[1] Mr Livingston and Mr Tuiavii face charges of driving with excess breath alcohol. The machine used to administer the relevant evidential breath test to Mr Livingston was a Seres Ethylometre 679T machine and gave a result of 833 micrograms per litre of breath. In Mr Tuiavii’s case the machine used was a Lion Intoxilyzer 5000 machine (model 6600) and gave a result of 948 micrograms of alcohol per litre of breath. The legal limit is 400 micrograms per litre of breath. Both Mr Livingston and Mr Tuiavii wish to dispute the reliability of their evidential breath tests on the basis of the evidence of an electronics expert, Mr Churchill. Mr Churchill has identified what he considers are faults in the particular devices used, which he believes raise doubt as to the accuracy of the results obtained. He, however, deposes that he needs access to the manufacturer’s technical manual in order to provide a full analysis.
[2] The background to these appeals is the decision of this Court in Attorney-General v Otahuhu District Court and Brown [2001] 3 NZLR 740. In that case Mr Brown, who was in a similar position to the appellants in this case, sought discovery from the police of the manufacturer’s manual for the Intoxilyzer device used to administer his evidential breath test. This Court held that, as the manual was held by the Institute of Environmental Science and Research Ltd (ESR), an independent legal entity, it was not within the power or control of the police as prosecuting authority. Therefore the police could not be compelled to discover it. In addition, it was held that there is no statutory authority or implied power to make pre-trial rulings on discoverability of documents held by third parties in a summary case. Neither does the New Zealand Bill of Rights Act 1990 (Bill of Rights) require such a pre-trial criminal discovery process. The Court held that the appropriate procedure, if the defence wishes to have a document admitted in evidence, is to issue a summons under s20 of the Summary Proceedings Act 1957.
[3] This course was followed by Mr Tuiavii and Mr Livingston. Mr Tuiavii issued a summons to the ESR and Mr Livingston to Dr Gainsford, a consultant to ESR, to give evidence concerning the evidential breath test devices and in particular to produce the manufacturers’ manuals in respect of the devices. The ESR and Dr Gainsford applied to the High Court to have the witness summonses in relation to Mr Tuiavii, Mr Livingston and a number of others in a similar position set aside. By judgment of 28 June 2002 Williams J granted these applications. Mr Tuiavii and Mr Livingston appeal against that decision.

Facts

[4] It is now standard practice in relation to blood-breath alcohol offences for the police to obtain from the ESR the following documents:

[a] the blood/breath alcohol correlations for each device. When drivers take an evidential breath test and subsequently elect or are required to undergo a blood test, the results of the evidential breath tests are then compared with those blood tests.

[b] the operation manual for the device.

[c] the maintenance and calibration logbook.

[d] the test result card for the individual case.

[5] These documents are disclosed by the police to any defendant who requests disclosure. The manufacturers’ instrument manuals for the devices are not provided to the police by the ESR and are not disclosed. These manuals, according to Dr Gainsford, relate to the devices in general rather than to specific machines and contain circuit diagrams and procedures for setting the electronic controllers within the instrument. In his view the manuals do not provide any information relating to the performance of a specific device on a particular day. Dr Gainsford deposed that these manuals contain trade secrets and confidential intellectual property and that they have been provided to ESR on the basis of strict confidentiality and only because ESR is a government agency. The manuals remain the property of the respective manufacturers and both have reserved the right to demand their return.
[6] There was no evidence from the manufacturers themselves but Dr Gainsford attached to his affidavit of 21 September 2001 letters from both of the manufacturers at issue indicating that they considered the manuals to be confidential and that they would require their return should disclosure be ordered. Affidavits were filed by the New Zealand distributors confirming that they had been instructed to recover the manuals should disclosure be ordered.
[7] We note here that Mr Churchill has indicated that he is prepared to provide any undertakings required by the Court including the signing of a confidentiality or non-disclosure agreement. He also said there is no real risk that his possession of the manuals would jeopardise the manufacturers’ commercial secrets. His company is wholly owned by himself and his wife, employs no staff and does not occupy commercial premises. Additionally he has no business or trade associations that could result in a conflict of interest or compromise the security of the information. The affidavit filed by the New Zealand distributor of the Intoxilyzer 5000 stated however that, even if Mr Churchill did comply with undertakings, it was inevitable that the information contained in the manuals would enter into the public arena as a result of such disclosure.
[8] Dr Gainsford deposed that, if the instrument manuals were withdrawn, then ESR would no longer be able to service or repair the evidential breath test devices. He believed, if that were to happen, that a substantial proportion of the devices would become inoperable within 12 months. Mr Churchill in cross-examination disagreed with Dr Gainsford that servicing would cease without the manuals. Mr Churchill contended that the manuals would only rarely be necessary for servicing as ESR staff would be expected to possess the requisite knowledge to deal with most routine service issues.
[9] In cross-examination Dr Gainsford was asked about alternatives to the ESR servicing role but indicated that in his view any alternatives, such as sending the devices offshore for servicing, would not be practical.
[10] We note too that the Minister of Police on 19 October 2001 claimed public interest immunity in relation to the instrument manuals. He directed that the ESR withhold the manuals sought (unless a court orders otherwise) on the ground that disclosure would be injurious to the public interest. We also note that Parliament on 19 December 2001 amended the Land Transport Act 1998 to remove the defence of error or possible error in the result of an evidential breath test.
[11] We now turn to the facts of the individual cases. As indicated above, the device at issue in Mr Tuiavii’s case was a Lion Intoxilyzer 5000 machine, model 6600. In his affidavit of 8 October 2001, filed in respect of Mr Tuiavii’s case, Mr Churchill deposed that he had been given a copy of the operation manual for the device together with 140 photocopied pages of the calibration and maintenance logbook covering the period 15 June 1989 to 6 November 2000 and a copy of the result card of 23 September 2000 for Mr Tuiavii. He deposed that he had not had access to the annual calibration certificates in the case of the device at issue, although there may be none. He had also not had access to the blood breath correlations for the machine but he would not be able to interpret the information from such correlations without taking advice from an independent expert in that field.
[12] He deposed that, from his examination of the operation manual and the logbook provided in respect of the device, he is of the opinion that the service checks of January 2000, August 2000 and October 2000 in particular “raise a doubt as to accuracy of the ostensible result of above 600 micrograms” in Mr Tuiavii’s case. The reason for this opinion is that the October 2000 service (one month after Mr Tuiavii’s test on 23 September 2000) showed the “analogue to digital reference was reset because between December 1999 and October 2000 it had drifted down 25 millivolts from 1.474 V to 1.449 V.” He went on to say that the voltage is required to be maintained to an accuracy of +/-1mV and that “the direction of drift would have made the machine read high”. He expressed the opinion that the drift was almost certainly present at the time of Mr Tuiavii’s test and that this critical calibration voltage had not been checked at the service checks in January or August 2000.
[13] In terms of why he considered reference to the manufacturers’ manual necessary in Mr Tuiavii’s case he said that reference to the manual:

