Journal of the New Zealand Medical Association Vol 132 | No 1490 | 22 February 2019

Smoke in our eyes: the Sense Partners’ evaluation of the legalisation of cannabis in New Zealand

Multimorbidity and multiple social disadvantage in a New Zealand high-needs free primary healthcare clinic population

Electronic whiteboards A response to: The ‘elephants in the room’ for New Zealand’s improve the acute surgical health system in its 80th anniversary year: general practice patient admission process charges and ownership model Publication Information published by the New Zealand Medical Association

NZMJ Editor NZMA Chair Professor Frank Frizelle Dr Kate Baddock

NZMJ Production Editor NZMA Communications Manager Rory Stewart Diana Wolken

Other enquiries to: NZMA To contribute to the NZMJ, fi rst read: PO Box 156 www.nzma.org.nz/journal/contribute The Terrace Wellington 6140 © NZMA 2019 Phone: (04) 472 4741

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 2019 New Zealand subscription rates Overseas subscription rates Individuals* $317 Individual $442 Institutions $549 Institutions $591 Individual article $30 Individual article $30

*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 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 2 www.nzma.org.nz/journal CONTENTS

EDITORIAL 42 6 Multimorbidity and multiple Smoke in our eyes: the Sense social disadvantage in a Partners’ evaluation of the New Zealand high-needs free legalisation of cannabis in primary healthcare clinic New Zealand population: a cross-sectional study Chris Wilkins, Marta Rychert, Sharmaine Sreedhar, Lauralie Richard, Jose S Romeo, Steve Randerson Tim Stokes ARTICLES LETTER 10 52 Electronic whiteboards improve A response to: The ‘elephants in the acute surgical patient the room’ for New Zealand’s health th admission process system in its 80 anniversary Yi Hai Li, Eve Fitzgerald, Ross Roberts year: general practice charges and ownership model 17 Kate Baddock, Jan White, Lesley Clarke Post-operative mortality rates for neck of femur fracture at 55 Waitemata District Health Board Getting surgical antibiotic Reuben J Kirk, Carlene MM Lawes, prophylaxis right, lessons from William Farrington, Peter Misur, the National Orthopaedic Surgical Matthew L Walker, Michal Kluger, Site Infection Improvement Min Yee Seow, Penny Andrew Programme: a call for action! Arthur J Morris, Sally A Roberts, 26 Nikki Grae, Deborah Jowitt Patterns of referral and uptake of BReast CAncer (BRCA) gene 59 testing of eligible women with Smokefree signage at railway ovarian cancer in New Zealand stations: a survey of 54 stations in Katherine Fraser, Ai Ling Tan, 11 local government areas Carrie Innes, Rosalie Stephens, Nick Wilson, George Thomson Amanda Tristram, Simone Petrich, Caroline Lintott, Peter H Sykes, METHUSELAH Kimberley Gamet, Alice Christian, Bryony Simcock 62 Effect of aspirin on cardiovascular 36 events and bleeding in the healthy The cost of teri unomide in the elderly treatment of relapsing-remitting 100 YEARS AGO multiple sclerosis J Alasdair Millar 63 Etiology and Treatment of In uenza

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 3 www.nzma.org.nz/journal SUMMARIES

Electronic whiteboards improve the acute surgical patient admission process Yi Hai Li, Eve Fitzgerald, Ross Roberts This study is a survey of registrars (junior surgical staff) working in the Department of Surgery at Christchurch Hospital before and after introduction of an electronic whiteboard for managing the admission of acute (emergency) patients. The study has shown increased effi ciency of the process following introduction of the electronic whiteboard. Signifi cantly, the registrars report increased satisfaction with the process and furthermore they reported that they were more aware of incoming patients. Other important outcomes of the study were that there did not seem to be any increased concern about patient or data security when this information was conveyed electronically versus telephonically. The electronic whiteboards are now being used in more areas as other services realise their benefi ts. Post-operative mortality rates for neck of femur fracture at Waitemata District Health Board Reuben J Kirk, Carlene MM Lawes, William Farrington, Peter Misur, Matthew L Walker, Michal Kluger, Min Yee Seow, Penny Andrew Our study investigated patient death rates following surgery for hip fracture at Waitemata DHB between 2009 and 2016. We found that 23.8% of patients died within one year of surgery; however death rates have gradually reduced in recent years. We also found that decreased time from hospital presentation to surgery was linked to improved survival in patients with hip fracture. Patterns of referral and uptake of BReast CAncer (BRCA) gene testing of eligible women with ovarian cancer in New Zealand Katherine Fraser, Ai Ling Tan, Carrie Innes, Rosalie Stephens, Amanda Tristram, Simone Petrich, Caroline Lintott, Peter H Sykes, Kimberley Gamet, Alice Christian, Bryony Simcock Women with specifi c variants of the BRCA (BReast CAncer) gene are at higher risk of devel- oping a specifi c type cancer of the ovaries known as high-grade serous cancer (HGSC). Knowing whether a woman with HGSC has one of these mutations allows us to improve deci- sions around treatment and test family members, as it may mean they are at higher risk. Our study determined that 70% of New Zealand women diagnosed with HGSC between 2015–2016 were referred for genetic testing for these BRCA variants, this is higher than rates shown in other international studies. This data allows for further research into methods to improve the number of women being referred and thus able to access the benefi ts of the knowledge gained for themselves and their families.

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 4 www.nzma.org.nz/journal SUMMARIES

The cost of teri unomide in the treatment of relapsing- remitting multiple sclerosis J Alasdair Millar Terifl unomide is a new drug funded for the treatment of multiple sclerosis (MS). It is a deriv- ative of an older and cheaper drug, lefl unomide. The standard dose of lefl unamide will produce the same dose of terifl unomide in the body and hence it is likely that identical benefi ts in MS would be obtained with lefl unomide at lower cost. I scrutinised the minutes of eight drug regulatory bodies, including those of New Zealand and Australia, to determine whether the above possibility was minuted during consideration of terifl unomide subsidy. Three bodies (European Union, Canada and the US) discussed the relationship between the two drugs, especially with regard to toxicity, but no body minuted the possible use of lefl unomide or the pricing implications. The majority dealt with the applications in a routine manner. It is possible that an opportunity to decrease the cost of treating MS has been missed, though a new clinical trial comparing the two drugs would be required. Multimorbidity and multiple social disadvantage in a New Zealand high-needs free primary healthcare clinic population: a cross-sectional study Sharmaine Sreedhar, Lauralie Richard, Tim Stokes It is now common for patients to need care for several chronic medical conditions they may have. This is called multimorbidity. Multimorbidity increases with age, is more common and occurs earlier in those living in poorer areas and disproportionately affects Māori and Pacifi c people. Patients living in poorer areas will have diffi culty accessing traditional GP clinics, often because of the cost of GP appointments, and may use ‘third sector’ (non-government, non-profi t) GP clinics, which may provide care at lower cost or be free. This study found that patients attending one such free GP clinic in Dunedin had very high levels of multimorbidity and half had both physical and mental health problems. The needs of such patients cannot at present be met by the traditional model of New Zealand general practice.

