NZMJ, First Read: Dr Stephen Child

Journal of the New Zealand Medical Association Vol 128 | No 1422 | 25 September 2015 Updated New Zealand health system cost estimates: further improvements in the age of ‘big data’

Trust, transparency: and why we need them both

New Zealand’s growing thirst OBITUARY: Analysis of the Auckland 2014 for a sugar-sweetened measles outbreak indicates that beverage tax Sir Patrick William Eisdell Moore adolescents and young adults Rethinking the conceptual could benefit from catch-up toolkit of organ gifting (2 May 1930 – 7 March 2015) vaccination Publication Information published by the New Zealand Medical Association

NZMA Chairman To contribute to the NZMJ, first read: Dr Stephen Child www.nzma.org.nz/journal/contribute

NZMJ Editor Other enquiries to: Professor Frank Frizelle NZMA PO Box 156 NZMA Communications Manager The Terrace Sharon Cuzens Wellington 6140 Phone: (04) 472 4741 NZMJ Production Jeremiah Boniface © NZMA 2015

To subscribe to the NZMJ, email [email protected]

Subscription to the New Zealand Medical Journal is free and automatic to NZMA members. Private subscription is available to institutions, to people who are not medical practitioners, and to medical practitioners who live outside New Zealand. Subscription rates are below. All access to the NZMJ is by login and password, but IP access is available to some subscribers. Read our Conditions of access for subscribers for further information www.nzma.org.nz/journal/subscribe/conditions-of-access If you are a member or a subscriber and have not yet received your login and password, or wish to receive email alerts, please email: [email protected] The NZMA also publishes the NZMJ Digest. This online magazine is sent out to members and subscribers 10 times a year and contains selected material from the NZMJ, along with all obituaries, summaries of all articles, and other NZMA and health sector news and information.

Subscription rates for 2015 New Zealand subscription rates Overseas subscription rates Individuals* $290 Individual $402 Individual article $25 Institutions $543 Individual article $25

*NZ individual subscribers must not be doctors (access is via NZMA Membership) New Zealand rates include GST. No GST is included in international rates. Note, subscription for part of a year is available at pro rata rates. Please email [email protected] for more information. Individual articles are available for purchase by emailing [email protected]

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 2 www.nzma.org.nz/journal CONTENTS

EDITORIAL 53 7 Analysis of the Auckland 2014 Trust, transparency: and why we measles outbreak indicates that need them both adolescents and young adults Stephen Child, Sanji Gunasekara could benefit from catch-up vaccination 10 Gary Reynolds, Cassandra Dias, Simon Administrative health data in Thornley, Ronald King, Anne Morrison, New Zealand: we have come Angela Matson, Richard Hoskins so far; where are the next opportunities? VIEWPOINT Wing Cheuk Chan 63 ARTICLES Rethinking the conceptual toolkit of organ gifting 13 Rhonda M Shaw Updated New Zealand health system cost estimates from health CLINICAL CORRESPONDENCE events by sex, age and proximity 69 to death: further improvements Conjunctival squamous cell in the age of ‘big data’ carcinoma Tony Blakely, June Atkinson, Giorgi Ali Mahdavi Fard, Leili Pourafkari, Kvizhinadze, Nhung Nghiem, Heather Nader D Nader McLeod, Anna Davies, Nick Wilson 70 24 A case of perforated chronic Predictors of vitamin D status in idiopathic megacolon pregnant women in New Zealand Benjamin Cribb, Rukshan Ranjan,

Alec J Ekeroma,Carlos A Camargo Jr, Nigel Henderson Robert Scragg, Clare Wall, Alistair Stewart, Ed Mitchell, Julian LETTERS

Crane, Cameron C Grant 73 35 Out of touch? The shortcoming The community pharmacy-based of New Zealand’s amended Sale anticoagulation management and Supply of Alcohol Act (2012) service achieves a consistently high for the Rugby World Cup (2015) standard of anticoagulant care Benjamin Riordan, Tamlin Conner, Paul Harper, Ian McMichael, Dale Jayde Flett, Damian Scarf Griffiths, Joe Harper, Claire Hill 75 45 Accident Compensation Prevalence of human Corporation claim status and papillomaviruses in the mouths benefit type is associated with of New Zealand women low back pain outcomes Rebecca Lucas-Roxburgh, Jackie Jon Cornwall, Achim Elfering, Rebecca Benschop, Magdalena Dunowska, J Crawford, Markus Melloh Matthew Perrott

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 3 www.nzma.org.nz/journal CONTENTS

78 OBITUARY Fluoxetine-induced phenytoin 88 toxicity: a clinical reminder about Sir Patrick William Eisdell Moore the perils of polypharmacy Sean Lance, Ian Ternouth 80 92 New Zealand’s growing thirst for a Methuselah sugar-sweetened beverage tax Gerhard Sundborn, Simon Thornley, 93 Bodo Lang, Rob Beaglehole 100 years ago: Measles in adults Farquhar Matheson, M.R, Glasgow, 83 House Surgeon, Wellington Hospital Is our focus on pharmaceutical company influence too narrow? 95 Erratum 128:1421 Lance Gravatt 85 Comment on: Getting serious about protecting New Zealand children against unhealthy food marketing Katherine Rich

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 4 www.nzma.org.nz/journal SUMMARIES

Predictors of vitamin D status in pregnant women in New Zealand Alec J Ekeroma, Carlos A Camargo Jr, Robert Scragg, Clare Wall, Alistair Stewart, Ed Mitchell, Julian Crane, Cameron C Grant New Zealand has a sun avoidance health policy and minimal dietary vitamin D fortification. Vitamin D deficiency was present in 109/259 (42%) of pregnant women in a south Auckland cohort. Of those enrolled in winter (June-August)/spring (September-November), vitamin D deficiency was present in 43% of European, 67% of Māori, 80% of Pacific and 59% of women of other ethnic groups. Supplementation for all pregnant women during winter/spring could be an appropriate intervention for prevention of vitamin D deficiency during pregnancy in New Zealand.

The community pharmacy-based anticoagulation management service achieves a consistently high standard of anticoagulant care Paul Harper, Ian McMichael, Dale Griffiths, Joe Harper, Caire Hill People taking the blood thinning drug, warfarin, require regular blood tests to maintain safe control. In the past this service was provided by doctors using laboratory based blood tests. In 2012 a service was introduced to enable trained pharmacists to provide warfarin management through community pharmacies. Our audit confirms that pharmacists provide a safe, efficient and convenient service for patients and that the standard of anticoagulant control is above international recommendations and has remained high in spite of a five-fold increase in the number of patients using the service.

Analysis of the Auckland 2014 measles outbreak indicates that adolescents and young adults could benefit from catch-up vaccination Gary Reynolds, Cassandra Dias, Simon Thornley, Ronald King, Anne Morrison, Angela Matson, Richard Hoskins A single child with measles at a high school would almost certainly cause a serious outbreak, because immunity in that age group is well below the national average. The rate of immunity in New Zealand is close to 95%, which suggests a high level of herd immunity, but the level of immunity among secondary school age children is between 65 per cent and 80 per cent. One of the reasons is that a controversial study linking the measles mumps and rubella vaccine to autism – later proven to be false - persuaded many parents not to vaccinate their children at that time. That means they remain vulnerable today, especially while gathered together at school. They would benefit greatly from national, targeted vaccination catch-up.

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 5 www.nzma.org.nz/journal SUMMARIES

Updated New Zealand health system cost estimates from health events by sex, age and proximity to death: further improvements in the age of ‘big data’ Tony Blakely, June Atkinson, Giorgi Kvizhinadze, Nhung Nghiem, Heather McLeod, Anna Davies, Nick Wilson This study benefited from the relatively high quality and comprehensiveness of health cost data. It showed how widely this expenditure varied across the life course – with relatively more being spent in the last year of life (though this varied widely by age of death). This analysis has benefited from quality improvements in cost data and methods refinements but yet further improvements in coming years are likely. This is particularly so with access to additional data sources and with the move towards better integration of “big data” in the New Zealand health sector.

Prevalence of human papillomaviruses in the mouths of New Zealand women Rebecca Lucas-Roxburgh, Jackie Benschop, Magdalena Dunowska, Matthew Robert Perrott Human papillomavirus (HPV) infection in the mouth is associated with an increased risk of developing HPV-positive head and neck cancer. This study of 234 young women found HPV in the mouths of 7 women. Of those positive for HPV, two were positive for an HPV type able to cause cancer. The other five women were positive for a non-cancer causing HPV. No associations were found between putative risk factors (smoking, alcohol consumption, and the number of sexual partners) and the presence of oral HPV infection.

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 6 www.nzma.org.nz/journal EDITORIAL

Trust, transparency: and why we need them both Stephen Child, Sanji Gunasekara

Perhaps the culture of accountability that we are relentlessly building for ourselves actually damages trust rather than supporting it. Plants don't flourish when we pull them up too often to check how their roots are growing: political, institutional and professional life too may not go well if we constantly uproot them to demonstrate that everything is transparent and trustworthy." – Onora O’Neill1

alls have been made recently for the profession. On the contrary, the more we publication of outcome data from try to increase the ‘accountability’ of the Cindividual surgeons.2 While the intent profession, the more we reduce the moti- of this is laudable—wanting to be transpar- vation of altruism, which is designed to put ent about care and to assist patient decision patients’ interests above our own interests. making—the implementation may not only On a wider scale, many ‘accountability’ present difficulties but is a direct affront measures—such as rewarding with to the trustworthiness component of our financial incentives or developing clinical doctor-patient relationship. pathways and then punishing or rewarding The New Zealand Medical Association to reduce variation—are developed because recognises the value of reliable, accurate confusion has evolved between measures data in informing clinicians, and in influ- that are supposed to reflect accountability, encing planning and investment in health and doctors being trusted to deliver the systems. As a professional body, we want to best possible care to our patients in the ensure the development of robust systems most efficient manner. Of course, key to that will generate meaningful data. We this argument is that we, as doctors, must agree with the underlying principles of the remain ‘trustworthy’ and to honour our Medical Council’s view that—if gathered commitment to professionalism. accurately, used correctly and explained Our concerns with the public reporting well—qualitative data down to individual of outcome/performance data are based clinician level could benefit clinicians, on a number of issues with a sampled few 2 administrators and regulators. outlined below; In other words, the concept of • variation in the type and quality of ‘transparency’ is laudable but that of outcome data collected: Reporting ‘accountability’ may have unintended poor quality outcome data without consequences. identifying the necessary limitations We remain unconvinced of the perceived will lead to meaningless debate and benefits of releasing these data to the public generate spurious conclusions about and in particular, any moves to report ‘raw’ apparent variations. The consequences or partially-adjusted performance/outcome of false identification of poor perfor- data. While the role of data in clinical mance can negatively affect clinicians improvement and Quality Assurance (QA) is and patients. incontrovertible, the public release of such • multiple confounding factors. Failure data is not. to adequately take these confounding ‘Blaming and shaming’ individual factors into account could lead to surgeons will not improve patient care misleading information that does not nor increase the trustworthiness of the reflect the actual competence of any

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 7 www.nzma.org.nz/journal EDITORIAL

individual named as lead clinician. If • limitations of observational data: reported in its raw format to the public, Much of the evidence referred to here unadjusted outcome data could also is non-experimental observational undermine confidence in the public data and, despite risk adjustment, health system. These factors4 include: confounding is inherent—which is • patient factors (eg, age, co-morbidities, precisely the problem with the use of ethnicity, socioeconomic deprivation, surgeon-specific data; the low-grade health literacy, diagnostic validity, evidence may be useful for gener- complexity/severity of condition) ating hypotheses but is inadequate for robustly testing these. A risk • system factors (eg, diagnostic/inter- assessment-based system might allow ventional facilities, healthcare team comparison of actual outcome against factors, supervision, resources vs predicted outcome over time, but still competing demands, management and understates the complexity, and may in 3,5 governance) itself compound the issue. • clinician factors (eg, volume of proce- • undermining the environment 6 dures, training, experience, and that is most conducive for quality case-mix). improvement: The best environment • failure to recognise the impact for quality improvement is one where of systems, processes and infra- clinicians feel safe to disclose adverse structure: Post-operative outcomes events and near misses, openly and (including mortality) relate to many frankly. Publicly reporting incomplete, factors other than the operative skill/ confounded outcome/ performance competence of the surgeon, including data could undermine the very envi- the quality of postoperative and ronment that is most conducive for nursing care, etc. quality improvement. • failure to take into account the • overlooking the performance contribution of the multidisciplinary outcomes of executive managers: team: Reporting outcomes for indi- Executive managers must be held as vidual surgeons ignores the effect of the accountable as clinicians for health multidisciplinary team and the context outcomes that matter. Indeed at Mid in which surgery is done.6 A lack of Staffordshire,5 failings in hospital evidence of poor performance is not management and overall systems evidence of acceptable performance. were found to be major contributors to the adverse health outcomes. The • promoting the practice of risk- medical profession might, even- averse medicine: Fears of a negative tually, become more comfortable with outcome can lead to a range of negative publicly reporting on individual clini- impacts—from unnecessary inves- cians’ outcomes when there is similar tigations (and associated costs) to reporting of: reduced learning opportunities for trainees, to a disincentive to provide • individual managers’ outcomes—over treatment for the most seriously ill time and regardless of location patients. In the UK, for example, publi- • health funders, for the populations cation of surgeon-specific data (SSD) they serve—measuring opportunity has coincided with a drop in both costs and health benefits forgone by the proportion and variety of cases the mix of decisions made to fund some performed by trainees, suggesting things and not others. that the publication of SSD provided a • failure to recognise professionalism disincentive for consultants to provide and the doctor-patient trust surgical training.7 The best surgeons relationship: Discussion on this may take on the highest risk cases, issue should explicitly recognise the which may have worse outcomes. professionalism that underpins the Public reporting may lead to the autonomy granted to the medical selection of lower risk patients.8 profession via the original societal

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 8 www.nzma.org.nz/journal EDITORIAL

contract. Analogies between doctors Overall, it is much more appropriate and pure service providers are to look at a unit, service or DHB-level unhelpful and fail to acknowledge analysis, which includes the total team the unique nature of the doctor- effect and not just one participant in patient trust relationship. Providing the care. raw performance data to the public, Thought also needs to be given to what a without context, is contrary to the patient can realistically do with clinician- professionalism that is at the core of specific outcome/performance data, being a doctor. particularly as most New Zealanders rely on • cost considerations: Any costs of the public hospital system. The notion that generating and managing performance/ such information can facilitate a patient’s outcome data are likely to come out of informed choice of clinician is neither Vote Health. Could the limited health tenable nor ethical (if the only choice dollar and resources be put to better use available entails a choice between public to improve the health of New Zealanders and private health care). and to reduce health disparities? What patients and the public want to • advantages of reporting at an know is that the doctors caring for them aggregate level: There are compelling are competent; public release of inade- arguments to focus on performance at quate data inadequately explained will not relatively aggregated levels rather than achieve that and has potential to “disturb at the level of individual practitioners. the roots while checking the tree”.

Competing interests: Nil Author information: Stephen Child, Director Clinical Education Training Unit, Auckland DHB; Chair, New Zealand Medical Association. Sanji Gunasekara, Policy Manager, New Zealand Medical Association. Corresponding author: Dr Stephen Child, Auckland City Hospital, Private Bag 92024, Auckland 1023. [email protected] URL: www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2015/vol-128-no-1422-25- september-2015/6659

REFERENCES: 1. Reith Lectures 2002, A for monitoring the results of surgeon-specific data Question of Trust - Onora of medical care. JAMA. reporting on surgical O’Neill (5 Lectures); 2005. 1989 Aug 18;262(7):925–30. training. Ann R Coll Surg Available from http:// Available from http://www. Engl. 2007 Nov;89(8):796-8. bufvc.ac.uk/dvdfind/ prgs.edu/content/dam/rand/ 8. Shahian DM, Edwards index.php/title/av36152 pubs/notes/2009/N3038.pdf FH, Jacobs JP, et al. Public 2. Better Data – the benefits 5. Delamothe T. Government's reporting of cardiac to the profession and initial response to Mid surgery performance: Part the public. Andrew Staffordshire report. BMJ. 1--history, rationale, conse- Connolly, Chairman, 2013 Apr 9;346:f2209; quences. Ann Thorac Surg. Medical Council of New Pollock AM, Price D. Mid 2011 Sep;92(3 Suppl):S2–11. Zealand, 25 March 2015. Staffordshire should lead 9. Ministry of Health. 3. Richard R. All changed, to a fundamental rethink Protected Quality Assur- changed utterly: British of government policy. BMJ. ance Activities under medicine will be trans- 2013 Apr 8;346:f2190. the Health Practitioners formed by the Bristol case. 6. Walker K, Neuburger J, Competence Assurance Act BMJ 1998;316(7149):1917– Groene O, et al. Public 2003. July 2014. Available 1918. Available from http:// reporting of surgeon from http://www.health. www.ncbi.nlm.nih.gov/ outcomes: low numbers govt.nz/publication/protect- pmc/articles/PMC1113398/ of procedures lead to false ed-quality-assurance-ac- 4. Tarlov AR, Ware JE Jr, complacency. Lancet. 2013 tivities-under-health-prac- Greenfield S, et al. The Nov 16;382(9905):1674–7 titioners-competence-as- Medical Outcomes Study. 7. Khan OA, Iyengar S, surance-act-2003 An application of methods Pontefract DE, et al. Impact

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 9 www.nzma.org.nz/journal EDITORIAL

Administrative health data in New Zealand: we have come so far; where are the next opportunities? Wing Cheuk Chan

n this issue of the Journal, Blakely et al Examples of research or clinical audit provide an update on their New Zealand opportunities in the future that could be Ihealth system cost estimates based on explored via data linkage of New Zealand health events that were captured in the administrative data may include (and not New Zealand administrative data by age, be limited to): sex and proximity to death.1 Consistent 1. Describing possible treatment gaps in with previous published studies,2,3 their a patient subgroup where there may work demonstrates that the captured health be clinical safety concerns or oppor- system costs were skewed to the last year tunity for quality improvement: eg, of life. Furthermore, Blakely et al included are there people with a mechanical an updated discussion on the strengths and heart valve who are not on any form limitations of the use of the New Zealand of anticoagulation for more than a administrative data.1 year? We should celebrate the fact that New 2. Undertaking a retrospective cohort Zealand has a national unique patient study to describe outcomes of patient identifier: the National Health Index (NHI) subgroups who had a procedure in for all health care users in New Zealand, hospital where benefits and risks something that many developed countries were less certain: eg, what is the cannot claim to have. There are ongoing 2-year case fatality rate for people processes in place to ensure duplicated over 80 years of age and who had a NHIs are cleaned or merged as appro- right hemicolectomy for colorectal priate.4 The unique NHI has enabled many cancer by cancer stage? linkage studies to be carried out at the 3. Facilitating an audit of a part of a population level in New Zealand.1-3,5 The treatment pathway: eg, what propor- ongoing improvement of NHI coverage and tions of patients with breast cancer data quality over time across a number of died prior to completing chemo- administrative health datasets has opened therapy by age and stage? up a number of new opportunities, not 4. Describing hospital events and costs just for ‘professional researchers’, but also associated with an entire interven- for working clinicians who would like tional pathway: eg, transcatheter to undertake clinical audits or reviews, aortic valve implantations are often particularly in regard to longer term health carried out in a tertiary facility, but outcomes, or identification of potential costs for the associated procedures, management gaps for improvement. Having such percutaneous coronary inter- a good visibility of longer term outcomes vention or pacemaker implants, for our local populations can be challenging might sit in other hospitals where for clinicians because New Zealand has a the tertiary facility may have limited very mobile population,6 and often there is visibility. no formal access to health care data outside 5. Describing practice patterns or one’s jurisdiction to capture the complete service provision at a population sets of subsequent health service events. level: eg, how many people with

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 10 www.nzma.org.nz/journal EDITORIAL

prostate cancer were managed by by the administrative data. One option to immediate radical treatment at diag- mitigate this limitation is to combine some nosis, active surveillance or watchful of the clinical data available from other waiting in the wider Auckland region sources and link them against the admin- in New Zealand? istrative data or use administrative data Blakely et al have acknowledged Craig to facilitate part of the clinical audit. For Wright and the New Zealand Ministry of example, alerting a clinician that there were Health (MoH) for the development of the other hospital events occurring outside the Health Tracker. The MoH’s Health Tracker index hospital of interest may be helpful. In has operationalised the concept of the situations where quantitative adjustments for confounding factors are not feasible, “health service utilisation” population into then making such limitations explicit in the a reality.5 Since the number of people in appropriate context is helpful to enable the the health service utilisation population reader make a more informed interpre- is very similar to the estimated resident tation of the findings: eg, over the counter population, this suggests the population medications available in supermarkets or coverage of the people who recently used pharmacies are not captured in the Phar- health services is excellent. The Health and maceutical Collection jointly owned by the Disability Intelligence team from the MoH MoH and PHARMAC. is expecting to lead ongoing updates of the Health Tracker in due course. While there are number of safeguards in place to ensure the administrative data While there are many exciting oppor- is accurate, some errors do occasionally tunities to use administrative data more occur. If the actual errors are confirmed, widely, potential interested users should they should be reported to the responsible be fully aware of the limitations related department so that the administrative data to the use of the administrative datasets, can be corrected, or a system improvement in particular in relation to cost estimates. can be put in place, so similar errors in the As noted by Blakely et al, virtually all of future may not recur. However, we also the administrative health datasets list have to be cautious of the fact that many prices or cost weights related to the events, claims of data inaccuracies, particularly rather than the actual costs of events. from inexperienced data users, could be District Health Boards (DHBs) are funded related to poor understanding on how the predominately on a capitation basis by the original data is captured, the purpose of population-based funded formula, and only the data collection, and how they should be some of the pricing recorded in the datasets used, rather than actual errors related to are used for reimbursement purposes the administrative data themselves. between DHBs when a patient attends a It can be very helpful to liaise with the facility outside the patient’s own domicile analytical, data quality and coding teams DHB. Some services are funded based on at the MoH to clarify coding practices, data inputs rather than based on outputs, such as enquiries and discuss how data should be mental health services; it is important this used. The MoH teams can be contacted via: is kept in mind when undertaking analyses [email protected] and coding_ involving health system costs. [email protected],9 Depending on The New Zealand MoH is currently under- the nature of the enquiry, a study protocol taking a review of the population-based is often required, including the research funding formula.7 The analytical insights questions to be answered and the proposed gained as part of the review will provide methods. If potentially identifiable data further understanding on how the admin- is requested, ethics approval may be istrative data could provide robust cost required from Health and Disability Ethics estimates in a pragmatic way. Committees. Data security and safe guards The other limitation of administrative to ensure privacy and confidentiality must data is that it is often difficult to make defin- be in place. itive claims of attribution between cause Finally, Blakely et al highlight the recent and effect because there are likely to be development of Integrated Data Infra- confounding factors that are not captured structure (IDI) by Statistics New Zealand.1

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 11 www.nzma.org.nz/journal EDITORIAL

The IDI combines information in an anon- exciting platform to undertake research to ymous way from a number of sectors, gain better understanding of the broader including health sector, along with data determinants of health, and more impor- from the Ministry of Education and the tantly, to develop potential pragmatic Ministry of Social Development. While IDI solutions that improve population health is at an early phase, it is potentially a very and equity.

Competing interests: Nil Acknowledgements: I would like to thank a number of Ministry of Health staff for cross checking a number of specific sections of this editorial within a very short time frame, and Doone Winnard from Counties Manukau District Health Board for commenting on an earlier draft of the editorial. Author information: Wing Cheuk Chan, Public health physician, Population Health, Counties Manukau District Health Board. Corresponding author: Wing Cheuk Chan, Counties Manukau District Health Board [email protected] URL: www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2015/vol-128-no-1422-25- september-2015/6660

REFERENCES: 1. Blakely T, Atkinson J, funded by Counties Population Mobility. Kvizhinadze G, et al. Manukau District Health Wellington: Statistics Updated New Zealand Board for people in the New Zealand; 2007. health system cost last year of life. The New 7. Topham-Kindley L. Fresh estimates from health Zealand medical journal look at DHB funding events by sex, age and 2011; 124(1335): 40-51. formula after seven-year proximity to death: 4. Statistics New Zealand. gap. New Zealand Doctor. further improvements IDI Data Dictionary: 3 December 2014:3. in the age of ‘big data’. Population cohort name 8. Ministry of Health. New Zealand Medical and demographics data Collections. 2015. http:// Journal 2015; 128(1422). (August 2015 Edition). www.health.govt.nz/ 2. Blakely T, Atkinson J, Wellington: Statistics nz-health-statistics/ Kvizhinadze G, Nghiem New Zealand; 2015. national-collections-and- N, McLeod H, Wilson N. 5. Chan WC, Jackson G, surveys/collections Health system costs by Wright CS, et al. The (accessed 21/09/2015). sex, age and proximity to future of population 9. Ministry of Health. death, and implications registers: linking routine Submitting a coding for estimation of future health datasets to assess query. 2015. http:// expenditure. The New a population’s current www.health.govt.nz/ Zealand medical journal glycaemic status for nz-health-statistics/classi- 2014; 127(1393): 12-25. quality improvement. BMJ fication-and-terminology/ 3. Chan WC, Jackson G, open 2014; 4(4): e003975. using-icd-10-am-achi-acs/ Winnard D, Anderson 6. Statistics New Zealand. coding-queries (accessed P. Healthcare services QuickStats About 21/09/2015).

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 12 www.nzma.org.nz/journal ARTICLE

ABSTRACT AIMS: We aimed to: (i) update previous health system cost estimates (Blakely et al NZMJ 2014;127(1393)) using updated costing data and more refined methods; and (ii) provide context around current developments in the improved networking of health information systems in New Zealand. METHODS: As per our previous work, national health event data were linked for hospitalisations, inpatient procedures, outpatient events, pharmaceuticals, laboratory tests, and primary care consultations for the whole country. For each health event a cost was assigned. Health expenditure by sex and age, and proximity to death (last 6 or 12 months of life), was then calculated. RESULTS: The updated and more accurate method allocated lower amounts of total public health expenditure than the previous work: $6.1, $6.0 and $6.7 billion dollars (inflation-adjusted to 2011 NZ$) in 2007/08, 2008/09 and 2009/10 financial years, respectively. But the latter is still only 52% of total health system costs ($6.7/$12.98 billion). Health system costs for people not within six months of death were similar to the previous work, except for being reduced in the most elderly age groups (range: $495 per person-year in 10–14 year old females; to $5,239 per person-year in 85–89 year old males). Costs in the last six months of life remained highly variable by age group (by a factor of 14 and being maximal at $23,400 or more among 1–4 year olds). The proportion of cumulative health expenditure in the last year of life declined with increasing age of death: eg, 47%, 25%, 13% and 6% for individuals aged 40, 70, 80 and 90 respectively. CONCLUSIONS: Health system costs vary markedly across the life course, and are skewed to the last year of life. This analysis has benefited from quality improvements in cost data and method refinements, but further improvements in coming years are likely. This is particularly so with access to additional data sources, and with the move towards better integration of “big data” in the New Zealand health sector.

e have previously published, in ment of health costings assigns individuals this journal, estimates of health to the date of discharge, a practice which Wsystem costs by sex, age and prox- is not problematic for short-stay events, imity to death, using rich New Zealand data but is problematic for long-stay events (eg, (‘Health Tracker’).1 Since then, there have multi-year admissions to hospital-level care been substantial improvements with the leading to death) if all that cost is then attrib- data sources, and ‘learnings’ about the reli- uted to the last year (or last six months) of ability of various facets of the data. Also, im- life. Additionally, we have identified an error provements to the allocation of person-time in our previous calculations of person-time and timing of cost occurrence have been which impacted on first year of life and last developed. For example, standard treat- year (and six months) of life costings.

