The High Court has allowed an appeal against a decision of the Queensland Court of appeal on negligence and manufacturer’s liability for defective goods in the context of a helicopter crash. The first respondent was severely injured in 2004 when a helicopter manufactured by the appellant crashed due to a loose bolt in the helicopter’s flexplate. The respondents claimed that the maintenance manual provided by the appellants gave inadequate instructions on the method for checking the tightness of the bolts, contrary to the law of negligence and ss 75AD and AE of the Trade Practices Act 1974 (Cth). The trial judge found that the manual was adequate in requiring torque seals to be placed over the bolts and inspected visually after every 100 hours of operation by a Licensed Aircraft Maintenance Engineer for damage to the seal indicating damage to the bolt. A majority of the QCA overturned that finding, holding that the manual was inadequate and should have recommended physically testing the bolt. Before the High Court the appellant sought to contend that the QCA erred in overturning the trial judge’s findings without demonstrating that they were improbable or contrary to compelling inference, that the method suggested by the QCA could be inferred from the manual by a maintenance engineer without it being specified, and that the QCA majority should not have held the appellant liable without considering whether its negligence or defective goods was causative of any loss.

The High Court (French CJ, Bell, Keane, Nettle and Gordon JJ) unanimously allowed the appeal, holding that the QCA erred in interpreting the evidence before the trial judge and erred in overturning the trial judge’s findings. In assessing the trial judge’s five possibilities of the presence and condition of the torque strip or its absence (see at [40]), the QCA concluded that there was only one possible finding: that the torque strip was in a condition that would not alert the engineers or pilots to a need to investigate further (at [42]). The High Court stated that (at [43], citations omitted):

The fact that the judge and the majority of the Court of Appeal came to different conclusions is in itself unremarkable. A court of appeal conducting an appeal by way of rehearing is bound to conduct a ‘real review’ of the evidence given at first instance and of the judge’s reasons for judgment to determine whether the judge has erred in fact or law. If the court of appeal concludes that the judge has erred in fact, it is required to make its own findings of fact and to formulate its own reasoning based on those findings. But a court of appeal should not interfere with a judge’s findings of fact unless they are demonstrated to be wrong by ‘incontrovertible facts or uncontested testimony’, or they are ‘glaringly improbable’ or ‘contrary to compelling inferences’. In this case, they were not. The judge’s findings of fact accorded to the weight of lay and expert evidence and to the range of permissible inferences. The majority of the Court of Appeal should not have overturned them.

On the evidence given at trial by a range of experts, it was open for the trial judge to first infer that a torque stripe was not applied to the bolt when it was incorrectly assembled (the first possibility, at [45]), that in the alternative that if a torque stripe was applied then it would soon be misaligned, which should have been clear to the engineer carrying out the inspection in accordance with the manual (the second, third and fourth possibilities, at [46]–[55]), and also to raise the possibility that the torque stripe had slipped (at [56]–[60]). The Court also noted that, contrary to the submissions of the respondent and the conclusions of the QCA, the trial judge had not made any inconsistent findings in explaining and assessing the alternative possibilities (see [65]–[70]). Consequently, the trial judge was correct to hold that the respondents had not demonstrated that the manual’s contents fell short of what was required to discharge the appellant’s duty of care here (see [78]). While it was not strictly necessary to consider the QCA’s reasoning on causation, the Court did so briefly and concluded that the QCA erred in its analysis of causation: even if it were possible that the torque stripe was incorrectly applied such that it could move with the bolt but not crack (a possibility not covered in the manual), that was not established to be any more likely to have occurred than the other possible problems that were canvassed by the manual, and thus it could not be concluded that failing to provide for that possibility was causative of the crash (see at [81]–[82]).