ACCIDENT COMPENSATION CORPORATION v NG AND “L” [2020] NZCA 274

ACCIDENT COMPENSATION CORPORATION v NG AND "L" [2020] NZCA 274 [7 July 2020]NOTE: DISTRICT COURT ORDER PROHIBITNG PUBLICATION OF THENAME AND IDENTIFYING DETAILS OF THE RESPONDENT L REMAINSIN FORCE. SEE [2017] NZACC 147.IN THE COURT OF APPEAL OF NEW ZEALANDI TE KŌTI PĪRA O AOTEAROACA125/2019[2020] NZCA 274BETWEEN ACCIDENT COMPENSATIONCORPORATIONAppellantAND BRENDA NG AND "L"RespondentsHearing: 29 April 2020Court: Kós P, French and Goddard JJCounsel: H B Rennie QC, A S Butler and N J Fenton for AppellantB J Peck, C L Kerr and E M Bransgrove for RespondentsJudgment: 7 July 2020 at 3 pmJUDGMENT OF THE COURTA The appellant's application for leave to adduce the affidavit evidence ofSpencer Beasley, Richard North, Andrew Connolly and Jonathan Wheeleris granted.B The respondents' application for leave to adduce the affidavit evidence ofFrances Matthews is granted.C The respondents' application for leave to adduce the affidavit evidence ofRuth Dyson is declined.D We answer the question of law submitted for determination by this Court:Was the High Court correct to conclude that the word "ordinary"in s 32(1)(c) of the Accident Compensation Act 2001 referring to an"ordinary consequence" means a consequence that is moreprobable than not?Answer: NoE The appeal is allowed and the decision of the High Court quashed.F The claims of the respondents are remitted to the District Court forreconsideration in accordance with this judgment.____________________________________________________________________REASONS OF THE COURT(Given by French J)[1] What does the phrase an "ordinary consequence" mean for the purposes ofthe treatment injury regime under s 32 of the Accident Compensation Act 2001?[2] In the High Court, Churchman J held that "an ordinary consequence" wasa consequence that was more probable than not.1 Or to put it another way, aconsequence that has a 50 percent or greater chance of occurring. Conversely, anyconsequence below that threshold, that is up to 50 per cent likely to occur, was not anordinary consequence.[3] The Accident Compensation Corporation (ACC) now challenges thatinterpretation by way of the case stated procedure.2 The issue is an important onebecause under s 32 cover for injury caused by medical treatment is not available ifthe injury was an ordinary consequence of the treatment.1 Accident Compensation Corp v Ng [2018] NZHC 2848 at [102]–[103] [High Court decision].2 Accident Compensation Act 2001, s 163(1).BackgroundThe legislative history[4] The legislative history leading to the enactment of s 32 featured inthe submissions of both parties and it is therefore helpful to provide a brief summary.[5] As originally enacted in 1972,3 New Zealand's no-fault accident compensationscheme provided cover for "personal injury by accident" without any specificreference to the adverse outcomes of medical treatment. Then shortly after itsintroduction, the legislation was amended by adding "Medical, surgical, dental, or firstaid misadventure" as an illustrative category of personal injury by accident but withoutany further definition.4[6] The scope of what came within "medical misadventure" was thus left tothe Courts to develop.[7] In the context of a scheme designed to provide cover for personal injury byaccident but not sickness or disease, the Courts recognised that some unexpected oraccident like event was required before there could be said to have been medicalmisadventure. "Medical misadventure" was held to encompass two concepts, medicalerror and medical mishap.5 The concept of medical error was familiar and needed noexplanation. As to what amounted to medical mishap and thus medical misadventure,this Court in Childs v Hillocks identified three key propositions:6(a) A totally unforeseen adverse consequence of medical treatment ismedical misadventure.(b) An adverse consequence of such treatment which is within the normalrange of medical or surgical failure attendant upon such treatment is notmedical misadventure.3 Accident Compensation Act 1972.4 Accident Compensation Amendment Act 1974, s 2(a)(ii). See the helpful discussion in StephenTodd Treatment Injury in New Zealand (2011) 86 Chi Kent L Rev 1169 at 1186.5 See Green v Matheson [1989] 3 NZLR 564 (CA); and Childs v Hillock [1994] 2 NZLR 65 (CA)at 72.6 Childs v Hillock, above n 5, at 72.(c) An adverse consequence of such treatment which is outside the normalrange of medical or surgical failure attendant upon such treatment ismedical misadventure.[8] The two-pronged approach developed by the Courts was subsequently adoptedby the legislature in the Accident Rehabilitation and Compensation Insurance Act1992 (the 1992 Act). Section 5 of that Act provided that "medical misadventure"meant "medical error or medical mishap".