This LawFlash provides a summary of the judgment of Lord Justice Flaux and Mr. Justice Butcher as it relates to the determination of the prevalence of coronavirus (COVID-19) under a particular policy. The court was somewhat hamstrung due to lack of expert evidence, but provides instructive guidance on how the issue of prevalence should be approached. The judges affirm that certain evidence may be relied upon, which need not be absolutely precise, although how such evidence is to be interpreted will depend on the facts in question.
This LawFlash is the second in a series following the judgment in the FCA’s Business Interruption Test Case. Previously we have discussed the “insured peril” under business interruption (BI) insurance policies, and the interconnected issues of causation and the “Trend Clause.”
PREVALENCE, TYPES OF EVIDENCE, AND METHODOLOGIES
Under policies containing “disease clauses” or “hybrid clauses,” the burden of proof falls with insured parties to prove the presence or occurrence (prevalence) of a notifiable disease (here, COVID-19) within a particular radius specified in the policy. This necessitates evidence of the prevalence of COVID-19 in that particular radius at a particular time. The question thus arises, what evidence can an insured party submit in order to discharge the burden of proof? Must this evidence be taken in isolation, or may a party adduce methodologies to interpret the evidence, in order to approximate actual prevalence?
Whilst the court did not hear expert evidence in relation to evidence of the prevalence of COVID-19 due to the expedited trial, it nevertheless was willing to hear arguments as to the type of evidence the insured may provide in order to satisfy the burden of proof, and arguments as to the assumptions which could be made in connection with such evidence. Two questions were considered: (1) what types of evidence may discharge the burden of proof, and (2) whether such evidence, assuming it is the best evidence, is sufficient in discharging the burden of proof.
Types of Evidence
Four types of evidence are featured in the case:
(i) Specific Evidence – where there is specific evidence of a case in an area, for example, “widespread reports that the care home down the road from the policyholder’s premises has been the location of an outbreak of COVID-19.” The specific evidence is likely to be restricted to a particular location.
(ii) NHS Deaths Data – data on deaths of individuals in hospitals in England that have previously tested positive for COVID-19. These are aggregated at the NHS Hospital Trust Level.
(iii) ONS Deaths Data – data on deaths occurring each week in England and Wales where the death certificate mentions COVID-19. These are aggregated at the level of local authority or health board.
(iv) Reported Cases – lab confirmed positive tests of COVID-19 across the United Kingdom. These are aggregated at national, regional, Upper Tier Local Authority, and Lower Tier Local Authority (LTLA) levels.
How the data is aggregated will determine the particular area that that set of data can be said to cover. The insurers conceded in specific scenarios that some forms of evidence will be sufficient to prove prevalence, such as where the specific region of data (e.g., the area covered by a particular LTLA) coincides completely within the specific policy area. However, there are many situations where the policy area does not correspond conveniently with the relevant region of data, or the data is restricted to a specific period of time. In an attempt to fill these situational gaps, two possible methodologies are accepted in principle.
(i) Averaging Methodologies e.g., “population weighting” – A ratio could be calculated between (1) the population of the policy area, and (2) the population of the relevant region of data, and from this the cases in a particular policy area could be extrapolated. An example best illustrates this methodology: “if there were 100 cases in Area A, which has a population of 50,000, and the relevant policy area, Area B, has a population of 5,000, then the weighted averaging methodology would extrapolate that Area B had 10 cases.”
(ii) Undercounting – This methodology reflects the understanding that the actual prevalence of COVID-19 is inevitably higher than the reported cases. An “undercounting ratio” could be used to determine actual prevalence. This could be calculated in accordance with reliable modelling, where the parties would have to prove the relevance of the modelling to the particular area in question.
The court categorises the above into two issues: inferences that can be drawn from (i) the timing of the presence of COVID-19, and (ii) the particular area in which COVID-19 can be deemed present.
Timing of presence: Inferences may be drawn in relation to a period of time in respect of the particular data in question, during which time COVID-19 could be deemed present. For example, considering the 7-12 day infectious period, an insured should be able to consider the reported cases on a particular date, and those cases on the 2-3 days either side of that date as indicating presence of COVID-19.
Geography of data: While accepting that the averaging methodologies outlined above may be used, the court could not go further than this. Expert evidence would be necessary in each case to determine the merits and weaknesses of each methodology or model relied upon, in relation to the specific area in question. This applies particularly to those cases where there “are more complex intersections between the reporting area and the relevant policy area,” for example varying population densities between areas.
Burden of Proof
Even when assuming the evidence above is the best available evidence, this may not be sufficient in satisfying the burden of proof as it may not be sufficiently relevant to the particular geographical location or timing in the particular case. The case of Equitas Ltd v R&Q Reinsurance Company (UK) Ltd[1] illustrates that in some circumstances, exact and absolute evidence may not be required to satisfy the burden of proof. In the present case, the insurers accept that absolute precision is not required. This does not mean however that the insurers nor the court will accept anything other than reliable and relevant evidence, through expert evidence and strict cross examination.
Thus, the courts are unable to determine whether the burden of proof is satisfied by the evidence and methodologies suggested in the test case, due to the absence of expert evidence and any substantive scrutiny. The judges refrain from the “artificial and potentially unhelpful exercise” of drawing conclusions on whether the burden of proof has been satisfied, as they are only able to rely upon hypothetical evidence, without reference to a particular set of facts. Excluding the concessions already made by the insurers in the particular factual scenarios referred to above, the hope is that the insurers will take a constructive and sensible approach to determine issues of prevalence as and when they arise, with respect to the particular facts at hand. Thus, whether the burden of proof is satisfied will depend upon the exact evidence and interpretive methodologies a party relies upon, and following sufficient scrutiny, a court will be in a position to determine the relevance of that evidence. From there, it will be for the courts to decide if the burden of proof is satisfied.
Trainee Nicholas Woolf contributed to this LawFlash.
[1] [2009] EWHC 2787 (Comm); [2010] 2 All ER (Comm) 855.
[View source.]