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 13 www.nzma.org.nz/journal ARTICLE

Given all these changes, our primary 2007 to June 2012. However, only the actual aim in this paper was to present updated 2007/08 to 2009/10 financial years were results for all the objectives and analyses used to estimate the costs, a restriction for in the previously published article, namely: two reasons. First, it is necessary to discard to estimate health system costs by demo- the most recent year of data for costs, as it graphics; for people within and not within is not possible to ‘know’ whether someone six or 12 months of death; to estimate what is within a year of death. Second, it became proportion of health system costs over a apparent that for earlier and later years, person’s life occur in the last year of life; data were not complete on all health events and to determine how much impact using and costs. costs stratified by proximity to death has on Each health event was then assigned a future national health expenditure projec- cost weight or unit price: casemix-funded tions (in the face of population growth, hospitalisations (using Ministry of Health aging and increasing longevity). cost weights per event5; primarily medical/ Our second aim was to provide addi- surgical events); community laboratory tional context to the current developments tests; non-admitted patient events (eg, in access to health data in New Zealand. outpatients and emergency departments); Third, we aimed to outline the next steps to community pharmaceuticals dispensed further improve the quality of New Zealand (including patient contribution); expected health system costing data for research. general practice costs (ie, using the capitation funding formula) and some actual By way of background, we note that general practice consultations (when not 83% of all health system expenditure in an enrolled patient in a capitated practice New Zealand is estimated to be publi- (ie, the general medical subsidy)). Goods cally funded,2 with many of the health and sales taxes were excluded as this is a events (eg, hospitalisations and outpa- transfer payment. All costs were inflation- tients) not involving any fee-for-service. adjusted to 2011 New Zealand dollars. The remainder of the costs are private and include out-of-pocket payments and Data not (as of early 2015) included in co-payments in primary care and for health our Health Tracker analysis included: lead insurance. There is growing research maternity carer-provided care; rest-home interest in understanding health system and hospice care; mental health care; dental costing at the national level in New Zealand health care outside of hospitals; patient (eg, for all costs1 and for cancer costs3), but transport (eg, National Travel Assistance); also at the district health board (DHB) level care directly funded by Accident Compen- (eg, Chan et al4). sation Corporation (ACC); and community physiotherapy. For the purposes of our objectives, missing rest-home and hospice Methods care means costs proximal to death will be We repeat below the basic methods underestimated (although these data should detailed in our previous paper analysing become accessible for research in coming cross-sectional health system cost data years; see Discussion). (albeit with some minor changes) and follow this with a list of more substantively Data management and person- updated and refined methods. time allocation We used tabular analyses on the 2007/08 Linked administrative health care to 2009/10 data, calculating summed and datasets with costs attributed average costs per person-year in each strata The New Zealand health system has had a of interest: sex by five-year age group by unique identifier of high quality since about financial year (2007/08, 2008/09, 2009/10) 1990 (the National Health Index [NHI], iden- and whether within six or 12 months of tifier). The following datasets were linked death or not. We censored people at death. using a unique identification number based Immigration data were not linked in with on the NHI identifier to create a record for these files, preventing correct censoring each New Zealander of all publically funded for emigration, but data were restricted health care events (eg, hospital admission, to individuals who were both listed as a and laboratory test) occurring from July New Zealand resident on the NHI, and

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 14 www.nzma.org.nz/journal ARTICLE

had a record in at least one of the data Updated data and methods since sources used (including being enrolled the previous analyses with a primary health organisation) in the Since the previous analyses,1 we made particular financial year. Finally, we calcu- a number of data and methods enhance- lated person-time weighted average costs ments. These are detailed in full in the over 2007/08 to 2009/10. Appendix, but include: the use of financial Estimating future health years instead of calendar years; a revised expenditure core population of New Zealand residents; We assembled Statistics New Zealand more accurate allocation of costs by timing; (SNZ) population count projections from revised restrictions when considering 2011 to 2041 for the median growth casemix funding; changes to the cost weight scenario, for males and females by five used for the calculation of hospitalisation year age-group (using the table builder costs; along with a number of other fairly at: http://www.stats.govt.nz/). The total minor improvements. population is projected to increase by 25% over these 30 years, but by over 150% Results for ages 75 and older. We then estimated The health system costs associated with future sex by age-specific mortality rates, individual health events included in Health by applying a 2% per annum mortality Tracker in 2011 NZ$, summed to $6.1, $6.0 rate reduction to the single year of age and $6.7 billion in 2007/08, 2008/09 and mortality rate from the 2010–12 SNZ life- 2009/10 financial years, respectively. But the tables. (A 2% per annum mortality rate latter still remained only 52% of total health reduction equates to gains of 2.5 years in system costs (6.7/12.98 billion). Pooling life expectancy per decade seen in recent these years, the per person-year costs by decades in New Zealand, a pattern that sex, age and proximity to death are shown seems likely to continue6). in Table 1 and Figure 1. Costs per person- We also undertook sensitivity analyses year disregarding proximity to death varied using SNZ projected mortality counts and approximately 10-fold from $535 for 10–14 rates with death data obtained from the year olds (sexes combined) to $5,600 for projection from 2011 (base year) to 2061.7 85–89 year olds (Table 1). (These SNZ projections equate to 1.3% The median values for the 21 age groups, (85–89 year old males) to 3.4% (55–59 year regardless of proximity to death, were old males and females) annual percentage $1,518 per year for males and $1,457 per changes in mortality rates, for 35+ year year for females. Removing person-time olds.) We then applied health system for people within six or 12 months of death costs for health events per person-year, did not alter these costs much at young both ‘simplistically’ using health system ages (due to death being rare), but did costs observed in 2009–11 not stratified quite considerably reduce the costs among by proximity to death, and then using the very old. For example, the cost per costs separately for people within six or person-year among 90–94 year olds (sexes 12 months of death. We did not add in any combined), regardless of proximity to death, trend data for changes in either health was $5,600, but was reduced to $4,629 if not service usage by age-group or changes in in the last six months of life. health costs over time given the very large Considering the assigned costs among uncertainties involved. For example, for the people within six months of death, these latter there is uncertainty around the future costs varied only three-fold between age ability of PHARMAC to keep constraining groups (in the under 95-year-old popu- costs of pharmaceuticals, vaccines and lation). That is, there was less percentage devices; and also the variable potential or relative variation by age in costs during performance of the New Zealand economy the last six months of life. A similar pattern (which provides the resources to fund was apparent also for costs in the last 12 health services), given its dependence on months of life (Table 1). commodity prices for exports and on inter- Differences in costs over the life course national tourism levels. between males and females showed a

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 15 www.nzma.org.nz/journal ARTICLE

Table 1: Estimated health system costs attributable to specific health events for New Zealand citizens from Health Tracker (per person- year and per death event during 2007/08 to 2009/10, in 2011 NZ$ inflation-adjusted values).*

Per person-year Per death Age Regardless of proximity Not in the last 6 months Not in the last year In the last 6 months of life In the last year of life group to death of life of life in years Female Male Both Female Male Both Female Male Both Female Male Both Female Male Both sexes sexes sexes sexes sexes <1 4,189 4,890 4,550 4,116 4,766 4,451 4,102 4,751 4,437 17,662 24,092 21,311 21,074 27,170 24,534 1–4 1,009 1,166 1,089 1,003 1,158 1,083 1,001 1,155 1,080 21,082 25,138 23,367 28,748 33,087 31,193 5–9 517 589 554 515 587 552 513 586 551 20,556 17,521 18,915 37,616 23,554 30,011 10–14 498 569 535 495 565 531 493 563 529 20,396 23,311 21,970 31,407 38,236 35,096 15–19 777 609 692 773 601 686 772 597 683 12,965 11,213 11,774 16,238 16,161 16,185 20–24 1,006 632 824 1,000 626 818 998 624 816 20,219 6,397 10,149 26,445 8,516 13,383 25–29 1,096 650 891 1,090 641 883 1,087 636 880 18,557 11,565 14,106 26,119 17,277 20,490 30–34 1,253 713 1,008 1,243 701 997 1,239 694 992 23,670 14,029 17,524 34,263 21,253 25,970 35–39 1,193 784 1,004 1,175 769 988 1,164 764 980 24,714 12,386 17,515 39,219 17,035 26,265 40–44 1,079 964 1,025 1,051 936 997 1,039 923 985 25,713 18,107 21,431 37,511 26,732 31,443 45–49 1,221 1,186 1,205 1,187 1,144 1,166 1,168 1,123 1,147 22,857 18,613 20,391 35,249 27,977 31,023 50–54 1,457 1,518 1,486 1,400 1,449 1,424 1,369 1,418 1,392 23,892 19,757 21,511 37,070 29,280 32,584 55–59 1,777 1,944 1,859 1,681 1,830 1,754 1,635 1,772 1,702 25,000 21,488 22,996 37,233 32,873 34,745 60–64 2,219 2,489 2,351 2,089 2,319 2,201 2,030 2,237 2,131 22,392 20,070 21,034 33,065 30,305 31,450 65–69 2,985 3,450 3,211 2,788 3,175 2,976 2,693 3,064 2,874 21,067 19,747 20,290 31,843 28,553 29,907 70–74 3,640 4,393 3,999 3,350 3,994 3,657 3,226 3,812 3,504 18,948 18,070 18,443 28,002 27,270 27,581 75–79 4,234 5,177 4,669 3,893 4,589 4,213 3,769 4,386 4,051 14,244 16,012 15,251 20,585 22,892 21,899 80–84 4,834 5,791 5,238 4,344 4,989 4,615 4,199 4,729 4,420 11,606 12,900 12,277 16,517 18,504 17,548 85–89 5,213 6,318 5,600 4,551 5,239 4,790 4,384 4,926 4,570 8,773 10,620 9,541 12,521 15,131 13,607 90–94 5,241 6,537 5,600 4,413 5,204 4,629 4,248 4,897 4,422 6,626 8,348 7,178 9,417 11,729 10,158 95+ 4,533 6,238 4,902 3,712 4,738 3,930 3,647 4,442 3,814 4,283 6,345 4,770 5,952 8,791 6,623 Notes: Bolded values show the higher costs when comparing females relative to males and vice versa. *But due to current data limitations this is still for only 52% of total health system costs ($6.7/$12.98 billion), see Discussion.

Figure 1: Estimated health system costs attributable to specific health events for New Zealand citizens during 2007/08 to 2009/10, disregarding proximity to death, and separately for within and not within six months of death.*

*But due to current data limitations this is still for only 52% of total health system costs ($6.7/$12.98 billion), see Discussion.

mixed picture (Table 1). Costs were higher To indicate the distribution of costs for males up to age 15 years, then costs at different points in the life course, were higher for females up to age 50 Figure 2 presents the cumulative health years (no doubt partly due to obstetric- system costs (in 2011 NZ$) for deaths at and women’s health-related costs). There different ages. This analysis is artificial were quite large sex differences in costs as it assumes a steady state (as per 2011) in the last six months of life (often higher for life span, treatment effectiveness and in females). This probably reflects higher costs throughout the life course with no incidence rates of sudden death in males discounting. (This ‘artificiality’ is, however, without preceding chronic illnesses (eg, similar to the way period life expectancy occupational injuries and suicide). is calculated, whereby mortality rates

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 16 www.nzma.org.nz/journal ARTICLE

Figure 2: Estimated cumulative health system costs from specific health events for New Zealand citizens (in 2011 NZ$) for deaths at different ages (see text for simplifying assumptions). Labels above each bar show the percent in last year of life.*

* But due to current data limitations this is still for only 52% of total health system costs ($6.7/$12.98 billion), see Discussion. observed at one point in time are assumed Regardless of the scenario, not to apply to a synthetic population over accounting for proximity to death overes- their lifetime.) Even so, these results give timates these future health system costs. some indication of how the proportion of That is, not allowing for deaths in the assigned health expenditure in the last year future being ‘pushed out more’ to older of life declines as the age at death increases, ages resulted in overestimated costs. This eg, 76% for death at age 10 years, 42% at age overestimate was by 1.3% to 4.5% by 2041, 50 years, 25% at age 70 years, 13% at age 80 depending on scenario. years, and 6% at age 90 years. Table 2 shows the estimated future health Discussion system costs with and without accounting for proximity to death (both 6- and What is new about these updated 12-month scenarios), and for the ‘simplistic’ results? scenario of ongoing 2% per annum This current work has produced updated reduction in mortality rates into the future values for all of the analyses in our 1 for all sex by age groups and the more previous study published in this journal. sophisticated SNZ estimates of mortality The general patterns are similar to that counts and rates. previously published, with two exceptions. Firstly, the costs being studied within the Figure 3 shows these results for last six months of life now decline from the period of 12 months proximal to around age 60 years for both sexes (rather death and the 2% per annum reduction than continuing to increase with age as scenario. It is important to note that these suggested previously; Figure 1). Secondly, estimated future health system costs are only based on: expenditure as captured the costs regardless of impending death by assigned health events (the total health now plateau in the 80+ age-group (rather spending is more than this individu- than continuing to increase; Figure 1). ally-linked data—see the Methods and These new results arise from the methods Discussion sections); demographic projec- improvements around attributing costs tions (eg, productivity and expectation proximal to the time of death (see Methods). trends are not included). Thus, interpre- These results are also more consistent tation should be more on the relative with our knowledge of how the health patterns, not the absolute dollar amounts. system in New Zealand typically operates

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 17 www.nzma.org.nz/journal ARTICLE

Table 2: Estimated future health system costs* (in 2011 NZ$ million) for the total New Zealand population with and without accounting for proximity to death, for varying scenarios of future mortality rates and six versus 12-month proximity to death.

2011 2021 2031 2041 a. Assuming 2% per annum reduction in mortality rates uniformly for all sex by age groups i. By 6-month proximity to death Accounting for proximity to death 6,569 7,690 8,907 9,878 Not accounting for proximity to death 6,569 7,725 8,988 10,009 % overestimate due to not accounting 0.0% 0.5% 0.9% 1.3% ii. By 12-month proximity to death Accounting for proximity to death 6,767 7,866 9,056 9,986 Not accounting for proximity to death 6,767 7,982 9,327 10,432 % overestimate due to not accounting 0.0% 1.5% 3.0% 4.5% b. Assuming SNZ projected future mortality rates uniformly by sex by age groups i. By 6-month proximity to death Accounting for proximity to death 6,568 7,683 8,896 9,862 Not accounting for proximity to death 6,568 7,721 8,981 9,992 % overestimate due to not accounting 0.0% 0.5% 1.0% 1.3% ii. By 12-month proximity to death Accounting for proximity to death 6,764 7,854 9,043 9,987 Not accounting for proximity to death 6,764 7,978 9,321 10,418 % overestimate due to not accounting 0.0% 1.6% 3.1% 4.3% * But due to current data limitations this is still for only 52% of total health system costs ($6.7/$12.98 billion) in 2011, see Discussion.

Figure 3: Estimated future health system costs (in 2011 NZ$) for the total New Zealand population. Labels are percentage overestimates when not accounting for proximity to death.

Note: Using 2007/08 to 2009/10 data for health system costs for health events from Health Tracker (Table 1), inflation- adjusted to 2011 NZ$. Mortality rates in 2011 sourced from Statistics New Zealand (SNZ) 2010–12 official life-tables, with age-specific mortality rates reduced by 2% per annum into the future. As noted elsewhere, this analysis is still for only 52% of total health system costs.

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 18 www.nzma.org.nz/journal ARTICLE

ie, less intensive provision of health That said, this Health Tracker services for the very old for whom such dataset is an extremely rich dataset interventions as major surgery might be for analyses, and will continue to less appropriate and in some cases for improve in the future. As examples, whom palliative care might be the service it is already contributing data for being provided. other work by the National Health Updated consideration of study Committee on a high-level scan of health spending in order to select limitations domains of health service use for Researchers and policymakers can have further work on prioritisation,8 and confidence about the general patterns for Treasury projections of future suggested by these results, but they should health expenditure.2 remain particularly cautious about the accuracy of all the specific values reported 2. Primary care costs are very simply in this study. Furthermore, some of our assigned on a per capita basis (consid- analyses (as shown in Figure 2), involve ering age, sex and ethnicity) to the assumptions eg, a steady state in costs New Zealand population using the throughout the life course (which is country’s health system’s capitation formula. Therefore, our analyses in contrast to the historical pattern of will tend to underestimate costs near increases in health costs in high-income death if primary care utilisation countries). Below we provide an updated increases near death. But, given list of the other major limitations with these that primary care expenditure is cost estimates identified to date: not a large component of end-of-life 1. Using the linked administrative care in New Zealand,4 this probably datasets in Health Tracker, $6.7 would not cause much of an under- billion (2011 inflation adjusted NZ$) estimation in the costs in the last six of almost exclusively Government months of life. health expenditure was attributed 3. Current Health Tracker data also to individual patient events in 2009. include very little privately funded But this is still only just over half health expenditure. But, given that (52%) of the combined Vote:Health 83% of all health system expenditure ($12.98 billion, nominal) appropri- is estimated to be publically funded ation in 2009–10. One reason is that in New Zealand,2 this limitation is not we have restricted hospitalisation too severe. costs to only those that are casemix- funded (as detailed earlier). Also, 4. This study did not estimate costs important components of Vote:Health by ethnicity, since it seems likely expenditure are not yet available to that any such cost differences will us, including data on maternity care, be due to conflated differences immunisation, cervical screening, in need, access and utilisation of specific programmes (eg, diabetes health services, and as such requires care improvement package, perfor- separate and careful analysis and mance-related payments), and more interpretation. But, given the impor- importantly for this study, Disability tance of health inequalities in New Support Services and other funding Zealand society, this should be a covering rest home and palliative priority area for future work. care. The latter will have resulted in 5. Many of the datasets used in this us underestimating some of the near- analysis use ‘prices’ that are poten- end-of-life costs. Nevertheless, due tially charged by agencies to funding to capital and ‘back office’ expen- bodies and which do not necessarily diture on administration, not all of represent the actual cost of the health Government funding can readily be event. For example, many community attributed to individual patient events lab contracts are bulk-funded and (eg, over 10% of public funding goes so the prices are only indicative and to prevention and public health potentially have not been updated for services, and health administration2). a number of years. This pricing issue

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 19 www.nzma.org.nz/journal ARTICLE

may mean that true costs are actually Successful international examples of this higher where costs go up, but where move from ‘bespoke’ to ‘fully integrated prices charged stay the same and and wider access’, include the “ScottisH become out-of-date. But if lower costs Informatics Programme (SHIP)” (http:// are achieved (eg, from operational www.scot-ship.ac.uk/) and the Ontario efficiencies obtained) then prices Institute for Clinical Evaluation Sciences charged might be sometimes higher (http://www.ices.on.ca/). The New Zealand than the true costs. Ministry of Health has already migrated 6. Our modelling around future health health data to the SNZ IDI, and is working costs does not fully address any through the feasibility of placing most future compression or expansion national health data collections in the IDI of morbidity that is not captured to facilitate research (personal communi- directly by proximity to death. cation, Jackie Fawcett, Ministry of Health, Further details of this are in the December, 2014). To support this initiative, Appendix. a Virtual Health Information Network (VHIN) has been established with joint What is the context for further membership of university academics and developments? public sector researchers. One of the goals While we have made use of Health of the VHIN is to encourage sharing of Tracker in this study, we have also been knowledge about health data (eg, meta- considering big picture issues for the New data) and analytical approaches (eg, Zealand health system.6 One of us (TB) has rationale of the methods, definitions, cross- been engaged in national- level ‘big data’ checking with clinicians, data management and health systems discussions. We are and analysis code) between researchers in also aware that DHB level interest in ‘big a collaborative model that enhances the data’ is growing (eg, Counties Manukau productivity and accuracy of New Zealand DHB has been using Ministry of Health health research using routine data. Formal data to inform health service and policy structures for such a VHIN are evolving (eg, development for a number of years, see: health researchers using the IDI may be http://www.countiesmanukau.health. strongly encouraged, or even required, to nz/about-us/performance-and-planning/ make available to other researchers their health-status-documents/). programing code). But now there is a likely imminent What all these developments mean step-change in data access occurring in is improvements in the type of analysis New Zealand, with moves to place routine presented in this article are imminent— health data (eg, mortality, hospitalisation, and will allow for both improved research laboratories, etc) linkable via the NHI into by New Zealand researchers and more the SNZ Integrated Data Infrastructure nuanced decision-making by policy makers (IDI).9 In the last decade, research groups and planners. such as PREDICT/VIEW at the University Regarding cost data specifically, we of Auckland (https://www.fmhs.auckland. foresee two parallel streams to improve the ac.nz/en/soph/about/our-departments/ quality of data for modelling in the next five epidemiology-and-biostatistics/research/ years. First, the ‘bottom-up’ costing in this view-study/research.html; accessed paper can be complemented by allocating 11 February 2015) and BODE3 at the remaining Vote:Health (and Vote:ACC) University of Otago (www.otago.ac.nz/ across individuals under plausible bode3; accessed 11 February 2015) have scenarios. For example, taking the total received copies of multiple routine admin- maternity care budget, and allocating it istrate health datasets for their dedicated across woman pro-rata to age-specific birth research purposes—what is sometimes rates. Second, and a ‘longer-term’ option called ‘bespoke’ linkage. But in the age which should replace the previous blended of ‘big data’,10 research productivity can ‘bottom-up’ and ‘top-down’ approach, is to be enhanced (and duplication and errors continue to work on the individual-level minimised) by allowing researchers easier data and costing rules, and ‘liberate’ them and fuller access to fully integrated data. for researcher use (cost models are already

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 20 www.nzma.org.nz/journal ARTICLE

in use within the Ministry, for example)— Furthermore, we show (as before) that perhaps through the SNZ IDI and VHIN. projections of future health system expenditure are slightly overestimated when not Conclusions accounting for proximity to death in costs. Further health data improvements in Health system costs are large in New coming years are likely with access to Zealand, vary across the life-course, and are additional data sources (eg, from Disability skewed to the last year of life (eg, around Support Services) and as New Zealand 25% of costs being in the last year of life of continues to move towards better inte- a 70-year-old). This analysis has benefited gration of ‘big data’ in the health sector. It from quality improvements in cost data and is often said that “New Zealand has some methods refinements relative to the previ- of the best health data in the world”. The ously published work on health costs in goal now should be to better harness these New Zealand. Nevertheless, the patterns in data for informing policy and undertaking the costs are largely unchanged from previ- world-class research, matching and even ously—except the decline in some costs exceeding the potential realised in other with age among older New Zealanders. jurisdictions such as Scotland and Ontario.

Appendix: additional methods details

Updated data and methods since the previous analyses Since the previous analyses,1 the following data and methods enhancements were performed: 1. Years covered: The financial years 2007/08, 2008/09 and 2009/10 are now used instead of calendar years 2007, 2008, 2009. 2. Core population: A revised core population of New Zealand residents was used. Previously the population inclusion criteria was broader and included those not listed as New Zealand resident if they had health system records with points of contact three or more months apart in any one year (including enrolment with a primary health organisation). 3. Cost allocation: More accurate allocation of costs by timing was performed. That is according to usual ‘administrative’ practice, costs were previously assigned to the end date of an event regardless of duration of event. That led to skewed costs in the last six months of life. In the current analyses, we re-allocated costs evenly over the duration of each event in time. 4. Use of casemix-funding: Another modification was made after more exhaustive examination of the input datasets. Hospitalisation costs were restricted to casemix- funding only as cost weights applied to non-casemix-funded events are unlikely to accurately reflect the true (opportunity) cost to the New Zealand health sector. Furthermore, without this restriction there was a risk of double counting the costs of some events where they appear in two datasets (eg, emergency department events in the National Non Admitted Patients Collection and those in the NMDS). We therefore decided to only allocate costs to those events that we have high confidence in: casemix-funded hospitalisations; community laboratory tests; non-admitted patient events (eg, outpatient and emergency department events); community pharmaceuticals dispensed (including patient contribution); general practice consultations (both that calculated based on the capitation funding formula routinely used in New Zealand, and fee-for-service when not an enrolled patient in a capitated practice). Restricting to these files resulted in around 50% of all Vote:Health funding being allocated to event- based expenditure. The ‘missing’ 50% includes the files not yet linked (maternity, rest-home, community mental, dental and physiotherapy care; see above) and inpatient events excluded from casemix-funding (which include but are not limited to

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 21 www.nzma.org.nz/journal ARTICLE

inpatient mental health events; events directly funded by ACC; events where the admitted person was a boarder; cancelled treatments; some transplant events; some spinal injuries; some same day chemotherapy for cancer events; some same day litho- tripsy, colposcopies, cystoscopies, colonoscopies, gastroscopies, and bronchoscopies; and some same day blood transfusions). 5. Calculation of hospitalisation costs: In addition to the exclusion of non-casemix- funded hospitalisations, the cost weight used for the calculation of hospitalisation costs changed from using one cost weight for all years, to using the cost weight used for funding in the specific financial year (ie, a different cost weight for each year, along with the appropriate year’s unit price). The reasons for this were pragmatic, but also because it reflects how events were actually costed that year. 6. Other improvements: A number of other fairly minor improvements were made (eg, with the calculations of PHO data), the details of which are available from the authors on request. The SAS code used in our analyses is available by contacting the authors (if not available on our website: http://www.otago.ac.nz/bode3). Compression or expansion of morbidity: implications for modelling of future costs Our modelling around future health costs does not fully address any future compression or expansion of morbidity that is not captured directly by proximity to death. For example, the diabetes epidemic may increase morbidity (and demand for health services) if our society is less successful at reducing incidence than we are at keeping people alive with diabetes, thereby seeing an expansion in morbidity (diabetes disease severity held constant, and likewise other causes of morbidity held constant). Conversely, if New Zealand society successfully controls obesity trends, this may reduce morbidity prevalence (through diabetes, but also cardiovascular disease, musculoskeletal and other impacts). Deter- mining past trends in compression or expansion in morbidity is challenging,11,12,13 let alone estimating future trends. That said, we suggest that one method to include in future expenditure projections is to use disability-adjusted life expectancy (DALE; as estimated in the recent New Zealand Burden of Disease study14), assume the same ratio of DALE to life expectancy (DALE:LE) in the future, and then back estimate by what percentage the prevalent years of life lived with disability (pYLDs) would need to change in the future to keep the DALE:LE ratio constant (or whatever other ratio is considered plausible). The percentage change in pYLDs across all sex by age groups necessary to generate the desired DALE:LE ratio in the future can then be used as a proxy for morbidity change, and therefore rescaling of the costs not within the last six or 12 months of life shown in Table 1.

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 22 www.nzma.org.nz/journal ARTICLE

Competing interests: Nil Funding: The Burden of Disease Epidemiology, Equity and Cost-Effectiveness Programme is funded by the Health Research Council of New Zealand (10/248). Acknowledgment: Access to Health Tracker data was provided by the Ministry of Health. In particular, we acknowledge the efforts and foresight of Craig Wright in initially assembling the data and managers within the Ministry who had the foresight to ensure this data source was created and made available. We thank Bronwyn Croxson and Dr Wing Cheuk Chan for helpful comments on early versions of this work and the two anonymous journal reviewers for very helpful comments. Author information: Tony Blakely, Research Professor, Department of Public Health, University of Otago, Welling- ton; June Atkinson, Data Analyst, Department of Public Health, University of Otago, Well- ington; Giorgi Kvizhinadze, Postdoctoral Fellow, Department of Public Health, University of Otago, Wellington; Nhung Nghiem, Research Fellow, Department of Public Health, University of Otago, Wellington; Heather McLeod, Professor, School of Management Studies, University of Cape Town, South Africa; Anna Davies, Research Fellow, Department of Public Health, University of Otago, Wellington; Nick Wilson, Associate Professor, Department of Public Health, University of Otago, Wellington. Corresponding author: Professor Tony Blakely, Department of Public Health, University of Otago, Wellington, PO Box 7343, Wellington, New Zealand. [email protected] URL: www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2015/vol-128-no-1422-25- september-2015/6662

REFERENCES: 1. Blakely T, Atkinson J, hospitals, including 28 August, 2014). Kvizhinadze G, Nghiem wiesnz11 methodology 10. Fung IC, Tse ZT, Fu KW. N, McLeod H, Wilson N. and casemix purchase unit Converting Big Data Health system costs by allocation. Wellington: into public health. sex, age and proximity to Ministry of Health, 2013. Science (New York, death, and implications 6. Woodward A, Blakely T. NY 2015;347:620. for estimation of future The Healthy Country? 11. Graham P, Blakely T, Davis expenditure. N Z Med A History of Life and P, Sporle A, Pearce N. J 2014;127:12-25. Death in New Zealand. Compression, expansion, 2. New Zealand Treasury. Auckland: Auckland or dynamic equilibrium? Health projections and University Press, 2014. The evolution of health policy options for the 2013 7. Statistics New Zealand. expectancy in New long-term fiscal state- National Population Zealand. J Epidemiol ment. Draft paper for the Projections: 2011(base)– Community Health long-term fiscal external 2061. 19 July 2012. URL: 2004;58:659-66. panel. Wellington: New http://www.stats.govt. 12. Christensen K, Doblham- Zealand Treasury, 2012. nz/browse_for_stats/ mer G, Rau R, Vaupel JW. 3. Blakely T, Atkinson J, population/esti- Ageing populations: the Kvizhinadze G, Wilson mates_and_projections/ challenges ahead. Lancet N, Davies A, Clarke P. NationalPopulationPro- 2009;374:1196-208. Patterns of cancer care jections_HOTP2011.aspx. 13. Statistics New Zealand & costs in a country with 8. National Health Ministry of Health. Longer detailed individual data. Committee. Strategic Life, Better Health? Trends Med Care 2015;53:302-9. Overview: Cardiovas- in health expectancy in 4. Chan WC, Jackson G, cular Disease in New New Zealand, 1996–2006. Winnard D, Anderson Zealand (Working Draft). Wellington: Statistics P. Healthcare services Wellington: National New Zealand, 2009. funded by Counties Health Committee, 2013. 14. Ministry of Health. Manukau District Health 9. Statistics New Zealand. Health Loss in New Board for people in the Integrated Data Infrastruc- Zealand: A report from last year of life. N Z ture. http://www.stats. the New Zealand Burden Med J 2011;124:40-51. govt.nz/browse_for_stats/ of Diseases, Injuries 5. Ministry of Health. New snapshots-of-nz/ and Risk Factors Study, Zealand casemix frame- integrated-data-infra- 2006–2016. Wellington: work for publicly funded structure.aspx (Updated Ministry of Health, 2013.

