PRODUCT liability law has struggled to develop a test for identifying when products are defective under the Consumer Protection Act 1987 (“CPA”). In Wilkes v Depuy International Ltd. [2016] EWHC 3096 (QB), Hickinbottom J. offered the most prolonged reflection on product defect since A v National Blood Authority [2001] EWHC 446 (QB), and rejected much of the framework of NBA. However, Wilkes provides little guidance regarding when products should be identified as being defective, reinforcing the need for a more deeply grounded approach.
The claimant in Wilkes suffered a fracture of an artificial hip replacement three years after its surgical implantation. The fracture occurred due to mechanical fatigue (at [109]). Though the parties disputed which variables contributed to the fatigue (at [110]–[111]), Hickinbottom J. concluded that the account of the failure incorporating a broader range of factors (including, for example, the weight of the patient) was more convincing (at [112]).
The claimant sought damages under the CPA from the manufacturer of the hip replacement, alleging the device's design posed too great a risk of fracture (at [116]). The disputed provision of the CPA, s. 3(1), identifies a defect wherever “the safety of the product is not such as persons generally are entitled to expect”. Fault or misconduct by the producer is irrelevant, and the enquiry is wholly concerned with consumer expectations regarding the product's safety (at [63]–[65]). In concluding that the hip replacement was not defective, Hickinbottom J. asserted that risk-benefit analysis is central to determining such expectations. Such a balancing assessment is particularly relevant where a product can only possess certain beneficial features if it also possesses prospectively harmful attributes, including an increased risk of failure (at [125]). The case indicates that consumers who make a risk-benefit trade-off in selecting a product ought to enjoy (or bear) the consequences of their decision (at [66], [70]).
Wilkes suggests that the status of products as defective (or non-defective) is in part a function of consumer preference. By this logic, if consumers are aware that, even when produced to specification, a product may inflict harm, and still use it, it strongly indicates that the product should not be deemed defective. Other cases suggest that consumer preference can play a critical role in ascertaining if a product is defective under the CPA; for example, because consumers enjoy hot beverages, that a hot beverage can harm by scalding does not make it defective. B. v McDonald's Restaurant Ltd. [2002] EWHC 490 (at [80]). However, relying on consumer behaviour to identify whether a product is defective threatens to exacerbate the circularity of defectiveness as a concept (at [68]). A test with such character further forces judges to make ad hoc, product-by-product determinations regarding the rational conduct of hypothetical consumers.
Wilkes rejects as unhelpful much of the reasoning in NBA which provides firmer structure. While acknowledging that failure to satisfy design specifications may be relevant to determining whether a product is defective, Hickinbottom J. deemed that “the categorisation of defects into ‘standard’/‘non-standard’, as a classification, is unnecessary and undesirable”. (at [94]). However, the standard/non-standard distinction could streamline assessing consumer expectation, since consumers presumably do not anticipate receiving substandard products. Hickinbottom J. also rejected Burton J.’s suggestion in NBA that “avoidability” lies outside the scope of the defect analysis, instead concluding that avoidability is best treated as “a matter of degree” (at [86]). Excluding avoidability as a factor prevents intrusion of producer fault into the product-oriented consideration of safety (at [85]; compare NBA at [50]). Yet Hickinbottom J. preferred to retain avoidability as a factor to make the risk-benefit analysis as comprehensive as possible (at [88]).
Even if this substantive criticism of NBA is valid, Wilkes deconstructs the most well-established framework while advancing no helpful alternative. Wilkes thus exemplifies the lack of definitiveness in assessment of product liability defects. For example, in Tesco Stores v Pollard [2006] EWCA Civ 393, Laws L.J.’s conclusion that a non-defective screw-top bottle was within what “persons were generally entitled to expect” was an almost purely factual judgment (at [18]). McDonald’s legal conclusion was likewise a granular factual assessment that consumers expect beverages to be hot (at [80]).