[9] "Medical error" was defined in the 1992 Act as the failure of a registered healthprofessional to observe a standard of care and skill reasonably to be expected inthe circumstances. The definition further provided that it was not medical error solelybecause desired results are not achieved or because subsequent events show differentdecisions might have produced better results.7[10] "Medical mishap" was defined as an adverse consequence of treatment by aregistered health professional properly given if the likelihood of the adverseconsequence of the treatment occurring was rare and the adverse consequence wassevere.8 What was meant by "rare" for this purpose was defined in the followingterms:9the likelihood that treatment of the kind that occurred would have theadverse consequence shall be rare only if the probability is that the adverseconsequence would not occur in more than 1 percent of cases where thattreatment is given.Where the likelihood that an injury would occur is in the ordinary course rare,but is not rare having regard to the circumstances of the particular person, itshall not be medical mishap if the greater risk to the particular person injuredwas known to that person prior to the treatment.[11] These definitions were continued in substantially similar terms in two lateriterations of the accident compensation legislation, namely the Accident Insurance Act1998 and the Injury Prevention, Rehabilitation, and Compensation Act 2001.107 Accident Rehabilitation and Compensation Insurance Act 1992, s 5(1).8 Section 5(1).9 Section 5(2) and (3).10 Accident Insurance Act 1998, ss 35–37; and Injury Prevention, Rehabilitation, and CompensationAct 2001, ss 32–34.[12] However, in early 2002, the then Minister for ACC initiated a review of thecover provisions for medical misadventure. The aim of the review was to identify andaddress issues of unfairness and inconsistency. A consultation document was dulyprepared, and submissions sought from stakeholders and the general public.11[13] Following the consultation process, the Office of the Minister for ACCprovided a Conclusions and Recommendations paper to the Cabinet SocialDevelopment Committee.12 The paper identified the existence of core problemsrelating to the cover definitions. In particular, it was stated that the medical mishapcriteria (rare and severe) were too arbitrary, often bearing little relation tothe circumstances of the patient and resulting in claimants unfairly missing out oncover.13 As regards medical error, the chief concerns were that requiring claimants toestablish fault was contrary to the ethos of a non-fault system and was hinderingthe claims process because of the attendant costs and delay.14[14] The consultation document proposed various options for reform, all of whichit was said would have the effect of slightly increasing the numbers of New Zealanderseligible for cover. The options included retaining the current legislative frameworkbut replacing the rarity threshold with an individualised risk assessment and no longerrequiring attribution of error to individual health practitioners. Another optionidentified was to provide cover for injuries in the treatment process that werepreventable. The third option was to amend the legislation by replacing the conceptof medical misadventure with a new "treatment injury" regime providing cover forunintended injury in the treatment process. The cost to the scheme of the increase incover resulting from option three was estimated to be $8.69 million per annum.15[15] It was this third option that found favour with the government of the day andin August 2004 the Injury Prevention Rehabilitation and Compensation AmendmentBill (No 3) 2004 creating a new treatment injury regime was introduced into the11 Lisa Rice Review of ACC Medical Misadventure — Consultation Document (AccidentCompensation Corporation and Department of Labour, 2003).12 Medical Misadventure Review — Conclusions and Recommendations (Office of the Minister forACC, 2004).13 At [4] and [22].14 At [4] and [27].15 At [69].House. The Bill's explanatory note recorded that the new regime would mean thearbitrary decisions required under rare and severe criteria would no longer exist.[16] The Bill did not originally contain any reference to "ordinary consequences".Instead, the relevant provision excluded from cover any injury that was an "anticipatedpart, or consequence of, the treatment".16 Following submissions on the Bill,the Department of Labour recommended that the Select Committee reconsider the useof "anticipated".17[17] According to ACC, this advice was prompted by numerous submissions fromthe public critical of the use of the word "anticipated". Submitters considered the wordwas ambiguous and in the view of the Chief Medical Officers of the District HealthBoards would effectively import a 50 per cent threshold cut-off for cover.18The Auckland District Health Board called for "anticipated consequence" to bechanged to "ordinary consequence" to ensure that the consequence of treatment in anygiven situation was measured objectively based on the ordinary outcome.19[18] When the Select Committee reported back to Parliament on the Bill, itrecommended that the words "anticipated part, or consequence, of the treatment" bereplaced with "necessary part, or ordinary consequence, of the treatment".20 The onlyexplanation given for the recommendation was that the replacement words wouldprovide greater clarity.