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 23 www.nzma.org.nz/journal ARTICLE

Predictors of vitamin D status in pregnant women in New Zealand Alec J Ekeroma, Carlos A Camargo Jr, Robert Scragg, Clare Wall, Alistair Stewart, Ed Mitchell, Julian Crane, Cameron C Grant

ABSTRACT INTRODUCTION: Newborn vitamin D status is largely determined by maternal vitamin D status during pregnancy. New Zealand has a sun avoidance health policy and minimal dietary vitamin D fortification. Vitamin D deficiency (serum 25-hydroxyvitamin D (25(OH)D) concentration <50nmol/L) is present in 57% of a sample of newborns from Christchurch and Wellington. To inform vitamin D supplementation policy, our aim was to describe the frequency of, and factors associated with, vitamin D deficiency during pregnancy. METHODS: We enrolled an ethnically diverse sample of pregnant women from a community maternity clinic in South Auckland, New Zealand, with serum 25(OH)D concentration measured at 27 weeks gestation. We examined the associations of enrolment season, maternal demographics, health, sunlight exposure and vitamin D intake with vitamin D deficiency. RESULTS: Vitamin D deficiency was present in 109/259 (42%). Enrolment season (P<0.001) and ethnicity (P=0.003) were independently associated with the odds of vitamin D deficiency, but not sunlight exposure or dietary vitamin D intake. Of those enrolled in winter (June–August)/spring (September–November), vitamin D deficiency was present in 43% of European, 67% of Māori, 80% of Pacific and 59% of women of other ethnic groups. CONCLUSIONS: These findings suggest that New Zealand’s targeted strategy for vitamin D supplementation may miss up to 42% of women with vitamin D deficiency in our population. Supplementation for all pregnant women during winter/spring could be an appropriate intervention for prevention of vitamin D deficiency during pregnancy in New Zealand.

he importance of vitamin D in pre-eclampsia (OR 1.79, 1.25-2.58), and maintaining calcium homeostasis is small for gestational age (OR 1.85, 1.52- Twell defined, and those with severe 2.26).3 A meta-analysis of 24 studies showed deficiency suffer bone disease, such as a serum 25(OH)D concentration <50 nmol/L rickets. The role of vitamin D in pregnancy was associated with an increased risk of and its effects on maternal, foetal and gestational diabetes (OR 1.38, 1.12–1.70), neonatal health is still unclear, although pre-eclampsia (OR 2.09, 1.50–2.90), preterm earlier systematic reviews concluded there birth (OR 1.58, 1.08–2.31) and small for ges- was insufficient evidence to suggest lower tational age (OR 1.52, 1.08–2.15).4 Although vitamin D status during pregnancy was not establishing causation, the consistency associated with adverse pregnancy out- of the associations of low vitamin D status comes.1,2 However, two recent meta-anal- with poorer pregnancy outcomes that these yses showed low vitamin D status is recent meta-analyses showed, are of po- associated with an increased risk of poor tential policy relevance in countries where pregnancy outcomes.3,4 A meta-analysis of vitamin D deficiency is prevalent. 31 observational studies that used a range Vitamin D deficiency is common in of serum 25-hydroxyvitamin D (25(OH)D) New Zealand. Two New Zealand studies cut-off values from 20 to 80 nmol/L showed of pregnant women have shown a high a lower 25(OH)D concentration during preg- prevalence of vitamin D deficiency.5,6 The nancy was associated with an increased risk 2008/2009 New Zealand Nutrition Survey of gestational diabetes (OR 1.49, 1.18-1.89), reported that one-third of 2,204 women of

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 24 www.nzma.org.nz/journal ARTICLE

child-bearing age had vitamin D deficiency,7 Ethical approval was obtained from the as defined by the New Zealand Ministry of regional MoH ethics committee (Northern Health (MoH) as a serum 25(OH)D concen- X committee, Application number tration <50 nmol/L.8 In a sample of 929 NTX09/11/101) and written informed healthy newborns born in Christchurch consent obtained from all participating and Wellington between 1997 and 2001, 350 women. Women were recruited from a (57%) had serum 25(OH)D < 50nmo/l.9 community-based primary care maternity o Studies performed in New Zealand clinic in Auckland (latitude 36 S) New have shown associations of low sunlight Zealand, from April 2010 to July 2011. exposure, higher latitude, winter and spring Study participants seasons and non-European ethnicity to be Women were eligible for enrolment associated with an increased risk of vitamin if their estimated gestation was 26 to 30 D deficiency.6,10,11 These associations are weeks and they had a singleton pregnancy. consistent with those identified in other We excluded pregnant women taking studies, which also showed, in addition, a vitamin D supplementation that exceeded lack of vitamin D supplementation, greater 200 IU/day, those with a history of renal body mass index (BMI),12 smoking, multi- stones or hypercalcaemia, or with any parity, skin colour13 and a range of variables serious complication of pregnancy at the describing socioeconomic disadvantage.14 time of enrolment. The MoH guidelines recommend vitamin Data collection D supplementation be considered for Face-to-face interviews were completed high-risk women, whom they defined as with women at enrolment with the data those that are dark skinned, avoid sunlight, collected allowing description of maternal or are resident at more southern latitudes demographics and health, pregnancy 8 during the winter months. Specifically, health, dietary sources of vitamin D and routine measurement of serum 25(OH) sunlight exposure. We defined enrolment D concentration during pregnancy is seasons as summer (December–February), discouraged in current New Zealand autumn (March–May), winter (June– clinical practice, therefore the decision August) and spring (September–November). making regarding vitamin D supplemen- Ethnicity was defined as the mother’s tation is based on the presence or absence self-prioritised ethnicity. of risk factors. A screening tool, based on Dietary vitamin D intake was esti- predictors of vitamin D deficiency, would mated using an interviewer-administered potentially assist clinicians in deciding who semi-quantitative food frequency question- to supplement. Ideally, a clinical screening naire (FFQ). The FFQ captured the major test should: address a significant public sources of vitamin D in the New Zealand health condition for which treatment is food supply derived from the New Zealand available; have a high sensitivity and spec- Food Composition Database.17 The FFQ ificity; be safe, inexpensive and widely contained 12 food items using standard available; and lead to an improvement in food serving sizes. Frequency of recent health outcomes.15 consumption of each item was described Our aim is to identify and quantify the across an 8-point frequency range (never, risk factors of vitamin D deficiency in an < monthly, 1–3 times per month, weekly, ethnically diverse sample of pregnant 2–4 times/week, 5–6 times/week, daily, ≥ 2 women in New Zealand. We sought to times per day). Frequencies were converted develop a simple predictive tool for vitamin into daily intake, for example a response D status based on the identified risk factors. of 2–4 times per week was converted to 0.4 servings/day (ie, 3 times per week). Methods The vitamin D content of each food Study design and setting was obtained from the New Zealand This study was completed within a Food Composition Database.17 Any foods randomised trial of vitamin D supplemen- not in this database were obtained from tation during pregnancy and infancy, the the AUSNUT2007 food database.18 When findings of which have been reported.16 estimating participant vitamin D intake

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 25 www.nzma.org.nz/journal ARTICLE

standard serving sizes were used, for variables significantly associated with example each serving of yoghurt was 150g vitamin deficiency. A simple predictive tool and each serving of milk 250ml. These are for vitamin D deficiency was developed recommended standard serving sizes for which included the variables identified with the New Zealand population. Brands of this multivariate model. specific foods known to be fortified with Associations were described using odds vitamin D were also captured (margarine, ratios (OR) and 95% confidence intervals yoghurt, milk). Frequency of consumption (CI). A two-tailed P value of <0.05 was of each food item in the FFQ was converted considered statistically significant. to intake of vitamin D in µg/day. Exposure to ultraviolet B radiation Results was estimated indirectly by determining Enrolment of the 259 pregnant women the number of self-reported hours spent was evenly distributed by season (summer outdoors in the sun during the preceding 59 (23%), autumn 71 (26%), winter 67 (27%), four weeks. In New Zealand adults, spring 62 (24%)). The sample was ethni- reported sunlight exposure correlates with cally diverse (European 40 (15%), Māori 62 serum 25(OH)D concentration.19 (24%), Pacific 113 (44%), and all other ethnic Venous blood samples were obtained groups 44 (17%)). Median (interquartile from the pregnant women at enrolment range) age was 28 (22–31) years and median and serum 25(OH)D concentration was body mass index (BMI) 32 (27–38) kg/ measured using isotope-dilution liquid m2. Fifty-six percent of the women had a chromatography-tandem mass spec- tertiary education. The mean (SD) serum trometry in a Vitamin D External Quality 25(OH)D concentration was 63 (35) nmol/L. Assurance Scheme-certified laboratory.20,21 Vitamin D deficiency was present in 109 Statistical analysis (42%) women, while severe vitamin D Statistical analysis was performed using deficiency (serum 25(OH)D<25 nmol/L) was SAS version 9.3 (SAS Institute Inc, Cary, present in 28 (11%). NC, USA). The sample was described using In unadjusted analyses, vitamin D status proportions and means with standard devi- varied with season, ethnicity, BMI, and time ations or, for non-normally distributed data, spent outdoors (Table 1). More specifically, medians with interquartile ranges. External there was an increased odds of vitamin D prioritization was used for respondents deficiency for women enrolled in winter with more than one prioritized ethnicity, (OR=16.7) or spring (OR=9.9); for women of with the priority order being: Māori, Pacific, Pacific (OR=5.5) or of other (OR=10.0) ethnic Asian and European/Other.22 groups; and a decreased odds for women For the purpose of this study, vitamin D who used sunscreen (OR= 0.5) (Table deficiency was defined as serum 25(OH) 2). Sunscreen used varied by ethnicity D concentration <50 nmol/L and severe (European 65%, Māori 41%, Pacific 32%, vitamin D deficiency as 25(OH)D <25 Other 47%, P=0.003). nmol/L. For outcome analyses we cate- When these unadjusted analyses were gorised serum 25(OH)D concentrations restricted to the non-European women into four groups: <25; 25 to <50; 50 to<75; (n=219) the increased risk associated and ≥75 nmol/L. Using the chi-square and with enrolment during winter (OR=18.6, Wilcoxon rank-sum test, we determined 95%CI 3.5–342.5) or spring (OR=9.7, unadjusted associations of variables with 95%CI 1.7–183.1) compared with summer vitamin D status as defined by these four persisted. While sunscreen use was no groups and with vitamin D deficiency. longer associated with a reduced odds Multivariable logistic regression was used of vitamin D deficiency (OR=0.7, 95%CI to determine the independence of associa- 0.3–1.7) avoidance of sun exposure during tions of variables with vitamin D deficiency. summer was (OR=0.37, 95%CI 0.1–0.9). Variables were entered into the model, In the multivariable analysis, the vari- if the p-value in univariate analyses was ables independently associated with the <0.15 and were removed from the model odds of vitamin D deficiency were season of sequentially until the model only contained enrolment and maternal ethnicity (Table 3).

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 26 www.nzma.org.nz/journal ARTICLE

Table 1: Participant characteristics, by maternal serum 25-hydroxyvitamin D concentration during pregnancy.

Serum 25-hydroxyvitamin D concentration in nmol/L, Variable n (column %) or median (IQR*) (n=259 unless otherwise stated) P value† <25 25 to <50 50 to <75 ≥75

n=28 n=81 n=61 n=89

Enrolment season <0.001 Summer 1 (4) 6 (7) 15 (24) 37 (42) Autumn 3 (11) 11 (14) 20 (33) 37 (42) Winter 15 (53) 36 (44) 9 (14) 7 (7) Spring 9 (32) 28 (35) 17 (29) 8 (9)

Maternal demographics and health and household demographics

Age in years 28 (24–30) 28 (23–32) 28 (23–32) 26 (21–31) 0.50

Ethnicity 0.008 European (E) 1 (4) 8 (10) 10 (16) 21 (24) Māori (M) 4 (14) 18 (22) 15 (25) 25 (28) Pacific (P) 14 (50) 46 (57) 26 (43) 27 (30) Other (all others not E, M, or P) 9 (32) 9 (11) 10 (16) 16 (18)

Education 0.96 Primary 5 (18) 11 (13) 7 (11) 16 (18) Secondary 8 (29) 24 (30) 18 (30) 24 (27) Tertiary 15 (53) 46 (57) 36 (59) 49 (55)

Body mass index in kg/m2 (n=257) 31 (26–38) 34 (28–40) 32 (27–38) 31 (25–37) 0.04

Doctor-diagnosed diabetes (n=256) 0.72 Yes 0 (0) 1 (1) 2 (3) 2 (2) No 28 (100) 77 (99) 59 97) 87 (98)

Doctor-diagnosed hypertension 0.94 Yes 1 (4) 5 (6) 3 (5) 4 (4) No 27 (96) 76 (94) 58 (95) 85 (96)

Cigarette smoker pre-pregnancy (n=258) 0.70 Yes 10 (36) 37 (46) 27 (44) 34 (39) No 18 (64) 44 (54) 34 (56) 54 (61)

Number of people in house 4 (3–7) 5 (4–7) 4 (3–6) 4 (3–6) 0.27

Pregnancy health

First pregnancy 0.44 Yes 18 (64) 59 (73) 43 (70) 70 (79) No 10 (36) 22 (27) 18 (30) 19 (21)

Gestation in weeks (n=256) 28 (27–29) 27 (27–29) 27 (26–28) 27 (26–28) 0.68

Current cigarette smoker 0.13 Yes 2 (7) 19 (23) 8 (13) 20 (22) No 26 (93) 62 (77) 53 (87) 69 (78)

Dietary sources of vitamin D

Dietary vitamin D in ug/day 2.5 2.6 3.0 2.9 0.80 (1.7–3.5) (1.7–4.2) (1.9–4.0) (1.6–3.8)

Vitamin D supplement (n=232) 0.70 Yes 1 (4) 3 (4) 4 (8) 3 (4) No 24 (96) 71 (96) 47 (92) 79 (96)

Sunlight exposure

Minutes/day outdoors past 4 weeks 31 (20–60) 49 (21–94) 45 (20–94) 77 (31–120) 0.004

Sunscreen if outside in summer 0.10 Yes 9 (33) 26 (32) 29 (48) 43 (48) No 18 (67) 55 (68) 32 (52) 46 (52)

Avoids sun exposure 10.00–16.00 hours in summer 0.09 Yes 12 (43) 50 (63) 42 (69) 60 (67) No 16 (57) 30 (37) 19 (31) 29 (33)

Head covering worn (n=195) 0.38 No head covering or hat 12 (86) 43 (86) 41 (85) 63 (76) Head covering or hat‡ 2 (14) 7 (14) 7 (15) 20 (24)

* Interquatile range † From chi-Square, Fisher’s Exact or Wilcoxon rank-sum test ‡ Includes scarf/head covering, burka, cap, winter pull-over beanie or cowboy hat/hat with rim

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 27 www.nzma.org.nz/journal ARTICLE

Table 2: Unadjusted associations between participant characteristics and maternal serum 25-hydroxyvitamin D concentration during pregnancy.

Variable Associations with serum 25(OH)D <50 nmol/L

Odds ratio (95%CI) P value*

Enrolment season 0.004 Summer 1.0 Autumn 2.6 (0.3, 52.5) Winter 16.7 (3.2, 307.6) Spring 9.9 (1.8, 184.7)

Maternal demographics and health and household demographics

Age in years 1.0 (0.9, 1.1) 0.88

Ethnicity 0.06 European 1.0 Māori 2.7 (0.4, 53.6) Pacific 5.5 (1.1, 101.6) Other 10.0 (1.8,189.7)

Education 0.90 Primary 1.3 (0.4, 3.6) Secondary 1.1 (0.4, 2.6) Tertiary 1.0

Body mass index in kg/m2 1.0 (1.0, 1.1) 0.06

Doctor diagnosed diabetes Yes 0.4 (0.0, 2.4) 0.35 No 1.0

Doctor diagnosed hypertension Yes 1.2 (0.4, 3.7) 0.76 No 1.0

Cigarette smoker pre-pregnancy Yes 1.1 (0.7, 1.8) 0.73 No 1.0

Number of people in house 0.13 7 or more 1.9 (1.0, 3.8) 4 to 6 1.1 (0.6, 2.1) Less than 3 1.0

Pregnancy health

First pregnancy Yes 0.8 (0.5, 1.4) 0.40 No 1.0

Gestation 1.2 (1.0, 1.4) 0.09

Current cigarette smoker Yes 1.0 (0.6, 1.9) 0.90 No 1.0

Dietary sources of vitamin D

Dietary vitamin D ug/day 1.0 (0.9, 1.1) 0.94

Sunlight exposure

Average time per day outdoors past 4 weeks Less than 60 minutes 1.6 (1.0, 2.6) 0.07 60 Minutes or more 1.0

Sunscreen if outside in summer Yes 0.5 (0.3, 0.9) 0.01 No 1.0

Avoids sun exposure 10.00–16.00 hours in summer Yes 0.6 (0.4, 1.1) 0.08 No 1.0

Head covering worn No head covering or hat 1.4 (0.4, 9.2) 0.68 Head covering or hat† 1.0

* From univariable logistic regression analysis † Includes Scarf/head covering, burka, cap, winter pull-over beanie or cowboy hat/hat with rim NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 28 www.nzma.org.nz/journal ARTICLE

Table 3: Multivariable associations between participant characteristics and maternal serum 25-hydroxyvitamin D concentration during pregnancy.

Associations with vitamin D deficiency as defined by Variable a serum 25(OH)D <50 nmol/L

Odds ratio (95%CI) P value*

Enrolment season Summer 1.0 Autumn 2.2 (0.8, 6.3) 0.13 Winter 31.1 (11.9, 92.4) <0.001 Spring 13.5 (5.4, 38.2) <0.001

Ethnicity European 1.0 Māori 2.8 (1.0, 8.3) 0.051 Pacific 5.8 (2.3, 15.8) <0.001 Other 5.3 (1.8, 17.2) 0.004

* From multivariable logistic regression analysis

Table 4: Diagnostic performance of screening tests for vitamin D deficiency in pregnant women based upon season or upon season and ethnicity.

Screening test based upon season Season when pregnant Vitamin D deficiency* Present Absent

Winter/spring 88 41

Summer/autumn 21 109

Screening test performance All ethnic groups Sensitivity = 81%, specificity = 73%, PPV = 68%, NPV = 84%

European Sensitivity = 66%, specificity = 61%, PPV = 43%, NPV =100%

Māori Sensitivity = 82%, specificity = 76%, PPV = 67%, NPV = 89%

Pacific Sensitivity = 85%, specificity = 76%, PPV = 80%, NPV = 82%

Other Sensitivity = 56%, specificity = 73%, PPV = 59%, NPV = 70%

Screening test based upon season and ethnicity Season when pregnant Vitamin D deficiency* and maternal ethnicity Present Absent

Winter/spring and non- 79 29 -European ethnicity

Summer/autumn or 30 121 European ethnicity

Screening test Sensitivity = 73%, specificity = 81%, PPV = 73%, NPV = 80% performance

* Serum 25-hydroxyvitamin D <50 nmol/L PPV = positive predictive value NPV = negative predictive value

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 29 www.nzma.org.nz/journal ARTICLE

There was increased odds of vitamin D defi- have found associations between some of ciency associated with enrolment in women these factors and vitamin D deficiency in of Māori, Pacific or other ethnic groups (in pregnancy,27, 28 these associations were not comparison with European women); and consistently present in studies performed in winter or spring (compared with summer). Australia.23, 29 For example, a study done in A prediction model based on winter/spring Canberra and Campbelltown identified four seasons alone (Table 4) had comparable factors predictive of vitamin D deficiency: performance as a screening test for vitamin season: ethnicity: BMI; and vitamin D supple- D deficiency as one that used both season mentation.23 Our study design prevented us and maternal ethnicity. In comparison with from assessing the contribution of vitamin a model that included both season and D supplement use to vitamin D status during ethnicity, the sensitivity for a model based pregnancy. In New Zealand, only 36% of solely upon season was 81% (95% CI 72–88) pregnant women take any vitamin or mineral vs. 73% (63–81) and specificity 73% (65–80) supplements, many of which are inadequate vs. 81% (73–87). sources of vitamin D.30 Avoiding sunlight exposure in summer was not significant a predictor for the Discussion whole sample but it remained significant In this multi-ethnic, New Zealand-based for the non-European participants (Table pregnant cohort, the prevalence of vitamin 2a and 2b), although this difference was no D deficiency (defined as 25(OH)D concen- longer significant after adjusting for season tration <50 nmol/L) was 42%. The only and ethnicity. Although there is a general factors independently associated with consensus that vitamin D levels increase vitamin D deficiency were ethnicity, with with increased sunlight exposure,31 it is not women of Māori, Pacific, and other ethnic possible to make a single recommendation32 groups all at increased risk compared with and there is a variable response to UVB women of European ethnicity; and season radiation, with some individuals having low with women tested during winter or spring serum 25(OH)D concentrations despite high being at increased risk compared with those levels of sun exposure.33 tested during summer. The associations we observed of season Our study, the third to describe vitamin and ethnicity with vitamin D status are D status in a pregnant population in New consistent with those described previously Zealand, is the largest and the only one in New Zealand studies of women of child that has enrolled an ethnically diverse bearing age and of newborns.7,9,34 The New population over the full calendar year and Zealand guidelines support the identifi- the first to assess the factors associated cation and supplementation of pregnant with vitamin D deficiency. Our findings women at risk of vitamin D deficiency, confirm the higher prevalence of vitamin based on the presence of specific factors: D deficiency in darker-skinned women, “dark skin, living in the lower regions of as reported in a study of 228 South Asian New Zealand in winter and completely women living in Auckland6, and the high avoiding sunlight.”8 The guideline did not prevalence in winter and spring as reported identify ethnicity per se as a risk factor, in the same paper6, and elsewhere. 23-25 nor is season emphasised except with Vitamin D deficiency was also found to be reference to those living in more southern prevalent (87%) in a study which enrolled a parts of New Zealand. Ethnicity defines multi-ethnic sample of 90 pregnant women groups of individuals who differ with from a single primary care practice in respect to skin colour, socio-economic Wellington city.5 Our study confirms that status and cultural behaviour, which in the vitamin D requirements in pregnancy turn can determine the amount of sun cannot be met by the current vitamin D exposure and the types of food consumed. content of the New Zealand diet, particu- Classifying skin pigmentation more simply 26 larly for Pacific and Māori women. as non-European ethnicity will simplify for We found no association between serum women and for health workers the identi- 25(OH)D levels and age, parity, smoking, BMI, fication of those women at greater risk of or dietary vitamin intake. Whereas others vitamin D deficiency.

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 30 www.nzma.org.nz/journal ARTICLE

Routine measurement of serum 25(OH) Limitations of our study includes its D concentration is only recommended in modest sample size and enrolment from symptomatic women during pregnancy in a single metropolitan centre, albeit of an New Zealand,35 which differs from advice ethnically and socioeconomically diverse in a position statement for Australian sample. We did not seek to enrol a sample and New Zealand infants which states, that was representative of the New Zealand “pregnant women, especially those who maternity population, so caution is required are dark-skinned or veiled, should be when generalising from the study region to screened” to prevent infant vitamin other New Zealand regions. 36 D deficiency. This statement reflects If pregnant women were to require current clinical practice in Australia vitamin D supplementation, a daily dose where women with risk factors have of 400 to 600 IU/day is recommended.8 screening blood tests.37 A recent study However, a recent study had shown that from Australia had however advocated a dose of 2,000 IU/day from 13 weeks for change. Using only three predictor gestation was needed to protect 98% (44 factors (non-European ethnicity, BMI neonates) from vitamin D deficiency.40 We >30kg/m2 and all women in winter) it was showed, in the randomised trial in which shown that 91% of those with vitamin D these women were enrolled, that daily deficiency would be identified and unnec- vitamin D supplementary doses of 1,000 essary testing avoided in 58%.23 IU or 2,000 IU are safe as defined by serum Supplementation based upon a simple calcium concentration measurement.16 Not predictor model using ethnicity (Māori, all the vitamin preparations prescribed Pacific and Other) and season (winter and in pregnancy in New Zealand contain spring) (Table 4) would result in 72% of the recommended dose of vitamin D women with vitamin D deficiency being and not all women have access to the supplemented, but 19% of those who were recommended supplements due to cost. vitamin D sufficient also being supple- Our findings support vitamin D supplemented. Our preferred predictor model is mentation for every pregnant woman in for the supplementation of all pregnant winter and spring. Even if issues of access women during winter and spring, which to appropriate vitamin D supplementary would result in 81% of those with vitamin doses were addressed, poor adherence D deficiency being supplemented and 27% to nutritional guidelines in pregnancy is of those who were vitamin D sufficient also common41 and may limit the success of any being supplemented. The predictor model universal supplementation programme. could potentially have an even higher We acknowledge that there are no data on predictive value (PV) amongst similar the cost-effectiveness of supplementation populations living at more southern without testing in winter.32 latitudes in New Zealand. The predictive Although missing 15% of those with value of the model may not be as high in severe vitamin D deficiency is a concern, population areas where the majority of the the costs involved with measuring serum population are European. Our prediction 25(OH)D levels on all pregnant women model, has a similar predictive value to in New Zealand (as recommended in the that recently reported from Australia,23 but position statement by Paxton et al 2013)32 has a better PV to that of Jensen et al for a is prohibitively expensive and acknowl- sample of pregnant women in Norway— edged by the Ministry of Health as not perhaps reflecting the more limited dietary being feasible.8 Our proposal for vitamin sources of vitamin D in New Zealand D supplementation of all pregnant women compared with Norway.38,39 A study that during 6 calendar months of the year aimed to identify risk factors for vitamin D (from June to November inclusive) is deficiency in children aged 12–22 months easier to remember and administer than living in Dunedin (latitude 45oS) confirmed current MoH guidelines, which make no a strong seasonal variation but lacked the clear recommendation on supplemen- diversity of sample needed to investigate tation other than to identify groups of the relationship between ethnicity and pregnant women, who—because of their vitamin D status.10 skin pigmentation, sunlight avoidance

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 31 www.nzma.org.nz/journal ARTICLE

behavior or underlying conditions—will be at greater risk of being vitamin D defi- Conclusion ciency.8 The current New Zealand MoH Vitamin D deficiency was present in recommendations potentially miss 42% 42% of this ethnically diverse sample of pregnant women living in Auckland, New of vitamin D deficient pregnant women. Zealand. Winter/spring seasons and non-Eu- Our recommendation does not remove ropean ethnicity were the only independent the responsibility of the clinician caring risk factors for vitamin D deficiency. Given for a pregnant women to acknowledge the high prevalence of vitamin D defi- that she may be at higher risk of more ciency in our population and the limited severe vitamin D deficiency due to her number of factors associated with vitamin skin pigmentation, lifestyle or presence D deficiency, we recommend that a simple of liver or kidney disease, and to manage approach would be for all pregnant this appropriately at an individual level,8 women to receive vitamin D supplemen- with such management likely to include tation during winter and spring, although measurement of 25(OH)D levels and poten- it is acknowledged that data on costs and tially larger doses of vitamin D. benefits of various options is lacking.

Competing interests: The study was performed in South Auckland and the Department of Paediatrics, Faculty of Medical and Health Sciences, University of Auckland. Author information: Alec J Ekeroma, Department of Obstetrics and Gynaecology, Faculty of Medical and Health Sciences, University of Auckland, New Zealand; Carlos A Camargo Jr, Emergency Medicine, Massachusetts General Hospital, Harvard Medical School, Boston, 02114, USA; Robert Scragg, School of Population Health, University of Auckland, Auckland; Clare Wall, Department of Nutrition, University of Auckland, Auckland; Alistair Stewart, Epidemiology & Biostatistics, University of Auckland, Auckland; Ed Mitchell, Department of Paediatrics: Child & Youth Health, University of Auckland, Auckland 1142, New Zealand; Julian Crane, Department of Medicine, University of Otago, Wellington, New Zealand; Cameron C Grant, Department of Paediatrics: Child & Youth Health, University of Auckland, Auckland 1142, New Zealand. Corresponding author: Alec Ekeroma, c/o Pacific Women’s Health Research Unit, Department of Obstetrics and Gynaecology, Middlemore Hospital, University of Auckland, Private Bag 93311, Auckland, New Zealand [email protected] URL: www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2015/vol-128-no-1422-25- september-2015/6663

REFERENCES: 1. Nassar N, Halligan G, 2. . Cochrane Database of Fraser WD. Maternal vita- Roberts C, Morris J, Ashton Systematic Reviews 2012. min D status and adverse A. Systematic review of pregnancy outcomes: 3. Aghajafari F, Nagule- first-trimester vitamin a systematic review sapillai T, Ronksley PE, D normative levels and and meta-analysis. The Tough SC, O’Beirne M, outcomes of pregnancy. Journal of Maternal-Fetal Rabi DM. Association Am J Obstet Gynecol & Neonatal Medicine between maternal serum 2011; 205: 208.e1-7. 2013; 26(9): 889-99. 25-hydroxyvitamin D level 2. De-Regil LM, Palacios 5. Judkins A, Eagleton C. and pregnancy and neona- C, Ansary A, Kulier R, Vitamin D deficiency tal outcomes: systematic PenaRosas PJ. Vitamin in pregnant New review and meta-analysis D supplementation for Zealand women. NZMJ of observational studies. women during pregnancy. 2006; 119(1241): http:// Cochrane Database of BMJ 2013; 346: f1169. journal.nzma.org.nz/ Systematic Reviews Issue 4. Wei S-Q, Qi H-P, Luo Z-C, journal/119-1241/2144/.

