects in clinical
toxicology.


Poisons Centre data remain an important and
sometimes unique source of information, particularly on rare kinds of poisoning.
Whereever prospective cohort studies and RCTs
are not to be expected in the future, there is an
obligation for Poisons Centres and clincal toxicologists to collect data on such cases accurately,
carefully, and as completely as possible.
Brent J. Clin Toxicol 2005; 43: 881-6
Swiss Toxicological Information Centre
4
EAPCCT 2008: Causality Assessment in Poisoning
Rationale
Quality of data on human poisoning
 exposure uncertain
 no experimental setting
 by history only (patient‘s, bystanders‘)
 supported by the observed toxic effect

Having a measure on the likelyhood of exposure would be a substantial improvement of
the data quality.
 information about the exposure itself
 assessment of the toxic effect in the view of the
exposure (causality)
Swiss Toxicological Information Centre
5
EAPCCT 2008: Causality Assessment in Poisoning
Rationale
Causality assessment is necessary
 for statistical purposes
 for epidemiological purposes
 for toxicology databases
 for publication (case reports and case series)
 for the generation of data on prior probabilities
for Bayesian statistics in the diagnostic process
Whyte IM. Clin Toxicol 2002; 40: 211-2
Whyte IM et al. Clin Toxicol 2002; 40: 223-30
Swiss Toxicological Information Centre
6
EAPCCT 2008: Causality Assessment in Poisoning
Rationale
Link between severity grading and causality assessment
 The EAPCCT (together with the IPCS and the
European Commission) has developed a
standard severity grading system, the PSS.

Severity grading implies that the symptoms
described are related to the toxic exposure (i.e.
there is a causal relationship between these
symptoms and the exposure)

It is nothing than consequent now to continue in
agreeing on a standard system of causality
assessment.
Swiss Toxicological Information Centre
7
EAPCCT 2008: Causality Assessment in Poisoning
Standardisation
Standardized causality assessment
 is aimed at decreasing ambiguity of the data


plays a key role in data exchange
limits the drawing of erroneous conclusions
... is therefore a major factor of data quality.
Meyboom RHB et al. Drug Saf 1997; 17: 374-89
Agbabiaka TB et al. Drug Saf 2008; 31: 21-37
Swiss Toxicological Information Centre
8
EAPCCT 2008: Causality Assessment in Poisoning
History
Sir Austin Bradford Hill (1965):
 Strength of the association
 Consistency of the observed association
 Specificity
 Temporality (chronology)
 Biological gradient
 Plausibility
 Coherence
 Experiment
 Analogy
Hill AB. Proc R Soc Med 1965; 85: 295-300
Swiss Toxicological Information Centre
9
EAPCCT 2008: Causality Assessment in Poisoning
Causality in ADR
Assessment of causality is routine in pharmacovigilance (spontaneous reporting).
Categories of methods
 Opinion of experts, clinical judgement or
global introspection
(n=4; 12%)
 algorithms or standardized assessment
methods
(n=26; 76%)
 Probabilistic or Bayesian approaches (12%)
Meyboom RHB et al. Drug Saf 1997; 17: 374-89
Agbabiaka TB et al. Drug Saf 2008; 31: 21-37
Arimone Y et al. Eur J Clin Pharmacol 2005; 61: 169-73
Swiss Toxicological Information Centre
10
EAPCCT 2008: Causality Assessment in Poisoning
Causality in ADR
Determining factors in causality assessment
 temporal sequence (chronology, temporality)
 time to onset
 previous information on the drug
 background epidemiological and clinical information
 dose relationship (e.g. overdoses)
 response pattern
 characteristics and mechanisms of the ADR
 rechallenge - dechallenge
 alternative aetiologies (differential diagnoses)
 concomitant drugs
 analytical confirmation
Wiholm BE. Drug Inf J 1984; 18: 267-9.
Agbabiaka TB et al. Drug Saf 2008; 31: 21-37
11
Swiss Toxicological Information Centre
EAPCCT 2008: Causality Assessment in Poisoning
Causality in ADR
Classification of events, degrees of causality
 definite / confirmed / certain
 causative
 probable / likely
 possible
 non-assessable / unclassifiable
 unclassified / conditional
 unlikely / coincidental / doubtful / remote /
unlikely