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 5 www.nzma.org.nz/journal EDITORIAL

Smoke in our eyes: the Sense Partners’ evaluation of the legalisation of cannabis in New Zealand Chris Wilkins, Marta Rychert, Jose S Romeo, Steve Randerson

he New Zealand coalition government referendum, the public will be bombarded has announced it intends to conduct a with estimates of a rather dubious nature Tnational referendum on the personal from both sides of the debate, including use of cannabis at the next general election from a burgeoning commercial cannabis in 2020,1 signalling the potential for a major industry seeking to maximise the profi t change in policy direction with signifi cant potential of any future market.10 In this health implications. Proponents of canna- context, the media and public may come to bis law reform have articulated a range of rely on the Sense fi gures as an independent reasons for the legalisation of cannabis, source based on economic analysis—as has including the potential for new commercial already been seen.11 opportunities, addressing disproportionate We have three principal concerns conviction rates for Māori, reducing drug about the Sense evaluation of a regulated enforcement costs, promoting regional eco- commercial cannabis market: (1) the belief nomic development, capturing lost tax reve- that the price of cannabis can be raised in nue, improving access to drug treatment and a commercial cannabis market to restrain 2–5 undermining criminal gangs. It is import- any increase in demand; (2) the assumption ant that the public health consequences of that a commercial cannabis market will not different reform options are also given a result in any increase in health and social high priority in the ongoing debate. harm; and (3) the projected level of social Consequently, it was initially heartening benefi ts from additional spending on drug to see the economic consultancy fi rm “Sense education and prevention. Partners” utilise an established Drug Harm The fi rst issue relates to how a 6 Index methodology to evaluate the impact commercial cannabis market will impact of several drug policy reform options, levels of cannabis use, and most impor- including a legal regulated commercial tantly adolescent and high frequency use, 7 market for cannabis. The Sense report’s as this type of use is most associated with conclusion that the legalisation of cannabis harm.12 The Sense report authors concede in New Zealand will confer a net social that the legalisation of cannabis “could” benefi t of $10–$53 million per year (without increase the demand for cannabis,7 but 7 any additional health intervention) suggests propose that a tax on cannabis sales can such a reform is both desirable and risk be utilised to raise the price of cannabis to free. While we are on record as being restrain any increase in use: “The proposal broadly supportive of cannabis law reform, is to set the levy on cannabis sales such including decriminalisation and sales by that market prices increase modestly, to “not for profi t” trusts with obligations to minimise any lift in [cannabis] demand 8–9 return earnings back to the community, (but ensure incentives are such that supply we believe the Sense evaluation does not comes into the formal economy)”.7 The provide an adequate picture of the potential Sense authors conclude “based on tentative risks of a commercial (albeit regulated) evidence and effective use of price signals, cannabis market. we assume as a base case that there will be It is important to raise these concerns at no increase or decrease in harm from drug this time as in the run up to the national use as a result of legalisation”.7

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 6 www.nzma.org.nz/journal EDITORIAL

Unfortunately, empirical experience to recent establishment of legal regimes.15–18 date shows that the legalisation of cannabis All of the studies to date have focused on has resulted in substantial declines in the the two US states that were the fi rst to price of cannabis despite the imposition legalise cannabis and have established of signifi cant excise taxes. For example, regulated commercial markets (ie, Colorado the price of legal cannabis in Washington and Washington State). Cerda et al investi- State has fallen by 25% each year since gated the impact of cannabis legalisation on the opening of legal retail outlets in 2014, adolescent cannabis use and found mixed despite the imposition of a 37% sales tax on results (ie, increased use in Washington cannabis products.13 Similarly, in Colorado, State and no change in Colorado), perhaps the average retail price of cannabis per gram refl ecting the more commercial medicinal has declined by 48% since the opening of cannabis market that existed in Colorado legal retail outlets in 2014, again despite the preceding legalisation of recreational levying of a combined 25% tax on cannabis use.15 More recent studies have also found products (ie, 15% excise and an additional no change in adolescent cannabis use in 10% sales tax).14 Economic analysis suggests Colorado following legalisation.16–17 Wang et that (everything else remaining the same) al found signifi cant increases in hospitalisa- legal production of cannabis will result tions, emergency department visits and calls in substantial declines in price over time to the poison centre in relation to cannabis as producers draw on cost savings from following the commercialisation of legal economies of scale and modern agricultural medicinal and then recreational cannabis production practices.13 This suggests it will use in Colorado.18 take a very high tax rate and/or minimum Legal medicinal cannabis regimes have price on cannabis, and continual revision of operated in a number of US states for consid- these tax rates, to maintain the legal price of erably longer than the legal recreational cannabis and counter the effi ciency gains of regimes, and consequently studies of the legal cannabis production. impact of these medicinal regimes provide Our second concern is the Sense report some insight into the longer-term impact of authors’ assumption that a regulated greater legal availability as they introduce commercial cannabis market will not result legal use and often retail sales, albeit in any increase in cannabis-related harm. within a medical framework. A number of The Sense authors do acknowledge that studies have found no difference in rates of any increase in cannabis use could result adolescent cannabis use in US states with in signifi cant additional health and social medicinal cannabis regimes compared to costs: “In the absence of harm reduction those without medicinal regimes.19 However, measures, a 10% increase in [cannabis] use studies of the impact of medicinal cannabis represents nearly $100 million of additional regimes on adult cannabis use in the US personal and community harm according have found higher rates of use, daily or near to the Drug Harm Index”.7 Yet, given their daily use, and incidence of cannabis depen- earlier assumption concerning the ability dence in states with medicinal cannabis to constrain any increase in use following regimes.19–22 These fi ndings, along with the legalisation with a price rise via a tax, they previously reported increases in hospital- include only a relatively small increase isations, emergency department visits and in health spending of $1–$10 million per poison service calls following legalisation, year (ie, to cover brief health interven- suggest that any balanced evaluation of the tions by primary care nurses).7 The fi rst impact of a regulated commercial cannabis part of the above quote (“In the absence of market should include the possibility of harm reduction measures…”) also appears increasing adult cannabis use and related to suggest they believe harm reduction health costs. programmes can be used to largely mitigate Our third concern is the projected social the harm from any increase in cannabis use. benefi ts from increased spending on There is currently only a small number of drug education and prevention under the published studies on the impact of the legali- cannabis legalisation option. According to sation of recreational cannabis use on levels the Sense evaluation, additional spending of use and harm, refl ecting the relatively of $9 million per year on drug education

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 7 www.nzma.org.nz/journal EDITORIAL

and prevention will yield social benefi ts of benefi ts. This general conclusion concerning $19–$64 million per year7 based on bene- modest effects is qualifi ed by the obser- fi t-cost ratios “found in the literature”.7 vation that some prevention programmes These estimated social benefi ts actually have no evidence of benefi t, whereas others lift the Sense evaluation of cannabis show evidence of some effectiveness, and legalisation from a fi scal loss of (-) $9–$11 that replication evidence for effective million per year to a net social benefi t of programmes is inconsistent”.24 7 (+) $10–$53 million per year. The Sense The Sense report highlights the projected authors only cite one reference to support $191–$249 million that could be collected in these projected social benefi ts—a discussion tax and licensing fees from legal cannabis by a leading drug policy researcher on why sales,7 but this must be balanced against research has had so little impact on drug the additional health and social costs 23 policy (from 2001). This article provides from increased cannabis use and depen- no fi ndings concerning the benefi t-cost dency that the taxpayer, not the cannabis ratios of drug education and prevention. It industry, will be required to pay for under actually laments the “lack of good quality a commercial market approach. Those research and demonstration” of the effec- who currently disproportionately bear tiveness of drug prevention programmes, the harm of alcohol and tobacco use are and is sceptical about the effectiveness also likely to disproportionately bear the of some widely used drug prevention harm of commercially available cannabis, 23 programmes in the US at the time. A including Māori, high-risk youth, those more recent review of drug prevention suffering mental illness and lower socio-eco- programmes by leading drug policy scholars nomic groups. It is important that as New concludes: “Evidence-based school or family Zealanders are asked to consider different oriented prevention programmes are on cannabis law reform options, they are average a relatively low cost activity that provided with rigorous and balanced evalu- can yield modest, but potentially important ations of likely outcomes.