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 32 www.nzma.org.nz/journal ARTICLE

6. von Hurst PR, Stonehouse tion-based cohort study. 22. Ministry of Health. W, Coad J. Vitamin D Clinical endocrinology Presenting Ethnicity: status and attitudes 2013; 79: 333–41. Comparing prioritised and towards sun exposure 14. Naugler C, Zhang J, Henne total response ethnicity in South Asian women D, Woods P, Hemmelgarn in descriptive analyses living in Auckland, New BR. Association of vitamin of New Zealand Health Zealand. Public Health D status with socio-demo- Monitor surveys. Nutr 2010; 13(4): 531-6. graphic factors in Calgary, Wellington; 2008. 7. Ministry of Health. Alberta: an ecological 23. Perampalam S, Ganda K, Vitamin D Status of New study using Census Chow KA, et al. Vitamin D Zealand Adults. Findings Canada data. BMC Public status and its predictive from the 2008/09 New Health 2013; 13(316). factors in pregnancy in 2 Zealand Adult Nutrition 15. National Advisory Australian populations. Survey. Wellington: The Australian & New Committee on Health and Ministry of Health, 2012. Zealand journal of Disability. Screening to obstetrics & gynaecology 8. Ministry of Health. Improve Health in New 2011; 51(4): 353-9. Companion Statement on Zealand: Criteria to assess Vitamin D and Sun Expo- screening programmes. 24. Schneuer FJ, Roberts CL, sure in Pregnancy and Wellington; 2003. Guilbert C, et al. Effects Infancy in New Zealand. of maternal serum 16. Grant C, Stewart A, . Wellington; 2013. 25-hydroxyvitamin D Scragg R, et al. Vitamin concentrations in the first 9. Camargo CA, Ingham T, D During Pregnancy and trimester on subsequent Wickens K, et al. Vitamin Infancy and Infant Serum pregnancy outcomes in D status of newborns in 25-Hydroxyvitamin D an Australian population. New Zealand. Brit J Nutr Concentration. Pediatrics The American journal 2010; 104(07): 1051-7. 2014; 133: e143-353. of clinical nutrition 10. Houghton LA, Szymlek- 17. Crop and Food Research. 2014; 99(2): 287-95. Gay EA, Gray AR, Ferguson New Zealand Food Compo- EL, Deng X, Heath A-LM. 25. Li W, Green TJ, Innis SM, sition Database products. Predictors of vitamin D et al. Suboptimal vitamin Christchurch; 2008. status and its association D levels in pregnant with parathyroid hormone 18. Food Standards Australia women despite supple- in young New Zealand New Zealand. AUSNUT ment use.(QUANTITATIVE children. Amer J Clinic 2007 Database Files. In: RESEARCH)(Report). Nutri 2010; 92(1): 69-76. Zealand FSAN, (ed.); 2007. Canadian Journal of Public 11. Nessvi S, Johansson L, 19. Scragg R JR, Holdaway Health 2011; 102(4): 308. Jopson J, et al. Association IM, Lim T, Beaglehole 26. Shrapnel W, Truswell of 25-hydroxyvitamin R. . Myocardial infarc- S. Vitamin D deficiency D3)levels in adult New tion is inversely in Australia and New Zealanders with ethnicity, associated with plasma Zealand: What are skin color and self-re- 25-hydroxyvitamin D3 the dietary options? ported skin sensitivity to levels: a community-based Nutrition & Dietetics sun exposure. Photochem- study. Int J Epidemiol 2006; 63(4): 206-12. 1990; 19(3): 559-63. istry and photobiology 27. Bodnar L, Catov J, Simhan 2011; 87(5): 1173-8. 20. Maunsell Z, Wright DJ, H, Holick M, Powers R, 12. Levy MA, McKinnon Rainbow SJ. Routine Roberts J. Maternal Vita- T, Barker T, et al. isotope-dilution liquid min D Deficiency Increases Predictors of vitamin D chromatography-tandem the Risk of Preeclampsia. status in subjects that mass spectrometry assay J Clin Endocrinol Metab consume a vitamin D for simultaneous measure- 2007; 92: 3517-22. supplement. European ment of the 25-hydroxy 28. Williams D, Fraser A, journal of clinical metabolites of vitamins Fraser W, et al. Asso- nutrition 2015; 69(84-89). D2 and D3. . Clin Chem ciations of maternal 13. Andersen LB, Abrahamsen 2005; 51: 1683-90. 25-hydroxyvitamin B, Dalgard C, et al. Parity 21. Lewis J, Elder P. Serum D in pregnancy with and tanned white skin 25-OH vitamin D2 and D3 offspring cardiovascular as novel predictors of are stable under exag- risk factors in childhood vitamin D status in early gerated conditions. Clin and adolescence. Heart pregnancy: a popula- Chem 2008; 54(11): 1931-2. 2013; 99: 1849-56.

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 33 www.nzma.org.nz/journal ARTICLE

29. Bendall A, Costa C, Woods 2007; 92(6): 2130-35. et al. Development and C, Howat P. Vitamin D 34. Rockell JE, Green TJ, Skeaff Validation of a Vitamin levels in pregnant women CM, et al. Season and D Status Prediction booking for antenatal ethnicity are determinants Model in Danish Preg- care in Far North Queens- of serum 25-hydroxyvi- nant Women: A Study land. Australian and tamin D concentrations of the Danish National New Zealand Journal of in New Zealand children Birth Cohort. PLoS ONE Obstetrics and Gynaecol- aged 5-14 y. J Nutr 2013; 8(1): e53059. ogy 2012; 52(4): 391-4. 2005; 135(11): 2602-8. 39. Calvo MS WS, Barton CN. . Vitamin D intake: 30. Morton SMB, Atatoa 35. Ministry of Health. a global perspective of Carr PE, Bandara DK, al. Companion Statement current status. . J Nutr e. Growing Up in New on Vitamin D and Sun 2005; 135: 310-6. Zealand: A longitudinal Exposure in Pregnancy study of New Zealand chil- and Infancy in New 40. March KM, Chen NN, dren and their families. Zealand. Wellington: Karakochuk CD, Shand Report 1. Before we are Ministry of Health, 2013. AW, Innis SM, et a. born Auckland:. Univer- Maternal vitamin D3 36. Munns C, Zacharin MR, sity of Auckland; 2010. supplementation at 50 Rodda CP, et al. Prevention μg/d protects against low 31. Holick MF. Vitamin D and treatment of infant serum 25-hydroxyvitamin deficiency. New Engl J and childhood vitamin D D in infants at 8 wk of age: Med 2007; 357(3): 266-81. deficiency in Australia and a randomized controlled New Zealand: a consensus 32. Paxton GA, Teale GR, trial of 3 doses of vitamin statement. Med J Aust Nowson CA, et al. Vitamin D beginning in gestation 2006; 185(5): 268-72. D and health in pregnancy, and continued in lactation. infants, children and 37. Nowson CA, McGrath JJ, J Amer Soc Nutri 2015: 1-9. adolescents in Australia Ebeling PR, Haikerwal A, 41. Morton S, Grant CC, Wall C, and New Zealand: a al e. Vitamin D and health Atatoan Carr PE, Bandara position statement. Med J in adults in Australia and DK, al. e. Adherence to Aust 2013; 198(3): 142-3. New Zealand: a position nutritional guidelines 33. Binkley N, Novotny R, statement. Med J Aust in pregnancy: evidence Krueger D, et al. Low 2012; 196(11): 686-7. from the Growing Up Vitamin D Status despite 38. Bjørn Jensen C T-LA, in New Zealand birth Abundant Sun Exposure. Vadgard Hansen L, Strøm cohort study. Public J Clin Endocrinol Metab M, Odgaard Nielsen N, health nutrition 2014.

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 34 www.nzma.org.nz/journal ARTICLE

The community pharmacy- based anticoagulation management service achieves a consistently high standard of anticoagulant care Paul Harper, Ian McMichael, Dale Griffiths, Joe Harper, Claire Hill

ABSTRACT AIM: To ensure that the Community Pharmacy-Based Anticoagulation Management Service (CPAMS) in New Zealand has continued to deliver a high standard of anticoagulant care as the service has grown to provide warfarin supervision to over 4,000 patients. METHODS: A clinical audit of patients managed through CPAMS over two years from 1 January, 2013. Anticoagulant control was assessed by measuring the time in therapeutic range (TTR), proportion of high and low INR results and incidence of reported bleeds. Compliance with the service was evaluated by monitoring the frequency of testing and the interval between tests. RESULTS: There has been only a modest change in the TTR from 76.4% to 74% during the audit period, despite the growth in patient numbers from 850 to 4,350. There was no change in the proportion of INR results above 4.0. Bleeding was reported in less than 4% of visits and 82% of bleeds were minor. 75% of patients attended for INR testing on the expected date, and only 3.3% were more than 2 weeks overdue. The interval between tests remained constant at approximately 19 days. CONCLUSION: CPAMS provides safe reliable anticoagulant care with a consistently high level of anticoagulant control.

n New Zealand approximately 35,000 tient-testing and decision support software people regularly take warfarin and more (DSS). The aim was to reduce the general Ithan 80% have their treatment super- practitioners’ workload by making use of vised by their own general practitioners. other skilled health professionals and to see There is no standardised procedure for war- if this type of service could achieve a safe farin management, with no regular audit level of anticoagulant control. process or national reporting. The level of The pilot study included 690 patients anticoagulant control achieved by general managed through 15 pharmacies over 9 practitioners is unknown, but an audit in months. It achieved a high level of antico- 2004 of a large cohort of patients managed agulant control with the INR maintained through primary care showed that the INR in the therapeutic range at 78.6% of the was maintained within the therapeutic time3. The study also received positive range only 55% of the time.1 International feedback from the patients, pharmacists guidelines recommend the time in the ther- and general practitioners involved.4 On the apeutic range (TTR) should be greater than basis of these findings, a service funded by 60%.2 In 2010, a proposal was put to Health the Ministry of Health was introduced into Workforce New Zealand to pilot an alter- approved pharmacies from late 2012. native method of warfarin management Over the last 2½ years, the service supervised by pharmacists using near-pa- has grown steadily, is now available in

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 35 www.nzma.org.nz/journal ARTICLE

140 pharmacies and provides warfarin asked about missed medication, changes supervision for over 4,500 patients. to medication, episodes of bleeding since One potential risk of expanding the the previous test and hospital admissions; service is that the level of anticoag- this was recorded on the computer system. ulant control could fall. It is recognised Since April 2014, the user has been able to that the outcome of pilot studies and record the severity of the bleed as minor clinical trials is not always maintained (gum bleeding, spotting from the nose, when the service is extended into more minor bruising), moderate (blood in the general use. In the Community Pharmacy bowel motions, haematuria, bruising >4cm) Based Anticoagulation Management or major (bleeding requiring attention Service (CPAMS) pilot, the pharmacists in hospital, intracranial bleeding, major involved were highly motivated, had a gastro-intestinal bleed, major urinary tract well-established relationship with their bleed or any bleed requiring a blood trans- local general practitioners and were fusion). The service was supported with able to provide close supervision with close collaboration between the pharmacist appropriate collaboration with medical and the referring doctor. A mechanism staff during the trial period; such close was in place at each pharmacy to ensure monitoring may not be possible in all that results outside the INR range 1.5 to pharmacies long-term. There was also 4.0 could be discussed with a supervising potential for patient selection bias, as doctor at the patient’s own general practice. patients were referred at the discretion of the local general practitioners. Although Assessment of anticoagulant the entry criteria for patients were wide control and allowed for unstable patients to be Time in therapeutic range (TTR) was 5 included, there was the possibility that the calculated using the method of Rosendaal, more complex cases were not referred as which uses linear interpolation between this service was new and seen by some as successive INR values; the time in days experimental. above, within and below the therapeutic range was calculated for the total popu- The aim of our audit was to see if the lation for each month by calculating the CPAMS service has continued to provide sum of all results up to each time point. The a high level of anticoagulant control as it TTR was also calculated for each patient has grown and to ensure that it maintains using all their available INR results from appropriate INR testing frequency and good November 2011 to February 2015. adherence to testing on time. The proportion of INR results between 4.0 to 4.5, 4.5 to 5.0, 5.0 to 5.5, 5.5 to 6.0 and Methods above 6.0, and the proportion of INR result Data for the audit were collected from at 1.5 or lower were calculated each month, the decision support software database based on INR tests performed in that (INR Online Ltd, Palmerston North, New month. The number of episodes of bleeding Zealand). All patients on warfarin between reported by the patients was calculated

1 January, 2013 and 31 December, 2014, each month and reported as a percentage managed through CPAMS were included in of the total number of tests performed each the audit. To enable accurate calculation month. From April, 2014, the severity of the of TTR, the INR results from 1 November bleed was also recorded. 2011 to 28 February 2015 were collected for the audit patients. The following data were Test intervals and adherence recorded for each patient: date of birth; to testing gender; reason for anticoagulant therapy; At each INR test the interval since the and length of treatment. The following data previous test was recorded and the mean were collected each time an INR test was interval calculated each month. Adherence performed: date of the test; the INR result; to testing was calculated by measuring the the DSS recommended dose; the given difference between the recommended test dose; the DSS recommended date of the date and the actual test date and recorded next test; and the pharmacists selected date as on time, 1 to 3 days late, 3 to 7 days late, 7 of next test. At each test the patient was to 14 days or more than 14 days late.

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 36 www.nzma.org.nz/journal ARTICLE

Figure 1: Growth of patient numbers and pharmacies providing the service.

Table 1: Diagnostic groups.

Diagnosis Female Male Total Percent Mean age (yrs) Atrial fibrillation 1,515 2,178 3,693 63.00% 71.8

Deep vein thrombosis 247 319 566 9.60% 60

Mechanical heart valve 242 448 690 11.70% 59.2

Pulmonary embolus 226 218 444 7.60% 61.5

Other 179 294 473 8.10% 62.6

Total 2,409 3,457 5,866 100.00% 67.65

Statistical analysis cations for treatment were atrial fibrillation The assessment of trends for each (63%), venous thromboembolic disease parameter was measured using the (DVT & PE) (17.2%) and mechanical heart Mann-Kendall test; p<0.05 was regarded valves (11.7%); 59% of patients were male. as a significant trend. The comparison of The median age for males was 69.7 years the incidence of bleeding at various TTR (mean 67.3 years) and for females 71.6 ranges was calculated using chi-squared. years (mean 68 years). The mean age was The comparison of the mean time between higher in patients with atrial fibrillation intervals at various TTR ranges was than the other patient groups. evaluated using t-test. p<0.05 was regarded as significant. Assessment of Results stability of control A total of 5,866 patients on warfarin have To assess the stability of control we been managed through CPAMS during measured the time in the therapeutic the audit period. The number of patients range, proportion of tests above 4.0 and actively on treatment has grown steadily below 1.5 and the incidence of reported over the 2-year audit period, from 850 bleeds each month. (January 2013) to 4,350 patients (December Time in the therapeutic range: At the start 2014). By December 2014, the service was of the audit period, 32 pharmacies were provided in 126 pharmacies (Figure 1). providing the service to 853 patients. The Demographics TTR for this population was 76.4% based The reason patients were taking warfarin on 17,000 INR results. As the size of the is shown in Table 1. The most common indi- patient population increased there was

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 37 www.nzma.org.nz/journal ARTICLE

Figure 2: The time above, within and below the therapeutic range each month.

Figure 3: Patients were grouped by deciles based on their TTR. The figure shows the number of patients in each group at the end of the audit period.

Figure 4: Percentage of tests with an INR above 4.0 each month.

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 38 www.nzma.org.nz/journal ARTICLE

Figure 5: Percentage of tests with an INR of 1.5 or less each month.

Figure 6: Proportion of tests where the patient reported bleeding since the previous test.

Figure 7: Relationship between the time in range (patients grouped in deciles based on TTR) and the reported incidence of bleeding. Bleeding rate falls significantly between TTR 50% and 100%; P<0.001.

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 39 www.nzma.org.nz/journal ARTICLE

Figure 8: Mean interval between INR tests + 1 SD.

Figure 9: Relationship between time in range (patients grouped in deciles based on TTR) and mean test interval (+ SEM).

Figure 10: Proportion of tests performed on the expected test date or later than the expected date.

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 40 www.nzma.org.nz/journal ARTICLE

a fall in the TTR to 74.8% (45,000 results) control improved, with a mean of 23 days at 12 months and 74% (124,000 results) at for patients with a TTR of 80 to 90% and 29 24 months (significant decreasing trend, days with a TTR of 90 to 100% (Figure 9). p<0.05). The change in the TTR was largely Adherence: Adherence to testing was due to an increase in the proportion of assessed by monitoring the proportion time below the therapeutic range, which of patients who attended for testing on increased from 12.6% to 16% (p<0.05). The the expected test date and the proportion proportion of time above the therapeutic who were late for testing. 75% of tests range remained around 10% for the whole were performed on the expected date, a audit period (Figure 2). further 7.8% were only one day late. A The time in range was calculated for total of 93.8% of tests were performed each patient at the end of the audit period. within a week of the expected test date and The distribution of results is shown in only 3.3% were more than 2 weeks late. Figure 3; 75.4% of patients had a TTR Adherence remained stable throughout the greater than 60%. audit period (Figure 10). Proportion of INR results outside the therapeutic range: Between 3% and 4.7% of Discussion INR results had a value of >4.0 each month. The main finding of our audit is that This did not change significantly over the community pharmacy-based anticoagulant audit period (no significant trend, Mann- service consistently delivers high quality Kendall test) (Figure 4). Between 4.1% and anticoagulant care with the time in thera- 6.3% of INR tests were less than or equal to peutic range above 74%. This is in spite of the 1.5 (Figure 5). The results show a significant fact that the patient population has grown positive trend. approximately five-fold during the audit Proportion of bleeds reported by the period (Figure 1), and that the service is now patient at each visit: The patient was asked supervised by more than 250 trained pharma- about bleeding events since the previous cists. Our results also show that the frequency test at each visit. The user could not of testing remained remarkably constant, and proceed with the INR test until this had patient adherence to testing is consistently been answered on the computer system. high. However along with this growth, there The software was changed in April 2014 has been a trend showing a modest fall in to allow the user to record the severity of anticoagulant control, with the TTR dropping the bleed. The proportion of tests where from 76.4% to 74% over two years. bleeding was reported, ranged from 2.1% The TTR is the most widely used to 4.9% each month. The trend was not measure of anticoagulant control, with uniform, with more bleeds reported during several studies showing a correlation the months September to December. There between the time in range and the inci- is no significant trend during the first 15 dence of both bleeding and thrombosis. months (only bleeding reported). The rate A TTR around 75%, as achieved by the of reporting was higher after the pharma- CPAMS programme, would be expected cists were able to report the bleed severity; to achieve good clinical outcomes with a 82.3% of bleeds are minor, 15.6% moderate low incidence of warfarin complications. and 2.1% major (Figure 6). Figure 7 shows Two large clinical trials6,7 showed that the relationship between anticoagulant patients on warfarin with good control control and the incidence of bleeding calcu- (INR within range 75% of the time) had lated at the end of the audit period. an annual mortality of 1.69% and major bleeding events of 1.58%, whereas the Assessment of stability of testing poorly controlled patients (INR within frequency and adherence to the therapeutic range <60% of the time) testing had an annual mortality of 4.2% and Test interval: The mean interval between major bleeding of 3.85%. A TTR of 74% is each test was calculated each month. This significantly higher than the results from remained constant at approximately 19 a meta-analysis of anticoagulant clinics days throughout the audit (Figure 8). The in the USA (TTR 63%)8 and well above the interval increased significantly as the INR level of control reported in this country

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 41 www.nzma.org.nz/journal ARTICLE

(TTR 55%)1. Our results are similar to those lation, varied experience of the supervising achieved by a pharmacy-based service in pharmacists, and varied adherence to the Canada (TTR 73%)9. decision support software. It is likely that Although the TTR is a useful measure of the proportion of more complex cases has the quality of anticoagulant management, increased as the service has expanded. there is some evidence that the variability Patients were referred to the service at the of INR results is also linked to adverse discretion of their doctor, and in some prac- events.10 There is no standardised method tices only stable patients were submitted of recording variability,11 but moni- initially; only when the doctors were toring the proportion of INR results at the comfortable with the service were the more extremes of measurement can give an complex cases referred. The level of control indirect assessment. Over the audit period fell most during the first 12 months of the we found no significant change in the audit, and stabilised during the second proportion of INR results above 4.0 (Figure year—suggesting that control had reached a 4) and only a slight increase in INRs below more steady state (Figure 2) 1.5. Although the trend reached statistical Another concern with an expanded significance, the proportion of low INR service and a more diverse group of patients results reported each month only increased is that it could become less efficient, with by about 1 % over the whole audit period patients requiring more frequent testing and (Figure 5). worse adherence to the expected test dates. The most direct measure of the compli- However, our results show that the interval cations of warfarin is the incidence of between tests remains remarkably stable bleeding. In our audit we have tried to with mean interval of approximately 19 days capture this by asking the patients about with a similar distribution of results (Figure bleeds since their previous test. This has 8). The consistent test interval is, in part, some limitations as it is dependent on due to the use of DSS as this recommends the pharmacists asking the question and a test date at each visit and the pharma- recording the answer on the computer cists followed this 70% of the time. There is system, however the software is designed no recognised standard test frequency for to ensure that a question about bleeding warfarin management. International guide- must be answered before the INR test can lines advise that testing every 4 to 6 weeks be performed. During the first 15 months of is appropriate for stable patients, but more the audit the software only allowed the user frequent testing is necessary for those with to record the presence of bleeding without less stable control. Studies have shown that 12,13 giving details of the severity; during this more frequent testing improves control, period, the number of tests where bleeding however the benefit of frequent testing has was reported fluctuated from 2% to 3.5% to be balanced against the practicality of each month with no overall trend. For delivering a manageable service. We believe the last 8 months, pharmacists were able that a mean test interval around 20 days has to add details of severity and a slightly been appropriate for our service, with the higher rate of bleeding was reported. It is testing interval extended to around 4 weeks difficult to know if this increase is clini- for well-controlled patients (Figure 9). cally significant, as the different criteria for Another consistent finding is the reporting may have influenced the pharma- adherence to testing, with 75% of patients cists’ decision to report a bleed. Anecdotally, attending for their INR tests on the pharmacists reported that not all minor expected date and only 3.3% more than 2 bleeds were reported prior to the change. weeks overdue. The robust recall system Of note, the rate of reported bleeds in our which identifies patients overdue for study correlates with the TTR supporting testing and delivers automated e-mail the hypothesis that there is a relationship reminders to the patients contributes to between anticoagulant control and bleeding the high adherence rate. complications (Figure 7). The use of DSS for warfarin dosing, The gradual fall in the TTR over time test dates and reminders has helped to could be due to a number of factors, maintain the consistent service but several including changes in the patient popu- other factors are likely to contribute to

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 42 www.nzma.org.nz/journal ARTICLE

the success of the programme. One key a simpler process and they feel more component is the streamlined process involved with their care. In conclusion, our which enables patients to have an INR results confirm that a community pharma- test, see their previous results, have a cy-based anticoagulant service can deliver consultation with their pharmacist and a safe and reliable service and further get immediate treatment advice in a single expansion would be appropriate to offer visit. The feed-back from patients is that the same level of care to a larger number of this has made their warfarin management patients on warfarin.

Competing interests: Paul Harper reports he is a share holder and director of INR Online Ltd, a patient management system to assist in Anticoagulation Management. Dale Griffiths reports some funding from Health Workforce New Zealand during the conduct of the study and personal fees from Pharmaceutical Society of New Zealand, as a member of the Society’s National Executive, outside the submitted work. Joe Mr. Harper reports he is a shareholder and director of INR Online Limited, a patient management system to assist in Anticoagulation Management. Author information: Paul Harper, Clinical Haematology, Palmerston North Hospital; Ian McMichael, Pharmacist, Pharmacy 547, Hamilton; Dale Griffiths, Pharmacist, West View Pharmacy, Auckland; Joe Harper, INR Online Ltd, Auckland; Claire Hill, Devon Medical Centre, New Plymouth. Corresponding author: Paul Harper, Clinical Haematology, Palmerston North Hospital, Palmerston North, New Zealand. [email protected] URL: www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2015/vol-128-no-1422-25- september-2015/6664

REFERENCES: 1. Young L, Ockelford P, a collaborative model in compared with warfarin Harper P. Audit of commu- primary care. Int J Pharm in patients with non- nity-based anticoagulant Pract. 2015 ;23:173-81. valvular atrial fibrillation monitoring in patients 4. Shaw J, Harrison J, (SPORTIF III): randomised with thromboembolic Harrison J. A community controlled trial. Lancet. disease: is frequent pharmacist-led antico- 2003;362:1691-8. testing necessary? agulation management 7. Akins PT, Feldman HA, Internal Medicine Journal. service: attitudes towards Zoble RG et al. Secondary 34(11):639-41, 2004 a new collaborative stroke prevention with 2. Ageno W, Gallus AS, model of care in New ximelagatran versus Wittkowsky A, et al. Zealand. Int J Pharm warfarin in patients with Oral anticoagulant Pract. 2014;22:397-406. atrial fibrillation: pooled therapy: Antithrombotic 5. Rosendaal FR, Canne- analysis of SPORTIF III Therapy and Prevention gieter SC, van der Meer and V clinical trials. of Thrombosis, 9th ed: FJ, Briët E. A method to Stroke. 2007;38:874-80. American College of Chest determine the optimal 8. Baker WL, Cios DA, Physicians Evidence- intensity of oral antico- Sander SD, Coleman CI. Based Clinical Practice agulant therapy. Thromb Meta-analysis to assess Guidelines. Chest. 2012; Haemost. 1993;69:236-9. the quality of warfarin 141Suppl 2:e44S-e88S 6. Olsson SB, executive control in atrial fibrilla- 3. Harrison J, Shaw J, Harri- steering committee of tion patients in the United son JE. Anticoagulation SPORTIF III Investigators. States. J Manag Care management by commu- Stroke prevention with Pharm 2009; 15:244–252. nity pharmacists in New the oral direct thrombin 9. Young S, Bishop L, Twells Zealand: an evaluation of inhibitor ximelagatran L et al. Comparison of

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 43 www.nzma.org.nz/journal ARTICLE

pharmacist managed Outcomes. 2014;7:664-9. oral anticoagulation: a anticoagulation with usual 11. van Leeuwen Y, Rosen- systematic review and medical care in a family daal FR, Cannegieter SC. meta-analysis. Lancet. medicine clinic. BMC Prediction of hemorrhagic 2006;367:404-11. Fam Pract 2011; 12: 88. and thrombotic events in 13. Horstkotte D, Piper C, 10. Razouki Z, Ozonoff A, patients with mechanical Wiemer M. Optimal Zhao S, et al. Improving heart valve prostheses Frequency of Patient quality measurement for treated with oral anti- Monitoring and Intensity anticoagulation: adding coagulants. J Thromb of Oral Anticoagulation international normalized Haemost. 2008;6:451-6 Therapy in Valvular ratio variability to percent 12. Heneghan C, Alonso-Coello Heart Disease. J Thromb time in therapeutic range. P, Garcia-Alamino JM, Thrombolysis. 1998; Circ Cardiovasc Qual et al. Self-monitoring of 5 Suppl 1(3):19-24.