exclude / negative / unrelated
Wiholm BE. Drug Inf J 1984; 18: 267-9.
Agbabiaka TB et al. Drug Saf 2008; 31: 21-37
12
Swiss Toxicological Information Centre
EAPCCT 2008: Causality Assessment in Poisoning
Causality in ADR
Limitations and weaknesses
 High inter-rater variability:








Miremont (1994): Physicians tend to assign very high scores
to suspected ADRs. Agreement of methods: 6%
Blanc (1979): Overall inter-rater agreement on a VAS was
low (κ=0.20).
Some depend grossly on raters‘ knowledge
Some are organ-specific
No „gold standard“ algorithm
Not all suitable to assess drug-drug interactions
Either superficial or very time-consuming
Data to compute prior odds often unavailable
Miremont G et al. Eur J clin Pharmacol 1994; 46: 285-9
Blanc S et al. Clin Pharmacol Ther 1979; 25: 493-8
Arimone Y et al. Eur J Clin Pharmacol 2005; 61: 169-73
Benahmed S et al. Eur J Clin Pharmacol 2005; 61: 537-41
13
Swiss Toxicological Information Centre
EAPCCT 2008: Causality Assessment in Poisoning
Causality in ADR
Consequences of limitations and weaknesses
 None of the assessment systems has ever been validated


(i.e. shown to consistently and reproducibly produce a fair
approximation of the truth).
Causality assessment has therefore limited scientific
value. It neither eliminates nor quantifies uncertainty but,
at best, categorises it in a semiquantitative way.
Standardized causality assessment has not been able to
neutralize the inherent limitations of spontaneous reporting systems (i.e. uncertainty regarding the causal involvement of the drug, and underreporting).
Meyboom RHB et al. Drug Saf 1997; 17: 374-89
Swiss Toxicological Information Centre
14
EAPCCT 2008: Causality Assessment in Poisoning
Causality in ADR
Perspecvtives (2008)
 The idea of creating standardized causality assessment



systems to provide reliable and reproducible measures of
the relationship-likelihood in suspected cases of ADR
seems unfeasible, since no single method has achived
this to date.
The differences in ADR causality criteria and the unavoidable subjectivity of judgements may be responsible for
the lack or reproducibility of most methods.
So far, no ADR causality assessment method has shown
consistent and reproducible measurement of causality.
Therefore, no single method is universally accepted.
Agbabiaka TB et al. Drug Saf 2008; 31: 21-37
15
Swiss Toxicological Information Centre
EAPCCT 2008: Causality Assessment in Poisoning
Causality in Poisoning
Differences and similarities to adverse drug reactions
 Spontaneous reporting similar in pharmacovigilance and




Poisons Centres
Incomplete data frequent
Uncertainty of exposure more important in poisoning
Uncertainty of dose and differential diagnoses more
important in poisoning
Concept of dechallenge and rechallenge not feasible in
toxicology
Swiss Toxicological Information Centre
16
EAPCCT 2008: Causality Assessment in Poisoning
Causality in Poisoning
Confirmation system by von Clarmann (1982)
Level of presumptive
evidence
Toxin
1
Exposure
1
Effect
1
Level of confirmation
Toxin
1
Exposure
1
Effect
1
Level of independent
confirmation
Toxin
1
Exposure
1
Effect
1
Additional evidence
Exclusion of
other causes: 1
→ Score: 1-10
von Clarmann M. Rote Liste 1982, p. 95-6
Swiss Toxicological Information Centre
17
EAPCCT 2008: Causality Assessment in Poisoning
Exposure-Effect Relationship
1. Exposure assessment
 confirmed
 likely
 unlikely
particularly important
in asymptomatic cases
2. Causality assessment
 likely
 unlikely
 conditional
 none
 not assessable
Swiss Toxicological Information Centre
feasible only
in symptomatic cases
18
EAPCCT 2008: Causality Assessment in Poisoning
Likelihood of exposure
Exposure is...
 confirmed
if...
analytical detection of substance
(= objective measure)
 likely
observed exposure by others