is of enormous value when interpreting the data gathered regarding the performance of the machine at any particular time or over any period for which performance or calibration data is available. Reference to the manual allows failure modes to be analysed and component or subassembly weaknesses to be identified and possible future failures predicted with considerable accuracy.

[14] We comment here that future failures would seem irrelevant to whether the device was working correctly at the time of Mr Tuiavii’s test. He went on to say, however, that:

Maladjustment of ‘electronic controllers’ or operating parameters that are out of specification can also be identified by reference to the manufacturer’s specification as laid down in the manual. The manual information is vital and lack of independent access poses a significant barrier to effective analysis of machine performance.

The technical manual is provided by the manufacturer to allow the performance of their product to be maintained within design parameters. Amongst other important information it will contain the adjustment procedures and critical parameters of voltage and current that must be maintained within specified limits to ensure correct and reliable functioning of the machine.

[15] He also referred to the need to have access to the manual to understand the role of the software, which controls a number of functions in the device. Details of the modifications to that software were crucial in his view to the evaluation of the service record for the device. However, Mr Churchill did not explain why this information was relevant to the actual fault he identified as having been present.
[16] Dr Gainsford, in his second affidavit of 7 November 2001, deposed that it is not possible to infer that there was a drift in the analogue to digital reference between December 1999 and 26 October 2000 as the analogue to digital reference was measured only after the device had been serviced which involved unplugging connectors, cleaning optical components and preventive replacement of other components. He stated that it is possible that the difference could have been as a result of this servicing. He deposed to his belief that the device was working correctly at the time of Mr Tuiavii’s case because Mr Tuiavii’s result card showed no significant change in internal standards, which he would have expected if the device had experienced a change in its analogue to digital reference of 25mV. He also noted that, in the period between 23 December 1999 and 1 October 2000 (the period referred to by Mr Churchill), there were 53 blood-breath correlations recorded for this device and in each case the evidential breath test was consistent with the subsequent blood alcohol test.
[17] Dr Gainsford also deposed that it is not correct that the Intoxilyzer 5000 device is required to be maintained to an accuracy of +/-1mV. The analogue to digital converter in the Intoxilyzer is not even sensitive to reference variations of that size. Mr Churchill in his affidavit in reply of 30 November 2001 attempted to refute this on the basis that the voltage is as a matter of practice set by ESR technicians to a precision of +/-1mV and from other material he has seen this is a required standard.
[18] Moving now to the case of Mr Livingston, the device in question was a Seres Ethylometre 679T. In his case Mr Churchill, by affidavit of 26 July 2001, identified two alleged defects in the particular device used for Mr Livingston’s evidential breath test. Mr Churchill expressed the opinion that both of these defects may have compromised the outcome of that test. The first defect identified was that, in comparison with other Seres 679T machines of similar age, the device in question had experienced an abnormally high level of instability in voltage supplied to the central processing unit. He said that the records show that ESR failed to check this voltage for a period of several years and subsequently failed to maintain it within the limits prescribed by the manufacturer. He went on to say that the performance and accuracy of the Seres device relies on a number of calibration adjustments, some of which are interdependent. The adjustments, among other things, control the duration and volume of the breath sample and the operating conditions for the analytical cell at the heart of the breath-alcohol measurement system.
[19] The second defect identified was that the components of the machine required to transport the breath sample and deliver it to the analytical cell were damaged in such a way as to compromise the delivery of the correct sample size and volume. This in his opinion meant that the outcome of any test could not be relied upon as a true and accurate measure of the subject’s breath-alcohol content.
[20] Mr Churchill also expressed the view that the maintenance records show that ESR failed to maintain and repair the machine in a professional manner and that the machine had been returned to service on several occasions without conclusive repair. He acknowledged that the Seres 679T is claimed to be a self-testing device. He expressed the view, however, that those testing capabilities “cannot be expected to overcome the deficiencies in the condition of this particular machine or the lack of stability and consistency in its operational history. The final monitoring layer remains human.” He also stated that the self-testing functions are constrained, both in terms of the range or number of functions monitored and the timing of those checks. There are, he said, several significant operational aspects of the Seres 679T that are not subject to monitoring and these functions are capable of causing errors in breath-alcohol readings. He deposed that he needed access to the manual to identify the functions that were not included in those procedures and to relate this information to the maintenance and calibration records.
[21] Dr Gainsford did not in his affidavits deal specifically with the concerns raised by Mr Churchill in relation to the device used in testing Mr Livingston. He did, however, in cross-examination deal with some similar issues in relation to a device used on another of the parties. He said that the Seres device was designed to ensure that the flow rate does not go below a certain level. If it drops below that level then the device will not give a reading. In addition, he said that the flow meter controls the breath sample and it is not in any way involved in the analysis of the breath sample. Mr Churchill in his oral evidence said it was not always the case that the device will not give a reading. He said the electronic controls do not lock into any particular setting, but they are monitored in relation to a reference to see that they stay within acceptable limits. If any of those standards exceeds the given reference to an extent set down by the manufacturer as their operating criteria, at that point the machine will usually cease to operate or put up a warning message for some corrective action that it may be possible to take. Neither party has referred to the blood-breath correlations for the device used for Mr Livingston’s evidential breath test.
[22] Finally we note that, when asked in cross-examination whether he was able to say that there was a reasonable possibility in relation to any of the evidential breath tests involved that the result should in fact have been under 400, Mr Churchill replied that it was impossible for him to answer. He indicated that, for the reasons set out in his various affidavits, he considers that there was a reasonable doubt as to the accuracy of the device on the day of the readings in question but, without more information (presumably without the manuals), he could not necessarily substantiate that doubt in all cases. He said that, if more information were available, in some cases his doubt may be removed and in some confirmed.