Competing interests: Nil. Author information: Chris Wilkins, SHORE & Whariki Research Centre, College of Health, Massey University, Auckland; Marta Rychert, SHORE & Whariki Research Centre, College of Health, Massey University, Auckland; Jose S Romeo, SHORE & Whariki Research Centre, College of Health, Massey University, Auckland; Steve Randerson, SHORE & Whariki Research Centre, College of Health, Massey University, Auckland. Corresponding author: Associate Professor Chris Wilkins, SHORE & Whariki Research Centre, College of Health, Massey University, Auckland. [email protected] URL: http://www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2019/vol-132-no-1490- 22-february-2019/7811

REFERENCES: 1. Binding referendum on referendum-on-legalising- from: http://nzier.org. legalising cannabis for cannabis-for-personal-use- nz/static/media/fi ler_ personal use to be held at to-be-held-at-2020-election public/96/47/9647693f-7878- 2020 election (18 Dec 2018). 2. NZIER. The high cost of 4fa3-b0fe-a4092dfd329e/ Available from: http:// (not) stopping people nzier_insight_61- www.radionz.co.nz/news/ getting high. NZIER high_cost_of_not_letting_ political/378523/binding- Insight 61. 2016: Available people_get_high_-_fi nal.pdf

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 8 www.nzma.org.nz/journal EDITORIAL

3. New Zealand Drug advertising campaign (17 17. Harpin SB, Brooks-Rus- Foundation. Whakawātea Dec 2018). Available from: sell A, Ma M, James KA, te Huarahi A model drug http://www.nzherald. Levinson AH. Adoles- law to 2020 and beyond. co.nz/business/news/ cent Marijuana Use http://www.drugfoun- article.cfm?c_id=3&- and Perceived Ease of dation.org.nz/assets/ objectid=12177738 Access Before and After uploads/2017-uploads/ 11. Legalised cannabis market Recreational Marijuana Model-drug-law//Whakawa- could grow NZ’s coffers Implementation in tea-te-Huarahi-July2017.pdf by $240m a year in tax (7 Colorado. Substance 4. Fyers A. Could our farmers Nov 2018). Available from: Use and Misuse, 2018; cash in on legal cannabis? http://www.stuff.co.nz/ 53(3):451–456. (6 July 2017). Available national/health/108411295/ 18. Wang GS, Hall K, Vigil from: http://www.stuff. legalised-cannabis-market- D, Banerji S, Monte A, co.nz/business/farming/ could-grow-nzs-coffers- VanDyke M. Marijuana cropping/93226526/ by-240m-a-year-in-tax and acute health care could-our-farmers-cash- 12. Fergusson DM, Boden JM, L contacts in Colorado. Prev in-on-legal-cannabis John Horwood JL. Psycho- Med. pii: S0091-7435, 5. Hayes S. Sir Richard social sequelae of cannabis 2017; (17)30120-2. Branson: Farmers should use and implications for 19. Hasin, D.S. US Epidemi- swap cows for cannabis policy: fi ndings from the ology of Cannabis Use (27 March 2017). Available Christchurch Health and and Associated Problems. from: http://www.newshub. Development Study. Soc Neuropsychopharmacology co.nz/home/money/2017/03/ Psychiatry Psychiatr Epide- Reviews. 2018; 43:195–212. sir-richard-branson- miol. 2015; 50:1317–1326. 20. Wen H, Hockenberry JM, farmers-should-swap- 13. Caulkins JP. Recognizing Cummings JR. The effect of cows-for-cannabis.html and regulating cannabis medical marijuana laws on 6. McFadden Consultancy. as a temptation good. adolescent and adult use Research Report: The International Journal of of marijuana, alcohol, and New Zealand Drug Harm Drug Policy. 2017; 42:50–6. other substances. J Health Index 2016. Wellington: 14. Marijuana Prices in Denver Econ 2015; 42:64–80. Ministry of Health, 2016. and Colorado: Winter 21. Martins SS, Mauro CM, 7. Sense Partners. Estimating 2017/2018 and Spring 2018 Santaella-Tenorio J, Kim the impact of drug policy Update (16 May 2018). JH, Cerda M, Keyes KM, options Moving from a Available from: http://www. et al. State-level medical criminal to a health-based coloradopotguide.com/ marijuana laws, mari- approach Final Report colorado-marijuana-blog/ juana use and perceived 31 October 2018. http:// article/marijuana-prices- availability of marijuana www.drugfoundation. in-denver-and-colorado- among the general U.S. org.nz/assets/uploads/ winter-20172018-and- population. Drug Alcohol Cost-benefi t-analy- spring-2018-update/ Depend 2016; 169:26–32. sis-drug-law-reform.pdf 15. Cerda M, Wall M, Feng 22. Hasin DS, Sarvet AL, Cerdá 8. Wilkins C. After the T, Keyes KM, Sarvet A, M, Keyes KM, Stohl M, legalisation of cannabis: Schulenberg J, et al. Galea S, et al. US adult illicit the Cannabis Incorporated Association of state cannabis use, cannabis Society (CIS) regulatory recreational marijuana disorder, and medical mari- model for recreational laws with adolescent mari- juana laws: 1991–1992 to cannabis in New Zealand. juana use. JAMA Pediatr, 2012–2013. JAMA Psychi- New Zealand Medical 2017; 171:142––149. atry 2017; 74:579–588. Journal 2016; 129. 16. Brooks-Russell A, Ma M, 23. Reuter P. “Why does 9. Wilkins, C. A “not-for-prof- Levinson AH, Kattari L, research have so little it” regulatory model Kirchner T, Goodell EMA, impact on American for legal recreational Johnson RM. Adolescent drug policy?” Addiction. cannabis: Insights from Marijuana Use, Marijua- 2001; 96(3):373–376. doi: the regulation of gaming na-Related Perceptions, and 10.1080/0965214002005437 machine gambling in New Use of Other Substances 24. Babor T, Caulkins J, Fischer Zealand. International Before and After Initiation B, et al. Drug Policy and Journal of Drug Policy, of Retail Marijuana Sales the Public Good. (2nd 2018; 53:115–122. in Colorado (2013–2015). edition). Oxford: Oxford 10. Kiwi startup Helius Prevention Science. 2018, University Press; 2018. Therapeutics launches 24 July. doi: 10.1007/ nationwide cannabis s11121-018-0933-2

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 9 www.nzma.org.nz/journal ARTICLE

Electronic whiteboards improve the acute surgical patient admission process Yi Hai Li, Eve Fitzgerald, Ross Roberts

ABSTRACT BACKGROUND: Electronic whiteboards have largely replaced the use of traditional whiteboards in many hospital departments. They are used to electronically record and display a variety of patient information to streamline the admission process and the quality of handover between relevant sta . We assessed the impact of such a system upon the patient admission process in a busy general surgery department. METHODS: A survey of 12 qualitative questions was completed by surgical registrars working within a general surgery department in 2013 prior to the introduction of electronic whiteboards and again in 2016 a er introduction. The questions compared the satisfaction of the admission process before and a er its introduction. RESULTS: There was an improvement in sta satisfaction with the admissions process a er the introduction of electronic whiteboards (78% vs 9% high level of satisfaction, p<0.05). A statistically non-significant rise was also seen in individual areas of the admissions and handover process. No change was seen in sta attitude to security of patient details (50% vs 42% high level of satisfaction, p=0.671). CONCLUSION: Electronic whiteboards assist in the process of admitting patients to a general surgical department. This strengthens the case for the introduction of electronic whiteboards across a range of hospital departments.