reliable history from patient
realiable
 possible
indirect evidence of exposure

no evidence of exposure
unlikely
 no exposure
excluded by negative analytics
Swiss Toxicological Information Centre
19
EAPCCT 2008: Causality Assessment in Poisoning
Degrees of Causality
A causal relationship between exposure and effect is...
 likely
adequate chronology
typical or expected symptoms
no other causes

possible
adequate chronology
typical symptoms but possible other
causes

conditional
adequate chronology
atypical symptoms and
no other cause
 unlikely

no adequate chronology and/or
atypical symptoms
other causes present
not assessable no symptoms, insufficient information
Swiss Toxicological Information Centre
20
EAPCCT 2008: Causality Assessment in Poisoning
Proposed Algorithm
temporal sequence
adequate?
NO
unlikely
(toxicokinetics!)
YES
NO
effect typical/expected?
NO
described in literature or
pharmacology (mechanism)
YES
other causes
absent or unlikely?
other causes
absent / unlikely
YES
NO
possible
YES
likely
none
Swiss Toxicological Information Centre
conditional
„new effect“
21
EAPCCT 2008: Causality Assessment in Poisoning
10 Year Experience
Swiss Toxicol. Information Centre (1997-2006)
Degree of causality
confirmed
likely
possible
conditional
unlikely
not assessable
none
No.
4875
27680
2095
435
970
1844
941
Percent
10.1%
57.5%
4.3%
0.9%
2.0%
3.8%
2.0%
S.D.
0.8%
0.5%
2.1%
3.2%
1.8%
3.2%
1.7%
asymptomat
TOTAL
9247
48162
19.2%
0.8%
Swiss Toxicological Information Centre
22
EAPCCT 2008: Causality Assessment in Poisoning
Discussion



Consequent causality assessment in Poisons
Centres to their cases should be added to standard features of data handling.
One important requirement would be routine
collection of follow-up data.
A Bayesian approach would be preferable, but is
unrealistic as the effort to obtain and calculate
the prior odds would be immense. Furthermore
these prior odds would not necessarily be applicable to different geographical places. Therefore
an algorithm-based approach may be more feasible.
Buckley NA et al. Clin Toxicol 2002; 40: 213-22
Swiss Toxicological Information Centre
23
EAPCCT 2008: Causality Assessment in Poisoning
Discussion (2)



A long as prospective studies and RCTs are not
available in certain fields of clinical toxicology,
Poisons Centre data remain important sources
of information.
This does not mean that not every effort should
be taken to perform such trials.
Collecting data in Poisons Centres must not be
a reason to prevent or impede efforts to perform
high quality research.
Greller HA. Clin Toxicol 2004; 42: 129-30
Buckley NA et al. Lancet 1996; 347: 1167-9
Whyte IM. Clin Toxicol 2002; 40: 211-2
Swiss Toxicological Information Centre
24
EAPCCT 2008: Causality Assessment in Poisoning
Discussion (3)



