Product liability litigation and science

United Kingdom

Sir--Your well intentioned proposals to reform the system of compensation for alleged drug-induced injury are indeed "naive" Footnote1. There is a contradiction in your acceptance that science is rarely "clear cut" and your suggestion that the resolution of claims should be based on "expert consensus". Science and law serve different functions and thus have different traditions of reasoning. Science is truth seeking; it is uncertain and is subject to constant revision. Law dispenses justice; it imposes certainty and finality. Causation in science is probabilistic and universal; causation in law is deterministic and particular. Science concerns the uncertainty of truth; law has devised a fiction of certainty: that which is more probable than not, is certain Footnote2. Any determination based on probabilistic evidence is imperfect, but not necessarily "doomed to failure".

The oral contraceptive pill litigation Footnote3. raised a question: "Can x cause y?". It is both a scientific hypothesis and a legal issue of generic causation. The question was answered by assessment of epidemiological evidence in accordance with legal criteria. To resolve the inherent conflict, the law resorted to a transposed conditional based on the doubling of risk: it treated the probability of the data (relative risk of 2, aetiological fraction of 50%) as the probability of the truth of the allegation (balance of probability). Footnote4

The pill claim failed because the court found it lacked merit. Even if the court rejected the Cox's regression analysis, it would have found that the "true" relative risk was 1·7 (ie, under 2). The court stated also: "Epidemiology has been in the van of developing medical knowledge . . . pending fuller biological understanding, epidemiology should be able to light the way to an understanding of factors or exposure, which may cause such conditions to occur."3 These words apply to all claims in which generic causation is in issue, especially when there is no published evidence, whether biological or epidemiological, to support causation.

The court is a lay tribunal and so needs expert assistance. However, it is familiar with scientific notions of causation, and is used to assessment of complex data. Some scientists confuse evidence and hypothesis. The court has to distinguish between mere belief and probative evidence when assessing expert testimony. The court recognises the post hoc ergo propter hoc (after it, so because of it) fallacy, which forms the sole basis of so many drug claims.

The pill case represents yet another expensive failed group action. It accords with the negligible success rate of legally aided pharmaceutical litigation in Britain. Your concerns about such claims are better directed at their funding by legal aid.

CMS Cameron McKenna acted for two of the defendants in the oral contraceptive pill litigation. Anthony Barton Taken from Lancet (16 November 2002) 3 60; 16 12

Footnote 1 Editorial. Epidemiology on trial. Lancet 2002; 360: 421

Footnote 2 Barton A, MacRae K. Causation and statistics. In: Powers M, Harris N, eds. Clinical negligence, 3rd edn. London: Butterworths, 2000.

Footnote 3 XYZ and others -v- Schering HC and others. [2002] EWHC 1420 (QB)

Footnote 4 Kapshandy TE. Proof of causation. Lancet 1992; 339: 876