[19] The new injury treatment regime was duly enacted in 2005. The salientprovisions are ss 32, 33 and 34. Section 32 which is at the heart of this appeal states:16 Injury Prevention, Rehabilitation, and Compensation Amendment Bill (No 3) 2004 (165-2), cl 13.17 Department of Labour "Submission to the Health Select Committee on the Injury Prevention,Rehabilitation, and Compensation Amendment Bill (No 3) 2004 at 24.18 Ken Mackinnon "Submission to the Health Committee on the Injury Prevention, Rehabilitationand Compensation Amendment Bill (No 3) 2004" at 3; and District Health Board's Chief MedicalOfficers "Submission to the Health Committee on the Injury Prevention, Rehabilitation, andCompensation Amendment Bill (No 3) 2004".19 Auckland District Health Board "Submission to the Health Committee on the Injury Prevention,Rehabilitation, and Compensation Amendment Bill (No 3) 2004".20 Injury Prevention, Rehabilitation and Compensation Amendment Bill (No 3) (165-2) (selectcommittee report) at 3.32 Treatment injury(1) Treatment injury means personal injury that is—(a) suffered by a person—(i) seeking treatment from 1 or more registered healthprofessionals; or(ii) receiving treatment from, or at the direction of, 1 ormore registered health professionals; or(iii) referred to in subsection (7); and(b) caused by treatment; and(c) not a necessary part, or ordinary consequence, ofthe treatment, taking into account all the circumstances ofthe treatment, including—(i) the person's underlying health condition at the timeof the treatment; and(ii) the clinical knowledge at the time of the treatment.(2) Treatment injury does not include the following kinds of personalinjury:(a) personal injury that is wholly or substantially caused bya person's underlying health condition:(b) personal injury that is solely attributable to a resourceallocation decision:(c) personal injury that is a result of a person unreasonablywithholding or delaying their consent to undergo treatment.(3) The fact that the treatment did not achieve a desired result does not,of itself, constitute treatment injury.(4) Treatment injury includes personal injury suffered by a person asa result of treatment given as part of a clinical trial, in thecircumstances described in subsection (5) or subsection (6).(5) One of the circumstances referred to in subsection (4) is where theclaimant did not agree, in writing, to participate in the trial.(6) The other circumstance referred to in subsection (4) is where—(a) an ethics committee—(i) approved the trial; and(ii) was satisfied that the trial was not to be conductedprincipally for the benefit of the manufacturer ordistributor of the medicine or item being trialled; and(b) the ethics committee was approved by the Health ResearchCouncil of New Zealand or the Director-General of Health atthe time it gave its approval.(7) If a person (person A) suffers an infection that is a treatment injury,cover for that personal injury extends to—(a) person A's spouse or partner, if person A has passedthe infection on directly to the spouse or partner:(b) person A's child, if person A has passed the infection ondirectly to the child:(c) any other third party, if person A has passed the infection ondirectly to that third party:(d) person A's child or any other third party, if—(i) person A has passed the infection directly to his or herspouse or partner; and(ii) person A's spouse or partner has then passedthe infection directly to the child or third party.[20] The effect of s 32 is that cover for treatment injury arises where a person suffersa personal injury which satisfies four pre-requisites:(a) the injury was suffered while seeking or receiving treatment;(b) was caused by the treatment;(c) was not a necessary part or ordinary consequence of the treatmenttaking into account all the circumstances of the treatment includingthe patient's underlying health conditions and the clinical knowledge atthe time of the treatment; and(d) is not excluded from cover by s 32(2). (Section 32(2) excludes coverwhen the injury is wholly or substantially caused by the patient'sunderlying health condition, or is wholly attributable to a resourceallocation decision or is the result of the patient unreasonablywithholding or delaying consent.)