Causality assessment in Poisons Centres has
its place mainly for the generation of „epidemiological“ data, and for hypothesis generation,
rather than data on treatment effects.
Causality assessment will be a necessity for
common data collection.
Only cases with sufficient causality (i.e. a likely
relationship between exposure and effect)
should be reported or published.
Swiss Toxicological Information Centre
25
EAPCCT 2008: Causality Assessment in Poisoning
Finis
[email protected]
Swiss Toxicological Information Centre
26
EAPCCT 2008: Causality Assessment in Poisoning
References
1. Agbabiaka TB, Savovic J, Ernst E. Methods for causality assessment of adverse drug
reactions. A systematic review. Drug Saf 2008; 31: 21-37.
2. Hill AB. The environment and disease: association or causation? Proc R Soc Med 1965;
85: 295-300.
3. Wiholm BE. the Swedish drug-event assessment methods. Special workshop – regulatory. Drug Inf J 1984; 18: 267-9.
4. Miremont G, Haramburu F, Bégaud B, Péré JC, Dangoumau J. Adverse drug reaction:
Physician‘s opinions versus a causality assessment method. Eur J Clin Pharmacol
1994; 46: 285-9.
5. Blanc S, Leuenberger P, Berger JP, Brooke EM, Schelling JL. Judgements of trained
observers on adverse drug reactions. Clin Pharmacol Ther 1979; 25: 493-8.
6. Karch FE, Lasagna L. Towards the operational identification of adeverse drug reactions.
Clin Pharmacol Ther 1977; 21: 247-54.
7. Kramer MS, Leventhal JM, Hutchinson TA, Feinstein AR. An algorithm for the
operational assessment of adverse drug reactions: I. Background, description, and
instructions for use. JAMA 1979; 242: 623-32.
Swiss Toxicological Information Centre
27
EAPCCT 2008: Causality Assessment in Poisoning
References
8.
Benahmed S, Picot MC, Hillaire-Buys D, Blayac JP, Dujols P, Demoly P. Comparison
of pharmacovigilance algorithms in drug hypersensitivity reactions. Eur J Clin
Pharmacol 2005; 61: 537-41.
9.
Brent J. 2005 Louis Roche Lecture. Professional societies and evidence-based
clinical toxicology. Delivered at the XXV International Congress of the EAPCCT, Berlin,
Germany. Clin Toxicol 2005; 43: 881-6.
10. Greller HA. How to position our practice. Clin Toxicol 2004; 42: 129-30.
11. Isbister GK. Data collection in clinical toxinology: Debunking myths and developing
diagnosic algorithms. Clin Toxicol 2002; 40: 231-7.
12. Buckley NA, Whyte IM, Dawson AH. Diagnostic data in clinical toxicology – Should
we use a Bayesian approach? Clin Toxicol 2002; 40: 213-22.
13. Buckley NA, Karalliedde L, Dawson A, Senanayake N, Eddleston M. Where is the
evidence for treatments used in pesticide poisoning? Is clinical toxicology fiddling
while the developing world burns? J Toxicol Clin Toxicol 2004; 42: 113-6.
14. Whyte IM. Introduction: Research in clkinical toxicology – The value of high quality
data. Clin Toxicol 2002; 40: 211-2.
Swiss Toxicological Information Centre
28
EAPCCT 2008: Causality Assessment in Poisoning
References
15. Whyte IM, Buckley NA, Dawson AH. Data collection in clinical toxicology: Are there
too many variables? Clin Toxicol 2002; 40: 223-30.
16. Hoffman RS. Dies consensus euqal correctness? Clin Toxicol 2000; 38: 689-90.
17. Buckley NA, smith AJ. Evidence-based medicine in toxicology: Where is the
evidence? Lancet 1996; 347: 1067-9.
18. Arimone Y, Bégaud B, Miremont-Salamé G, Fourrier-Réglat A, Moore N, Molimard M,
Harambouru F. Agreement of expert judgement in causality assessment of adverse
drug reactions. Eur J Clin Pharmacol 2005; 61: 169-73.
19. Meyboom RHB, Hekster YA, Egberts ACG, Gribnau FWJ, Edwards IR. Causal or
casual? The role of causality assessment in pharmacovigilance. Drug Saf 1997; 17:
374-89.
20. Auriche M, Loupi E. Does proff of causality ever exist in pharmacovigilance? Drug Saf
1993; 9: 230-5.
21. Edwards IR, Biriell C. Harmonisation in pharmacovigilance. Drug Saf 1994; 10: 93102.
Swiss Toxicological Information Centre
29
EAPCCT 2008: Causality Assessment in Poisoning
References
22. Meyboom RHB, Egberts ACG, Edwards IR, Hekster YA, de Koning FHP, Gribnau
FWJ. Principles of signal detection in Pharmacovigilance. Drug Saf 1997; 16: 355-65.
23. Neubert A, Dormann H, Weiss J, Criegee-Rieck M, Ackermann A, Levy M, Brune K,
Rascher W. Are computerized monitoring systems of value to improve
pharmacovigilance in pediatric patients? Eur J Clin Pharmacol 2006; 62: 959-65.
24. Hauben M, Reich L, Gerrits CM, Younus M. Illusions of objectivity and a
recommendation for reporting data mining results. Eur J Clin Pharmacol 2007; 63:
517-21.
25. von Clarmann M. Rote Liste 1982. p. 95-6.
* * *
Swiss Toxicological Information Centre
30
EAPCCT 2008: Causality Assessment in Poisoning
A.B. Hill on causality (1965)
1. Strength of the association:
 Size of the effect