he volume of clinical work in hospital among relevant staff, and improving the departments is increasing year on effi ciency of patient fl ow through a depart- Tyear.4,7–9 With an increased workload ment.1,4,6,9,12 Electronic whiteboards are used comes the need to fi nd effi ciencies in patient by various members of the multidisciplinary care, including documentation and commu- team, including doctors, nurses, physiother- nication. A number of technological ap- apists and occupational therapists, and have proaches to improving effi ciency have been proven to be effective in patient care.1,2 suggested, including the use of electronic The introduction of these systems 1–2,10,11 whiteboards. Electronic whiteboards clearly incurs a fi nancial cost. As such it is have been introduced in many hospitals important that the benefi ts of such a system over recent years, typically replacing tradi- be evaluated, and we sought to determine tional handwritten dry-erase whiteboards how the introduction of electronic white- in areas of high patient turnover such as the boards within a surgical department 1,2,5,9 emergency department. Usually, they in- affected the patient admissions process. volve a method of entering clinical details to a central location whereupon this informa- FloView tion is available across one or more display Christchurch Public Hospital in New screens in the department. The benefi ts that Zealand built the ‘FloView’ electronic electronic whiteboards offer include in- whiteboard system for use in the general creased accuracy of documentation and the surgery admission unit in 2013. The system ability to access patient information from allows authorised users to enter information a number of locations simultaneously.2–4 relevant to the patient from any networked This has the combined potential of reducing computer within the Canterbury District medical errors, improving communication Health Board domain, and some other users to view it, including the emergency

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 10 www.nzma.org.nz/journal ARTICLE

department staff. This patient information is results returned by surgical registrars. displayed on a central screen in the general The answers were collated and are shown surgery admitting unit to inform all team in Tables 1–3. The total number of each members of the status of current patients response to each question was recorded, and to alert the team of any new referrals. along with the percentage of respondents A basic version of an electronic whiteboard who chose each answer. had previously been implemented in the Analysis of qualitative survey responses emergency department. required conglomeration of the data into Notes were previously handwritten or two answer groups. The responses to each documented on a physical whiteboard question were assessed individually. As seen and suffered from two major drawbacks. in the tables, the answers to each question Handwritten notes are prone to being lost, were grouped into ‘high’ satisfaction misplaced or thrown away and can only (those responses in the top two satisfaction be viewed from one location, necessitating brackets) and ‘low’ satisfaction (those excessive time communicating the same responses in the bottom three satisfaction information by telephone to other staff. groups). The relative ratio of responses in Mislaid handover notes also represent a the high satisfaction group was compared signifi cant patient privacy risk, especially between the two surveys from 2013 and if alternatives exist. The FloView system 2016. A Chi-squared test was conducted to was introduced to ease the fl ow of patients return a p-value for the difference in satis- during the admission process and the faction between the two surveys. A p-value quality of handover of patient details from of less than 0.05 was chosen to represent either the emergency department or direct statistical signifi cance of a difference general practitioner referrals. between the data sets. Method Results Data were collected from general surgery In 2013, 12 surveys were completed (n= registrars working in the department of 12). In 2016, 14 surveys were completed general surgery at Christchurch Public (n=14). This represents all of the available Hospital. The registrars were asked to registrars working within the department. complete a survey of 12 questions regarding The returned answers are displayed in their experience of the patient admis- Tables 1–3. It can be appreciated from the sions process. Survey questions covered a survey results that in the second survey variety of topics, including ease of identi- (following introduction of the electronic fying patients, patient fl ow and security of whiteboards), every question shows an patient details. A list of the questions can increase in the numbers of respondents be found in Tables 1–3. Printed surveys choosing a high satisfaction response. What were provided to every registrar within the differs is the degree of statistical signifi cance department. The surveys were completed of these improvements. We present here and returned anonymously. four specifi c question results which we feel The fi rst survey was performed in 2013 merit further discussion. prior to the introduction of electronic white- Anticipating workload boards within the department. A second When accepting a patient referral into survey consisting of identical questions a department it is essential that all team was provided three years later in 2016 to a members can effi ciently anticipate the level (now different) group of registrars who had of workload. This process is facilitated by been working with the electronic white- access to a patient’s working diagnosis, board system. All questions were qualitative age and comorbidities. Following the in nature and graded from ‘high’ to ‘low’ introduction of the FloView electronic white- consistently with fi ve possible responses board, there is an increase in the number of to each question. Surveys were completed surgical registrars reporting that they can over a one-week period. A number of acute usually accurately or very accurately predict surgical nurses were also involved but, for their level of workload for a new patient accurate comparisons between the two (78% vs 42% reported high satisfaction, surveys, we have only discussed here the p=0.054) (Table 1).

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 11 www.nzma.org.nz/journal ARTICLE

Table 1: Impact on registrar workfl ow.

2013 2016 p-value 1.) Time taken for phone calls: how much time would you spend making/receiving phone calls per patient accepted? High satisfaction <5 mins 2 (17) 6 (43)

5–6 mins 6 (50) 7 (50)

Low satisfaction 7–8 mins 1 (8) 1 (7) 0.091

9–10 mins 0 0

>10 mins 3 (25) 0

2.) Phone calls: how disruptive to other tasks you may be doing is accepting/organising admission of an acute inpatient? High satisfaction Not disruptive at all 0 0

Mildly disruptive 1 (8) 3 (23)

Low satisfaction Moderately disruptive 2 (17) 3 (23) 0.315

O en disruptive 6 (50) 3 (23)

Always disruptive 3 (25) 4 (31)

3.) Time spent on non-patient workload: how would you rate the amount of time spent on non-patient contact work compared to direct patient care? High satisfaction Minimal 0 0

Moderate 5 (42) 5 (36)

Low satisfaction Significant 5 (42) 7 (50) 0.756

Too much 2 (17) 2 (14)

Excessive 0 0

4.) Ability to anticipate workload from handover: how accurately can you anticipate your workload for each patient from current handover system? High satisfaction Very accurately 0 2 (14)

Usually accurately 5 (42) 9 (64)

Low satisfaction Sometimes accurately 5 (42) 2 (14) 0.054

Usually inaccurately 2 (17) 1 (7)

I never know 0 0

5.) Ease of ward rounds: how easy is it to know who and where all the patients that need to be seen are? High satisfaction Usually easy 1 (8) 3 (21)

Very easy 6 (50) 9 (64)

Low satisfaction Average 3 (25) 1 (7) 0.117

Hard 2 (17) 0

Awful 0 1(7)

6.) Awareness: how aware are you of the progress of a patient through the admission process? High satisfaction Totally aware 2 (17) 5 (36)

Usually up to date 5 (42) 5 (36)

Low satisfaction Somewhat aware 3 (25) 3 (21) 0.484

Not really aware 2 (17) 1 (7)

Not at all aware 0 0

Registrar survey answers in 2013 (n=12) and 2016 (n=14). Total number of answers with (%) of respondents. Chi-square p value for di erence between high and low satisfaction groups.

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 12 www.nzma.org.nz/journal ARTICLE

Table 2: Impact on handover.