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 44 www.nzma.org.nz/journal ARTICLE

Prevalence of human papillomaviruses in the mouths of New Zealand women Rebecca Lucas-Roxburgh, Jackie Benschop, Magdalena Dunowska, Matthew Perrott

ABSTRACT AIM: Human papillomavirus (HPV) in the oral cavity has been retrospectively associated with an increased risk of developing HPV-positive head and neck squamous cell carcinoma (HNSCC). The aim of this study was to determine the prevalence of oral HPV infection in a local population of New Zealand women aged 18 to 25 years, including determination of HPV genotypes, and to assess potential risk factors for oral HPV infection using participant questionnaire responses. METHODS: Oral brushings and questionnaire responses were collected from 234 women recruited from sexual health and student health centres. Questions covered age, ethnicity, sexual partners, alcohol consumption and smoking. PGMY primers were used for HPV detection by PCR, and results confirmed by sequencing and the cobas® 4800 HPV system. RESULTS: The prevalence of HPV infection was 3.2% of 216 women (95% CI: 1.6%–6.5%). Samples from two women (0.9%, 95% CI: 0.3%–3.3%) contained oncogenic HPV, and another five (2.3%, 95% CI: 1.0%–5.3%) were positive for HPV 13. No significant associations were found between putative risk factors and the presence of oral HPV infection. CONCLUSION: The prevalence of HPV in the oral cavity of New Zealand woman was comparable to results of other studies, but showed an unusual distribution of HPV types. The comparatively high detection rate of HPV 13 suggests that further work into clinical significance of oral HPV 13 infection is warranted.

ead and neck squamous cell carci- with an increased risk of HPV-positive nomas (HNSCCs) are the sixth most HNSCC,6,7 the pathogenesis of HPV-positive common cancers worldwide, with cancers of the oral cavity is not completely H 8 an estimated incidence of 405,000 cases and understood. Risk factors for HPV-positive 200,000 deaths annually.1 These cancers HNSCCs include sexual behaviours, such as comprise tumours of the oral cavity, oro- an increased number of sexual partners and pharynx, hypopharynx, and larynx.2 Head practising oral sex.6,9,10 HPV-positive HNSCCs and neck SCCs are a heterogeneous group are most common in the oropharynx, and of cancers showing two distinct pathways in particular the tonsils.6,11 HPV-positive to malignancy. The first pathway involves HNSCCs appear to show some similarities smoking and alcohol as risk factors, and the with cervical cancer, including a pre-req- second is mediated by infection with a high uisite for a persistent infection with a risk human papillomavirus (HPV) type.3 The high-risk HPV type for a varied period of incidence of HPV-positive HNSCC in devel- time before the development of cancer.1 As oped countries has increased significantly is the case with cervical HPV infection, only in the past few decades.4,5 For example, data a very limited number of primary HPV infec- from the US have shown a 225% increase in tions persist and develop into malignancy.5 cases of HPV-positive HNSCC between 1984 The prevalence of HPV in the mouths 1 and 2004. of healthy individuals varied from 0% to Although detection of HPV DNA in the 81% in overseas-based studies,.12,13 Both oral cavity has been found to be associated low- and high-risk HPV types were detected

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 45 www.nzma.org.nz/journal ARTICLE

in the oral cavity.14-16 High-risk HPV types, researcher, oral brushings were collected such as 16 and 18, have been shown to be by self-sample. The left and right buccal associated with the development of oral surfaces and the base of the tongue were cancer, with over 90% of HPV-positive brushed using a cytobrush (Thermo- HNSCC linked to infection with HPV 16.3,8,11 Fisher). Brushes were vigorously rinsed Infection with low-risk HPV types, such in ThinPrep® vials containing 20 mL as 6 and 11, is causally associated with Preservcyt® (CYTYC Corp). Samples were respiratory papillomatosis, a condition stored at room temperature until further characterised by the growth of multiple processing. After mixing, 10 mL of each papillomas, usually in the larynx.17 sample was centrifuged at 13,000xg for 10 There is currently no data on the preva- minutes. The pellet was resuspended in lence on HPV in the oral cavity among New 200 µL of phosphate buffered saline pH Zealanders. Thus, the aim of this study was 7.0 (PBS), then transferred to a microtube for DNA extraction using High Pure™ to determine the prevalence of oral HPV PCR template preparation kit (Roche infection in a local population of New Zealand Diagnostics) as per the manufacturer’s women aged 18 to 25 years, including deter- instructions. The quality and quantity mination of HPV genotypes, and to assess of extracted DNA was assessed using a potential risk factors for oral HPV infection NanoDrop™ spectrophotometer (Thermo using participant questionnaire responses. Scientific). Methods Screening PCR with PGMY primers Initially, PCR with PGMY09 and PGMY11 Study design and recruitment primers18 was used for detection of HPV A cross-sectional design was used for this DNA in test samples. Each 20 µL reaction study. Eligible participants were women contained 2 mM MgCl2, 0.3 mM of each aged 18 to 25 years. Written consent was dNTP, 0.2 µM of each primer, one unit obtained from all participants prior to their of Platinum® Taq, and 3 µL of template participation. The study was approved DNA in 1 x reaction buffer. The PCR was by the Massey University Human Ethics carried out using a SensoQuest® lab cycler Committee Southern A (approval 13/12). (SensoQuest GmbH). An initial denatu- Recruitment took place between June and ration at 95 °C for 2 minutes was followed November, 2013, at sexual health and by 40 cycles of denaturation (95 °C for 5 student health centres in the Manawatu seconds), annealing (55 °C for 5 seconds), region. All participants were asked to and elongation (72 °C for 10 seconds for 30 complete a questionnaire which encom- cycles, increased to 30 seconds for the last passed demographic information and 10 cycles), followed by a final elongation at putative risk factors for oral HPV infection, 72 °C for 7 minutes. Positive (3 µL cobas® including smoking, alcohol consumption, positive control, and 3 µL HPV 16 DNA that and the number of sexual partners. Infor- had been extracted from a cobas® HPV 16 mation on whether the participants had positive cervical specimen) and negative been vaccinated with HPV vaccine Gardasil® (3 µL cobas® negative control, and 3 µL was also collected. Alcohol consumption water) controls were included in each run. was measured in standard drinks per week, PCR products were visualised following as defined by the New Zealand Health electrophoresis through a 1.0% agarose gel Protection Agency (330 mL can of beer, a containing 0.5 µg/mL ethidium bromide 100 mL glass of wine, or a 30 mL glass of in 0.5% Tris-borate-EDTA (TBE) buffer straight spirits). Being a current or previous at 100 volts for 45 minutes. Any sample smoker was defined as smoking more than that produced a 450 base pair band was 20 cigarettes in a lifetime. A sexual partner considered a suspect positive. was defined as a partner of either sex, and Beta-globin PCR the definition of sexual activity included The presence of amplifiable DNA in vaginal, anal, and oral sex. samples negative for HPV DNA by PGMY Sample collection and processing PCR was assessed by PCR targeting a human After thorough explanation of the beta-globin gene using the PCO4 and GH20 procedure by a registered nurse or the primers.19 Each 20 µL reaction contained

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 46 www.nzma.org.nz/journal ARTICLE

1.5 mM MgCl2, 0.2 mM each dNTP, 0.5 µM and incubated at 37 °C for approximately of each primer, 0.4 units of Platinum® Taq, 3 hours. DNA from PCR reactions that and 3 µL of template DNA in 1 x PCR buffer. produced product of the expected size PCR was performed as described for the (616 bp) was prepared using ExoProStar™ PGMY PCR. Positive (3 µL cobas® positive (GE Healthcare, Life Sciences) as per the control) and negative (3 µL water) controls manufacturer’s instructions, and sent were included in each run. A sample was for sequencing using the Big Dye Termi- considered positive if a product of the nator 3.1 chemistry at the Massey Genome expected size (268 bp) was visible on a gel. Service (Massey University). Initially, Samples that did not produce the expected two colonies were sequenced from each band were removed from the analysis. ligation reaction. If neither of these two Roche cobas® 4800 testing colonies contained HPV sequences, an additional three colonies were sequenced, Suspect positives based on PGMY PCR for a total of 5 colonies. The sequences were sent for testing using the Roche cobas® were compared with other sequences 4800 system, which detects 14 high-risk HPV available in GenBank using Basic Local types (16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, Alignment Search Tool (BLAST).21 59, 66, and 68).20 The remaining ~10 mL of oral brushings from each suspect positive Data analysis was centrifuged at 13,000xg for 10 minutes, Initial exploration of data was by the pellet was then resuspended in 1.5 mL summary statistics, tables and plots. Preservcyt®, and this preparation was sent Descriptive analysis was performed for for cobas® testing at a commercial labo- each variable using statistical software ratory (MedLab Central, Palmerton North) R 3.0.2 (R development core team 2011, according to standard protocols. Types 16 R foundation for statistical computing, and 18 were reported individually while the Vienna, Austria). Categories were collapsed other detectable HPV types were reported for analysis to avoid issues associated with as ‘other high risk type/s detected’. data scarcity for the variables ethnicity, alcohol consumption, and number of sexual Confirmatory testing partners. Associations between cate- All suspect positives that tested negative gorical variables and the detection of HPV using cobas® system were re-tested were explored using univariable logistic using PGMY PCR. Samples that did not regression. Variables were allowed to enter produce the expected band on re-testing a multivariable model if the Likelihood were considered negative for HPV DNA. Ratio Test (LRT) was statistically significant The identity of the 450 base pair bands at a p-value ≤0.20. produced by the remaining samples was confirmed by sequencing. DNA was purified from the gel using a freeze-squeeze Results procedure. Briefly, the excised band was A total of 234 participants were recruited placed in a 1.5 mL microtube on the top as part of this study. Participants were from of a filtered pipette tip, which had been sexual health (n=55) and student health cut short to fit into the tube. The tube was (n=179) centres. Each participant provided snap-frozen in liquid nitrogen, then centri- an oral brushing and at least partially filled fuged at 13,000xg for 3 minutes. The filtrate in the questionnaire. containing DNA was used for sequencing, A summary of the results obtained at either directly or after cloning into a each stage of the testing process is shown pCR4-TOPO vector using TOPO TA cloning in Figure 1. A total of 49 samples were kit for sequencing with Top10 competent considered suspect positives from the cells (Life Technologies Inc.), according to initial PGMY PCR. Of the 185 PGMY negative the manufacturer’s instructions. samples that were subjected to the beta- The colonies were screened for the globing PCR, 18 samples were excluded due presence of the 450 bp insert using PCR to a lack of amplifiable DNA. This resulted with M13 primers. Up to 10 colonies were in 216 samples for inclusion in the analysis. randomly picked into 10 µL Luria Bertani After all confirmatory testing was (LB) broth containing 50 µg/mL ampicillin, completed, seven of 216 samples were

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 47 www.nzma.org.nz/journal ARTICLE

Figure 1: Flow chart of testing methods and results obtained from each step.

1 High risk type from cobas® testing was a non-16/18 and could be any of the following high risk types; 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66, or 68.

considered positive for HPV. Thus, the alcohol consumption, and the number of prevalence of oral HPV infection among sexual partners) and the presence of oral sampled women was 3.2% (95% CI: 1.6%– HPV infection (including all types, an onco- 6.5%). An oncogenic HPV type was found genic type, or type 13) were found. in two samples (prevalence: 0.9%; 95% CI: 0.3%–3.3%), and HPV 13 was found in Discussion five samples (prevalence of 2.3%; 95% CI: The oral HPV prevalence obtained in 1.0%–5.3%). this study, although comparable to results A summary of questionnaire responses of other published studies,14,16 showed an is shown in Table 1. Questions one to unusual distribution of HPV types—namely three were answered by all participants. the comparatively high prevalence of Question four was only applicable to those HPV 13 and the lack of any HPV 16 among answering yes for question three (n=134). HPV-positive samples. The latter has been This question had eight missing responses. reported as the most prevalent HPV type Questions five to seven had 10, 9, and 11 in a number of other oral HPV prevalence missing values, respectively. studies of healthy individuals.14,15 Our study Of the 216 study participants, 196 (91%, found no associations between previously 95% CI: 86%–94%) were eligible for the free reported risk factors for oral HPV—namely Gardasil® immunisation based on their smoking, alcohol consumption, and the date of birth. Of those eligible, 68 (35%, 95% number of sexual partners—and the CI: 28%–42%) were not vaccinated. The detection of oral HPV. The lack of asso- majority (n=122) of the vaccinated women ciation may be due to the low oral HPV had completed the course of three injec- prevalence seen in this study. tions, while 12 participants received one or HPV 13 is typically transmitted through two injections. Around two thirds (n=82) of oral contact, and the shared use of objects the vaccinated women received the vaccine contaminated with saliva.8,22 HPV 13 is before becoming sexually active. The considered an non-oncogenic HPV type, as remaining third (n=44) were vaccinated after infection with this type is not associated they had commenced sexual activity. No with malignancy.23 However, detection significant associations (at p<0.2) were found of HPV types 13 and 32 in the oral cavity between putative risk factors (smoking, have been linked to development of focal

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 48 www.nzma.org.nz/journal ARTICLE

Table 1: Questionnaire responses from study participants.

Question Variable Level Overall HPV-positive HPV negative number number (%) number (%) number (%) 1 Age Born before 1990 20 (9) 0 20 (100) Born during / after 1990 196 (91) 7 (4) 189 (96) Missing 0 0 0

2 Ethnicity NZ European 153 (71) 4 (3) 149 (97) Māori 38 (18) 1 (3) 37 (97) Other 25 (11) 2 (8) 23 (92) Missing 0 0 0

3 Vaccination status Not vaccinated 82 (38) 3 (4) 79 (96) Fully vaccinated 122 (56) 4 (3) 118 (97) Partially vaccinated 12 (6) 0 12 (100) Missing 0 0 0

4 Vaccine and sexual debut1 Before sexually active 82 (61) 1 (1) 81 (99) After sexually active 44 (33) 3 (7) 41 (93) Missing 8 (6) 0 8 (100)

5 Smoking Never smoked 157 (73) 6 (4) 151 (96) Current smoker 28 (13) 1 (4) 27 (96) Previous smoker 21 (10) 0 21 (100) Missing 10 (4) 0 10 (100)

6 Alcohol consumption Doesn’t drink 47 (22) 2 (4) 45 (96) 1-5 drinks per week 128 (59) 5 (4) 123 (96) >6 drinks per week 32 (15) 0 32 (100) Missing 9 (4) 0 9 (100)

7 Sexual partners 0-5 partners 135 (63) 3 (2) 132 (98) 6-10 partners 34 (16) 2 (6) 32 (94) >11 partners 36 (17) 2 (6) 34 (94) Missing 11 (4) 0 11 (100) 1 Question only applicable to those answering yes for initial vaccination question (n=134)

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 49 www.nzma.org.nz/journal ARTICLE

epithelial hyperplasia (FEH) or Heck’s five HPV 13 cases were of non-European disease.24 Although FEH lesions are benign, decent (one Guatemalan, one Pacific they can persist for many years leading Islander and one New Zealand Māori), to cosmetic problems, which to date have even though non-Europeans constituted limited treatment options.24 FEH is char- only 29% of the study participants. The acterised by numerous painless papules two remaining HPV subtypes detected in on the lips, tongue, and buccal mucosa.24 the study were from New Zealand Euro- Heck’s disease has been described predom- peans. However, given the low numbers inantly in people of the Eskimo, American of HPV 13 positive samples, the apparent Indian, and Latin American ethnicity, and predisposition to HPV 13 infection among is strongly associated with poverty and non-Europeans should be interpreted poor living conditions.25 FEH and Heck’s with caution. In addition, as we did not disease are usually seen in children or perform an oral examination, it is unclear adolescents.25 It is unclear what proportion if any participant had visible oral lesions of oral HPV 13 infections progress to FEH suggestive of FEH. or how prevalent HPV 13 is in healthy The detection of oral HPV in this study adults. We have identified five cases of employed a multi test approach. Of the HPV 13 infection among 216 healthy 49 suspect positives in the initial PGMY females. By comparison, the prevalence PCR, only five turned out to contain HPV of oral HPV 13 infection among 689 Dutch sequences. This is likely a reflection of men with high-risk sexual behaviours was the degeneracy of HPV primers and low 0.1 % (95% CI: 0%–0.8%).26 Thus, it seems annealing temperature employed in that the prevalence of HPV 13 infection in the assay. While both are necessary for our study (2.3%) was comparatively high. detection of multiple HPV types, they also There are several possible reasons for facilitate non-specific amplification in the differences between our data and the samples negative for target HPV sequences. Dutch data, based largely on differences Our results underscore the importance in the study populations. The majority of confirmatory testing when assessing of our study subjects were university PCR data, particularly if they have been students, who were therefore likely to be generated using degenerate primers. living in student flats or hostels, both of The prevalence of oral HPV is also of which could be considered inferior living interest with regard to the prevention of situations. Although the living condi- HPV-positive HNSCC through vaccination.28,29 tions for the Dutch participants were not The quadrivalent cervical cancer vaccine specified, the comparatively high prev- (Gardasil®) was introduced in New Zealand alence of HPV 13 among subjects of the in 2008 and was initially offered free to any current study may fit with the reported female born after 1990. The vaccine protects association between FEH and poor living against HPV types 16, 18, 6 and 1128 and is conditions.25 Secondly, our sampling popu- now routinely offered to 12 year old girls as lation consisted exclusively of females, part of the national immunisation schedule. in contrast to the Dutch study, where the However, coverage rates are variable and population consisted exclusively of males. the vaccine’s introduction was controversial, Although there are no data available for largely due to the sexually transmitted HPV 13 prevalence among woman from nature of HPV.30 Results of a Costa Rica-based other countries, FEH was reported to be case control study showed that vaccination more common in females than in males.24 with Gardasil® resulted in a 93% (95% CI: Lastly, our study population may have 63%–100%) reduction in oral HPV infection included females genetically predisposed between the test and control groups.29 We to HPV 13 infection. A genetic predis- were not able to assess the effect of the position among selected non-European Gardasil® vaccine on oral HPV infection populations for FEH has been described by among New Zealand woman due to the low others.24,25 Published data on FEH in Māori prevalence of oral HPV infection with only and Pacific people are limited to a case one sample positive for HPV vaccine type report of Heck’s disease in a Polynesian (HPV 18) detected in the current study. The child in 1966.27 In our study, three of the HPV 18 positive sample was obtained from a

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 50 www.nzma.org.nz/journal ARTICLE

woman who had received all three doses of coverage of at least 70% as used in models of the Gardasil® vaccine, but after the onset of vaccine effectiveness and cost efficiency.32 sexual activity. Although this is only a single The results from this study have provided case, the finding supports the importance of initial data on oral HPV infection among vaccination prior to exposure to HPV. young women in New Zealand. The The vaccine coverage in the current study comparatively high detection rate of HPV population was 56% (95% CI: 50%–63%) 13 suggests that further work to determine which is consistent with between 48% the clinical implications of this infection and 52% coverage for the same age group is warranted, especially in the Māori and reported nationally.31 Thus, one of the chal- Pacific Islander groups who tend to be lenges for successful immunisation with over represented with respect to poverty- Gardasil® may be achieving the desired associated diseases.

Competing interests: Nil Acknowledgements: We wish to thank Massey Medical, UCOL Student Health, Radius Medical, and The Sexual Health Centre Palmerston North Hospital for their involvement in participant recruitment and sample collection. Thank you also to the study participants. We would also like to acknowledge LabPlus Auckland for the donation of the PGMY primers, MedLab Central for the donation of clinical samples, Gavin Thomas for performing the cobas® testing, and Roche Diagnostics for the discount on DNA extraction and cobas® testing kits. Author information: Rebecca Lucas-Roxburgh, PhD student, Institute of Veterinary, Animal and Biomedical Sciences, Massey University; Jackie Benschop, Senior Lecturer, Molecular Epidemiology and Public Health Laboratory, Institute of Veterinary, Animal and Biomedical Sciences, Massey University; Magdalena Dunowska, Senior Lecturer, Institute of Veterinary, Animal and Biomedical Sciences, Massey University; Matthew Robert Perrott, Senior Lecturer, Institute of Veterinary, Animal and Biomedical Sciences, Massey University. Corresponding author: Rebecca Lucas-Roxburgh, PhD student, Institute of Veterinary, Animal and Biomedical Sciences, Massey University, Palmerston North, New Zealand. [email protected] URL: www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2015/vol-128-no-1422-25- september-2015/6665

REFERENCES: 1. Miller DL, Puricelli MD, Head and Neck Cancer: the risk factor profiles for Stack MS. Virology and Role of Human Papilloma- human papillomavirus molecular pathogenesis virus. Clin Oncol (R Coll type 16-positive and of HPV (human papil- Radiol). 2010;22(7):538-46. human papillomavirus lomavirus)-associated 4. Chenevert J, Chiosea type 16-negative head and oropharyngeal squamous S. Incidence of human neck cancers. J Natl Cancer cell carcinoma. Biochem papillomavirus in Inst. 2008;100(6):407-20. J. 2012;443:339-53. oropharyngeal squamous 7. Mork J, Lie AK, Glattre E, 2. Sturgis EM, Cinciripini cell carcinomas: now et al. Human papilloma- PM. Trends in head and and 50 years ago. Hum virus infection as a risk neck cancer incidence Pathol. 2012;43(1):17-22. factor for squamous cell in relation to smoking 5. Pytynia KB, Dahl- carcinoma of the head prevalence: an emerging strom KR, Sturgis and neck. N Eng J Med. epidemic of human EM. Epidemiology of 2001;344(15):1125-31. papillomavirus-associ- HPV-associated orophar- 8. Gillison ML, Castellsague ated cancers? Cancer. yngeal cancer. Oral X, Chaturvedi A, et al. 2007;110(7):1429-35. Oncol. 2014;50(5):380-6. Eurogin Roadmap: 3. Evans M, Powell NG. The 6. Gillison ML, D’Souza G, comparative epidemi- Changing Aetiology of Westra W, et al. Distinct ology of HPV infection

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 51 www.nzma.org.nz/journal ARTICLE

and associated cancers L. National prevalence of update. J Oral Pathol of the head and neck oral HPV infection and Med. 2013;42(6):435-42. and cervix. Int J Cancer. related risk factors in the 26. Mooij SH, Boot HJ, 2014;134(3):497-507. U.S. adult population. Speksnijder AG, et al. 9. Deschler DG, Richmon Oral Dis. 2012;19(1):106. Six-month incidence and JD, Khariwala SS, et al. 17. Gillison ML, Alemany L, persistence of oral HPV The “New” Head and Snijders PJ, et al. Human infection in HIV-negative Neck Cancer Patient- papillomavirus and and HIV-infected men who Young, Nonsmoker, diseases of the upper have sex with men. PloS Nondrinker, and HPV airway: head and neck One. 2014;9(6):e98955. Positive: Evaluation. cancer and respiratory 27. Hettwer KJ, Rodgers MS. Otolaryngol Head Neck papillomatosis. Vaccine. Focal epithelial hyper- Surg. 2014;151(3):375-80. 2012;30Suppl5:F34-F54. plasia (Heck’s disease) in 10. Jordan R, Gillison M, van 18. Gravitt PE, Peyton a Polynesian. Oral Surg Zante A. Oropharyngeal CL, Alessi TQ, et al. Oral Med Oral Pathol. Carcinoma: A Unique Improved amplification of 1966;22(4):466-70. genital human papilloma- Human Papillomavi- 28. D’Souza G, Dempsey viruses. J Clin Microbiol. rus-Associated Tumor of A. The role of HPV in 2000;38(1):357-61. the Head and Neck. Pathol head and neck cancer Case Rev. 2011;16(4):173-5. 19. Saiki RK, Gelfand DH, and review of the HPV 11. Marur S, D’Souza G, Stoffel S, et al. Primer- vaccine. Prev Med. Westra WH, Forastiere Directed Enzymatic 2011;53Suppl1:S5-S11. Amplification of DNA AA. HPV-associated 29. Herrero R, Quint W, with a Thermostable DNA head and neck cancer: Hildesheim A, et al. Polymerase. Science. a virus-related cancer Reduced prevalence of epidemic. Lancet Oncol. 1988;239(4839):487-91. oral human papilloma- 2010;11(8):781-9. 20. Martin IW, Steinmetz HB, virus (HPV) 4 years after 12. Esquenazi D, Bussoloti Lefferts CL, et al. Evalu- bivalent HPV vaccination Filho I, Carvalho Mda G, ation of the Cobas 4800 in a randomized clinical Barros FS. The frequency HPV Test for Detecting trial in Costa Rica. PloS of human papillomavirus High-Risk Human Papil- One. 2013;8(7):e68329. findings in normal oral loma-Virus in Cervical 30. Osazuwa-Peters N. mucosa of healthy people Cytology Specimens. Human papillomavirus by PCR. Braz J Otorhino- Pathogens. 2012;1(1):30-6. (HPV), HPV-associated laryngol. 2010;76(1):78-84. 21. Altschul SF, Gish W, Miller oropharyngeal cancer, 13. Terai M, Hashimoto K, W, et al. Basic local align- and HPV vaccine in Yoda K, Sata T. High ment search tool. J Mol the United States--do prevalence of human Biol. 1990;215(3):403-10. we need a broader papillomaviruses 22. González-Losa MR, vaccine policy? Vaccine. in the normal oral Suarez-Allén RE, Canul- 2013;31(47):5500-5. cavity of adults. Oral Canche J, et al. Multifocal 31. hpv immunisation Microbiol Immunol. epithelial hyperplasia coverage by ethnicity 1999;14(4):201-5. in a community in the vaccination and eligibile Mayan area of Mexico. Int 14. Kreimer AR, Bhatia RK, birth cohort-feb2014. J Dermatol. 2011;50:304-9. Messeguer AL, et al. Oral Ministry of Health Human Papillomavirus 23. Cuberos V, Perez J, Lopez Website. Accessed Decem- in healthy individuals: a CJ, et al. Molecular and ber 3 2014. http://www. systematic review of the serological evidence of the health.govt.nz/system/ literature. Sex Transm epidemiological associa- files/documents/pages/ Dis. 2010;37(6):386-91. tion of HPV 13 with focal hpv_immunisation_cover- 15. Rautava J, Willberg epithelial hyperplasia: age_by_ethnicity_vacci- J, Louvanto K, et al. a case-control study. J nation_and_eligibile_ Prevalence, genotype Clin Virol. 37:21-6. birth_cohort-feb2014.pdf distribution and 24. Bennett LK, Hinshaw M. 32. Marra F, Cloutier K, Oteng persistence of human Heck’s disease: diagnosis B, et al. Effectiveness papillomavirus in oral and susceptibility. Pediatr and cost effectiveness of mucosa of women: a six- Dermatol. 2009;26(1):87-9. human papillomavirus year follow-up study. PloS 25. Said AK, Leao JC, Fedele vaccine: a systematic one. 2012;7(8):e42171-e. S, Porter SR. Focal review. PharmacoEconom- 16. Sanders A, Slade G, Patton epithelial hyperplasia – an ics. 2009;27(2):127-47.

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 52 www.nzma.org.nz/journal ARTICLE

ABSTRACT AIM: To analyse the epidemiology, serology and vaccine effectiveness in a recent New Zealand measles outbreak that started in Auckland, from December, 2013 to June, 2014, to guide further preventive measures. METHOD: Cases had a clinically compatible illness, which was either confirmed by PCR or serology, or were linked to a laboratory confirmed case. RESULTS: A total of 113 cases with 3,113 contacts were traced and managed in the Auckland region. Thirteen overseas acquired cases, produced a total of 98 locally acquired secondary cases, (plus two cases with unknown travel history). The majority of cases occurred in adolescents and young adults; 68/113 cases (60.1%) were aged 10 to 19 years. Among cases, 38.9% (44/113) were unimmunised, and 31.8% (36/113) had unknown immunisation status. A further 15.0% (17/113) of cases had received one or two doses of measles, mumps, rubella (MMR) vaccine. Of the contacts who underwent serological testing for immunity (n=735), the lowest levels of serological immunity were observed in people aged 10 to 24 years. Vaccine effectiveness was calculated for the 15–24 year age cohort at 92% (95%CI; 82–97). CONCLUSION: Results suggest that an adolescent catch-up immunisation programme would prevent further outbreaks of imported measles.

easles is a highly contagious viral declared “measles free” by the World Health exantham that historically—before Organization (WHO) in March 2014, New Mthe introduction of the vaccine in Zealand has not yet attained this milestone. 1969—infected 95–98% of children by age Measles became notifiable to public health 18 years.1 Despite 46 years of an effective authorities in New Zealand in 1996. vaccine, measles and its complications The MMR vaccine has high efficacy continue to cause global mortality and in clinical studies.4 Since a live vaccine morbidity. The burden of disease is largely requires no formal boost, the second in developing countries, where measles dose is added to improve efficacy from is the most common vaccine-preventable 90 to 95%.4 Measles vaccine was intro- cause of paediatric death.2 The incidence of duced in 1969 to New Zealand, with the disease in the western world has declined second dose added in 1992. Under the markedly in association with mass vacci- New Zealand national immunisation nation campaigns.3 Many countries have schedule, children receive MMR vaccine eliminated indigenous cases of measles at 15 months and 4 years. The rationale using a two-dose measles-mumps-rubella for this timing is to minimise interference (MMR) vaccine regime. While Australia was by circulating maternally-derived anti-

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 53 www.nzma.org.nz/journal ARTICLE

Table 1: The criteria for measles diagnosis for “confirmed” and “probable” cases. All others were deemed “not a case”.

Confirmed Probable Not a case Laboratory results consistent Fever ≥38°C No clinically compatible illness with measles and a generalised maculopapular OR rash or Fever ≥38°C and: and a generalised • cough or Laboratory results not maculopapular rash • coryza or consistent with measles and: • conjunctivitis or infection • cough or • Koplik’s spots • coryza or or • conjunctivitis or Note: • Koplik’s spots • No link to confirmed case Diagnosed vaccine reaction and link to confirmed case • No relevant laboratory results

bodies in the infant, with published reports was assigned to a public health nurse (PHN) indicating that vaccine effectiveness for the collection of all clinical information increases steadily after 6 months, reaching (diagnosis and clinical course) and missing its maximum of 95–98% at 15 months details (immunisation history and status). when the vaccine is given. High coverage Each notification was assessed for case levels are required for the prevention of status, as “confirmed”, “probable” or “not outbreaks, with modelling studies showing a case”, using the criteria in Table 1. A a vaccine uptake of over 95%5 to eliminate confirmed case required either a clinically the disease. Historically, the MMR vaccine compatible illness with links to a confirmed has suffered from unjustified controversy case, or the isolation of measles virus from about its safety (starting in 1998),6,7 so a clinical specimen, such as nasopharyngeal large cohort of adolescents are likely to be or throat swab polymerase chain reaction partially or completely unimmunised. (PCR), or serological presence of measles Many measles outbreaks continue to immunoglobulin M (IgM), or rise in measles occur in New Zealand and the western immunoglobulin G (IgG) antibody titre world, which are usually the result of between acute and convalescent sera. A imported disease. In New Zealand, 248 “probable” case had clinically compatible cases were notified in 2009 and 597 cases illness with no laboratory testing. All in 2011.8 The most recent New Zealand notified suspect cases that did not fulfil outbreak started in Auckland in December, these criteria were classified “not a case”. 2013, then spread to the Waikato and ended The outbreak coordinator initiated June, 2014. This study analyses the epidemi- case management, contact tracing and ology, immunology and vaccinology of this management for households, and identified measles outbreak to identify factors that high-risk contacts of cases. Contacts were may help to restrain further outbreaks. identified from ED, and GP waiting rooms, institutions such as schools and sports clubs, Methods air travel and other significant sources of The setting: Auckland city has a popu- exposure. Quarantine and self-isolation was lation of 1.4 million, with port and airport recommended until serology or PCR results connections to the rest of the world. could be assessed and post exposure prophy- Previous analysis of the Auckland Regional laxis was given for high-risk household Public Health Service (ARPHS) database contacts where vaccine was contraindicated (unpublished results) demonstrated an or pregnant contacts. indigenous measles-free population, with Laboratory analysis: Clinically compatible most cases imported from Asia. measles illness was confirmed by laboratory The method: All measles notifications to PCR testing if symptoms had arisen within the ARPHS were processed direct from the the previous five days. Thereafter, anti- laboratory or via telephone contact with measles IgM and IgG serology were used to hospitals, emergency departments (ED), and confirm measles illness (Table 2). Serology general practitioners (GP). Each notification was also obtained for exposed contacts to

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 54 www.nzma.org.nz/journal ARTICLE

Table 2: Measles ten-day infectious periods and timing of investigations. Day -5 -4 -3 -2 -1 0 1 2 3 4

Symptoms Prodrome Rash Rash (5 days) onset (5 days)

Investigation PCR PCR PCR PCR PCR PCR or Serology timing serology

establish their immune status. Serological from the Philippines. Soon after this, on analysis was qualitative, an optical density 27 December, the next case became ill–a result below 0.9 was negative, between 12-year-old from the Philippines visiting 0.9–1.09 equivocal and a result over 1.10 an international camp. Overall, 113 cases considered positive and “serologically were reported, 13 of which were imported immune”. Where possible, at least one from overseas and 9 were associated with sample from each of the linked chains of travel to the Philippines. Of the 18 (16%) transmission was genotyped at the National local cases that did not appear to be directly Measles Reference Laboratory to identify linked to imported cases, 16 (89%) were measles strains. later linked to known imported cases by Statistical analysis: Exploratory data genotyping. Two cases were not linked analysis was carried out in R (version by contact or genotyping to known cases, 3.1.2). The ‘lattice’ and ‘ggplot2’ libraries with one genotype unknown and one were used to produce the plots. The ‘binom’ different genotype to any known imported package exact method was used to calculate case. While the remainder (n=82; 73%) 95% confidence intervals. Age-specific were linked to imported cases, a total of 98 incidence rates were calculated by ethnic (88%) were locally acquired (Figure 1). The group from the Statistics New Zealand 2013 numbers of cases peaked in March, 2014, estimated resident population, derived with a temporal association of local cases from census data. Vaccine effectiveness with cases acquired overseas. (VE) was calculated using the screening The majority of cases occurred in adoles- method,9 in which the odds ratio is calcu- cents and young adults, with 41/113 cases lated to compare the proportion of cases (36.3%) occurring in 15 to 19 year olds, and that are vaccinated with the proportion 27/113 (23.4%) cases among 10 to 14 year in the general population (this approxi- olds (Figure 2). The majority of the cases mates to the relative risk). The difference self-identified as New Zealand European between 1 and the odds ratio is the vaccine (58/113; 51%), with Asian (19/113; 17%), effectiveness: Pacific (14/113; 12%) and Māori (13/113; 12%) also affected. Age- and ethnic-specific incidence rates (Figure 3) showed the highest population levels of illness in Where: Europeans (or other ethnicity) aged 15 to 24 P(vaccinated = YES| case = YES) is the years (n=30), and Pacific children aged 0 to probability of being vaccinated against 4 years (n=7). In the Pacific children aged measles, among cases. 0 to 4, 6 of 7 cases occurred in infants aged P(vaccinated = YES| case = NO) is the less than 1 year of age. probability of being vaccinated against Among the cases, 44/113 (38.9%) occurred measles, in the general population. in the unimmunised, 36/113 (31.8%) had unknown immunisation status, 16/113 Results (14%) occurred in infants too young to The measles outbreaks started on receive the vaccine, and 17/113 (15%) had 17 December, 2013, with the last case received one or two doses of the vaccine, occurring on 2 June, 2014. The first yet still developed the disease (16/113 had reported case was an 18-year-old, who had both vaccine doses; 14%). Of note, Māori recently visited Sydney, and the measles rates of disease were lower than European strain was later genotyped and linked to and other groups, except children aged a fellow traveller at the same dance event under 4 years.