High Court judgment

[23] Williams J’s first conclusion was that the instrument manuals were at best only of peripheral, secondary or background relevance. The manuals did not assist with demonstrating performance of a particular breath-testing device on the day and time at which the evidential breath test was carried out. They would provide only background assistance for Mr Churchill in confirming his views as to unreliability of the results as he would be able then to assess against a “known rather than an inferential standard.” The judge noted that Mr Churchill had been able to provide evidence for defendants in the past without access to the manuals. He also expressed the view (although obviously not definitive) that the various matters raised by Mr Churchill had been satisfactorily explained by Dr Gainsford.
[24] His second conclusion was that the witness summonses had been issued for an improper purpose, namely to obtain discovery and inspection of documents held by third parties rather than to have the documents produced in court. He said that the description of the contents meant that production of the relevant instrument manual as an exhibit is unlikely of itself to be of much assistance to a court in deciding the case.
[25] Williams J then turned to s35 of the Evidence Amendment Act (No 2) 1980 (EAA). He held that there was a special relationship between ESR and Dr Gainsford and the manufacturers. He then referred to the importance of retention of the instrument manuals in New Zealand to the breath alcohol-testing regime and to road safety in New Zealand. He found that the evidence had shown that the manufacturers’ expressed intention to withdraw the manuals was sincere. There was a contractual obligation of non-disclosure and the Court should be reluctant to make an order which will require breach of a contract by one of the parties to it. Weighing these factors he held that the public interest in the preservation of confidences between ESR and Dr Gainsford on the one hand and the manufacturers on the other outweighed the likely significance (modest in any event) of evidence contained in the instrument manuals in relation to the prosecutions of the respondents.
[26] Williams J found it unnecessary, given the above, to consider the Minister’s certificate that it would be injurious to the public interest in enforcement of the breath alcohol provisions of the Land Transport Act 1998 for ESR and Dr Gainsford to disclose the instrument manuals. He did, however, at para [74], reiterate his finding that:

the evidence demonstrates that the blood and breath alcohol testing regimes would be jeopardized in the medium to long term were the instrument manuals to be withdrawn from New Zealand.

[27] For completeness he then considered the undertakings as to confidentiality offered by Mr Churchill. He concluded that such undertakings would in the circumstances have been inadequate. He compared this situation to the usual case where the commercial information is in issue only in respect of that particular piece of litigation and where involvement of the expert is limited to that one case. In this case he noted that Mr Churchill appeared to be one of the few experts involved in this type of work in New Zealand and was the expert involved in all the cases that were before the Court. Although there were variations between those cases and others in which Mr Churchill would become involved the manuals would be the background against which the individual circumstances of all cases would be measured. He then said that Mr Churchill had given little thought as to how he could write reports or give evidence for those who had retained his services without breaching the confidentiality undertakings or how he could avoid cross-fertilisation between cases when he had been given access to the instrument manual for one case only. He remarked that the evidence had suggested a keen interest on the part of Mr Churchill’s international colleagues in discovering the contents of the instrument manuals. He concluded that no form of confidentiality undertaking is likely to be effective beyond the short term in maintaining confidentiality and that this provided a further reason for setting aside the witness summonses.