[21] This Court has had occasion in previous cases to consider various aspects ofthe treatment injury provisions, in particular the meaning of "treatment" and the phrase"caused by treatment."21 The present case is however the first time it has consideredthe meaning to be attributed to the words "not an ordinary consequence".22[22] Finally, for completeness, we note that in March 2010 the Injury PreventionRehabilitation and Compensation Act 2001 was renamed the Accident CompensationAct 2001. For ease of reference, throughout the rest of this judgment we refer tothe latter as "the 2001 Act".This proceeding[23] The interpretation issue relating to the meaning of "ordinary consequence"came before Churchman J in the context of three test cases. Only two of those casesare part of the appeal in this Court.23 In those two cases, ACC had declined cover fortreatment injury on the grounds the injury in each case was an ordinary consequenceof the treatment. This had been reversed on appeal in the District Court.24 ACC thenobtained leave to appeal to the High Court under s 162 of the 2001 Act on variousquestions of law all centring on the correct interpretation of "ordinary consequence".25[24] The first case concerned a patient Ms Ng. She suffered a significant strokefollowing emergency surgery to treat three anterior choroidal artery aneurysms.A neurosurgeon advising the Corporation stated that the rate of such a complicationfollowing surgery can be as high as 16 per cent or up to 22.6 per cent where clippingwas involved as it was in Ms Ng's operation. Her treating neurosurgeon advised that,21 See Aitkinson v Accident Compensation and Insurance Corp [2002] 1 NZLR 374 (CA); AccidentCompensation Corp v Ambros [2007] NZCA 304, [2008] 1 NZLR 340; McEnteer v AccidentCompensation Corp [2010] NZCA 126, [2010] NZAR 301; Cumberland v Accident CompensationCorp [2013] NZCA 590, [2014] 2 NZLR 373; and Adlam v Accident Compensation Corp [2017]NZCA 457, [2018] 2 NZLR 102.22 The term was considered in Roche Products (New Zealand) Ltd v Austin [2019] NZCA 660 butthe delineation of the boundaries of ordinary consequence was left to await the outcome of thisappeal (at [33]). The approach taken in Roche Products was differently expressed but is consistentwith the approach taken in this judgment.23 The third case did not turn on the meaning of "ordinary consequence": HK v AccidentCompensation Corp [2018] NZACC 85 (DC).24 Ng v Accident Compensation Corp [2017] NZACC 48 (DC) [Ng District Court decision]; and L vAccident Compensation Corp [2017] NZACC 147 (DC) [L District Court decision].25 Accident Compensation Corp v L [2018] NZACC 84 (DC); Accident Compensation Corp v Ng[2018] NZACC 46 (DC); and HK v NZACC 106 (DC).when treating aneurysms, he expects complications in the region of 10–15 per centand additionally a mortality rate of 10 per cent.[25] ACC's Complex Claims Panel concluded that "[c]onsidering [Ms Ng's]specific situation and presentation on an established background risk of between 16%to 22% this would be considered an ordinary consequence of treatment".[26] In the District Court, Judge Mathers acknowledged that statistics can assistan individual assessment but considered the percentages did not themselves answerwhether the stroke was an ordinary consequence. She also stated it was "too simplisticto simply consider whether it is more or less likely, or say a 50/50 basis" and thata "good dose of common sense" was vital.26 She concluded on the evidence thatthe stroke was an unpredictable and untoward complication of the operation andtherefore was not an ordinary consequence.27[27] That result was upheld in the High Court but on the basis a stroke with anincidence rate in the order of 16–22 per cent, being less than 50 per cent, was not anordinary consequence.28 It was not an outcome that was more probable or more likelythan not.[28] The second test case concerned Ms L. She developed right leg paralysis andother neurological deficits following surgery to remove a spinal malformation.Expert evidence as to the incidence rate of such an adverse consequence ranged from10–38 per cent.[29] In holding the injury was not an ordinary consequence, Judge Powell rejectedevidence of a 38 per cent incidence rate as being too high.29 He relied on other expertevidence to the effect the minimum risk of neurological deficit would be 10 per centbut that it was more likely than not that Ms L would not deteriorate followingsurgery.3026 Ng District Court decision, above n 24, at [40].27 At [43].28 High Court decision, above n 1, at [108]–[110].29 L District Court decision, above n 24, at [33].30 At [35]–[36].[30] This conclusion was upheld in the High Court. It is implicit in theDistrict Court decision that had the Judge accepted an incidence rate of 38 per cent hemay not have considered the injury to be an ordinary consequence.Under the formulation adopted by Churchman J however, an incidence rate of38 per cent being less than 50 per cent would not equate to an ordinary consequence.31[31] In adopting an interpretation of "more probable than not", Churchman Jinvoked the established principle that where language used in the accidentcompensation legislation is ambiguous, a generous and unniggardly interpretationfavourable to claimants is to be preferred.32 He considered this required him to adoptan interpretation that focused on the synonyms of the word "ordinary" such as"normal", "usual", "standard", "typical", or "expected" (rather than the antonyms suchas "extraordinary").33 The synonyms were in his view consistent with the meaning"more probable than not".34[32] As