Wilkes thus formalises a test that is so “flexible” (at [78], [96]), and that considers so many circumstances (for example, compliance with regulatory standards) as contributory but non-determinative, that courts have little meaningful guidance. The emphasis upon risk-benefit analysis is characteristic of this unstructured multi-factorial approach, as courts must simply weigh competing factors intuitively.
Following Wilkes, an assessment of product defect resembles an inverted assessment of breach in negligence: rather than asking if the tortfeasor behaved unreasonably, the court enquires if reasonable consumers would use the product. This approach encounters significant problems. It deviates from the ostensible intention of the CPA and the underlying EC Council Directive 85/374/EEC (the “Directive”) to protect consumers through a strict liability regime. It also creates a puzzle of statutory interpretation, given that CPA, s. 5(1) suggests any person harmed by a product may recover, as opposed to merely the purchaser or consumer, and s. 3 refers to the safety expectations of persons, not merely purchasers. Yet it is purchasers (or, in the case of sophisticated medical products, their professional intermediaries (at [107]–[108]), who are positioned to perform the risk-benefit analysis of a product. Wilkes’ formula thus inadvertently privileges the expectations of purchasers, rather than, as the statute dictates, considers the expectations of safety from persons generally. Finally, given the thin case law on what qualifies as a defective product under the CPA, producers have little guidance on when they can expect their products to be deemed defective.
Comparing the factual contexts of Wilkes and NBA, however, suggests that there may be a deeper unspoken logic driving the differing outcomes. The product defect in NBA – tainting of transfused blood – was such that conduct by the victim could not have mitigated the harmful effects, and no reasonable person would expect to receive tainted blood without warning. Conversely Wilkes is rife with allusions to factors under consumer control that increased the (unavoidable) possibility of product failure. If treated purely as a question of product defect, however, the legal outcomes should have been the same – both recipients of blood transfusions and those who undergo medical device implantations might be aware of some risk, but also would expect the product to inflict no harm if it performs as expected. The real question is where the costs of product failure should fall. Wilkes suggests that if consumers can make predictive choices regarding outcomes from the use of the product, individual harmed consumers should bear the cost of risks associated with the product.
While employing reasonableness of consumer choice to query the defectiveness of products is alien to the relevant legislation, such a consideration can be incorporated into a firmer legal test that honours the intentions of the CPA and the Directive. Any appropriate test for defectiveness must allow recovery for harm caused by products as a strict matter. To craft a test that cleaves to this feature of product liability while incorporating information from consumer behaviour, it is helpful to enquire into the consequence of classifying an instance of harm as flowing from a defect: it distributes the economic effect of the instance of harm across all consumers of the product, rather than placing it all upon any person unfortunate enough to suffer harm. Such a spreading of consequences occurs when, in response to being held liable for harm caused by a defective product, a producer either modifies its practices to avoid the defect or raises the cost of the product (or, if the defect cannot be cured in a cost-effective manner, withdraws it from the market, either by choice or by producer bankruptcy). Identification of a “defect”, therefore, should be understood as a cost-spreading measure akin to producer-sponsored insurance. It ensures that a person who loses the game of “Russian roulette” (NBA at [65]) when harmed by a product does not solely bear the effects of such misfortune.
Courts, therefore, should decide if a product is defective by enquiring if it is appropriate to insulate a person who happens to be harmed by a product from his or her bad luck. This approach synthesises the virtues of NBA (its provision of analytical structure to courts and parties) and Wilkes (its simplicity and lack of superfluous categories), remains faithful to the CPA and the broader purposes of contemporary product liability, and provides courts with clearer guidance in making judgments about product defects. It serves the intention of the CPA by focusing on the nature of the product rather than the conduct of the parties, yet can incorporate considerations such as a consumer's role in bringing about the harm (as was apparently at play in Wilkes). Most importantly, it gives courts a touchstone regarding when a product should be deemed defective, thus facilitating consistency in the law and providing an alternative to frustratingly particularised judgments.