Summary of submissions

[28] Mr Wotherspoon submitted that the witness summons in Mr Tuiavii’s case should not have been set aside because the ESR did not make out its case based on the inherent jurisdiction of the High Court. In his submission, it is clear that there is a defence of error or possible error in the accuracy of an evidential breath testing result. The expert called on behalf of Mr Tuiavii has said that he requires access to the manual in order to further that defence. The ESR did not call any expert in electronics to refute this and therefore the Court should not have concluded that the evidence was only of marginal relevance. Unless it was clear that the manual could not possibly be admissible the District Court is the appropriate forum to determine admissibility, because such a decision should be taken on the basis of all the evidence. This was not a case of abuse of process that would justify setting the summonses aside because there is a clear relationship between the manual and Mr Tuiavii’s defence. In addition, the Court should not have anticipated s35 of the EAA as those issues too were for the District Court judge at trial. In his submission, it was in any event not proved that the production of the manual would put the ESR in breach of its legal obligations or that the manufacturers would in fact be entitled to withdraw the manuals. Finally Mr Wotherspoon submitted that, if public interest dictated that the manuals should not be released because the manufacturers would withdraw them, then the Court should issue a stay of proceedings or exclude the evidence obtained by the device on the basis that Mr Tuiavii would not be afforded his rights under the Bill of Rights, particularly s25(a), the right to a fair hearing, s25(e), the right to present a defence, and s25(f), the right to obtain the attendance and examination of witnesses for the defence under the same conditions as the prosecution.
[29] Mr Johnson submitted that Mr Churchill’s evidence showed two defects present at the time of Mr Livingston’s evidential breath test, both of which may have compromised the outcome of that test. Mr Churchill’s evidence also, in his submission, showed that access to the manual is essential for a full and comprehensive analysis of the forensic performance of the machine and therefore that the information sought is material to Mr Livingston’s defence. He submitted that the defence experts should have the same access to the manual as the experts for the Crown so that they were operating on a “level playing field”. In addition, access to the manuals would not be on a broad basis given the confidentiality undertakings and on Mr Churchill’s evidence would be unlikely to compromise any intellectual property rights of the manufacturers. As this is the case, Mr Livingston’s right to a fair trial should clearly outweigh any privacy and confidentiality concerns. Mr Johnson submitted that it could not be in the public interest to permit the continued use of malfunctioning inaccurate machines. There was no suggestion that the Crown might feel obliged to abandon the trial if disclosure is granted, and in any event only cases before the enactment of the Land Transport (Road Safety Enforcement) Amendment Act 2001 are affected.
[30] On behalf of the ESR and Dr Gainsford, Mr Squire QC submitted, first, that the manuals are irrelevant or insufficiently relevant for the purpose of establishing the evidential foundation necessary to put in issue the reliability of the respective evidential breath tests. Secondly, he submitted that the witness summonses were an abuse of process because they were issued for the purpose of obtaining discovery and inspection of the manuals and not for the purpose of obtaining their production as exhibits. Finally, he submitted that production of the manuals should be excused under s35 of the EAA, having regard to the special relationship of confidentiality existing between the ESR and the manufacturers and the public interest in promoting the enforcement of New Zealand’s drink driving laws and public health and safety.
[31] Mr Pike’s main argument for the Attorney-General (who had been granted leave to intervene) was that, as long as an approved evidential breath testing device was used in an approved manner and the result card disclosed a breath alcohol content of over 600 micrograms per litre, the result was conclusive evidence that the suspect had the breath-alcohol level specified. In his submission there is no defence in law to a breath-alcohol prosecution on the basis of the malfunction of a particular breath testing device. The decision in Police v McKay [1995] 1 NZLR 303 is not, as has been assumed, authority for the existence of such a defence and the statutory scheme is against such a defence existing. Although this Court accepted the existence of such a defence in R v Livingston [2001] 1 NZLR 167, the remarks were obiter and were not made after full argument. The Court in that case was concerned rather to ensure that the defence did not transform into a general attack on the reliability of each test result in a manner that would undermine the integrity of the legislative scheme. In addition, Mr Pike submitted that the amendments to the Land Transport Act 1998 enacted on 19 December 2001, while recognising that there is a defence of error or possible error of an evidential breath test device, are not legislative acceptance of the possibility of error in cases where the result is over 600 mcg with a defence of unreliability then applying.

Issues

[32] The issues arising therefore are:

[a] whether there is a defence based on possible malfunction of an individual device;

[b] if so, whether the instrument manual is relevant to such a defence;

[c] whether the witness summonses were issued for an improper purpose – i.e. for discovery and inspection;

[d] whether the summonses should have been set aside on the basis of s35 of the EAA;

[e] whether the decision as to admissibility should have been left to the District Court.

Is there a defence based on possible malfunction of a device?