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 55 www.nzma.org.nz/journal ARTICLE

Figure 1: Histogram which illustrates the number of measles cases overseas and locally acquired, by date of notification. A local peak is highlighted in March, 2014. There is a temporal association of local cases from people travelling to New Zealand from overseas.

Figure 2: A histogram, illustrating the numbers of confirmed and probable measles cases overseas and locally acquired by age, showing peaks in the under 1 and adolescent age groups.

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 56 www.nzma.org.nz/journal ARTICLE

Figure 3: Age and ethnic-specific population incidence rates (per 100,000) showing high rates of measles in adolescent European (or other) ethnicity and in Pacific infants under 4 years of age.

Of the total number of cases, almost a (n=735), those who tested negative (median quarter of cases (23%; 26) required hospital age: 28.8 years; n=81) or equivocal (median treatment. The majority (about two-thirds; age: 26.6 years; n=57) were generally 63/113) of the cases occurred in the younger than those who tested positive Waitemata District Health Board catchment (median age: 34.0 years; n=597). Lowest (North and West Auckland), with one-fifth levels of serological immunity were (21/113) occurring in Auckland District observed in people aged 10 to 24 years Health Board and one-quarter in Counties (Figure 4). Manukau (29/113). VE was calculated from the following The secondary cases linked by contact information: Of the total number of cases, arose primarily at one school (n=37), and 60/113 (53%) were unimmunised with 16 from household (n=29), healthcare (n=8; 3 aged less than 15 months (too young to from primary care and 5 from hospital) and immunise) and 44 known not to be immu- social contacts (n=8). nised, with 36 having an unknown vaccine During the outbreak, 3,113 contacts of the status. If the unknowns are assumed vacci- 113 cases were traced by ARPHS and quar- nated, the prevalence of vaccination among antined from school, work or large-scale cases is estimated at 53/113 (47%). Statistics social events. Of the contacts who had their New Zealand estimates that current immu- immune status determined either by history nisation coverage levels are 92% at 2 of MMR vaccination or serology (n=2,594), years;10 however, in the sample of contacts non-immune status was highest in under available to ARPHS during this outbreak, 1 year olds (68/75; 90.7%) and the 20 to 29 in which serology was tested, 83/737 (11%) year group (60/325; 18.5%). Adolescents had negative serology, indicating no, or aged 10 to 14 and 15 to 19 years also had inadequate, vaccination. If the prevalence high rates of non-immune status (22/149; of vaccination is assumed to be 90%, then 14.8% and 28/186 15.1%, respectively). the VE is conservatively estimated at 89% (if Of the contacts, who underwent sero- the 36 with unknown vaccination status are logical testing for immunity (n=737), two assumed to be vaccinated). If the unknown were aged less than 15 months and both vaccine status group are instead assumed were non-immune. Of the remainder to be non-vaccinated, then the VE calculates

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 57 www.nzma.org.nz/journal ARTICLE

Figure 4: Proportion of tested exposed contacts aged ≥15 months who were either serologically immune or equivocal, by age category (n=735; 597 positive, 57 equivocal and 81 negative).

as 98%, making the true value likely to lie data confirms the Auckland outbreaks to between 89 and 98%. be the result of a series of suboutbreaks Since it is recommended that age-spe- with 2 prevalent measles strains. Any cific VE is calculated,11 the 15 to 24 year connection with the recent prolonged 12,13 age group was selected for this analysis. Sydney outbreak was not confirmed, 12 This cohort accounts for the majority of with different predominant genotypes (B3) cases and likely to have statistical power. in Auckland and (D8) in Sydney. While New Other age groups had small numbers of Zealand has exported a relatively unimmu- cases, with little precision to undertake nised adolescent cohort to Sydney, it has this calculation. The VE for this age group clearly not exported measles. was calculated, using the immune status of The vast majority of cases were unim- contacts as a proxy for population immune munised or had unknown immune status, status. Among the 122 contacts who were although there was a 14% rate of measles serologically tested in this age group, 84% in cases having two documented MMR (106/122) were immune. Among cases of vaccinations. While the apparent vaccine this age, immune status was known in 33/50 failure rate appears high, the VE calculation individuals. In this restricted group, 22/33 confirms a high overall level of vaccine reported being unimmunised. Based on effectiveness in the adolescent cohort these counts, the VE for this age group was (92%,CI: 82–97%). These results suggest the 92% (95% CI:82–97%). reason for rapid expansion of cases were the unimmunised, low adolescent immunity Discussion and the highly contagious nature of the These results confirm the highly conta- disease. The majority of contacts of cases gious nature of measles in New Zealand, were exposed at school, at home or socially, with the appearance of a series of small with reassuringly few exposed in primary outbreaks contributing to the 2014 cases. care, hospitals, the workplace, early While the index case contracted measles childcare centres or while on flights. at an international dance event in Sydney, The VE results should be treated with many cases in this series of outbreaks caution, as the method relies on a number were associated with a single high school of assumptions, such as a sufficiently large in Auckland’s North Shore, where the case proportion of the population being unvac- had visited the Philippines and returned cinated for the disease to propagate, so that to school while contagious. Genotype there is a lack of herd immunity effect.11

NZMJ 25 September 2015, Vol 128 No 1422 ISSN 1175-8716 © NZMA 58 www.nzma.org.nz/journal ARTICLE

The method also assumes that vaccine 1. Sustained historic low coverage due coverage is relatively stable, and is subject to the controversy about the MMR to information bias and confounding, as and its safety through the late 1990 with all observational studies. and early 2000s.6 Measles elimination in 2. Waning immunity has been previ- 20 New Zealand ously reported. In a community with low incidence or in elimination Elimination of measles requires the stage of measles, vaccine-induced maintenance of a very low reproduction immunity decreased with time to number (R ) at <1 rather than the typical 0 below protective levels without measles R of 12–18, as likely occurred 0 natural boosting.4,21 in this outbreak—predominantly at the single high school. Maintaining a low 3. In New Zealand, systemic organisa- average number of secondary infections tional issues of the previous vaccine produced by an infective person reduces schedule resulted in suboptimal 5 coverage: rapid spread of measles. If R0 is maintained below 1, the number of cases i) The shift of MMR2 from 11 years will decrease and all endemic chains of to 4 years of age in 2001, despite a transmission will eventually succumb. school-based catch-up programme To achieve this sustained measles elim- for 5–10 year old children. ination requires very high immunity ii) Compromised vaccine quality, as levels. There are many examples of immu- this cohort predated an effective cold nisation programmes that have controlled chain in New Zealand established in measles transmission to the point of 2004. 14 elimination. Mathematical models The elimination of measles is a priority showed herd immunity and eventual and part of the New Zealand schedule elimination of disease for industrialised strategy. This data lends weight to a countries requires between 91 to 94% catch-up programme to improve coverage vaccine coverage for adults and 95% for in the adolescent and young adult cohort.19 secondary school age cohort.5 Models of disease transmission for New Zealand Vaccine catch-up campaigns This study was consistent with other suggest that coverage of greater than 90% measles outbreaks around the world. for both MMR doses would eventually An outbreak of measles occurred in eliminate outbreaks.15 The coverage in East Sussex, UK, from December, 1992, New Zealand is currently 93% for the first to February, 1993, with 66/96 cases at dose of MMR,10 although historically it the local school. The majority of cases has been significantly lower. A national occurred in the 11 to 17 year age group coverage survey in 1991 demonstrated (78%) with 74% of the students attending that less than 60% of children were fully the school fully immunised. While most immunised by two years of age, including cases were unimmunised,16/66 school only 42% of Māori children and 45% of cases occurred in the previously immu- Pacific children. Gains, albeit slow, were nised. VE was high and estimated at 92%. made over the following years, so that by Low vaccine uptake and waning antibody 2005, 77% of children were fully immu- levels in this age cohort contributed to nised by two years of age, including 69% the outbreak and reinforced the need 16 of Māori and 85% of Pacific children. for catch-up immunisations to reach this This has left a significant legacy cohort, in susceptible group of adolescents.19 their current adolescent or young adult In April 2013, a large outbreak was years, unprotected from measles. reported from Wales with 693 cases in Adolescent immunity to measles the Swansea area. The highest number As shown by previous studies,17,18,19 of cases was noted in the 10 to 14 year adolescent populations are most at risk of old age group,22 and a MMR catch-up disease, with the highest burden in teen- programme was implemented. Seroprev- agers aged 10 to 19 years. The reasons for alence of immunity to measles infection this are uncertain but might reflect: increased by 5 to 10% after the adolescent

catch-up campaign. Similar results from In the spring of 1985, an outbreak of campaigns among school-aged children measles occurred among adolescents have been observed in England,23 in Corpus Christi, Texas, even though Australia24 and Korea.25 vaccination requirements for the school attendance were strictly enforced. Only Vaccine failure 4.1% of the students at this school (74/1806) While natural infection confers life-long lacked detectable antibody to measles and immunity, vaccination gives protection more than 99% had records of vaccination. against measles transmission for a Of the seronegative students, all of the 14/74 20 prolonged time. MMR has been demon- who contracted measles had been vacci- strated to be protective for over 20 years nated. This demonstrates that outbreaks 26 and thought to be lifelong. While antibody can occur in secondary schools even when levels decline over time, secondary more than 99% of students have been vacci- vaccine failure due to waning protection nated and more than 95% were immune.29 is relatively rare. A UK study showed 3% of children who received MMR vaccine Conclusion between 12 and 18 months of age were Careful analysis of the New Zealand seronegative after 4 years.27 In Canada, outbreak suggests European adolescent 3.6% and 12% of children, vaccinated by and young adults, with a historically low different vaccine strains between 12 and 15 immunisation coverage, are most at risk months of age, were found to become sero- of measles. Along with prompt reporting 28 negative after 5 to 6 years of follow-up. of measles to initiate timely action and the This outbreak showed a failure rate of 14% quarantine of students without documented among cases with two documented MMR histories of immunisation during school vaccinations, though it was not possible outbreaks, the New Zealand population to determine if the failure was due to the and elimination effort could benefit from a vaccine (primary failure) or due to waning comprehensive catch-up programme in this immunity (secondary failure). age group.

Competing interests: Nil Acknowledgements: Thank you to the public health medical, nursing and administrative staff at Auckland Regional Public Health for their dedicated efforts to control this measles outbreak. Author information: Gary Reynolds, Auckland Regional Public Health Auckland District Health Board; Cassandra Dias, Auckland Regional Public Health Auckland District Health Board; Simon Thornley, Auckland Regional Public Health Auckland District Health Board; Ron King, Public Health Intelligence Team, Auckland Regional Public Health Service; Anne Morrison, Auckland Regional Public Health Auckland District Health Board; Angela Matson, Auckland Regional Public Health Auckland District Health Board; Richard Hoskins, Auckland Regional Public Health Auckland District Health Board. Corresponding author: Gary Reynolds, Auckland Regional Public Health, Auckland District Health Board [email protected] URL: www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2015/vol-128-no-1422-25- september-2015/6666

REFERENCES: 1. Langmuir AD. Medi- coverage http://www.stats. 18. Gay N, Ramsay M, Cohen cal importance of govt.nz/browse_for_stats/ B, et al. The epidemiology measles. Am J Dis Child snapshots-of-nz/nz-so- of measles in England 1962;103(3):224-6. cial-indicators/Home/ and Wales since the 1994 2. Murray CJ, & Lopez Health/childhood-im- vaccination campaign. AD. Mortality by cause munisation.aspx) Commun Dis Rep CDR for eight regions of the 11. Cohen AL, Taylor T Jr, Rev 1997;7:R17–21. world: global burden Farley MM, et al. An 19. Morse DE, OShea M, of disease study. Lancet Assessment of the Screen- Hamilton G, et al (1994). 1997;349:1269-76. ing Method to Evaluate Outbreak of measles 3. Simons E, Ferrari M, Fricks Vaccine Effectiveness: in a teenage school J, et al. Assessment of The Case of 7-Valent population: the need to the 2010 global measles Pneumococcal Conjugate immunise susceptible adolescents. Epidemiol. mortality reduction goal: Vaccine in the United Infect. 113 : 355-365. results from a model of States. 2012; PLoS ONE surveillance data. Lancet 7(8): e41785. doi:10.1371/ 20. Christenson B & Bottiger 2012; 379: 2173-8. journal.pone.0041785 M. Measles antibody: comparison of longterm 4. Davidkin I, Valle M. 12. Najjar Z, Hope K, Clark P, vaccination titres, early Vaccine-induced measles et al. Sustained outbreak vaccination titres and virus antibodies after of measles in New South naturally acquired two doses of combined Wales, 2012; risks for immunity to and measles, mumps and measles elimination in booster effects on the rubella vaccine: a Australia. West Pacific measles virus. Vaccine 12-year follow-up in Surveill Response 1994;12(20):129–33. two cohorts. Vaccine J.2014; 5(1):14-20. 21. Guris D, McCready J, 1998;16(20):2052–7. 13. Wood JG, Heywood AE, Watson JC, et al. Measles 5. Gay NJ. The theory of Menzies RI, et al. Predict- vaccine effectiveness measles elimination: ing localised measles and duration of vaccine implications for the outbreak potential induced immunity in design of elimination in Australia. Vaccine the absence of boosting strategies. J Infect Dis 2015;33:1176-1181. from exposure to measles 2004; 189(S1): S27-35. 14. Zahraei SM, Gouya MM, virus. Pediatr Infect 6. Tickner S, Leman Mokhtari A, et al. Success- Dis J 1996;15:1082–6. PJ, Woodcock A. ful control and impending 22. Iacobucci G. Wales Factors underlying elimination of measles in sets up drop-in vacci- suboptimal childhood the Islamic Republic of nation clinics to tackle immunisation. Vaccine Iran. J Infect Dis. 2011;204 measles outbreak. BMJ 2006; 24: 7030–7036. (Suppl1):S305-11. 2013; 346: f2452. 7. Ramsay ME. Measles: 15. Roberts MG, Tobias MI. 23. Pebody RG, Gay NJ, the legacy of low vaccine Predicting and preventing Hesketh LM, et al. coverage Arch Dis Child measles epidemics in Immunogenicity of 2013;98(10):752-754. New Zealand: applica- second dose measles– 8. Institute of Environmental tion of a mathematical mumps–rubella (MMR) Science and Research model. Epidemiol. Infect. vaccine and implications 2012. Annual summary of (2000), 124, 279–287. for serosurveillance. outbreaks in New Zealand 16. Turner N. The challenge of Vaccine 2002;20:1134–40. 2011 https://surv.esr.cri. improving immunization 24. National Centre for Immu- nz/PDF_surveillance/ coverage: the New Zealand nisation Research and AnnualRpt/AnnualOut- example. Expert Rev Surveillance of Vaccine break/2012/2012Out- Vaccines 2012;11(1):9–11. Preventable Diseases. breakRpt.pdf 17. Vandermeulen C, Roelants Australian Measles Control 9. Farrington CP. Esti- M, Theeten H, et al. Campaign 1998 Evalu- mation of vaccine Vaccination coverage in ation Report. Sydney: effectiveness using the 14-year-old adolescents: University of Sydney, screening method. Int J documentation, timeliness Royal Alexandra Hospital Epi 1993;22(4):742-746. and sociodemographic for Children; 1999. 10. Ministry of Health July determinants. Pediatrics 25. Kim S, Han H, Go U, et 2014 New Zealand MMR 2008;121(3):e428-434. al. Sero-epidemiology

of measles and mumps Capner P, et al. Antibodies years after immunization: in Korea: impact of to measles, mumps and effect of vaccine type and the catch-up campaign rubella in UK children 4 age at vaccination. Vaccine on measles immunity. years after vaccination 1995;13(16):1611–6. Vaccine 2004; 23: 290-297. with different MMR 29. Gustafson TL, Lievens AW, 26. Markowitz LE, Preblud SR, vaccines. Vaccine Fine PE, et al. Duration of 1995;13(9):799–802. Brunell PA, et al. Measles measles vaccine induced 28. Boulianne N, De Serres G, Outbreak in a fully immunity. Ped infect Ratnam S, et al. Measles, immunised secondary- dis J. 1990, 9:101-110. mumps, and rubella school population. N Engl 27. Miller E, Hill A, Morgan- antibodies in children 5–6 J Med 1987;316(13):771-4.

Rethinking the conceptual toolkit of organ gifting Rhonda M Shaw

ABSTRACT Organ transplantation is often spoken of as a “gift of life”. However, the concept of the “gift of life” has been subject to scrutiny over the course of the last two decades within social sciences and bioethics for failing to reflect the complexities of tissue and organ exchange. I suggest a new ethical model of organ donation and transplantation is needed to capture the range of experiences in this domain. The proposed model is both analytical and empirically oriented and draws on research findings linking a series of qualitative sociological studies undertaken in New Zealand between 2007 and 2013. The studies were based on document analysis, field notes, and 127 semi-structured in-depth interviews with people from different cultural and constituent groups directly involved in organ donation and transplantation. The aim of this viewpoint article is to disseminate to a New Zealand audience of healthcare professionals published research promoting an expanded conceptual toolkit of organ donation. The toolkit, which includes the concepts of unconditional gift, gift relation, gift exchange, body project and body work, is designed to provide an explanatory framework for organ donors and transplant recipients and to assist the development of ethical guidelines and health policy discourse.

rgan transplantation is often spoken physical and therapeutic work undertaken of as a “gift of life”, particularly on one’s own body and the bodies of others. Owhere the sale of body tissue and Aside from contributing to social science organs is illegal. However, over the course knowledge about organ donation, the aim of the last two decades, the concept of the of the toolkit is to help patients and families “gift of life” has been subject to critique. involved in organ transfer procedures In light of such criticism, I suggest that a make better sense of their experiences, new ethical model is needed to capture to aid healthcare professionals when the multivalent phenomenon of organ communicating information about organ gifting and to incorporate the range of donation and transplantation experiences experiences in this domain. The model I to prospective donors, donor families, propose is based on an expanded toolkit of and recipients and recipient families, and concepts that include unconditional gift, gift to assist in the development of ethical relation, gift exchange, body project, and guidelines and policy discourse. body work. In short, these ways of framing The proposed model is both analytical organ donation and transplantation can and empirically oriented, and links a series be defined as follows: an unconditional of qualitative sociological studies,1-6 all of gift refers to a one-way altruistic act; the which received research ethics approval gift relation refers to dyadic relationships (VUW 2-2007-SACS; MEC/08/03/027; AUTEC between intimates or people who are 08/179; VUW 16628/4/06/09; MEC/11/ familiar to one another; gift exchange is an EXP/089). These studies were informed anthropological concept that emphasises by debates in sociology and anthropology the social significance of giving, receiving, about gift theory and phenomenological and reciprocating; body project is a term accounts of organ exchange. The objective used by sociologists to refer to the use of of the studies was to examine how research one’s body as integral to the process of participants make moral decisions about identity-construction; and body work, a bodily exchanges, and to investigate the corollary sociological concept, refers to degree to which participant experience

aligns with existing moral vocabulary to necessarily subscribe to one framework describe the meanings of organ donation consistently or throughout their life and transplantation from a first-hand course. Individual and cultural views may perspective. As such, study participants vary depending on the body part being were invited to consider the salience of given or received, its symbolic significance, altruism and the metaphor of the gift as the specificities of the donor-recipient tools for making sense of their experi- relation, and the socio-cultural, moral, ences. Where vocabulary was out-of-step spiritual and sometimes even political with participants’ accounts, the aim was to views of the participants in relation to rethink it. bodily integrity and identity. Rather than suggesting a single model of organ The project donation based on a notion of the “gift of life”, the framework I propose comprises a The studies were undertaken in New range of explanatory concepts, which are Zealand and are based on document ideally flexible enough to include diverse analysis, field notes, and 127 semi-struc- understandings and experiences. Because tured in-depth interviews of 1 to 2½ hours the proposed framework is provisional, duration, conducted between 2007 and the creation of a meaningful model entails 2012. Three studies were undertaken by ongoing data collection, analysis and the author and examined the views of 19 hermeneutic reflection. Some concepts living directed kidney donors, six living may not translate well outside the context non-directed kidney donors, nine members of the global North, and some will be more of deceased donor families, and 27 trans- applicable to certain types of donation plant recipients of kidneys, lungs, and than others. hearts. Fifty-one background interviews with expert academics, stakeholders and healthcare professionals (HCPs) support Unconditional gift the project. These interviews include In this description, organ donation is discussions with 15 intensivists and donor conceptualised as a gift, given freely and and recipient coordinators, and 11 trans- without remuneration or external reward. plantation specialists (nine surgeons and The gift is assumed to have a clear-cut nephrologists and two nurses). Study meaning affixed to altruism, which is participants primarily self-identified as regarded as a disinterested one-way act. New Zealand European or Pākehā and The donated organ is given unconditionally were interviewed by the author. Fifteen by the living organ donor or deceased living organ donors, organ recipients and donor family. Some proponents of this whānau, 12 of whom identified as Māori, account maintain that there should be were interviewed in a connected study by no relationship between the donor and an associate investigator. This paper is an recipient or their respective families. abridged version of a much longer article This way of characterising organ donation by the author.7 The rationale for a summary has been adapted for use by both HCPs version of the article is to promote read- and transplant recipients. Some trans- ership among a New Zealand HCP audience. plant recipients use gift of life discourse to For an extended discussion of the research express their gratitude to deceased donor this paper draws on, and for details about families or to anonymous living organ study methods and supporting references donors. Although deceased donor families see the publications listed below.1-7 might act altruistically, they do not typically Empirical research shows that people describe donation in terms of a gift. do not think uniformly about organ Various HCPs, including transplantation donation in unquestionable moral terms specialists and surgeons, support the view as an altruistic gift. Rather, they under- of organ donation as an unconditional stand organ transplantation practices in gift. Some use the term gift to distin- a variety of divergent and overlapping guish deceased and live donation from ways. In practice, people hold a combi- commercialisation. They contend that gift nation of views about organ donation and language transmits the positive message transplantation simultaneously, and do not that donation is a noble and morally worthy

act. The term gift is thus an important part emotionally related donors often do not see of their linguistic toolkit. It is useful for their donation as a gift because they feel educational and promotional purposes, it they get something back from donating an can be used (in the right circumstances) organ, such as alleviating the suffering or with deceased donor families to thank them saving the life of a family member or loved for their donation, and in conversation with one, or improving the recipients’ health and transplant recipients in the clinical context. thereby making life better for the whole In addition, some HCPs maintain that family, themselves included. highlighting the magnitude of the gift serves to encourage transplant recipients to value Gift exchange the donated organ, and to show gratitude In jurisdictions where organ commerce is to the donors by taking responsibility for prohibited, lay accounts of organ donation looking after themselves post-transplant. and transplantation often come closest to a Viewing the gift as a one-way altruistic act gift exchange model. In this understanding, enables transplant recipients to accept the for social and cultural reasons, people often donated organ as their own, rather than feel morally obligated to reciprocate the thinking about the organ as an alien body gift of an organ in a meaningful way. A gift part from a foreign body. Being able to or gift exchange relationship is based on integrate and accept the newly donated giving, receiving and reciprocating; it is not organ is beneficial for the post-operative to be confused with pure altruism, which is emotional and psychic wellbeing of trans- a one-way transaction. plant recipients. Persons who subscribe to a gift exchange model of social life may encounter diffi- Gift relation culties in a closed donation system In this understanding, the gift relation governed by a social-organisational imper- involves face-to-face gift giving, often of an ative of confidentiality and anonymity. asymmetric nature, between people who For some transplant recipients, the need have prior personal or intimate relation- to express gratitude to the donor, donor ships with one another. In organ exchange family, or to the health system is linked to the archetypal gift relation occurs between strong affects. The desire for contact with a parent and child, although it may extend the donor or donor family can range from to living organ donation by donors who are curiosity, to desire for knowledge about biologically or emotionally related to the the donor, a desire to meet the donor, or recipient (sibling, spouse, or friend). This a desire to establish a relationship. For relation is unlike anonymous donation to a some transplant recipients, for whom stranger, in that non-directed donation entails integrating an organ from another person a cognitive recognition of the other and feels like incorporating a part of the donors’ detached reflection on the other’s suffering. genetic makeup and personhood, making If organ donation occurs as part of a gift connection with the donor or donor family relation, the donated organ often goes unrec- is important. ognised as a gift by the donor and recipient. Anxiety and ambivalence around In other words, the motivation that prompts donation and transplantation can be organ donation within some family and exacerbated when donation occurs across kinship groups, and between close friends, different kinds of body-subjects (eg, from is not articulated in the deliberate language male to female or young to old), or between of moral decision-making. This is because persons who subscribe to different belief biologically and emotionally related donors systems and worldviews. For some cultural often make instantaneous, ‘snap’ deci- groups, such as the indigenous peoples sions about donating before consciously of Aotearoa New Zealand, Australia, and weighing the risks, benefits, alternatives, and North America, connection may be espe- outcomes of their donative offer. They see cially important, as the sharing of biological their decision as ‘natural’ or ‘the right thing matter such as body parts implicates entire to do’. Their decisions are often based on kin networks, and has ramifications beyond normative expectations about their care- its effect on individual donors and recip- taking and caregiving roles. Biologically and ients. Organ donation and reception not

only creates intersubjective bonds between As with all projects, organ donation is people, it entails ongoing responsibilities something that the living donor plans. to those people and their families. To fail The intention to donate an organ, which to make connection with a donor or their may or may not be coherently expressed family, in the form of gratitude or some as part of an altruistic value system, kind of reciprocity, would be considered is carefully deliberated; from the first disrespectful and socially unacceptable. A inquiry about donation and the decision gift exchange framework may therefore be to donate, through to the medical tests preferred by some donors and recipients, donors are required to undergo and because it emphasises the interconnect- the eventual surgery. For non-directed edness between persons who give and donors who do not have an established receive body parts, and foregrounds the relationship with their recipient, the impact bodily exchanges can have for donation event has a finite end. That is, relational ontologies spanning beyond the despite their view that organ donation is interpersonal to the cosmological. a gesture of shared humanity, the matter of kinship and relatedness through the Body project bio-medically engineered transfer of their kidney to the body of another person is An alternative way of understanding not a key concern for them. living organ donation is to think about it as a body project. Classifying organ donation as a body project refers to the Body work way people purposively use and transform Body work refers to work undertaken their bodies in a variety of ways to commu- on one’s own body and to work performed nicate and express different ideas, values on the bodies of others. Organ donation is and beliefs. While living organ donors are body work that is embedded in a network not concerned to transform their bodies to of caring relationships that involve HCPs, enhance appearance and body image—as numerous stakeholders, donors, transplant is the case with most late twentieth and recipients, families, and friends. twenty-first century projects that involve Organ donation requires the interven- work on the body—they focus on what their tions of a hierarchy of HCPs, who work on body can do and how it can be used to facil- the bodies of deceased donors, living organ itate wellbeing for other people. In donative donors, and transplant recipients. This body projects, the donor makes an indi- work includes the activities of healthcare vidual decision to transform his or her body assistants who mediate between patients for the benefit of another person, with the and nurses to provide a range of services help of biomedicine and technology. in hospital, through to ICU physicians, This description is in keeping with specialists and transplantation surgeons, the perspective that many living who literally engage with and operate on non-directed donors hold. While the the physical bodies of patients in their care. desire to donate an organ is performative Body work involves work on the body- of the donors’ moral identity and sense self by donors and recipients to prepare for of social responsibility, insofar as it is and recover from organ transplantation. an action that is directed toward others, To donate and receive organs, donors and the donative decision may also be recipients have to be relatively fit and experienced by the living organ donor healthy, as well as being able to demon- as self-defining. From the non-directed strate to relevant HCPs the capacity to act donors’ perspective, organ donation is autonomously. Complying with clinical an extreme example of civic-mindedness and psycho-social assessment criteria and is consistent with a donation mind- takes effort. Living donors must learn the set. Although improved self-esteem may requisite communicative and moral reper- be a consequence of donation, studies tories to convey their donative conviction show that this is not a primary motivator. to HCPs. Being able to articulate altruistic Living organ donors may give time, blood, intent, in terms of “gift of life” rhetoric, is money, and offer services to others, and part of that package. Likewise, transplant often on a regular basis. recipients must adhere to extensive self-

care routines by taking responsibility for they involve nonpaid labour. This labour their diet, medication, and lifestyle, as part is increasingly outsourced to the home of their new postoperative health regimen. as a consequence of health philosophies Body work also involves the activities of and programmes supporting self-man- non-remunerated caregivers: a network of agement and due to economic reforms in invisible others made up of family members the formal health sector. The body work and friends, who help the living donor and concept enables us to take into account transplant recipient recover and heal post- the costs incurred by home productivity operatively. This work includes emotional, losses associated with organ donation psychological and moral support, accom- and transplantation, which need to be panying family members and friends to considered when evaluating the economic hospital visits, and providing care during effects of organ transfer for patients and convalescence (ie, housework, dependent those who care for them. Body work thus care, and miscellaneous daily activities). goes beyond the gift-commodity binary to Body work is important to consider permit discussion around compensation because it links all the participants involved for organ donors, recipients, and their in organ transplantation procedures. No respective families. matter which of the previous concepts donors, recipients, and HCPs prefer to use Conclusion to conceptualise organ transfer procedures, The expanded conceptual toolkit there is no question that they will all be presented here sketches different concep- involved in body work of some kind or tions of organ donation and transplantation, another. The danger of sentimentalising based on theoretical analysis and empirical organ transplantation as altruistic is that data drawn from sociological research. The it can conceal the mundane, physical and author’s view is that gift terminology in bodily aspects of this work and the value the context of organ donation and trans- associated with it. plantation is useful, but has limitations. Body work not only focuses on HCPs’ Principally, it does not take socio-cultural expertise, it acknowledges transplant and economic context into account, or recipients’ activities of self-care, and the different permutations of exchange rela- social care management and welfare of tions between donors and recipients and donors, those they care for, and those who their respective biographies. In contrast, care for them. It thus accounts for the the proposed toolkit incorporates some of complex relational and material aspects of the nuanced meanings that donors and organ donation and transplantation that recipients attribute to their decision-making involve labour on the body, without losing and experiences, as well as including the sight of its important metaphysical and perspectives of HCPs and analysts that may value dimensions. prove useful in the provision of pre- and Furthermore, the concept of body work postoperative care and informed consent may permit consideration of some of the procedures in both the clinical context and indirect costs associated with transplan- policy development. tation that are currently hidden because

Competing interests: Nil Acknowledgements: Thank you to the research participants who generously participated in this research programme. The author is grateful to the Marsden Fund and Kidney Health NZ for their support. Author information: Rhonda Shaw, Senior Lecturer, School of Social and Cultural Studies, Victoria University of Wellington, New Zealand. Corresponding author: Dr Rhonda Shaw, School of Social & Cultural Studies, Victoria University of Wellington, PO Box 600, Wellington 6140, New Zealand. [email protected] URL: www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2015/vol-128-no-1422-25- september-2015/6667

REFERENCES: 1. Shaw RM. The notion New Zealand organ dona- tative Health Research. of the gift in the donation system. Health: An 2015; 25, 5:600-11. tion of body tissues. Interdisciplinary Journal 6. Webb R, Shaw RM. Sociological Research for the Social Study of Whanau, whakapapa Online. 2008; 13, 6. Health, Illness & Medicine. and identity in experi- 2. Shaw RM. Perceptions 2012; 16, 3:295-310. ences of organ donation of the gift relationship: 4. Shaw RM. Live kidney and transplantation. Views of Intensivists and donation as body work. Sites. 2011; 8, 1:40-58. Donor and Recipient Critical Social Policy. Coordinators. Social 2014; 34, 4:495-514. 7. Shaw, RM. Expanding Science & Medicine. 5. Shaw RM, Webb R. the conceptual toolkit of 2010; 70, 4:609-15. Multiple meanings of organ gifting. Sociology 3. Shaw RM. Thanking and “Gift” and its value for of Health & Illness. reciprocating under the organ donation. Quali- 2015; 37, 6: 952-966.