regards legislative intent, the Judge held there was an intention to departfrom the "hard-edged" statistical threshold contained in previous legislation and thatthe 2001 Act was intended to extend the situations in which cover would beavailable.35[33] Dissatisfied with this outcome, ACC sought and obtained leave fromChurchman J to appeal to this Court under 163 of the 2001 Act by way of case statedon a question of law.36[34] The question of law in respect of which leave was obtained is:37Was the [High] Court correct to conclude that the word "ordinary" ins 32(1)(c) [of the Accident Compensation Act 2001] referring to an "ordinaryconsequence" means a consequence that is more probable than not?31 High Court decision, above n 1, at [121].32 At [101(j)–(k)]. See further Accident Compensation Corp v Mitchell [1992] 2 NZLR 436 (CA) at438; Harrild v Director of Proceedings [2003] 3 NZLR 289 (CA) at [39]; and J v AccidentCompensation Corp [2017] NZCA 441, [2017] 3 NZLR 804 at [52].33 At [61] and [101(m)].34 At [101(m).35 At [59] and [101(b)].36 Accident Compensation Corp v Ng [2019] NZHC 207.37 At [18].[35] Before turning to consider the question for determination, it is necessary firstto resolve a procedural issue regarding further evidence.Applications for leave to adduce further evidence[36] Both parties applied for leave to adduce further evidence for the appeal in thisCourt and both opposed the other's application.[37] ACC sought leave to adduce affidavit evidence from four medical experts,a paediatric surgeon, an oncologist, a general and colorectal surgeon and a plastic andreconstructive surgeon. Each affidavit identifies the unintended adverse consequencesthat may arise from treatments in their respective specialities, the incidence rate ofthose adverse outcomes and their practices in assessing those risks before embarkingon the procedure.[38] The common theme of the affidavits is that a 20 per cent to 50 per cent chanceof a significant adverse outcome is considered very high and that doctors will seek toavoid treatments that carry such high risks. Further, that it is only in rare or exceptionalcircumstances that medical procedures carrying risks of significant adverseconsequences exceeding 50 per cent would ever be undertaken. In the case ofpaediatric surgery, Professor Beasley stated that a surgery carrying such risk wouldonly be performed where the risk of death without treatment was high, and he "cannotthink of any elective procedure" that would be performed in such circumstances.[39] The affidavits do not address the factual circumstances of the cases of Ms Ngand Ms L nor do they address the merits or otherwise of the High Court decision underappeal. They are simply tendered for the purpose of explaining how medicalprofessionals approach a procedure that carries a risk of significant adverse effects toa patient, and the levels of risk of adverse outcomes associated with a range of medicalprocedures.[40] The respondents do not dispute the factual accuracy of the contents ofthe affidavits but oppose leave being granted on the grounds such extrinsic evidenceis not relevant to the interpretation exercise.[41] We disagree. In our view, the evidence provides the Court with a substantiallyhelpful insight into the social and medical context in which s 32 has to be construed.38We therefore grant leave.[42] Turning to the respondents' application for leave to adduce further evidence,this concerns two affidavits filed in response to the four affidavits from the specialists.[43] The first is an affidavit from the Hon Ruth Dyson who was the Minister forACC from 15 August 2002 to 5 November 2007. The purpose of the affidavit is toprovide Ms Dyson's view as the Minister responsible for introducing the currenttreatment injury provisions as to what was intended at the time of its drafting andenactment and whether this intention accords with the approach advocated by ACC inthis appeal.[44] In our view the affidavit is plainly inadmissible. It has never been the law thatthe meaning of a statute can be ascertained by reference to the subjective intentions ofthe Minister responsible for the Bill, let alone from that Minister's evidence about theirsubjective intentions given some 15 years later.39 Such evidence is quite simplyirrelevant. We add that it would be wrong in principle for the meaning of legislationto be ascertained by reference to material concerning legislative history that did notform part of the publicly available record of the process leading up to enactment ofthe legislation.