[33] We deal first with Mr Pike’s argument that there is no defence based on the malfunctioning of a particular device. If that is accepted then the manuals will of course be irrelevant and inadmissible on that ground alone and the witness summonses should clearly have been set aside.
[34] The first point is that we agree with Mr Pike’s submission, as did Williams J, that the case of Police v McKay [1995] 1 NZLR 303 has been misunderstood. The relevant passage in McKay is as follows (at 305):

There are two safeguards. First, the Minister may always revoke an approval if satisfied that a particular model of a device is defective or unreliable. Secondly, if there is in any case evidence raising a reasonable doubt as to the accuracy of the results produced by the model, when properly used as prescribed by the Breath Tests Notice, or the instructions with the machine, then that device cannot be treated as within the Ministerial approach. (emphasis added)

[35] In McKay the defendant had conceded that the breath test result was accurate. The point at issue was whether the model of the Seres breath testing device was an approved device for the purposes of the Breath Test Notice. The argument was that, because since the date of the Ministerial approval the French manufacturers had provided an upgraded EPROM for the model (a microprocessor controlling the software supplied as part of the machine), the Seres was a new device. The submission was that the Minister had not approved the upgraded device as a new device and the former Ministerial approval did not apply. It was held that the device was still an approved device and the fact that a manufacturer modifies a device from time to time does not convert it into a different kind of device.
[36] Mr Pike submitted, and we accept, that the comment in McKay about safeguards was concerned solely with the question of whether it would be inevitable that no matter what the kind of modification made, or the effects of any modification on the reliability of the kind of device already approved, there could be no challenge to the breath test result so long as the model designation of the device remained unaltered. The second safeguard was not related to a particular device but to the situation where a post-approval modification put the reliability of the devices in general at issue. It was treated as an implicit general limitation of the Minister’s approval.
[37] We now move on to examine the legislative framework. The legislation at issue in this case is the Land Transport Act 1998 before its amendment in 2001. Under s56(1) of that Act a person over 20 commits an offence if the person drives or attempts to drive a motor vehicle on a road while the proportion of alcohol in the person’s breath, as ascertained by an evidential breath test subsequently undergone by the person, exceeds 400 micrograms of alcohol per litre of breath. Section 56(2) deals with blood-alcohol offences in similar terms.
[38] Section 77(1) provides that there is a conclusive presumption that the proportion of alcohol in a defendant’s breath at the time of the alleged offence (ie the driving or attempted driving) was the same as the proportion of alcohol in the person’s breath as indicated by the evidential breath test. Section 77(2) deals with blood tests in similar terms. These subsections thus deem the blood/breath, or blood/alcohol results achieved through subsequent testing, to be the same as at the time of the offence. They do not explicitly provide that the breath test or the blood test is conclusive per se. Indeed a blood test is clearly not conclusive. The defendant must be provided with a sample of the blood taken in the test and it is open to the defendant to have that blood analysed privately. Mr Pike accepted that, in a case where the results obtained differed from those obtained by the Crown testing, there would be a defence available that the defendant’s analysis is correct.
[39] Mr Pike submitted, however, that evidential breath tests are conclusive. This is on the basis that the evidential breath testing equipment is approved as conclusive by a notice which has the status of subordinate legislation. The term “conclusive evidential breath-testing device” is defined in the Land Transport Act in s2 as:

an evidential breath testing device approved by the Minister of Police, by notice in the Gazette, as conclusive without the option of a blood test for the purposes of section 77.

[40] Section 77(5) provides that the result of a positive evidential breath test is admissible without the option of a blood test being given if the test indicates that the proportion of alcohol in the breath of the person who underwent the test exceeded 600 micrograms of alcohol per litre of breath. If the reading is in the range of 400-600 a blood test option is available but the breath result is admissible if the person does not take up that option.
[41] Mr Pike submitted that the scheme of the Act is that there is a very wide margin for possible error with the blood test safeguard up until the point when the result is 50% higher than the prescribed adult limit. He submitted that it would be against the scheme of the legislation if a person with a result in the 400-600 range could refuse the option of a blood test but still challenge the result on the basis of malfunction of a particular device on a particular occasion.
[42] He pointed out too that the specification for the conclusive devices sold to New Zealand includes provision for two breath samples to be given and for the device to produce no evidential result in cases where both breath samples are over 600, but vary by 15%. Mr Pike referred to Duell v Ministry of Transport [1993] 1 NZLR 13, which confirmed that this was the case, as did Dr Gainsford in his oral evidence. Dr Gainsford was referred to the case of one of the defendants in which the first reading was 552 and the second was 691, a 25% variance. He accepted that in this situation the machine is not programmed to issue an incomplete test result but the person would have the automatic right to elect to have a blood-alcohol test.
[43] Mr Pike also pointed out that both the Intoxilyzer 5000 and the Seres are programmed to default to a negative result of 400 micrograms per litre where the measured result is up to but not including 440 micrograms per litre. Dr Gainsford’s evidence was that there was a standard deviation in the manufacturer’s specification of 20 mcg per litre and the ESR had imposed upon themselves a specification of +/-2.5% which he described as a variation present in any measuring device. We note also that both devices are self-test devices. This means, as Dr Gainsford explained in his evidence, that once the device is in the field the internal standards lock into the calibration which has been set into the device, and the device will fail, or in other words not produce the breath test result, if any of those internal standards and a number of other checks and tests fail.
[44] We agree that the scheme of the 1998 Act suggests that the result of an evidential breath test by an approved device was intended to be assumed to be reliable. This is particularly the case where the blood test option was available and refused. The question is whether the result is to be assumed infallible or whether it is only to be deemed presumptively reliable.
[45] The absence of a defence of malfunction on a particular occasion was not made explicit in the legislation. Such a limitation on a person’s right to defend him or herself should normally be explicit rather than implicit. We consider it significant that the Land Transport Act was passed after the case of McKay and also after a number of cases based upon (admittedly a misunderstanding of) McKay, where the existence of a defence of malfunction of a particular device was assumed. Parliament must be taken to have been aware of these cases when it passed the 1998 Act and the fact that the defence was not excluded can thus be taken as Parliamentary recognition of it. The case of Allen v Police [1999] 1 NZLR which, Mr Pike submitted, had boosted the “McKay” defence, was decided in August 1998, some four months before the 1998 Act was finally passed.
[46] In Allen’s case Giles J said that, following McKay, it was clear that the reliability of any breath-testing device could be put at issue provided there was a proper evidential foundation. The Judge referred to an earlier case, Burrows v Ministry of Transport HC AK AP 149/91 13 August 1991, in which Barker J said (at p8-9), in the context of the failure of the prosecution to provide certification records, logbooks and repair records relating to a breath-testing device:

Also of relevance is the fact that the Intoxilyzer ... is deemed to be a conclusive device. Once its reading over 600 is accepted, the suspect has no right to demand a blood test. Consequently, defence counsel may feel under a duty to be more diligent in checking that a given device was properly working for any given evidential breath test.

[47] Giles J also noted an unreported decision of this Court, Lloydd v R CA378/94, 1 March 1995, a case heard soon after McKay. In that case, however, the Court merely noted what had been said in McKay but there was no argument on this point as the case was concerned only with the evidential foundation necessary to put reliability at issue.
[48] We also consider it significant that Parliament recognised the existence of a defence of malfunction in the 2001 amendments. The purpose of the Land Transport (Road Safety Enforcement) Amendment Act 2001, as set out in s3(a), is to improve road safety by:

Removing the defence of error or possible error in the result of an evidential breath test and ensuring that an evidential breath test is conclusive, but subject to safeguards to protect the rights of defendants.

In addition, s5(2) amended s64(4)(a) of the Land Transport Act (contained under the heading of “defences”) which now reads that it is no defence to proceedings for an offence against the Act that: “there was or may have been an error in the result of the breath screening test or evidential breath test” (emphasis added).

[49] The 2001 amendments introduced a new s75A which provides for a certificate of compliance to be provided in respect of a device. Under s75A(3) the certificate must be treated as conclusive evidence of the matters stated in the certificate and neither the matters stated in the certificate nor the manufacturer’s specifications for the device concerned may be challenged, called into question or put in issue in any proceedings in respect of an offence involving excess breath alcohol recorded by the device. These amendments were fully debated and Parliament added a new safeguard by providing the possibility of blood tests in cases where the result is over 600 as well as where it is under 600. That can be regarded as a quid pro quo for the removal of the ability to challenge the reliability of a breath testing device on a particular occasion.
[50] Finally this Court in R v Livingston [2001] 1 NZLR 167, at para [10] and Attorney-General v Otahuhu District Court and Brown [2001] 3 NZLR 740, at para [44] proceeded on the basis that such a defence existed. While these recent comments were obiter, they were considered judgments and we see no reason to depart from them.
[51] We thus conclude that there is, in respect of alleged offences prior to the commencement of the 2001 amending legislation, a defence of possible malfunction of a particular device on a particular occasion. In our view, however, that defence is a very narrow one indeed. In Livingston at para [9] this Court made it clear that the comments in McKay are limited to the particular context (ie where modifications have been made) and:

should not be read as permitting a general attack on the reliability of tests when an approved device has been operated in any approved manner. That would undermine the whole purpose of the statutory approval regime.

[52] We comment also that the Ministerial approval must be taken to assume maintenance broadly within the parameters set by the manufacturers and to assume normal operating breakdowns and repairs again broadly conducted in accordance with the manufacturer’s specifications. We use the word “broadly” to highlight that the Minister must have assumed that maintenance would be conducted by a professional body like the ESR and that any deviations from the manufacturer’s specifications would be based on that professional body’s experience and expertise. In our view also the Minister’s approval must further be taken to assume that faults will either be detected by the self-test mechanisms and no result obtained or, if a result is obtained, that no error would cause such a degree of erroneous result that a result of over 600 should in fact have been under 400. This follows from the safeguard of a blood test being provided for results under 600 only.
[53] The Court in Livingston accepted (at para[10]) that, if there is a specific complaint that the device did not for some particular reason produce a reliable result on the particular occasion, then it is open to a defendant to challenge the reliability of the particular result. However, it was made clear that, in order to have the point left to the jury or fact finder, the defendant must satisfy an evidentiary threshold and convince the judge that the jury could be left with a reasonable doubt about the reliability of the device on the particular occasion. We would add that the reasonable doubt must be a doubt not as to reliability per se but must also take into account the possible extent of the unreliability (i.e. take into account that the jury in order to acquit must consider that there is a reasonable possibility that the result should have been 400 or under). If there is an evidential foundation laid, then the Crown must prove reliability (in the sense set out above) beyond reasonable doubt. The Court went on (at para [13]) to comment that:

[w]ithout such foundation the presumptive reliability of a test which has been properly conducted with an approved device will prevail, as must have been the legislative intention.