Examples:
 Scrotal cancer from soot exposure in chimney


sweepers (Pott P, 1775)
Mortality from lung cancer in smokers
Mortality from cholera (Snow J, London 1855)
Hill AB. Proc R Soc Med 1965; 85: 295-300
Swiss Toxicological Information Centre
31
EAPCCT 2008: Causality Assessment in Poisoning
A.B. Hill on causality (1965)
2. Consistency:
 Has the effect been observed repeatedly?
By different persons, in different places,
circumstances and times?
Hill AB. Proc R Soc Med 1965; 85: 295-300
Swiss Toxicological Information Centre
32
EAPCCT 2008: Causality Assessment in Poisoning
A.B. Hill on causality (1965)
3. Specificity:
 The association is limited to specific exposures and the disease shows specific
features.
Hill AB. Proc R Soc Med 1965; 85: 295-300
Swiss Toxicological Information Centre
33
EAPCCT 2008: Causality Assessment in Poisoning
A.B. Hill on causality (1965)
4. Temporality:
 „Which is the cart and which is the horse?“
Hill AB. Proc R Soc Med 1965; 85: 295-300
Swiss Toxicological Information Centre
34
EAPCCT 2008: Causality Assessment in Poisoning
A.B. Hill on causality (1965)
5. Biological gradient:
 Dose-response effect
Hill AB. Proc R Soc Med 1965; 85: 295-300
Swiss Toxicological Information Centre
35
EAPCCT 2008: Causality Assessment in Poisoning
A.B. Hill on causation (1965)
6. Plausibility:
 Is the causation biologically plausibe?

Limitations
 „But this is a feature we cannot demand. What is

biologically plausible depends on the biological
knowledge of the day.“
„When you have eliminated the impossible,
whatever remains, however improbable, must be
the truth“ (Sherlock Holmes to Dr. Watson)
Hill AB. Proc R Soc Med 1965; 85: 295-300
Swiss Toxicological Information Centre
36
EAPCCT 2008: Causality Assessment in Poisoning
A.B. Hill on causation (1965)
7. Coherence:
 „... the cause-and effect interpretation [...]
should not seriously conflict with the generally
known facts of the natural history and biology
of the disease ...“
Hill AB. Proc R Soc Med 1965; 85: 295-300
Swiss Toxicological Information Centre
37
EAPCCT 2008: Causality Assessment in Poisoning
A.B. Hill on causation (1965)
8. Experiment:
 Does experimental evidence support the
cause-and-effect interpretation of our
observation?

Example:
 Reduction of the effect after the introduction of
preventive measures
Hill AB. Proc R Soc Med 1965; 85: 295-300
Swiss Toxicological Information Centre
38
EAPCCT 2008: Causality Assessment in Poisoning
A.B. Hill on causation (1965)
9. Analogy:
 Similar effects in similar situations and after
similar exposures
Hill AB. Proc R Soc Med 1965; 85: 295-300
Swiss Toxicological Information Centre
39