[45] The second affidavit is from a general practitioner (Dr Matthews) who has beeninvolved in several treatment injury cases as an expert witness. In addition to hermedical qualifications, she also has two law degrees. Her affidavit is tendered forthe purpose of identifying concerns about the interpretation of s 32 suggested by ACCbased on risk assessments.38 For the relevance of the wider social context of a provision see Commerce Commission v FonterraCo-operative Group Ltd [2007] NZSC 36, [2007] 3 NZLR 767 at [22].39 Application v Secretary of State for the Environment, Transport and the Regions, ex parte SpathHolme Ltd [2001] 2 AC 349 (HL) at 396 cited in Gibbs v Crest Commercial Cleaning Ltd (2005)2 NZELR 447 (EmpC) at [82]; Wilson v First County Trust Ltd (No 2) [2003] UKHL, [2004] 1AC 816; Skycity Auckland Ltd v Gambling Commission [2007] NZCA 407, [2008] 2 NZLR 182at [52]; and Diggory Bailey and Luke Norbury Bennion on Statutory Interpretation (7th ed,LexisNexis, London, 2017) at [24.3].[46] Dr Matthews comments on the way in which the medical community assessesrisk and opines that if the test for cover under the accident compensation scheme wereto involve the application of concepts of low, medium, and high risk as suggested byACC, that would be problematic. That is because, she says, each person and eachdiscipline has their own ideas about what constitutes risk and there "are only so manythings that treatment providers can warn about and quantify".[47] She also makes the point that consequences can vary greatly in degree ofseverity and further claims that "most of the time" medical practitioners do not knowif an adverse consequence will happen in advance.[48] The affidavit then goes on to evaluate the merits of the High Courtinterpretation which Dr Matthews supports, saying it provides clarity and is easy toapply.[49] We have reservations about some aspects of this affidavit, in particular itscritique of the High Court decision which is more in the nature of a submission thanexpert evidence. However, we consider the bulk of the affidavit to be relevant andhelpful and are therefore willing to grant the application to adduce it as furtherevidence on appeal.Arguments on appealThe appellant[50] ACC contends that although the Judge expressly acknowledged there was noevidence of a parliamentary intent in 2005 to make a paradigm shift so that all or nearlyall treatment injuries should be covered, that is precisely the effect of his decision.According to ACC, the decision means it will be underwriting medical treatment inNew Zealand regardless of medical error and regardless of the ordinariness in medicalterms of the consequences occurring or the nature of the consequences.[51] The decision only excludes from cover harm caused by medical treatmentswhere there was a 50 per cent or greater risk of such harm. Yet that is to exclude onlya tiny minority of harm resulting from all medical treatments because (a) medicaltreatments involving that degree of risk are in fact rarely — and in some contexts,including for example clinical trials, never — undertaken, and (b) in the case of sometreatments for example chemotherapy, the incidence of any particular harmful sideeffect (as distinct from the incidence of adverse side effects generally) will neverapproach 50 per cent and will therefore never be ordinary.[52] ACC accepts that a generous approach to interpretation is desirable but not atthe expense of the integrity of the scheme.[53] According to ACC's counsel Mr Rennie QC, the interpretation adopted bythe High Court is plainly wrong. It was, he argued, inconsistent with the plainlanguage of s 32(1)(c), the legislative scheme and history and the medical treatmentcontext to which the regime applies.[54] He submitted that in order to be consistent with those things, the phrase "notan ordinary consequence" should be interpreted by reference to categories of riskcommonly used by medical professionals. In support of that general submission, hereferred us to the following guidelines used by ACC.Whether an adverse consequence of treatment is "ordinary" is an objectivetest focused on the facts of the particular case. The factors that are relevant tothat test might include:(a) the likelihood at a general population level (or appropriate populationsegment) of the consequence occurring as a result of the treatment;(b) the particular circumstances of the claimant's case that makethe likelihood of an adverse consequence higher or lower, such as:(i) the claimant's underlying health factors (eg, age, smoking,health status);(ii) the clinical knowledge of the claimant's circumstances atthe time of treatment; and(iii) the circumstances in which the treatment occurred; and(c) the significance of the adverse consequence.The section requires a flexible and pragmatic assessment. Where appropriateand relevant data is available, statistical studies of medical risks may providea starting point for the assessment but they must be analysed and applied inlight of the individual claimant's particular circumstances. No single riskpercentage determines in all circumstances whether a significant adverseconsequence of medical treatment will be an "ordinary" consequence.Overall, when assessed objectively in the particular circumstances of the case:(a) an adverse consequence that was a low risk of occurring is unlikely tobe an ordinary consequence;(b) an adverse consequence that was a medium risk of occurring is almostcertainly an ordinary consequence, unless that adverse consequence isdisproportionate to the purpose of the treatment and the benefitexpected to be derived from it; and(c) an adverse consequence that was a high risk of occurring isan ordinary consequence. In the surgical context, a high risk ismedically assessed as one that occurs in about 10% or more of cases.