This article originally appeared in Issue 4 (March 2020).
We are all familiar with the challenge of proving that a given breach of duty was causally relevant to injury and all too often parties, or more specifically their expert witnesses, will revert to statistics to make good their case. After all, save in the most mundane case, it is highly unlikely that the precise factual scenario which forms the subject matter of the litigation will have been replicated elsewhere previously, still less reported. In those circumstances, there is an understandable, if superficial, attraction in relying on a reported statistical association between one event (the breach of duty) and another (the injury). Two recent cases, in which I have acted, have explored, and cast doubt on, the extent to which that approach can, of itself, satisfy the legal test of causation.
In Clements v Imperial College Healthcare NHS Trust  EWHC 2064, the Claimant was less than an hour old when she collapsed on her mother’s breast and suffered severe hypoxic brain damage. Her parents claimed that she had suffocated on her mother’s breast and that they should have been advised by the midwifery staff to keep Cerys’ nose clear and continuously to check that she was still breathing. To corroborate her case on causation, the Claimant relied on published studies which had found a statistical association between sudden unexpected postnatal collapse (‘SUPC’) and the baby being breastfed, or being skin to skin, at the time: the implication of that association being, of course, that the causal mechanism of collapse was suffocation through obstruction of the baby’s nose and mouth on the maternal breast. On that basis, the Claimant invited the court to draw the inference that in every case where, as here, the collapse occurred while the baby was on the mother’s breast, the probable cause was suffocation.
In truth, and as the Defendant sought to demonstrate, the literature was rather more nuanced. While it was true that, in many cases of SUPC, the baby was on the mother’s breast, that was hardly surprising given the high incidence of collapse in the first hour or so following delivery and the nearly universal practice of babies being skin to skin, if not actually breast feeding, at that time. Equally, the literature also reported a significant number of babies collapsing when suffocation could not possibly be the cause.
Having already found for the Defendant on breach of duty, Mrs Justice May went on to reject the Claimant’s case on causation. She found that:
“The precise cause or causes of extremely rare events of sudden untoward postnatal collapse in newborns remain unknown. Although in some cases asphyxiation by lying prone, or up against the mother’s side or breast in skin to skin contact has been posited as a cause, these are not the only hypotheses which have been put forward in the literature; in a significant minority of cases the circumstances in which the baby has been found are suggestive of no apparent cause at all. I have concluded that Cerys’ collapse remains unexplained; the evidence has not demonstrated on the balance of probabilities that she was suffocated by her mother’s breast.”
In other words, even where the Claimant could place her case within a cohort of apparently similar cases where the literature suggested a statistical association between the relevant event and the clinical outcome, she still had to prove that, in her particular case, the mechanism which she postulated was the actual cause of her injury.
Similar issues were explored in AB v East Lancashire Hospitals NHS Trust  EWHC 3542 (QB). The Claimant suffered a perinatal arterial ischaemic stroke (‘PAIS’) at about the time of her birth and sustained devastating injuries as a result. She alleged that she should have been delivered some weeks earlier and that, if she had been, the stroke would have been avoided. The Trust denied any negligence in her antenatal care and said that, even if she had been delivered earlier, she would still have suffered the stroke.
As in Clements, the Claimant relied on a number of case studies which purported to identify a significant statistical association between abnormal placental pathology and PAIS; the causal mechanism postulated in those studies was, plausibly, emboli from the placenta moving through the vascular system and causing a cerebral infarct. Here, it was common ground that the placenta was profoundly abnormal by the time of delivery and the experts even agreed that, if the cause of the PAIS were placental emboli, then earlier delivery would have avoided the same.
The Defendant argued that, while the literature had identified an association between placental dysfunction and PAIS, it had also done so in relation to a host of other factors. Equally, the pathogenesis of PAIS was, as the Claimant’s experts accepted, poorly understood and all of the possible associations remained only that. Crucially, perhaps, the Defendant was able to point to the fact that neither of the Claimant’s experts was able satisfactorily to explain the pathophysiological process by which there could be, as would have been the case here had the cause been placental embolus, a delay of up to nearly two days from the time of the PAIS to its clinical manifestation.
The court dismissed the claim, finding for the Trust on breach of duty and both medical and legal causation. Mrs. Justice Lambert found that:
“…as Mr Whitting submitted, a belief, however strong, of an association or possible causal link between placental pathology and PAIS which has not been scientifically demonstrated is not proof of causation in general and certainly not in the particular case. Likewise, a widely held hypothesis, or presumption, is not proof on the balance of probabilities of cause, or in the legal context, of causation.”
Both cases confirm that a statistical association is not, itself, proof of medical cause or, in a legal context, of causation.
John Whitting QC appeared for the Defendant in both cases.