[54] In accordance with these principles, the first consequential point is that any unreliability challenge as to the working of a particular device on a particular occasion cannot be a disguised challenge to the reliability of the type of device in general. The second point is that an adequate evidentiary foundation must be laid for any such challenge. This will not be an easy task, given the statutory scheme which provides for a presumption of reliability (although not infallibility) as a result of the Ministerial approval regime.
[55] The evidential foundation may be able to be laid if there is evidence of an outside event such as dropping the device just before the test (an example given in Livingston, para [10]). The other example given in Livingston is lack of proper maintenance. As indicated above, the Ministerial approval must be assumed to be given on the basis that maintenance is carried out by a professional body. Any lack of maintenance therefore must be of such proportions as to be outside any view of proper maintenance by such a body with practical experience of breath testing machines and also be of such a character that doubts could be raised as to reliability to the extent required to cause a true result of under 400 to be read as being above 600. The same applies in respect of alleged faults and alleged substandard repairs. As this Court said in Brown at para [45]:

Clearly, too, merely to suggest or even prove deficiencies in maintenance of the particular machine will not suffice. Those deficiencies must be shown to be capable of affecting the reliability of the particular result...In terms of the statutory scheme, unreliability is not lightly to be inferred.

Are the instrument manuals relevant?

[56] We now go on to examine whether the instrument manual is relevant to the appellants’ cases in the light of the conclusions above. In Mr Tuiavii’s case, the possible fault identified is a voltage drift. The question in relation to that is whether the device in question is sensitive to a drift of 25mV. That is a narrow question and it does not appear to us that access to the whole of the manual is necessary to answer it.
[57] We have already commented above on other reasons put forward by Mr Churchill for needing access to the manual in Mr Tuiavii’s case. The reasons have little or no relevance to the actual fault. We accept that Dr Gainsford in cross-examination said that it is vital for a person servicing, maintaining, calibrating and repairing a device to refer to the instrument manual. It would seem to us logically to follow that, if a general challenge to proper maintenance and repair were open, then it would be necessary to have access to the manual so as to be able to assess the departures from the standards set by the manufacturer and perhaps also to assess the consequences of any such departures. As we have already indicated, there is, however, no such general challenge available. There must be an actual fault identified, being a fault that is capable of possibly affecting reliability to the extent required. If that does not apply, then access to the manual can only be for a fishing expedition to try and identify a fault or for the purpose of mounting a general challenge on the reliability of the device in question. The witness summons process cannot be used for this purpose in the context of the breath-alcohol statutory scheme.
[58] We comment that in Mr Tuiavii’s case there also appears to be a real problem, at least on the current state of the evidence, in there being a sufficient evidential foundation for a defence based on malfunction of the device in question, even given the possible fault identified. This is because the blood/breath correlations for the period referred to by Mr Churchill strongly suggest that any fault had little or no effect on the operation of the device and certainly not such an effect as to cause a result of over 600 possibly to reduce to 400 or under. Mr Churchill was able to suggest that the fault may have affected the accuracy of the result but was not able to comment on the possible magnitude of any error, and there is no defence evidence challenging the blood/breath correlations or their interpretation.
[59] In the case of Mr Livingston, the only specific reason put forward for access to the manual is to identify the functions that were not included within the self-test functions. This appears to us to be a disguised attack on the device in general and therefore not legitimate. Mr Churchill also indicated that in his view maintenance and repairs had not been conducted in a professional manner. Apart from a concern identified as to failure to maintain voltage within the limits prescribed by the manufacturer, he is not more specific. What he appears to be attempting therefore is a generalised attack on maintenance and again that is not within the scheme of the legislation. In terms of the two faults identified, it is not clear why, or even if, Mr Churchill requires access to the manual (except to relate these faults to the self-test procedures). We comment too that, as in Mr Tuiavii’s case, it would appear that any need for voltage information at least (and possibly also information on sample sizes) does not require access to the whole of the manual.
[60] Given the limited nature of any defence of malfunction and the limited nature of the alleged faults identified, we are of the view that the instrument manuals are irrelevant to the appellants’ cases. Williams J was therefore correct to set the witness summonses aside.

Were the witness summonses issued for an improper purpose?