[55] Although in written submissions these guidelines were described as being"required" by s 32(1)(c), Mr Rennie advised in oral submissions that they were notbeing propounded as "hard line tests" of ordinary consequences but rather asevidencing ACC's practice. He hoped the Court might find them of value.The respondents[56] Counsel for the respondents, Ms Peck, sought to uphold the High Courtdecision and supported its reasoning.[57] In her submission, the Judge's interpretation provided "a coherent, predictableand principled" test that produced fair and socially acceptable results. It accorded withboth the legislative history, the search for fairness and efficiency as well asthe underlying purpose of the accident compensation scheme generally. Ms Peckcontended that "ordinary consequences" should be interpreted generously and in a waythat allows for administratively efficient consequences which the Judge'sinterpretation achieved. She argued that as a matter of language, the word "ordinaryconsequence" easily bears the meaning of an occurrence expected to occur inthe majority of cases[58] Ms Peck challenged ACC's "floodgate" argument and said it overlooked that"not an ordinary consequence" was only one of a number of pre-conditions to coverand that the vast majority of treatment injury claims that fail do so at the causationhurdle. According to ACC's own statistics, prior to the High Court decision only asmall percentage of total treatment injury claims (13 per cent) were filtered out at the"not an ordinary consequence" barrier. Lowering the threshold would not thereforeoverwhelm the scheme.[59] Ms Peck was also critical of the test/guidelines proposed by ACC which shedescribed as effectively re-introducing severity and rarity by stealth. Developing theconcerns raised by Dr Matthews in her affidavit, Ms Peck submitted it was aconceptual trap to conflate a treatment risk-benefit analysis with how society shouldrespond to adverse outcomes of medical treatment. Risk assessments were forwardlooking and designed for the purpose of giving advice to patients and ensuringinformed consent. "Consequences" in contrast were about something that has actuallyhappened.[60] Ms Peck emphasised the importance of having a clear yardstick so as tominimise the costs and delay of making a claim and achieve consistency.She described clarity as "essential" and asked "if not the balance of probabilities, thenwhat?"Analysis[61] The meaning of an enactment must be ascertained from its text in light of itspurpose.40[62] We have come to a clear view that as a matter of plain language and legislativepurpose, the phrase "ordinary consequence" as it appears in s 32(1)(c) is not capableof bearing the meaning "a consequence more probable than not."[63] Dictionary definitions of "ordinary" include "normal", "customary", "regular","usual", 'standard", "typical", or "expected". The Judge considered that these andother synonyms were consistent with his interpretation. However, as Mr Rennie pointsout, none of these words necessarily connote frequency arising more than half the timeand indeed words that do connote that concept such as "probable" "likely" or morelikely than not" are not among the synonyms of "ordinary."40 Interpretation Act 1999, s 5.[64] If Parliament had intended "ordinary" to mean "more likely than not" there arewell established phrases which it could and, in our view, would have used to expressthat idea. It did not.[65] We accept that the purpose of the 2005 amendment was to increase cover butit is not possible to discern from either the text or the legislative history an intentionto effectively cover almost all treatment injuries when the prior regime had donenothing like that. Had that been the intention it is reasonable to expect that wouldhave been expressly highlighted and given prominence as a significant feature ofthe change. What information there is on the record points only to the intendedincrease in cover being a modest one.[66] We take Ms Peck's point that most claims falter at the causation hurdle anyway.But Parliament chose to include "not an ordinary consequence" as an additionalpre-requisite, and therefore must be taken to have intended it would be meaningful.To set a statutory exclusion at 50 per cent in a context where that threshold will seldombe reached is however to render it meaningless. That is quite apart from the furtherobjection that Parliament having deliberately moved away from a test expressed inpercentage values, it should not be for the Court to re-introduce one.[67] What then is the correct interpretation of "not an ordinary consequence?"