2013 2016 p-value 1.) Time taken for sta handover: how long does it take to complete handover process at the end of your shi ? High satisfaction <2 mins 0 0

2–5 mins 0 1 (7)

Low satisfaction 5–10 mins 4 (33) 5 (36) Invalid

10–15 mins 2 (17) 5 (36)

>15 mins 6 (50) 3 (21)

2.) Ease of handover: how easy is it to complete handover process? High satisfaction Very straightforward 0 2 (14)

Usually straightforward 3 (25) 6 (43)

Low satisfaction Variable 6 (50) 5 (36) 0.098

O en di icult 3 (25) 1 (7)

Always challenging 0 0

Registrar survey answers in 2013 (n=12) and 2016 (n=14). Total number of answers with (%) of respondents. Chi- square p value for di erence between high and low satisfaction groups.

Patient flow system reduced the need for handwritten A smooth patient fl ow from admission and notes, it does not replace it altogether and assessment through to management and many of the potential lapses of security of discharge is essential in a busy department. patient details during the handover and Accurate record keeping of patient details admission process remain. Accordingly, staff can speed up this process and make patient in 2016 do not see an appreciable difference transit much more effi cient. Improvement in security of patient details compared to is seen in registrar assessment of patient their colleagues who worked prior to the fl ow after the introduction of electronic introduction of electronic whiteboards. It whiteboards (78% vs 46% reporting high is however important to note that both sets satisfaction, p=0.127) (Table 3). of registrars overwhelmingly felt that data were protected at least somewhat securely Overall satisfaction with admission (50% vs 42% reporting high satisfaction, process p=0.671) (Table 3). As well as questioning how registrars felt about specifi c aspects of the patient Discussion admission process, their overall satis- faction was evaluated. We see a signifi cant We surveyed two groups of surgical regis- improvement in overall satisfaction that trars both before and after implementation general surgery registrars have with the of electronic whiteboards to assist in the ease of the process of admitting patients to admissions process in the acute general hospital following the introduction of elec- surgery department of a major hospital in tronic whiteboards (78% vs 9% reporting New Zealand. The registrars were asked high satisfaction, p=0.001) (Table 3). about a range of topics, including ease of receiving a patient handover, safety Security of patient details of patient details and how easy it was to Due to the importance of patient confi - locate patients. Due to the low participant dentiality we also questioned how secure numbers, most of the results were not statis- patient details were felt to be under the tically signifi cant. However our results do electronic whiteboard system. While the show a general improvement in doctors’

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 13 www.nzma.org.nz/journal ARTICLE

Table 3: Impact on patient care.

2013 2016 p-value 1.) Communication/accuracy of patient details: how well is information communicated throughout admission? High satisfaction Very well 0 1 (7)

Well most of the time 5 (42) 9 (64)

Low satisfaction Variable 7 (58) 4 (29) 0.125

Poorly in general 0 0

Terrible 0 0

2.) Security of patient details: how secure do you think patient information is using the current handover process? High satisfaction Very secure 0 1 (7)

Reasonably secure 5 (42) 6 (43)

Low satisfaction Somewhat secure 3 (25) 5 (36) 0.671

Not very secure 3 (25) 2 (14)

Not at all secure 1 (8) 0

3.) Patient flow: how well does the system work regarding flow of patients from GP/ED through Surgical Assessment Review Area through to ward? High satisfaction Flow is never an issue 0 1 (7)

Not usually an issue 6 (46) 10 (71)

Low satisfaction Patient flow sometimes bad 4 (31) 3 (21) 0.127

Patient flow is usually bad 2 (15) 0

Patient flow doesn’t happen at all 0 0

4.) Satisfaction in the admissions process: how satisfied are you in the entire acute admission and handover process? High satisfaction 100% satisfied 0 1 (7)

Mostly satisfied 1 (9) 10 (71)

Low satisfaction Generally satisfied 6 (55) 1 (7) 0.001

Somewhat satisfied 2 (18) 2 (14)

Not at all satisfied 2 (18) 0

Registrar survey answers in 2013 (n=12) and 2016 (n=14). Total number of answers with (%) of respondents. Chi- square p value for di erence between high and low satisfaction groups. attitudes towards how the process of patient in the overall quality of patient admissions admission is managed. Some show an appre- and handover. This validates the responses ciable increase in doctor satisfaction, though to other questions where the trend is not a statistically signifi cant one. towards greater satisfaction with individual When asked how satisfi ed they were with aspects of the admissions process though the whole process of patient admissions, without such a high degree of statistical the doctors did have a favourable opinion signifi cance. This may be due in part to the with a statistically signifi cant improvement relatively small number of doctors involved

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 14 www.nzma.org.nz/journal ARTICLE

in surveys both in 2013 and 2016. This study The doctors’ attitudes to the safety of has shown that doctors working in acute patient details warrants addressing as demand feel that patient handover and the improving access to patient information acute admissions process is improved by the on a whiteboard also risks its visibility to use of electronic whiteboards. In the years non-relevant staff and public. In relation to subsequent to FloView’s introduction to the patient confi dentiality, patient identifi ers, general surgical department, it has been diagnoses and further comments can be deployed throughout all other inpatient masked if required, especially if the display wards, highlighting its perceived integrity to screen is located in a publicly accessible the care of patients by not only allied health, area. However, the doctors surveyed felt but also hospital management. that the security of patient data was no Staff have found that electronic white- better or worse than it had been prior to the boards help with patient fl ow and introduction of electronic whiteboards. This anticipation of workload by virtue of the highlights the ongoing need to be vigilant FloView being able to display patient with respect to the security of patient information to all relevant staff. Both details regardless of the medium used to the 2013 and 2016 cohort noted a high record them. degree of disruption from phone calls and It will be useful to repeat this assessment thus reduced patient contact time. The of the system as an ongoing project to disruption, and thus patient fl ow, could determine how satisfi ed doctors are with theoretically be reduced by having another the system that they use on a daily basis. staff member (for example a triage nurse) In addition, for services yet to implement adding new referrals onto the electronic such a system, it may also be useful to assess whiteboard by way of a virtual handover, whether the electronic whiteboard reduces thus allowing the on-call doctor to continue the waiting times of patients to be seen, as seeing patients, improving both staff and implied by the improvement in workfl ow. patient satisfaction.

Competing interests: Nil. Acknowledgements: The authors wish to thank Mr Graeme McQueen (CDHB Senior Business systems analyst) for his input on the structure and background of our electronic whiteboard system, FloView. They also thank Mr Bruce Dobbs (CDHB General Surgery Scientifi c offi cer) for assistance with statistical analysis and Dr David Cameron (General Surgery Registrar) for his advice on the study design. Author information: Yi Hai Li, Registrar, Department of General Surgery, Christchurch Public Hospital, Christ- church District Health Board, Christchurch; Eve Fitzgerald, Registrar, Department of General Surgery, Christchurch Public Hospital, Christchurch District Health Board, Christchurch; Ross Roberts, General Surgeon (Upper GI), Department of General Surgery, Christchurch Public Hospital, Christchurch District Health Board, Christchurch. Corresponding author: Dr Ross Roberts, 2 Riccarton Ave, Christchurch. [email protected] URL: http://www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2019/vol-132-no-1490- 22-february-2019/7812

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 15 www.nzma.org.nz/journal ARTICLE