[61] The ESR and Dr Gainsford submitted that a further reason for setting the witness summonses aside is that they were issued for an improper purpose, namely discovery and inspection rather than for the purpose of having the instrument manuals admitted in evidence. They referred to the case of DPP v McKeown [1997] 1 All ER 737, 742 (HL) where it was said that there was no provision for discovery of documents for a summary trial and that the statutory powers to issue witness summonses should not be used as a “disguised attempt to obtain discovery”.
[62] Mr Wotherspoon submitted that as long as a witness summons is issued for the purpose of adducing a document in evidence, even if this is conditional on seeing the contents, a summons should not be set aside. He submitted that there will inevitably in such cases be an element of discovery in any witness summons of this type and points to the decision of Phillips J in R v Clowes (1992) 95 Cr App R 440, 448; [1992] 3 All ER 440, 449 where he said:

But, where a document is likely to contain material evidence the precise nature of which is not known to a party, the issue of a witness summons by that party is likely to be motivated both by a desire to see the contents of that document and a desire, conditional perhaps upon the precise nature of those contents, to adduce the evidence before the jury.

[63] We do not accept that the witness summons procedure can be used as a means of obtaining discovery of documents held by third parties. It is implicit in the taking out of the summons that the documents ordered to be brought to Court are relevant and otherwise admissible in the proceedings. There is no justification for invoking that procedure to ascertain whether the documents are in fact relevant and otherwise admissible. The party issuing the witness summons must also desire to have the documents adduced in evidence. The desire to adduce them in evidence can be conditional in the sense described by Phillips J but, as we have stressed, the documents must be admissible. We have held that it has not been shown that the manuals are relevant to the appellants’ cases. This means that they could not be adduced in evidence. It follows that the witness summonses were not issued for a proper purpose.

Should the evidence have been excluded in terms of s35EAA?

[64] We have held the manuals are irrelevant to the appellants’ cases. This means that the witness summonses were properly set aside and thus strictly there is no need to consider s35 of the Evidence Amendment Act (No. 2) 1980. We do, however, make the following points.
[65] We deal first with Mr Wotherspoon’s submission that Williams J was wrong to rely on s35 as that section applies only at the trial. While this is strictly the case, we consider that it is a relevant consideration in deciding whether to set aside a witness summons that a witness may be excused at trial from producing a document on the grounds set out in s35.
[66] Even aside from s35, there is a balancing exercise that must be conducted when considering a challenge to a witness summons. This Court commented in Brown at para [51] on the considerations to be taken into account where access to documentary material is challenged. In such a case a court:

is required to weigh and balance the various factors raised to determine whether in the interest of justice access to the material should be provided and, if so, on what terms. It is not sufficient... that the material is likely to be relevant to the issues at the trial or that it may assist the defence. The importance of the material to a fair trial has to be such as to override all other public interest considerations, including privacy and property rights of the non-party in possession of the material and any confidentiality obligations and confidentiality rights of third parties.

[67] The Court in Brown also dealt with confidentiality issues, as recognised in s35. It stated that any discretion under that section has to be exercised in light of the right to a fair trial affirmed by s25 of the Bill of Rights but, where other competing interests are also guaranteed by the Bill of Rights, “the exercise of the discretion will weigh those rights of the non-party along with other public interest considerations, including, of course the fair trial rights affirmed by s25 ”(para [52]). Earlier, at para [46], the Court had said that the non-party holder of material has rights to privacy and property and referred to the right to be secure against unreasonable search or seizure under s21 of the Bill of Rights, which “reflects the amalgam of property, privacy and individual freedom values.”
[68] We comment too that there was evidence on which Williams J was properly able to find that a relevant special relationship exists between ESR and Dr Gainsford and the respective manufacturers. While the evidence as to the possible withdrawal of the manuals could have been stronger (in that sworn evidence from the manufacturers could have been introduced), there was sufficient evidence for Williams J to consider withdrawal a possibility. The consequences of this for the breath-alcohol regime would be severe, even if perhaps they may not be insurmountable. The breath testing regime is central to the road safety programme as it relates to control of drink driving which, as is well accepted, is a significant contributor to fatal and seriously injurious road crashes. The breath testing regime is less intrusive on the public than blood testing would be. As the substantial fall in deaths and serious injuries on the roads over recent years suggests, it is a highly mobile and efficient tool in the deterrence and detection of drink driving. Anything that may generally threaten or inhibit that road safety regime is unlikely to be in the public interest.

Should a decision on the witness summonses have been left for trial?

[69] There may be some cases where a decision on whether a witness summons should be set aside ought to be left to trial. Given the lack of relevance of the manuals in this case and the public interest factors discussed above, we do not consider that this is such a case.
[70] We observe that if at trial the ESR or its witnesses did actually seek to rely on the manual in the giving of their evidence, despite, as we understand the position, not having done so in the past, that might show that, contrary to the conclusion we have reached, the manual did have relevance to the particular case.

Result and costs

[71] For the reasons given above the appeals of Mr Tuiavii and Mr Livingston are dismissed.
[72] We are unsure whether the appellants are legally aided. If there is any question of costs the parties have leave to file memoranda. In the case of the Crown, the ESR and Dr Gainsford any such memoranda must be filed and served on or before 3 July 2003 and by Mr Tuiavii and Mr Livingston on or before 17 July 2003.

Solicitors:
Kensington Swan, Wellington for Respondents
Crown Law Office, Wellington for Attorney-General

NZLII: Copyright Policy | Disclaimers | Privacy Policy | Feedback
URL: http://www.nzlii.org/nz/cases/NZCA/2003/114.html