Conjunctival squamous cell carcinoma Ali Mahdavi Fard, Leili Pourafkari, Nader D Nader

69-year-old farmer presented with upper and lower parts of cornea was pres- a growing non-tender lesion in his ent (Figure 1). Conjunctival congestion was A right eye. He first noticed the lesion present and corkscrew-like vascular loops two years earlier, but hadn’t sought medi- were evident (Figure 2). Cytopathological cal attention. He was previously a smoker, examination from incisional biopsy showed but his past medical history was otherwise squamous cell carcinoma. Magnetic reso- unremarkable. His visual acuity was 6/12 nance imaging with intravenous contrast on the right eye and 6/9 in the left. On enhancement did not show gross orbital examination, a raised fleshy, papillomatous involvement. Surgical excision and topical elevated growth involving interpalpebral mitomycin therapy was discussed with the fissure and extending beyond the limbus to patient, but he refused surgical treatment.

Figure 1: Slit-lamp examination of the right eye showing Figure 2: Slit-lamp examination of the lesion, corkscrew-like the large mass vessels are present

Competing interests: Nil Author information: Ali Mahdavi Fard, Assistant Professor of Ophthalmology, Tabriz University of Medical Sciences, Tabriz, Iran; Leili Pourafkari, Assistant Professor of Cardiology, Cardiovascular Research Center, Tabriz University of Medical Sciences, Tabriz, Iran; Nader D Nader, Professor of Anesthesiology, University at Buffalo, Buffalo, NY Corresponding author: Nader D Nader MD, PHD, FACC, Professor of Anesthesiology, University at Buffalo, Buffalo, NY [email protected] URL: www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2015/vol-128-no-1422-25- september-2015/6668

A case of perforated chronic idiopathic megacolon Benjamin Cribb, Rukshan Ranjan, Nigel Henderson

61-year-old male, with long-standing megacolon develops from refractory abdominal distension, presented constipation from any cause. Idiopathic A with generalised abdominal pain, megacolon is a rare and poorly under- shortness of breath, and marked abdominal stood condition. This condition has many distension. The patient reported increasing similarities with other conditions asso- abdominal distension over a 6-week period ciated with chronic megacolon, such as with associated infrequent bowel motions. Hirschsprung’s disease, chronic idiopathic No other significant medical or surgical intestinal pseudo-obstruction and idio- history of note. pathic megarectum. However, idiopathic X-ray and CT imaging showed extensive megacolon is unique in that the onset colonic distension and a large volume of symptoms can occur in early or late pneumoperitoneum (Figures 1 and 2). childhood or in adult life.1 The aetiology of There was also marked elevation of the this condition may be different between right hemidiaphragm with associated those patients with an early onset of volume loss of the right lung. symptoms and those with a late onset.1 He proceeded acutely to theatre. The symptoms of idiopathic megacolon 1 An attempt at pre-oxygenation prior also differ between these two groups. to induction of anaesthesia resulted Patients with an adult onset do not always in respiratory distress with oxygen complain of constipation, though it is the 1,2 desaturation. Needle decompression of most common symptom. Abdominal the abdomen was required to relieve the pain, abdominal distension, palpable tense pneumoperitoneum. At laparotomy, abdominal mass, irregular bowel habit the patient was found to have a massively and sometimes overflow diarrhoea are the distended colon (see Figure 3), and a other most common symptoms in patients large amount of free intra-peritoneal with an adult onset of idiopathic mega- gas without gross faecal contamination. colon.1,2, 3 Faecal impaction is uncommon.4 Colonic micro-perforation was suspected Colonic perforation is a rare complication and subtotal colectomy with end in association with idiopathic megacolon. ileostomy performed. Histopathological Only one other case could be identified analysis showed stercoral ulceration from the literature.5 Pathological features in the descending colon (the presumed of idiopathic megacolon are thickening area of micro-perforation). There was of the enteric smooth muscle, but with an no evidence of aganglionosis or other intact enteric nervous system.4 intrinsic neuromuscular abnormality. Functional abnormalities of the enteric The patient had a protracted post- nervous system, the autonomic nervous operative course due to respiratory system, enteric smooth muscle, or the failure, which necessitated tracheostomy interstitial cells of Cajal, may play a role in insertion and ventilatory assistance. This the pathophysiology of this condition.4 gradually improved and the patient made Conservative management of patients a full recovery. with idiopathic megacolon is frequently The causes of chronic megacolon can ineffective.1 For the majority of patients be divided into congenital, acquired and with an onset of symptoms later in life, idiopathic causes. In general, chronic medical treatment is unsuccessful.

Surgery is reserved for those who fail proctocolectomy with ileal pouch recon- conservative management or who develop struction is recommended.3,6 However, in a complication necessitating urgent the acute setting of perforation there are surgical intervention, such as perforation few surgical options available, other than or ischemia. The surgical options for idio- subtotal colectomy with end ileostomy pathic megacolon in the elective setting formation, as performed in this case. have previously been evaluated.3,6 Subtotal Although faecal impaction with ster- colectomy with ileorectal anastomosis coral perforation is not uncommon, appears to be the optimum procedure in this case report of perforated idiopathic patients with a non-dilated rectum.4,6 In megacolon is a rare complication of this patients with a dilated rectum, restorative interesting condition.

Figure 1: Axial computed tomography image of Figure 2: Coronal computed tomography image the abdomen demonstrating a massive pneumo- demonstrating the marked colonic distension with peritoneum with the abdominal viscera pushed associated elevation of the right hemidiaphragm posteriorly. Note the marked wasting of the and reduced right lung volume. anterior abdominal wall.

Figure 3: The resected specimen demonstrating the generalised distended colon.

Competing interests: Nil Author information: Benjamin Cribb, Department of general surgery, Taranaki Base Hospital, New Plymouth, New Zealand; Rukshan Ranjan, Department of general surgery, Taranaki Base Hospital, New Plymouth, New Zealand; Nigel Henderson, Department of general surgery, Taranaki Base Hospital, New Plymouth, New Zealand. Corresponding author: Nigel Henderson, Department of general surgery, Taranaki Base Hospital, New Plymouth, 123 Vivian St, New Plymouth 4310, New Zealand. [email protected] URL: www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2015/vol-128-no-1422-25- september-2015/6669

REFERENCES: 1. Barnes PR, Lennard-Jones Hawley PR, et al. Colec- caval obstruction and JE, Hawley PR, et al. tomy for idiopathic colonic perforation. Case Hirschspung’s disease and megarectum and megaco- report. Acta Chir Scand idiopathic megacolon in lon. Gut 1991; 32: 1538-40. 1988; 154(10): 605-7. adults and adolescents. 4. Gattuso JM, Kamm MA, 6. Gladman MA, Scott SM, Gut 1986; 27: 534-41. Talbot IC. Pathology of Lunniss PJ et al. Systematic 2. Gattuso JM, Kamm MA. idiopathic megarectum review of surgical options Clinical features of and megacolon. Gut idiopathic megarectum 1997; 41: 252-7. for idiopathic megarectum and megacolon. Annals of and idiopathic megacolon. 5. Scott MJ. Idiopathic Gut 1997; 41: 93-9. megacolon presenting Surgery 2005; 241: 562-74. 3. Stabile G, Kamm MA, with fatal inferior vena

Out of touch? The shortcoming of New Zealand’s amended Sale and Supply of Alcohol Act (2012) for the Rugby World Cup (2015) Benjamin Riordan, Tamlin Conner, Jayde Flett, Damian Scarf

ndy Ellis’s kick found touch in the to lay out in clear detail how wrong-footed 80th minute to win the 2011 Rugby the government’s game plan really is. AWorld Cup final. Four years on and In their submission, the New Zealand rather than kicking alcohol into touch, our Medical Association made it clear that government continues to pass legislation restricting trading hours is one of the few that keeps it in the field of play. Case in measures that reduces alcohol use and point is the recent amendment made to the the harms associated with it. Indeed, in a Sale and Supply of Alcohol Act (2012), that review of the literature, Hahn, et al (2010), will allow licenced premises to open and demonstrated that increasing trading times sell alcohol during the Rugby World Cup by just 2 hours led to a significant increase 2015 matches. Far from sneaking through, in alcohol-related harm.1 But, even small and despite a large number of submissions changes appear detrimental. For example, from individuals in the health sector (eg, in central Amsterdam, extending trading New Zealand Medical Association, Public hours from 4am to 5am led to a 34% Health Association of New Zealand, The increase in ambulance attendances for National Public Health Alcohol Working alcohol-related harm.2 A similar increase Group), and various other groups (eg, Alco- in Norway led to a 16% increase in violent hol Action Tai Tokerau, National Commu- crime.3 Conversely, and closer to home, nity Action on Youth and Drugs, Waitakere reducing trading hours in parts of Australia Anti Violence Essential Services, Women’s led to a decrease in assaults.4 Refuge), the bill passed with overwhelming support (99 For vs 21 Against). Specifically, To add to the concern, and in contrast to the bill allows bars, pubs, and licensed clubs the studies mentioned above, our govern- to open and sell alcohol an hour before ment’s relaxed trading hours coincide with the start of a match, during the match, and our country’s major sport event. While the for half an hour after the match ends. On government appears to think that this is nights with multiple matches played in the perfect time to relax trading hours, one succession, licensed premises will be able to could argue it is the perfect time to do the continue selling alcohol between matches, exact opposite. Unsurprisingly, events are provided that the second match starts less associated with an increase in alcohol use than an hour after the first. This will allow and alcohol-related harm. Although we licensed premeses to open well outside of are unaware of any research focusing on normal operating hours and may lead to an the increase in New Zealander’s drinking increase in both alcohol use and alcohol-re- during the World Cup, there was a marked lated harm. The aim of the current piece is increase in alcohol-related emergency

department presentations during the To conclude, the governments previous 2011 World Cup.5 In other event-related inaction with regard to alcohol reform was contexts in New Zealand, we have found concerning enough,8 the fact that their most that individual’s consume around 200% recent action may increase alcohol use and more alcohol during events relative to alcohol related harm, is nothing short of the amount they drink during a standard alarming. drinking session.6,7

Competing interests: Nil Author information: Benjamin Riordan, Department of Psychology, University of Otago, Dunedin; Tamlin Conner, Department of Psychology, University of Otago, Dunedin; Jayde Flett, Department of Psychology, University of Otago, Dunedin; Damian Scarf, Department of Psychology, University of Otago, Dunedin. Corresponding author: Dr Damian Scarf, Department of Psychology, University of Otago, P.O. Box 56, Dunedin, 9054 [email protected] URL: www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2015/vol-128-no-1422-25- september-2015/6670

REFERENCES: 1. Hahn RA, Kuzara JL, Elder hours on violence. The ecological momentary R, et al. Effectiveness Norwegian experience intervention to reduce of policies restricting from 18 cities. Addiction. university student hours of alcohol sales 2012;107(3):530-7. alcohol consumption. in preventing excessive 4. Kypri K, Jones C, McElduff J Stud Alcohol Drugs. alcohol consumption and P, Barker D. Effects of 2015;76(4):525-9. related harms. Am J Prev restricting pub closing 7. Riordan BC, Scarf D, Med. 2010;39(6):590-604. times on night-time Conner TS. Is Orienta- 2. de Goeij MC, Veldhuizen assaults in an Austra- tion Week a gateway to EM, Buster MC, Kunst lian city. Addiction. persistent alcohol use in AE. The impact of 2011;106(2):303-10. university students? A extended closing times preliminary investigation. of alcohol outlets on 5. Gardener M, Parke J Stud Alcohol Drugs. alcohol-related injuries T, Jones P. Impact of in the nightlife areas of the New Zealand 2011 2015;76(2):204-11. Amsterdam: a controlled Rugby World Cup on 8. Kypri K, Connor J, before-and-after an Urban Emergency Maclennan B, Sellman evaluation. Addiction. Department. NZMJ. 2015 D. What became of 2015;110(6):955-64. July; 128 (1418): 80-84. New Zealand’s golden 3. Rossow I, Norström T. 6. Riordan BC, Conner TS, opportunity for liquor The impact of small Flett JAM, Scarf D. A law reform? Drug Alcohol changes in bar closing brief Orientation Week Rev. 2013;32(6):557-60.

Accident Compensation Corporation claim status and benefit type is associated with low back pain outcomes Jon Cornwall, Achim Elfering, Rebecca J Crawford, Markus Melloh

ompensation schemes for injury and new episode of LBP was undertaken. The injury recovery are important, as study was approved by the Lower South Cthere is an association between com- Regional Ethics Committee (LRS/08/03/008). pensation-related factors and poorer health Patients attending primary care practitioners outcomes following injury.1 There are few were recruited across New Zealand, and previous data investigating outcomes of sent questionnaires at weeks zero, three, six low back pain sufferers in relation to the and twelve, then six months. Questionnaires support received from the Accident Com- were based on the Multinational Muscu- pensation Corporation (ACC), or social wel- loskeletal Inception Cohort Study statement fare system in New Zealand, and no studies addressing risk factors for the development specifically examine low back pain (LBP) of persistent LBP. outcomes by benefit type. Recent publica- Variables of interest included function tions have shown that some injury or illness (Oswestry Disability Index), pain (visual outcomes are worse where the ACC do not analogue scale), physical and mental 2,3,4 provide financial support, while another health (Physical and Mental Component reports no difference in outcomes between Scale Short Form 12 Health Survey ACC and non-ACC supported patients receiv- Questionnaire), fear-avoidance beliefs 5 ing lumbar spinal fusion surgery. For LBP (Fear- Avoidance Beliefs Questionnaire), and treated non-surgically, our previous study helplessness (pain catastrophising scale). reports a negative correlation between Patients were grouped into those having ACC claim status (accepted, or not) and a LBP ACC claim accepted, and those that benefit status (on a benefit, or not); poorer did not. In total, 124 ACC claim-accepted outcomes were shown for individuals patients, and 188 ACC claim-not- accepted receiving a benefit and without an accepted were included; 168 patients completed all 3 ACC claim. What was unclear is whether surveys. ACC-claim-not-accepted patients specific benefit type (sickness [SB], unem- were further grouped by benefit groups (on ployment [UB], invalids [IB], domestic pur- or not on benefit), including DPB (n=12), SB poses [DPB]) is predictive of outcome in LBP (n=11), UB (n=6), and IB (n=4). Mean time patients without an accepted ACC claim; on benefits (baseline) was 423 in DPB, 203 we therefore examined the relationship days in SB, 216 days in UB, and 304 in IB. between benefit type and LBP outcomes for Numbers were not adequate to allow signif- those without accepted ACC claims for LBP. icance testing between groups; drop-out Details on our methodology have been accounted for a reduction in benefit partic- published previously.6 In brief, a prospective ipants from 33 to 18 (55%) at three months, cohort study of patients presenting with a and to 13 over six months.

Despite the small group numbers, some those with fewer resources, like money trends in LBP outcomes were apparent or secure social frameworks.6 Without across the different time frames for those work as a stimulus, motivation may be low in different benefit classes. Specifically, to actively engage in seeking and facili- trends highlighted the performance of UB tating improvement. Further, there are who were either worse or unchanged for all many factors that influence recovery from measures at six months, while every other LBP, including management of resources group improved across most measures. UB such as social support,8,9 employment,8,9 were the only group to worsen over time and treatment;10 lack of work prospects for functional limitation, mental health, may also have contributed to UB patients pain, and helplessness; at 6 months they poorer performance. were unchanged in fear avoidance beliefs Even though study numbers were about work and physical activity, and were limited, the existence of trends between worse for physical health (with SB). The the different benefit groups points to a best results over six months were observed pressing need to examine LBP outcomes for DPB (the only ones to improve in FABQ in non-ACC supported individuals to more Physical Activity) while SB and IB improved closely determine modifiable risk factors in most assessed categories. for poor outcome in those individuals on A possible explanation for UB poor benefits. In particular, the UB category, performance compared to other benefit because of the trends observed suggesting groups may include a lack of motivation their performance is worse than other for improvement; previous studies have benefit groups. Further data are required indicated that work participation and to support these preliminary findings, and resource provision have positive effects to explore the relationship between LBP and are predictive of outcomes for LBP outcome and benefit type for those people recovery,7 with musculoskeletal disorders with and without accepted ACC claims for being more difficult to cope with for their injury.1

Competing interests: Nil Author information: Jon Cornwall, Senior Lecturer, Graduate School of Nursing, Midwifery and Health, Victoria University of Wellington; Achim Elfering, Professor, Department of Work and Organizational Psychology, Institute for Psychology, University of Bern, Switzerland; Rebecca J Crawford, Senior Research Fellow, Centre for Health Sciences, School of Health Professions, Zurich University of Applied Sciences, Switzerland; Markus Melloh, Professor, Centre for Health Sciences, School of Health Professions, Zurich University of Applied Sciences, Switzerland. Corresponding author: Jon Cornwall, Senior Lecturer, Graduate School of Nursing, Midwifery and Health, Victoria University of Wellington. [email protected] URL: www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2015/vol-128-no-1422-25- september-2015/6671

REFERENCES: 1. Murgatroyd DF, Casey cord injury and the role Research in Occupational PP, Cameron ID, Harris of no-fault compensation: Stress and Well Being: IA. The effect of financial longitudinal study. Spinal New Developments in compensation on health Cord 2013;51:919–25. Theoretical and Concep- outcomes following 5. Montgomery AS, Cunning- tual Approaches to Job musculoskeletal injury: ham JE, Robertson PA. Stress,Vol. 8, Emerald. Systematic review. PLOS The influence of no 8. van Tulder M, Becker A, One. 2015; DOI:10.1371/ fault compensation on Bekkering T. Chapter 3. journal.pone.0117597 functional outcomes after European Guidelines for 2. McAllister S, Derrett S, lumbar spine fusion. the Management of Acute Audas R, et al. Do differ- Spine. 2015;40(14):1140-7. Nonspecific Low Back ent types of financial 6. Melloh M, Aebli N, Elfering Pain in Primary Care. Eur support after illness or A, et al. Development of a Spine J. 2006;15:S169–91. injury affect socio-eco- screening tool predicting 9. Kendall NAS, Thompson nomic outcomes? A the transition from acute BE. A pilot program natural experiment in to chronic low back pain for dealing with the New Zealand. Soc Sci for patients in a GP setting: comorbidity of chronic Med. 2013;85:93–102. Protocol of a multinational 3. Melloh M, Cornwall J, prospective cohort study. pain and long-term Crawford RJ, Elfering BMC Musculoskelet unemployment. J Occup A. Does injury claim Disord. 2008;9:167. Rehabil. 1998;8:5–26. status and benefit status DOI:10.1186/1471- 10. Airaksinen O, Brox JI, predict low back pain 2474-9-167 Cedraschi C. Chapter 4. outcomes? Aust Med 7. Grebner S, Elfering A, European Guidelines J. 2015;8(8):268-276. Semmer, NK. 2010. The for the Management of 4. Paul C, Derrett S, McAllis- Success Resource Model of Chronic Nonspecific Low ter S, et al. Socioeconomic Job Stress. In, Perreweé, Back Pain. Eur Spine outcomes following spinal P.L. & Ganster, C.D. (Eds.), J. 2006;15:S192– 300.

Fluoxetine-induced phenytoin toxicity: a clinical reminder about the perils of polypharmacy Sean Lance, Ian Ternouth

dosing and drug interactions. Following Case report non-linear kinetics, its half-life is widely JS is a 62-year-old patient who was variable, dependent on dose, absorption, admitted to a secondary care hospital from saturation of metabolic pathways, and resthome care with progressive reduction in enzyme induction/inhibition.1 Metabolism her Glasgow Coma Score (GCS). She had been is via the liver microsomal cytochrome commenced on fluoxetine 20mg OD (once P450 (CYP450) enzymes. Consequently, a daily) 11 days prior by the GP for low mood. number of drugs interactions exist. Along Her other medical history includes with this, phenytoin is a potent CYP450 difficult-to-control epilepsy, for which inducer, specifically CYP3A4 and CYP2C19.1 she takes phenytoin 300mg OD, clobazam It is predominantly protein-bound, so when 20mg BD (twice daily), and levetiracetam saturated, small changes in doses produce 1g BD. Including these, she was taking large variations in levels of the free drug 19 different medications for a number of (contributing to toxicity).1 other medical problems. Fluoxetine is similar, in that it is largely Examination was consistent with protein-bound (95%), and metabolised by phenytoin toxicity, with a markedly elevated CYP450 enzymes (predominantly CYP2C19 plasma phenytoin level of 244 umol/L and CYP2D6). The active form (norfluoxetine) (reference range 40–80 umol/L). With no has a long half-life (approximately 9 days).2 other reasons for a supratherapeutic level, The first report of an interaction between it was concluded that the commencement phenytoin and fluoxetine was in 1992 of fluoxetine was the likely cause for the by Jalil, who described two patients who grossly elevated phenytoin level. developed phenytoin toxicity after having Fluoxetine and phenytoin were withheld, had fluoxetine added.3 In 1997, Schmider et and she showed slow improvement in al4 further detailed this with in vitro testing her neurological state over the following of liver tissue from six healthy volunteers. three weeks. This correlated with a gradual This revealed the interaction was via inhi- decline in her plasma phenytoin levels bition of CYP2C9-mediated metabolism of over this period. She was subsequently phenytoin.4 discharged at her baseline, with the plan to Spina and Perucca6 outlined the clinically restart phenytoin alone, once the levels were relevant pharmacokinetic interactions of back within the normal range. SSRIs, citing the aforementioned research We discuss the interaction between these regarding the interaction between fluox- two commonly prescribed medications, etine and phenytoin via inhibition of along with wider discussion regarding CYP2C9.5 They also noted that the clinical polypharmacy. consequences are compounded by the long half-life of fluoxetine.5 Discussion This case also highlights the dangers of Phenytoin’s unique pharmacokinetic polypharmacy—something that has been prolife leads to problems associated with previously well demonstrated in elderly

and residential care patients, where along with drug monitoring when polypharmacy results in higher numbers of prescribing new medications. It also serves adverse reactions and complications.6,7 to highlight the caution needed when This case demonstrates the vast potential prescribing in patients already on multiple for drug interactions, and highlights the medications and should serve as a clinical need for consideration of drug choice, reminder of such.

Competing interests: Nil Author information: Sean Lance, Medicine, Taranaki Base Hospital, New Plymouth; Ian Ternouth, Cardiology, Taranaki Base Hospital, New Plymouth Corresponding author: Sean Lance, Medicine, Taranaki Base Hospital, New Plymouth. [email protected] URL: www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2015/vol-128-no-1422-25- september-2015/6672

REFERENCES: 1. Medsafe [Internet]. April 1 [cited 2015 June itors in vitro: Studies of Wellington: New Zealand 25]; [about 11 screens]. phenytoin p-hydroxyla- Medicines and Medical Available from: http:// tion. Br J Clin Pharmacol. Devices Safety Authority. www.medsafe.govt. Dec 1997; 44(5): 485-498 Datasheet: Dilantin; 2013 nz/profs/Datasheet/f/ 5. Spina E, Perucca E. Epilep- May 29 [cited 2015 June fluoxcaptab.pdf sia. Feb 2002; 43(2): 37-44 25]; [about 18 screens]. 3. Jalil P. Toxic reaction 6. Field TS, et al. Arch Intern Available from: http:// following the combine Med. 2001; 161(13): 1629 www.medsafe.govt. administration of fluoxe- 7. Lu WH, et al. Effect of nz/profs/Datasheet/d/ tine and phenytoin: two polypharmacy, poten- Dilantincapsusptab.pdf. case reports. J Neurol tially inappropriate 2. Medsafe [Internet]. Neurosurg Psychiatry. medications and anticho- Wellington: New Zealand May 1992; 55(5): 412-413 linergic burden on clinical Medicines and Medical 4. Schmider J, et al. Inhibi- outcome: a retrospective Devices Safety Authority. tion of CYP2C9 by selective cohort study. CMAJ. Mar Datasheet: Fluox; 2014 serotonin reuptake inhib- 2015; 187(4): 130-137

New Zealand’s growing thirst for a sugar-sweetened beverage tax Gerhard Sundborn, Simon Thornley, Bodo Lang, Rob Beaglehole

indings from two large-scale, na- The speed at which public opinion has tion-wide surveys indicate that the shifted (in favour of a SSB tax) indicates that Fmajority of New Zealanders are now New Zealanders are increasingly aware of supportive of a tax on sugar sweetened the harms SSBs pose to health, and that of beverages. children especially. Strong media attention A significant shift has occurred in around SSBs may have facilitated this New Zealanders’ appetite for a tax on change. Over the last two years, the harm sugar-sweetened beverages (SSBs), if the that SSB intake poses to health, and the funds collected are to be used to prevent notion of a tax on SSBs, have been regularly 3-7 childhood obesity. profiled in mainstream media. The public support for an SSB tax is also Two polls, 18 months apart, show a strong echoed by public policy makers in recent increase in support of a tax on sugary research. A study conducted by Signal et al,8 drinks. The first poll, from February, 2014, interviewed key health policy and decisions found that 44% of respondents supported makers, including politicians, bureaucrats, a tax on sugary drinks.1 The second poll, food industry and key stakeholders about carried out in June 2015 (funded by the the acceptability and feasibility of a number National Heart Foundation and the Cancer of fiscal policies aimed at improving Society of Auckland), showed that support nutrition. Out of six policy options (that had risen to 52%, provided funds be used to included either a fat tax; salt tax; removal of address childhood obesity2—this represents GST from fruit and vegetables; a combined an eighteen percent rise in favour of a tax. fat, salt, GST removal type policy; and an Interestingly, there was an even stronger SSB tax), an SSB tax was most frequently drop in those who opposed such a tax. selected by respondents as the most feasible Opposition to a SSB tax dropped by 35% and acceptable option.8 from 49% in February 2014 to 32% in June There is also strong political support 2015 (Table 1). for the need to address SSBs by political These findings indicate a significant shift parties outside of government. A policy in public attitude towards the taxation of brief, written by the New Zealand Beverage SSBs, because a quarter of respondents Guidance Panel titled Options to reduce have moved to a more supportive, or less sugar sweetened beverage consumption in opposed, stance about the introduction of New Zealand, outlined 20 initiatives relating an SSB tax. to how SSB intake could be reduced, with a

Table 1: Horizon Research polls relating to public opinion for a sugar-sweetened beverage tax, comparing results obtained in 2014 and 2015.