REFERENCES: 1. Wong HJ, Caesar M, Bandali whiteboard system. J 10. Cabrera R, Katz C, Akin- S, et al. Electronic inpatient Am Med Inform Assoc. lonu A, et al. Wait here whiteboards: improving 2008; 15:184–194. while I get my attending: multidisciplinary commu- 6. Steitz BD, Weinberg ST, Improving fl ow in resident nication and coordination Danciu I, Unertl KM. continuity clinic. Academic of care. Int J Med Inform. Managing and Commu- Pediatrics. Conference: 2009; 78:239–247. nicating Operational Annual Spring Meeting 2. Hertzum M. Electronic Workfl ow: Designing and of the Association of emergency-department Implementing an Electronic Pediatric Program Direc- whiteboards: a study of Outpatient Whiteboard. tors, APPD 2017. United clinicians’ expectations Appl Clin Inform. States: 2016. 17–18. and experiences. Int J Med 2016; 7:59–68. 11. Chaboyer W, Wallen K, Inform. 2011; 80:618–630. 7. Ministry of Health. 2016. Wallis M, McMurray AM. 3. Randell R, Greenhalgh J, Publicly funded hospital Whiteboards: one tool to Wyatt J, et al. Electronic discharges – 1 July 2013 to improve patient fl ow. Med Whiteboards: Review 30 June 2014. Wellington: J Aust. 2009; 190:137–140. of the Literature. Stud Ministry of Health. 12. Bice W. Patient safety: Heal Technol Inform. 8. Ministry of Health. 2015. improved pre-treatment 2015; 210:389–393. Factsheet: Short stay coordination and commu- 4. Boger E. Electronic emergency department nication through the use of tracking board reduces ED events. Wellington: an electronic whiteboard patient length of stay at Ministry of Health. in radiation oncology. Medical physics. Confer- Indiana Hospital. J Emerg 9. France DJ, Levin S, ence: 60th annual meeting Nurs. 2003; 29:39–43. Hemphill R, et al. Emergen- program of the American 5. Aronsky D, Jones I, cy physicians’ behaviors Association of Physicists Lanaghan K, Slovis CM. and workload in the in medicine, AAPM. 2018. Supporting patient care in presence of an electronic United states: 2018; 45:202. the emergency department whiteboard. Int J Med with a computerized Inform. 2005; 74:827–837.

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 16 www.nzma.org.nz/journal ARTICLE

Post-operative mortality rates for neck of femur fracture at Waitemata District Health Board Reuben J Kirk, Carlene MM Lawes, William Farrington, Peter Misur, Matthew L Walker, Michal Kluger, Min Yee Seow, Penny Andrew

ABSTRACT AIM: Mortality rates of up to 38% at one year have been reported following surgery for neck of femur fractures. The aim of this review is to evaluate the post-operative mortality rates and trends over time for patients with fractured neck of femur at Waitemata District Health Board. METHOD: A retrospective cohort study of all patients who received surgery following a neck of femur fracture at Waitemata District Health Board between 2009 and 2016. Inpatient data was retrieved from electronic hospital records and mortality rates from the Ministry of Health, New Zealand. Analyses included crude mortality rates and trends over time, and time-to-theatre from presentation with neck of femur fracture. RESULTS: A total of 2,822 patients were included in the study; mean age 81.9 years, 70.4% female and 29.6% male. Overall post-operative crude rates for inpatient, 30-day and one-year mortality were 3.7%, 7.2% and 23.8% respectively. Adjusted analyses showed a statistically significant decrease in mortality rates between 2009 and 2016 at inpatient (p=0.001), 30 days (p=<0.001) and one-year (p=<0.001) time periods. There was also a significant association between time-to-theatre and mortality at inpatient (p=0.002), 30 days (p=0.0001), and one-year (p=0.0002) time periods. CONCLUSION: Mortality rates following surgery for fractured NOF have significantly improved over recent years at Waitemata District Health Board. Reduced time-to-theatre is associated with decreased inpatient, 30-day and one-year mortality.

ractured neck of femur (NOF) is the age, frailty, high American Society of most common cause for orthopaedic Anesthesiologists (ASA) score and delayed admission in older adults, and is asso- time-to-theatre, are associated with F 4,5,7,9,11,12 ciated with signifi cant morbidity and mor- increased early mortality. A history of tality.1 The incidence of NOF fracture in New congestive heart failure, chronic obstructive Zealand is reported to be 218 per 100,000 pulmonary disease, dementia and malig- per year,2 and is rising in parallel with the nancy increases the one-year mortality ageing population. Due to signifi cant costs three-fold in NOF fracture patients.13 associated with acute management, interim Guidelines and recommendations have care placement, rehabilitation and outpa- been developed internationally aiming tient support services, NOF fractures repre- to reduce mortality rates following hip sent a signifi cant burden upon the national fracture, particularly in those who are most healthcare system. vulnerable.14–16 The American Academy High mortality rates following surgery of Orthopaedic Surgeons (AAOS) and the for NOF fracture have been well docu- United Kingdom’s National Institute for mented, and range between 5.3–12% at 30 Health and Care Excellence (NICE) have both days and 17–38% at one year.3–10 Certain provided comprehensive, evidence-based factors including male gender, increasing guidelines that emphasise the importance

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 17 www.nzma.org.nz/journal ARTICLE

of early surgery and coordinating care group. Mortality data was obtained from the through a multidisciplinary hip fracture Ministry of Health, and included inpatient programme to help patients recover more and outpatient date of death to one year, but quickly and regain their mobility.14,15 Many did not include cause of death. hospitals in New Zealand have implemented Mortality at each time point was analysed Enhanced Recovery After Surgery (ERAS) using logistic regression modelling and 17,18 protocols that embody these guidelines, included admission year (2009 to 2016 and there is a growing body of evidence inclusive) in the model. The consistency to suggest that ERAS protocols reduce of results was assessed by the following length of hospital stay and improve func- subgroups: gender, age groups and ASA. 19 tional outcome. The evidence to suggest The association between mortality and that ERAS protocols reduce mortality rates time-to-theatre was assessed using logistic 19,20 remains limited. The purpose of this regression modelling. Analyses were also retrospective audit is to assess mortality adjusted for gender, age groups and ASA. rates in patients following surgery for NOF There were no adjustments for multiplicity. fracture between 2009–2016 at Waitemata All statistical analyses were performed using District Health Board (Waitemata DHB), SAS version 9.4. Auckland, New Zealand. Secondary aims are to assess trends in mortality rates over time, and with respect to age, gender, ASA grade Results and time-to-theatre. Between January 2009 and December 2016, 2,822 patients underwent surgery for fractured NOF at Waitemata DHB. Patient Method demographics are presented in Table 1. Over This is a retrospective cohort study of all 70% of patients were female (70.4%) and the patients who had operative management mean age was 81.9 years. More than 70% of a fractured NOF under the Orthopaedic of patients had an ASA score of 3 or more. service at Waitemata DHB from January There were similar proportions of ASA 3 2009 to December 2016 inclusive. Waitemata scores for males and females, but more DHB encompasses two hospital sites (North Shore Hospital and Waitakere Hospital) Table 1: Demographics of Waitemata DHB pa- tients that underwent surgery for NOF fracture that provide public health services to more (2009–2016). than 580,000 residents of the North Shore, Waitakere and Rodney districts of Auckland, n (%) New Zealand. Inpatient data was retrieved from elec- Total patients 2,822 tronic hospital records at Waitemata DHB. Gender Female 1,988 (70.4) Patients were identifi ed using the Inter- national Classifi cation of Disease, Tenth Male 834 (29.6) Revision, Australian Modifi cation, Eighth Age (years) <60 113 (4) Edition (ICD-10-AM 8th Ed.) coding system.21 Data includes all patients admitted to Wait- 60–69 220 (7.8) emata DHB with a diagnosis of fracture of 70–79 548 (19.4) the hip using the ICD-10-AM 8th Ed. codes S72.00–S72.08 (fractures of the neck of 80–89 1,298 (46) femur) and S72.10–S72.2 (peritrochanteric ≥90 643 (22.8) fractures). Specifi c causes of hip fracture, such as pathological and periprosthetic frac- ASA score* 1 119 (4.2) tures, were not identifi ed in this study. Hip 2 689 (24.4) fractures sustained as an inpatient, as well as those in the community were included. 3 1,566 (55.5) Given mortality rates were being assessed 4 435 (15.4) in NOF patients following surgery, those patients who did not receive surgery were *An ASA score of 3 is defined as a patient with severe systemic disease that is not incapacitating, and an ASA excluded. There were insuffi cient numbers score of 4 is defined as a patient with severe systemic for subgroup analyses in the non-surgical disease that is a constant threat to life.22

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 18 www.nzma.org.nz/journal ARTICLE

Table 2: Post-operative NOF fracture crude mortality rates at Waitemata DHB 2009–2016.