[68] In our view, it should be interpreted as meaning an outcome that is outside ofthe normal range of outcomes, something out of the ordinary which occasionsa measure of surprise. That is an interpretation that we consider, as did the Court inChilds v Hillock, best captures Parliament's intent in the context of a scheme which isunderpinned by the concept of "personal injury by accident" and which does notprovide universal compensation for sickness or ill-health. So, for example, side effectsof chemotherapy of a nature and severity that are encountered reasonably often andoccasion no surprise are ordinary consequences of that chemotherapy even if (as willoften be the case) such side effects are not encountered in more than 50 per cent ofcases.[69] Whether an adverse consequence is inside or outside the normal range ofconsequences of the medical treatment given to a particular claimant is ultimately amatter of judgment for the decision maker. It is to be exercised on a case-specificbasis taking into account all the circumstances of the treatment and the particularclaimant. Thus, relevant circumstances will include not only the nature of the harmsuffered but also its duration and severity as well as any other circumstances pertainingto the patient which may have rendered them more or less susceptible to the adverseconsequence. The decision may be informed by medical studies including relevantstatistical analysis (subject to the reservations detailed below) as well as the clinicalexperience of the treating physician(s) and other specialists.[70] As raised with counsel during the hearing, we consider that some caution isrequired when drawing on statistical analysis contained in medical studies of the kindreferred to in the decisions below, and in the expert evidence before us. Many of thesestudies involve small numbers of cases, and often the results are not accompanied byany measure of their statistical significance. There may also be significant differencesbetween the group studied — patients in a particular hospital or on a specialisedprogramme for example — and the group of recipients of similar treatment(s) inNew Zealand. The way in which the treatment group is defined, and the way in whichadverse outcomes are defined, will often involve significant judgment. These factorsunderscore the problematic nature of a test based on statistical frequency alone,including the 2001 test of rarity and the Judge's preferred test, focused on whether theadverse consequence is more probable than not. The 2005 amendment deliberatelymoved away from a statistical assessment of risk to a test that requires the exercise ofjudgment.[71] We acknowledge the temporal distinction between risk and consequence ashighlighted by Ms Peck. Although risk assessments undertaken by doctors are basedon the frequency of past actual occurrences and therefore are clearly relevant, we agreethe focus should be on whether the outcome that occurred is within the range ofordinary consequences rather than whether the risk of the outcome was predicted inadvance of treatment in a particular claimant's case.[72] We also acknowledge that our interpretation does not provide the precision orcomprehensive guidance that counsel, especially Ms Peck, were seeking.However, Parliament has chosen to use an imprecise test and in our view the Courtwould be straying beyond its proper function to disregard that and superimpose astructure of its own creation. As noted by this Court in Vodafone New Zealand Ltd vTelecom New Zealand Ltd, the Court must guard against "taking an inherentlyimprecise word and 'by redefining it thrusting on it a spurious degree of precision'".41If the lack of precision in s 32 is problematic, it is for the legislature to resolve.Outcome[73] The appellant's application for leave to adduce the affidavit evidence ofSpencer Beasley, Richard North, Andrew Connolly and Jonathan Wheeler is granted.[74] The respondents' application for leave to adduce the affidavit evidence ofFrances Matthews is granted.[75] The respondents' application for leave to adduce the affidavit evidence ofRuth Dyson is declined.[76] We answer the question of law submitted for determination by this Court:Was the High Court correct to conclude that the word "ordinary" ins 32(1)(c) of the Accident Compensation Act 2001 referring to an"ordinary consequence" means a consequence that is more probablethan not?Answer: No[77] The appeal is allowed and the decision of the High Court quashed.[78] The claims of the respondents are remitted to the District Court forreconsideration in accordance with this judgment.41 Vodafone New Zealand Ltd v Telecom New Zealand Ltd [2011] NZSC 138, [2012] 3 NZLR 153 at[54] per Blanchard J citing R v Monopolies and Mergers Commission, ex parte South YorkshireTransport Ltd [1993] 1 WLR 23 (HL) at 29.[79] As regards costs, it was common ground that in the event ACC were successfulno costs should be awarded. We consider that is appropriate, this appeal being inthe nature of a test case.Solicitors:Accident Compensation Corp, Wellington for AppellantJohn Miller Law, Wellington for Respondents