2014 2015

Revenue use not Revenue used for childhood mentioned obesity prevention Support SSB tax 44% 48% 52%

Oppose SSB tax 49% 35% 32%

20% excise tax included. This document was Zealanders, particularly those who are most received on parliament grounds on 19 June, vulnerable. 2014, and formally endorsed by the Green, The increasing public dissatisfaction 9 Labour, and Māori parties. from greater awareness of the effect of A refreshing initiative that Minister SSBs on our children’s health is shared by Coleman recently introduced is the policy makers and several political parties. mandate that hospitals and district health This makes the political acceptability of boards (DHBs) lead public policy to restrict introducing an SSB tax a likely reality in SSB intake. A blanket ban on sugary drinks the future. in all hospitals and DHBs has recently The introduction of an SSB tax would be been announced under the leadership of another positive step toward addressing the Director-General of Health and Chief childhood obesity, making a strong Executive, Chai Chuah. We congratulate Mr statement that New Zealanders, as a society, Chuah for this bold initiative and Minister value child health over corporate profits. Coleman for creating an environment in which such decisions can be made. The Whether the current Minister of Health will exclusion of sugary drinks from hospitals extend his legacy to include the introduction will be a positive legacy for these leaders.10 of a SSB tax, or not, remains to be seen. It seems inevitable, however, that an SSB tax However, SSBs present a problem that will be a major part of reclaiming our chil- requires a more urgent and larger-scale dren’s health, considering the growing public solution. New Zealand has the third highest support for its implementation. The only rate of childhood obesity in the developed question that remains is when. world, and a recent study has found that high sugary drink intake is conservatively This paper has been prepared by FIZZ attributed to 561 deaths in Australasia per (Fighting Sugar in Soft-drinks) New Zealand; year.11,12 Equivalent to 40% of the annual a public health advocacy group established road toll. Thus, New Zealand urgently by researchers to reduce the consumption needs a policy that will address such health of sugar sweetened beverages in New issues in a manner that benefits all New Zealand to zero by 2025.

Competing interests: Nil Author information: Gerhard Sundborn, Epidemiology and Biostatistics, University of Auckland; Simon James Thornley, Consulting Epidemiologist, Auckland; Bodo Lang, Department of Marketing, Business School, University of Auckland; Rob Beaglehole, Principal Dental Officer, Nelson Marlborough District Health Board. Corresponding author: Dr Gerhard Sundborn, Epidemiology and Biostatistics, University of Auckland, Auckland, New Zealand [email protected] URL: www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2015/vol-128-no-1422-25- september-2015/6673

REFERENCES: 1. Horizon Research. Sugar-added drinks. 10 reasons to skip fizzy Perceived harm from June 2015. drinks. Casey Seidenberg. sugar use and policy 3. New Zealand Herald. Thursday Sep 10, 2015. options. February, 2014. Martin Johnston. NZ http://www.nzherald. https://www.horizonpoll. scientists take on sugar. co.nz/lifestyle/news/ co.nz/attachments/docs/ Feb 8, 2014. http://www. article.cfm?c_id=6&o- horizon-research-sugar- nzherald.co.nz/lifestyle/ bjectid=11510857 use-and-harm- questionna. news/article.cfm?c_ 5. New Zealand Herald. pdf id=6&objectid=11198199 Niki Bezzant: Bring on 2. Horizon Research. 4. New Zealand Herald. the sugar tax. Jul 5, 2015.

http://www.nzherald. Conference, Auckland nzdoctor.co.nz/news/2015/ co.nz/lifestyle/news/ 2015. http://www.ana. september-2015/11/ article.cfm?c_id=6&o- org.nz/news-and-events/ ministry-sets-deadline- bjectid=11475877 events/anas-6th-nation- for- dhb-sugar-ban.aspx al-nutrition-and- physi- 6. 3 NEWS. May 3, 2014. 11. OECD, OECD Obesity cal-activity-conference National News - The Update 2014. 2014. 9. New Zealand Beverage War on Sugary Drinks. www.oecd.org/health/ Guidance Panel. 2014. Lucy Warhurst. obesity- update.htm Policy brief: Options to 7. Radio Live Drive. reduce sugar sweetened 12. Singh G, Micha R, November 14, 2014. beverage (SSB) consump- Khatibzadeh S, Lim S, Duncan Garner. Fizzy tion in New Zealand. Ezzati M, Mozaffarian Drinks and Health. Pacific Health Dialog: D. Estimated Global, 8. Signal L, Watts C, Murphy Vol 20 Number 1. Regional, and National C, Ni Mhurchu C, Eyles H. 10. Ministry sets deadline Disease Burdens Related to Fat Tax and All That: the for DHB sugar ban. Sugar-Sweetened Beverage pros and cons of taxing Cliff Taylor. September Consumption in 2010. food. Presentation at ANA 11, 2015. http://www. Circulation. June 29, 2015.

Is our focus on pharmaceutical company influence too narrow? Lance Gravatt

he 4 September issue of the Journal This type of direct financial benefit is reports the findings of accuracy and relatively obvious. However, there is a less Tbias in pharmaceutical industry ad- obvious, indirect benefit that may be just as vertising in New Zealand, and is accompa- influential on the medical profession. In the nied by an editorial by Toop and Mangin.1 period 1990–1996, US physicians held 5,051 We would like to see the suggested medical-device patents, which represented transparency of the commercial rela- nearly 20% of all medical device patents tionship by the medical profession and granted during this period. More than 60% the pharmaceutical industry extended to of the doctor-inventors were practicing 6 include the medical-device industry. This healthcare practitioners. The authors does however beg the question is there of the research presented evidence that any evidence to suggest that we should doctor-inventors demonstrated bias in their be concerned about any commercial rela- publications. tionship between the medical profession New Zealand doctors are encouraged to and medical-device companies. practice an arm’s length, transparent, rela- In 2014, the US federal government tionship with the pharmaceutical industry, reported that drug and medical-device but should we be broadening our approach makers gave nearly USD $6.5 billion to US to include the medical-device industry too? doctors and teaching hospitals. What is not We do not have to look too far for reason to made clear is how much of this was paid by be potentially concerned. While transparent, medical-device companies.2 it is surprising to see that the University of In 2006, a renowned spinal surgeon Auckland School of Medicine Foundation appeared before the US Senate Committee Trustees include the CEO for the Medical seeking funds for research into a bone- Technology Association of New Zealand, graft device made by Medtronic, which which is the peak industry body for the was granted. What the surgeon did not medical device companies in New Zealand. disclose was that his trip to Washington was I ask myself how comfortable the medical paid by Medtronic, and he had been on the profession would feel if the CEO for Medicines Medtronic payroll for several years—to the New Zealand was a member of the HRC? tune of USD $1.14 million. Medtronic would New Zealand is one of the least corrupt directly benefit from the funding.3 countries in the world, and let’s work to In 2007, five medical-device companies keep it that way. Making one group the (dePuy Orthopaedics, Zimmer, Biomet, scapegoat for the ills of a profession Stryker and Smith & Nephew) entered into is not constructive. The bullying and a USD $311 million settlement with the sexual harassment reported by our junior Department of Justice to settle allegations of doctors is not industry-made and yet needs violating the law prohibiting kick-backs and attention. Let us all expect integrity and inducement agreements with orthopaedic transparency, not just from the pharma- surgeons.4 ceutical industry, which, as it happens, is a mere remnant in New Zealand. In 2015, NuVasive Inc settled a false claims case regarding certain spinal surgery devices with a USD $13.5 million federal payment.5

Competing interests: Nil Author information: Lance Gravatt, Chairman, Te Arai BioFarma Ltd Corresponding author: Lance Gravatt, Chairman, Te Arai BioFarma Ltd [email protected] URL: www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2015/vol-128-no-1422-25- september-2015/6674

REFERENCES: 1. Ma A, Parkin L. Makers Paid $6.49 Billion Com. November 14 2007. Randomised controlled to Doctors, Hospitals in 6. Department of Justice. trials cited in pharma- 2014. Wall Street Journal. Medical Device Manu- ceutical advertisements 3. Surgeons on Big Pharma facturer NuVasive Inc targeting New Zealand Payroll. Drug Watch, to Pay $13.5 Million to health professionals: Do Settle False Claims Act January 18, 2012. they support the adver- Allegations. July 30 2015. 4. Pringle E. Feds Crack tising claims and what 7. Chatterji A K et al. Physi- Down on Medical Device is the risk of bias? NZ cian-Industry Cooperation Implant for Profit Med J. 2015: 128 (1421). in the Medical Device 2. Loftus P, Walker J. Drug Industry- Part II. Industry. Health Affairs. and Medical-Device 5. Lawyers and Settlements. 2008: 27 (6); 1532-43.

Comment on: Getting serious about protecting New Zealand children against unhealthy food marketing1 Katherine Rich

ear Editor, parents in educating children to have a This Viewpoint suggests that an balanced diet and be healthy individuals; Devaluation of the “degree of exposure advertisements for treat food, snacks or of New Zealand children to unhealthy food fast food should not encourage children to marketing” is required. We welcome such consume them in excess; the quantity of the an evaluation, but suggest the Viewpoint is food depicted in the advertisement should seriously outdated. The current situation, not exceed serving sizes that would be which has been extant for five years and is appropriate for consumption by a person set to continue into the future, is as follows. or persons of the age depicted; benefits of foods for a nutritious diet should not Advertising in all media (print, radio, be exaggerated and should not imply that television, cinema, and websites) in New a single food should replace a healthy Zealand is subject to oversight by the diet, nor undermine the importance of Advertising Standards Authority (ASA). consuming a variety of food; children This is the agency that provides industry should not be urged in advertisements self-regulation of advertising for New to ask their parents, guardians or care- Zealand. It is unique in world terms in that givers to buy particular products for them; its membership comprehensively includes and advertisements soliciting responses all major New Zealand advertising, media incurring a charge should state, “Children and broadcaster organisations. The ASA ask your parents first” or similar words. develops codes of practice and main- In this context, all major publishers tains compliance with the codes through and broadcasters require advertise- voluntary commitments and a public ments to show evidence of assessment complaints process. for compliance with the Codes and will In 2010, the ASA published the Children’s publish only in line with the Codes. Code for Advertising Food 20102 as a In order to assess compliance, the companion to the Code for Advertising Food3 ASA invites complaints from the public and the Code for Advertising to Children4. (consumers, competitors, representative There is an explicit recognition in the organisations etc), assesses the complaints Children’s Code for Advertising Food of the and publishes its determinations. ASA need to protect children, of the need to decisions are available in a searchable support the food and nutrition policies of database on the ASA website. There Government and the Ministry of Health Food have been 9 complaints5 about adver- and Nutrition Guidelines and the need to tising food to children in the past 5 years, protect the health and wellbeing of children. representing around 0.2% of complaints A selection of requirements in the codes received over that period (there are reflects the measures applied: advertise- around fifty complaints a year received ments should not undermine the role of by the ASA relating to food and beverages

of around 700–800 complaints annually). and endorsed by joint Australian and The grounds for complaint about adver- New Zealand Governments, rates over 30 tising food to children covered sexism different breakfast cereals from Kellogg’s, (boys only), nutrition related, disturbing Sanitarium and Hubbard’s with 4.5 stars to children, timing of advertisement and (out of 5 stars), a number that increases trade name. None were upheld. to 70 different cereals rating 4 stars or Television watching in New Zealand for more. As well, a number of products from children aged 5–14 years is decreasing, a range of fast food outlets rank with 4 or evidenced by reporting by the Ministry of more stars. Health.6 The Ministry reports that while What is disappointing is that voluntary nearly half of children aged 5–14 years agreements, such as those signed by (53 percent) usually watched two or more major soft drinks companies not to sell hours of television a day in 2011/12, this into schools, are being circumvented was down from 57 percent in 2006/07. The by ‘third parties’ and the schools them- measures applied by Free-to-Air Television selves for continuing to accept soft drinks broadcasters’ in their policies and voluntary for sale. Support from Government for rules7 include no advertising in specific pre- healthy eating programmes in schools, and school television programming times and greater promotion of the Heart Founda- limited advertising in school-age children’s tion’s Fuelled4Life programme, would be television programming times. The times welcomed by industry. variously cover 0600–0950 and 1400–1700. Overall, this Viewpoint creates the This, and several developments in the past false impression that self-regulation has two decades, render research such as by failed. In our view, self-regulation is very 8 Wilson, 1999, as of historical interest only. effective and increasingly so in terms of Work by the same authors and others in driving measures, such as limited food 9 2014 show that 67% of selected packaged advertising to children, reformulating food manufacturers and soft drink manu- foods and conveying the healthfulness of facturers in New Zealand had a publicly foods to consumers. available policy related to marketing to children on their company website. Yours sincerely, It is of particular interest that the recently adopted Health Star Rating Katherine Rich system (HSR), which has been developed

Competing interests: Nil Author information: Katherine Rich, Executive, New Zealand Food & Grocery Council Corresponding author: Katherine Rich, Executive, New Zealand Food & Grocery Council [email protected] URL: www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2015/vol-128-no-1422-25- september-2015/6675

REFERENCES: 1. Vandevijvere S, Swinburn 13/303, 12/442, 11/269, children-an-tv/ 6 July 2015 B. Viewpoint: Getting 11/161, 11/022, 10/704 8. Wilson N, Quigley R, Serious about protecting 6. Activity Levels in New Mansoor O. Food ads New Zealand children Zealand. Ministry of on TV: a health hazard against unhealthy food Health. Viewed at http:// for children? Aus N Z marketing. NZ Med J 2015 www.health.govt.nz/ J Public Health, 1999 Jul 3; 128(1417):38-40. your-health/healthy-living/ Dec; 23(6): 647–50. 2. Children’s Code for food-and-physical-ac- doi: 10.1111/j.1467- Advertising Food 2010. tivity/physical- activity/ 842X.1999.tb01553.x Advertising Standards activity-levels-new-zea- 9. 9 Sacks G, Mialon M, Authority, 2010. land 6 July 2015. Vandervijvere S, Trevena 3. Code for Advertising 7. The Children’s Television H, Snowdon W, Crino M, Food. Advertising Policies: Advertising in Swinburn B. Compar- Standards Authority, n.d. Pre-School and School- ison of food policies Age Children’s Television and commitments on 4. Code for Advertising to Programming Times. marketing to children and Children. Advertising ThinkTV. Accessed at product (re)formulation Standards Authority, n.d. http://www.thinktv. in Australia, New Zealand 5. ASA Decision No’s co.nz/thinktv/stan- and Fiji, Critical Public 14/507, 14 253, 13/485, dards-and-regulations/ Health, Aug 2014. doi

Sir Patrick William Eisdell Moore (17 March 1918—18 June 2015) Otolaryngologist, Head and Neck Surgeon, Kt., OBE, MB. ChB, DLO (Eng), FRCS, FRACS

atrick Moore was born in Bristol on 17 consulting rooms as a general surgeon in March 1918. His father, Arthur Eisdell Symonds Street, Auckland. This was where PMoore (“Eisdell”), was in England on a Pat’s lifelong interest in, and love for, horses post-War surgical appointment. Eisdell had was first kindled. He was fascinated by the met Pat’s mother, Alice, a nurse from York- variety of draught horses operated by the shire, in 1915 when Eisdell was serving as a local merchants who lived in the neigh- field surgeon with the RAMC on the Western bourhood. It was also a time when the Front. Patrick was the first child of three— first stirrings of artistic talent took form, two sons and a daughter—and although as he drew and sketched on any blank or christened William Ernest Moore, the infant receptive surface—often to the chagrin of had been nicknamed “Pat” while still in his parents. utero in anticipation he (or she) would be Pat commenced his secondary education born on St Patrick’s Day. The name stuck at Auckland Grammar as an 11-year-old. when that prediction proved correct. It was While his love of literature and the classics not until he reached the age of 21 that he led him to top the country in English in formalised his adopted name by deed pole, his matriculation year, he struggled with and also changed Ernest to Eisdell. mathematics. After a year at Auckland After the war, Eisdell returned to New University College, Pat continued his Zealand with his new family and set up studies at Otago University. For his first

four years in Dunedin, Pat was a resident Pat’s surgical ambitions. On leaving the at Selwyn College, and was elected the Battalion at the cessation of hostilities, he house presidency for the last of these was farewelled by the Commanding Officer years. As a medical student, he played on with the words: “You were not the most the wing of the university senior rugby academic tākuta (doctor) we had. You may team. He supplemented his meagre student not necessarily have been the most brave, allowance by selling his sketches, cartoons, but you were, definitely, the most Māori.” caricatures, short stories and poems to He was subsequently made Patron, and a various publications, including Punch. At life member of the 28th Battalion, a recog- student parties, his musical talents as a nition which meant the world to him. pianist were in great demand, although In 1946, he returned to Auckland Hospital. he did occasionally lament his popularity A 3-month rotation in eye and ENT surgery on the keys by reason of the restrictions kindled an interest which, in 1947, led him to it necessarily imposed on his ability to become an eye and ENT registrar. During this socialise more widely. year, he was greatly influenced by the coun- After leaving Selwyn College, Pat moved try’s foremost ENT consultant, James Hardie to a boarding house in Cargill Street, where Neil, and Pat decided on a career in ENT. he met fellow resident Beth Beedie—a final In 1948, after a year working as an year physiotherapy student from a medical anatomy demonstrator in Dunedin, Pat, family in Dannevirke. After qualifying, she Beth, and their two sons, sailed to London, was posted to Hawke’s Bay and Pat moved to where he spent two years working and Auckland for his final year of medical studies. demonstrating at the Middlesex, training Their courtship flourished, albeit remotely. at the Royal National Throat, Nose and Pat graduated in 1943, and he and Beth Ear Hospital in Grey’s Inn and studying married, commencing life together in a at the College of Surgeons. He passed his small flat near Auckland Hospital. Pat primary and, at his first attempt, his English worked as a house surgeon to obtain full fellowship examination and subsequently registration and eligibility to re-enlist in the DLO. He then obtained a very busy ENT the army. As a medical student, he had registrar post in Northampton for two years, been commissioned in the Otago University under ex-patriot Australian, Charles Gledhill. Medical Corps, and in his final year worked In 1952, Pat and the family returned to as a resident army medical officer in the Auckland where he set up rooms as an Auckland region. After obtaining full regis- ENT consultant in Symonds Street, and the tration, he wasted no time in enlisting with family was expanded by the addition of two the 2nd NZEF and was posted overseas, more sons. Pat was appointed junior ENT leaving Beth and their infant son, Anthony. surgeon at the recently opened department Once overseas, he single-mindedly set at Greenlane Hospital. He rapidly rose about joining the 28th (Māori ) Battalion, to head of department. Under his lead- with whom he served throughout the ership and innovation, the department Italian campaign, rising to the rank of grew quickly. Facilities expanded as did the Captain. Tall, freckled and red-haired, number of staff, to include a team of GPs and he was the only Pākehā in the Battalion. specialists in related disciplines, including Culturally immersed, he became fluent in allergy, oral surgery and voice therapy. He Te Reo and Tikanga Māori, making lifelong pioneered the use of microsurgery in New friendships with his comrades and devel- Zealand, encouraged the innovative use of oping a sophisticated understanding, even homo- grafts and was the first in the world at that early time, of how a bicultural New to transplant appropriately prepared and Zealand should look. He actively applied sterilised tympanic membranes, initially in those principles of biculturalism throughout animals and later in human beings. Building the rest of his life. on this research, he established the Deafness During the Battle of Faenza he was badly Research Foundation in 1962, an organi- wounded. His right arm was saved from sation specifically created to assist clinicians amputation at considerable risk to his life, undertaking research into hearing and deaf- only because the surgeon was aware of ness-related problems.

Between 1965 and 1977, his continuing New Zealand: initially for adults and later engagement with Māori led him to make for children. He was quick to acknowledge regular voluntary visits to the East Coast the need for specialist auditory verbal of the North Island, where ear disease training for implanted children and and resulting hearing loss was endemic, persuaded philanthropists and friends to particularly amongst children. Through his support the establishment of the now highly energies, he sourced sophisticated surgical successful Hearing House. equipment and instructed local doctors Pat’s interests were not simply limited and nurses in its use. These efforts were to deafness and hearing. He served on the rewarded by a remarkable reduction in the Council of Auckland University, helped incidence of ear disease on the East Coast. establish Riding for the Disabled and the Encouraged by these results and conscious Coeliac Society. He was Master of the of the common and understandable reluc- Pakuranga Hunt for nine years and Pres- tance on the part of many parents to take ident and Emeritus Member of the New their children to hospital, Pat’s natural flair Zealand Hunts Association. for innovation led him to raise funds for the development of mobile ear clinics, which His mastery of prose, verse and sketching took clinical services into the community has left an enduring literal and pictorial under the banner of the Deafness Research record in the annals of the many institu- Foundation. Following success from this tions with which he has been involved. initiative in Northland, it was adopted in Perhaps the best known is his brilliant water Auckland and then, through the gener- colour caricature; a montage of the 1940 osity of the Variety Club, a fleet of mobile professors of the Otago Medical School. ear clinics allowed similar services to be This remarkable drawing, which has been extended across many other regions. reproduced in numerous publications, now hangs outside the Dean’s office. In 2004, his Pat also encouraged research into the autobiography was published. The title, So pharmacological treatment of tinnitus, Old, So Quick, was coined from a quote by the inclusion of ear and hearing problems Ogden Nash and the book’s flowing literacy in the Dunedin multi-disciplinary study, and he investigated the effects of hearing style, humour and self-deprecation earned it loss on prison inmates. He led the first universally positive reviews. ENT teaching programme in the Auckland Pat’s vision, enthusiasm and selfless Medical School and supported the estab- contribution to medicine and the wider lishment of a dedicated tertiary degree community was recognised by the Queen in course for the training of audiologists. the 1982 Royal Honours, with the award of As President of the then Otolaryngological on OBE, and in 1992 he was knighted for his Society of New Zealand, Pat organised the services to medicine and the community. first joint conference with the Australian Auckland Grammar honoured him in 2005, society, with in excess of 100 attendees from with an Augusta Fellowship as a distin- either side of the Tasman. This inaugural guished old boy. In 2007, Selwyn College meeting was New Zealand’s first ENT inter- elected him an Honorary Fellow. national conference. Pat’s funeral, in a packed St Mary’s-in- By the mid-1970s, Pat appreciated that Trinity, was a moving and evocative tribute hearing research was moving its focus from to an extraordinary New Zealander. His the external and middle ear to the inner plain coffin, draped in a New Zealand flag, ear. At the same time he realised the activ- was adorned by a magnificent korowai ities of the Deafness Research Foundation (feathered clock) woven by the widows of required a higher and more sustained the Battalion, in its colours of red, black and level of scientific input and engaged a post- white. To the haunting strains of a karanga graduate scientist, Peter Thorne, to build (call of welcome) he was carried in by a research team. Pat also monitored the representatives of the Battalion, various iwi evolution of cochlear implants and, once a and representatives of longstanding friends. multi-channel device had been perfected, At the end of the service, a rousing haka- persuaded benefactors to support the intro- performed by the Auckland Grammar kapa duction of a cochlear implant programme in haka group—paid a final farewell.

Sir Patrick is survived by his beloved wife Court Judge), Chris (a leading commercial Beth, their four sons, Anthony (a pathologist property lawyer), 12 grandchildren and practising in Australia), Tim (a radiologist eight great-grandchildren. practising in North America), Simon (a High

Author information: Ron Goodey FRACS, Simon Moore, Chris Moore URL: www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2015/vol-128-no-1422-25- september-2015/6676

Magnetic resonance and the prediction of dementia Dementia risk prediction has conventionally been based on sociodemographic, neuropsy- chological, health, lifestyle, physical function, and genetic variables. This report concerns a population-based cohort study which considers whether magnetic resonance imaging (MRI) would improve prediction of dementia. The study involved 1,721 men and women without dementia who underwent MRI at baseline and had known dementia status over 10 years of follow-up. During follow-up, 119 (6.9%) of the participants progressed to dementia. The researcher’s conclusion was that MRI did not improve the prediction of dementia. In an editorial on the topic, it is noted that as dementia cannot be reversed, stopped or slowed down, its early diagnosis may not be helpful. BMJ 2015;350:h2863 and BMJ 2015: 350:h2994

Ezetimibe added to statin therapy after acute coronary syndromes Statin therapy reduces low-density lipoprotein (LDL) cholesterol levels and the risk of cardiovascular events, but whether the addition of ezetimibe, a nonstatin drug that reduces intestinal cholesterol absorption, can reduce the rate of cardiovascular events further is not known. This randomised trial was designed to elucidate the issue. Over 18,000 patients who had been hospitalised with an acute coronary syndrome within the last 10 days were involved. They were randomised to receive simvastatin 40mg and ezetimibe 10mg or simvastatin alone. The primary end point was a composite of several unfavourable cardiovascular events or non-fatal stroke. At a median follow-up of 6 years, the LDL cholesterol levels were significantly lower in the combination group. There was also a significant lowering of cardiovascular events in the combination group, although it was modest with an absolute risk difference of 2%. Adverse events causing discontinuation of treatment occurred in approximately 10% in each group. N Engl J Med 2015;372:2387-97

Risk of serious infection in biological treatment of patients with rheumatoid arthritis Biological drugs are a new class of disease-modifying treatment options for rheumatoid arthritis that have been reported to show large clinical and radiographic improvements compared with traditional disease-modifying antirheumatic drugs (DMARDs), such as methotrexate. Such drugs include the Tumour Necrosis Factor Inhibitors etanercept, infliximab and adalimumab. Although they may be useful in the management of rheumatoid arthritis, there is concern that they may be associated with an increased risk of serious infection compared with DMARDs. This review of the topic includes data from 106 trials which involved the use of the biological treatments and compared the rates of serious infections between their use and the use of traditional DMARDs. The conclusions were that standard-dose and high-dose biological drugs (with or without traditional DMARDs) are associated with an increase in serious infection in rheumatoid arthritis compared with traditional DMARDs, although low-dose biological drugs are not. Clinicians should discuss the balance between benefit and harm with the individual patient before starting biological treatment for rheumatoid arthritis. Lancet 2015;386:258-65

URL: www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2015/vol-128-no-1422-25- september-2015/6677

Measles in adults Farquhar Matheson, M.R, Glasgow, House Surgeon, Wellington Hospital

HIS MEASLY JOKE. Mr Pip: Here’s a letter from our Jack,jwith the Expeditionary Force am sorry to say, dear, he has been attacked by German— Mrs Pip: Horrible! Mr Pip; Measles, (Observer, 07 November 1914). Alexander Turnbull Library, Wellington, New Zealand. http://natlib.govt.nz/records/27580652

easles is so frequently a disease of childhood that it is not often one sees many cases in adults. Since the commencement of the war, a number of adult cases have been Madmitted to Wellington Hospital, most of them being members of the Expeditionary Forces. An examination of these cases shows that the course of the disease is essentially the same in adults as in children. The rash stage was usually attended with a high temperature, readings of 104 to 105 degrees F. being common: This often terminated by crisis, the pulse rate, too, showing a remarkable fall. During the febrile stage epistaxis was a very common symptom, fairly persistent in some cases, but never alarming, and usually terminating without treatment. There was one case with gastro-intestinal symptoms lasting for 24 hours. As one would expect, complications affecting the respiratory tract were most frequently met with. Practically every case had bronchitis and laryngitis to a varying degree. There was one death from pneumonia. In this case, we had not merely the ordinary lobar type due to the pneumococcus; but there was a secondary infection, probably streptococcic, which made the case hopeless from the beginning. Another case developed pneumonia during convalescence, but he ran a normal course with a crisis on the eighth day. Apart from these, there were no cases which caused one any anxiety. With regard to the laryngitis and bronchitis, in many cases these were very distressing: coughing being incessant. We depended on the hospital linctus, containing tr. camp. co.,

squills and tolu, and on glyco heroin to control the coughing, but the results were not as satisfactory as one could wish. An experiment was made with guiacol carbonate gr. v. doses t.d.s., six patients being put on this treatment when admitted. The results were most gratifying; cough and laryngitis were markedly less in these cases. One hesitates to draw conclusions from such a small number of cases, but taking everything into consideration, guiacol seems to have been of real benefit in those cases in which it was tried. Further experience with this only seems to bear out its value in such cases.

NZMJ APRIL 1915

URL: www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2015/vol-128-no-1422-25- september-2015/6678

Erratum 128:1421

n an earlier version of the paper, Randomised controlled trials cited in pharmaceutical advertisements targeting New Zealand health professionals: do they support the Iadvertising claims and what is the risk of bias? by Alison Ma and Lianne Parkin (NZ Med J 128:1421; 22-29), a publication error occurred in the competing interests section, which read: “Richard Milne reports personal fees from Fisher Paykel Healthcare during the conduct of the study and personal fees from Fisher Paykel Healthcare outside the submitted work.” Richard Milne was not involved in the production of that paper, and an editorial oversight occurred. This has been remedied.