2009 2010 2011 2012 2013 2014 2015 2016 Total Total number 327 371 306 320 362 389 388 359 2,822 of post-op NOF patients

Post-op inpatient 17 17 12 12 17 10 15 4 104 deaths n (%) (5.2) (4.6) (3.9) (3.8) (4.7) (2.6) (3.9) (1.1) (3.7%)

Post-op deaths at 30 33 38 22 20 30 18 22 21 204 days n (%) (10.1) (10.2) (7.2) (6.3) (8.3) (4.6) (5.7) (5.8) (7.2%)

Post-op deaths at 82 109 69 71 87 91 80 82 671 one year n (%) (25.1) (29.4) (22.5) (22.2) (24) (23.4) (20.6) (22.8) (23.8%) males had an ASA score of 4 (20% compared that died by one year following surgery to 15% of females). were aged 80 years or older. Crude mortality Crude mortality rates following surgery rates increased with rising ASA score. for fractured NOF at Waitemata DHB from Patients with an ASA score of 3 or 4 had the 2009–2016 are presented in Table 2. Overall, highest mortality rates of 23.6% and 53.8% a total of 104 (3.7%) patients died in hospital at one year respectively. pre-discharge, 204 (7.2%) within 30 days of Between 2009 and 2016 inclusive, surgery (including inpatient deaths), and mortality rates following surgery for NOF 671 (23.8%) died within one year of surgery. fracture decreased at inpatient, 30-day Inpatient mortality rates for non-surgical and one-year time points (Table 2). There patients were 9.9%, and at 30 days they were was a statistically signifi cant decrease in 15.1%. Further subgroup analyses were mortality over time in analyses adjusted not possible due to small numbers. Crude for gender, age groups and ASA in each mortality rates for females were 5.5% and time period, with odds ratio (OR) ranging 21.2% for 30 days and one year respectively; from 0.86–0.92 (Table 4). At Waitemata DHB and for males were 11.4% and 29.8% for 30 the mean time from admission with NOF days and one year respectively. fracture to surgery decreased from 41 hours Crude mortality rates by age and ASA in 2009 to 31 hours in 2016. There was a score are presented in Table 3. At one year statistically signifi cant association between crude mortality rates ranged from 11.0% time-to-theatre and inpatient (p=0.002), in those aged 60–69 years to 41.8% in those 30-day (p=0.0001), and one-year (p=0.0002) aged ≥90 years. Over 80% of the patients mortality adjusted for gender, age groups

Table 3: Overall crude mortality rates by age and ASA score at Waitemata DHB 2009–2016.

30-day mortality (%) One-year mortality (%) Age* 60–69 2.6 11.0 (years) 70–79 4.1 16.0

80–89 8.6 26.9

≥90 12.7 41.8

ASA score 1 0.8 3.4 2 1.6 8.0

3 5.9 23.6

4 21.7 53.8

*Patients <60 years of age have been excluded due to small numbers (one death).

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 19 www.nzma.org.nz/journal ARTICLE

Table 4: Post-operative NOF fracture mortality over time at Waitemata DHB 2009–2016.

Outcome Analysis OR (95% CI) P Inpatient mortality Unadjusted 0.89 (0.82–0.97) 0.007 Adjusted* 0.86 (0.79–0.94) 0.001

30-day mortality Unadjusted 0.90 (0.85–0.96) 0.001 Adjusted* 0.87 (0.82–0.93) <0.0001

One-year mortality Unadjusted 0.95 (0.91–0.98) 0.004 Adjusted* 0.92 (0.88–0.96) <0.0001

*Adjusted for gender, age groups and ASA. and ASA. The decreasing mean time-to- The results of our study concur with theatre and decreased mortality rates at much of the literature, indicating that Waitemata DHB during the study period is mortality rates following surgery for NOF illustrated in Figure 1. fracture remain between 17–38% at one year.4–10 Previous studies investigating hip Discussion fracture mortality rates within New Zealand hospitals have presented similar results. In this large retrospective cohort study Wimalasena et al (2016) reported 30-day we have shown that overall, 23.8% of mortality of 5%, and 120-day mortality of patients died within the fi rst year following 15% at Auckland City Hospital.23 Davidson surgery for fractured NOF at Waitemata et al (2001) reported that from 329 patients DHB. Patients of male gender, ASA score in Christchurch hospitals, the one-year ≥3, and age 80 years and older, were less mortality rate was 26%.24 Walker et al likely to survive beyond one year following (1999) found in a range of New Zealand a surgical procedure for NOF fracture. Our hospitals that mortality rates following hip results demonstrated a statistically signif- fracture were 8% at 35 days and 24% at icant decrease in mortality rates at inpatient, one year.25 Many studies have investigated 30-day and one-year time periods from 2009 the relationship between demographic to 2016. Further to this, we have found a factors and mortality rates in hip fracture. signifi cant association between reduced Increasing age has been highly associated time-to-theatre and lower mortality rates at with increased mortality.11 Our results inpatient, 30-day and one-year timeframes.

Figure 1: Trends of mean time-to-theatre and inpatient, 30-day and one-year mortality rates (2009–2016).

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 20 www.nzma.org.nz/journal ARTICLE

found that mortality rates increased with aimed at improving patient outcomes. each decade of life after 60 years of age, The ERAS pathway initiated at Waitemata and unsurprisingly the highest mortality DHB aims to provide optimal analgesia rates affl ict those aged 90 years and older. from time of presentation, surgery within With respect to gender, rates of hip fracture 48 hours, perioperative antibiotics for have been reported as much higher in the infection prophylaxis, inpatient anticoagu- female population than in males.2 This is lation for thromboembolism prophylaxis, likely due to increased incidence of osteo- and early mobilisation after surgery. While porosis in females.2 More than 70% of the ERAS protocols have been associated with patients in our study were female. Inter- decreased morbidity and reduced length estingly, we found that males had higher of hospital stay in patients undergoing mortality rates (29.8% compared to 21.2% at major surgery,18–20 there is limited evidence one year) following surgery. Similar fi ndings to suggest they reduce mortality rates in have been reported by Berggren et al (2016) NOF fracture patients. Proudfoot et al and Roche et al (2005) who also found male (2017) found there to be no change in the mortality to be higher than their female average length of stay for fractured NOF counterparts.7,26 It is unclear why men patients following the implementation of have higher mortality rates following NOF ERAS protocols at 18 DHBs across New fracture, perhaps this is due to increased Zealand.18 A meta-analysis by Neuman et al comorbidity and increased prevalence of (2009) reported that deep vein thrombosis, post-operative complications. Kastanis et al pressure ulcers, surgical site infection and (2016) reported that ASA score is strongly urinary tract infection declined by using correlated with post-operative complica- ERAS protocols for hip fracture patients. tions, and that males exhibited greater However, the study did not demonstrate a ASA scores.27 In our study there are similar decrease in inpatient or 30-day mortality.20 proportions of ASA 3 scores across genders, Macfi e et al (2012) performed an audit but slightly more males are ASA 4 (20% on hip fracture patients that were either compared to 15% of females). Despite this, managed using an ERAS pathway or conven- international guidelines have not recom- tional care. The authors reported that while mended gender-specifi c strategies to address the ERAS group had fewer complications, this inequality. there was no signifi cant difference between Our study found there to be a statistically the two groups with regards to 30-day 19 signifi cant decrease in inpatient, 30-day mortality. and one-year crude mortality rates over Although time-to-theatre has been time at Waitemata DHB between 2009 and gradually declining over time, since the 2016. Two large systematic reviews have implementation of the NOF fracture ERAS investigated mortality trends over time in pathway at Waitemata DHB in late 2013 surgically treated hip fractures. The fi ndings we have observed a further decrease in are in contrast to our results. Mundi et al mean time-to-theatre from 40 hours in (2014) reviewed 70 randomised controlled 2013 to 31 hours in 2016. Adjusted analyses trials published between 1981 and 2012 that performed in our study confi rm the asso- measured mortality. They found mortality ciation between time-to-theatre and crude rates to be similar over a 31-year period.28 mortality to be statistically signifi cant. This Haleem et al (2008) performed a systematic supports the notion that reduced time-to- review of all articles published on outcome theatre provides a mortality benefi t for after hip fracture over a four-decade patients with NOF fractures at inpatient, period from 1959 to 1998, which preceded 30-day and one-year time periods. It could our study. They concluded that one-year be suggested that ERAS protocols may also mortality rates have remained essentially have a mortality benefi t, through their static, ranging from 22–29%.29 We speculate role in further reducing time-to-theatre; that decreasing mortality rates over time in however, we have not assessed this to prove our study may be contributed to by imple- cause and effect. Further to this, we have mentation of the ERAS pathway for NOF not addressed whether there is a potential fractures at Waitemata DHB in 2013. morbidity benefi t with fewer post-operative ERAS pathways are evidence-based, complications associated to reduced time-to- multimodal, accelerated clinical pathways theatre; or if there is an optimum “window

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 21 www.nzma.org.nz/journal ARTICLE

of time” from presentation to surgery, in with regard to preparation for theatre and respect of best morbidity and mortality pain management.33 At the beginning of benefi t. Pincus et al (2017) have reported 2018 a ward-based orthogeriatrian has been that among adults undergoing hip fracture appointed full-time to oversee the medical surgery, a wait time greater than 24 hours is management of NOF fracture patients both associated with increased 30-day mortality pre- and post-operatively at Waitemata and complications (myocardial infarction, DHB. In this integrated care model, patients deep vein thrombosis, pulmonary embolus benefi t by receiving specialised geria- and pneumonia).30 trician treatment in collaboration with the International guidelines recommend orthopaedic service. Grigoryan et al (2014) that surgery should occur within 48 hours performed a meta-analysis to determine of admission.14,15 However, the evidence if orthogeriatric models improve patient within the literature is mixed, and based on outcomes. Results of this study supported variable strength data.14 Studies by Major et orthogeriatric collaboration to reduce length al (2016) and Forni et al (2016) both found of stay and improve short-term mortality 34 no correlation between surgical intervention after NOF fracture repair. A study by Förch within 48 hours and survival rates.9,31 The et al (2017) found orthogeriatric care to be authors suggested that early surgery might successful in reducing short-term mortality actually have a negative impact on survival, without showing effect on one-year as more time is needed to medically stabilise mortality. However, the authors reported complex patients. Holt et al (2010) demon- that surviving patients seem to benefi t from 35 strated that patients with major clinical an improved functional outcome. There abnormalities whose surgeries were post- appears to be a paucity of studies providing poned to allow for correction of these evidence that the integrated care model has abnormalities were more likely to survive a long-term mortality benefi t. than those whose operations were not Our study has a number of strengths. The delayed. However, patients whose opera- simple methodology makes it easily repro- tions were delayed but abnormalities were ducible for future studies. While the study unable to be reversed had the poorest has a limited number of variables, the data survival.32 In a retrospective cohort study of presented is considered to be accurate. 42,230 patients, Pincus et al (2017) found a Further to this, the eight-year study period 24-hour wait time threshold: a wait time of captured high patient numbers, allowing greater than 24 hours was associated with for statistically signifi cant evaluation of a greater risk of 30-day mortality and other mortality trends over time. There are also complications; after this time, complications some limitations of this study. Firstly, retro- increased irrespective of the complication, spective cohort studies require accuracy follow-up period or subgroup assessed.30 of administrative data. It is possible that We suggest that individual patient factors ICD-10-AM 8th Ed. diagnosis or other codes, and the potential for reversibility of medical such as ASA, were assigned incorrectly in problems needs to have a structured, coordi- the clinical inpatient system potentially nated management plan using anaesthesia, affecting our data and rates. Secondly, we orthogeriatric medicine and surgical input have made no adjustment for factors such prior to fast-tracking NOF fracture patients as, type of operation and components used, to theatre. intra-operative complications and mode Anaesthetists and orthogeriatrians play of death (eg, accidental death). Thirdly, a crucial role in assessing and resolving there was only limited data on patient medical obstacles to timely surgery; and are comorbidities. integral to the multidisciplinary care of NOF In future, more accurate data collection fracture patients. The International Fragility and linking of databases would allow Fracture Network recently developed a inclusion of place of admission and place consensus statement on the principles of of discharge in analyses as an indication of anaesthesia for patients with hip fracture, functional status. Improved data collection recommending that anaesthetists should would also enable more detailed subgroup be involved in designing and implementing analyses, such as evaluating mortality rates formal hip fracture pathways, particularly in pathological and peri-prosthetic hip

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 22 www.nzma.org.nz/journal ARTICLE

fractures. Future research should seek to the impact of integrated anaesthetist and confi rm a clear reduction in mortality asso- orthogeriatrian models of care on morbidity ciated with decreasing time-to-theatre, and and long-term mortality benefi t. whether an optimal “window of time” exists offering the best mortality benefi t in respect Conclusion of age, gender and comorbidity status. In conclusion, crude mortality rates This will aid decision-making about which following surgery for NOF fracture remain patients to fast-track to theatre versus those above 20% at one year, but have gradually that will benefi t from delay to surgery for declined over recent years at Waitemata medical optimisation. Research is needed to DHB. Patients of male gender, ASA score identify the reasons for mortality overall, the ≥3, and age 80 years and older, have poorer specifi c causes of death (and hence potential survival rates following surgery for NOF reversibility), the effects of frailty, dementia fracture. ERAS protocols have seen a and delirium as well as the relatively high reduction in time-to-theatre, which remains male gender mortality rate. Finally, more closely associated with decreased inpatient, research is needed to further improve 30-day and one-year mortality. NOF fracture-specifi c ERAS pathways and

Competing interests: Dr Misur reports personal fees from Stryker Corporation outside the submitted work. Acknowledgements: We acknowledge and thank Delwyn Armstrong, Head of Analytics, and Monique Greene, Information Analyst, Waitemata DHB for data extraction; and Varsha Parag, Biostatistician, National Institute of Health Innovation, University of Auckland for statistical analysis. Author information: Reuben J Kirk, House Surgeon, Department of Orthopaedics, Waitemata District Health Board, Auckland; Carlene MM Lawes, Public Health Physician, Institute for Innovation and Improvement, Waitemata District Health Board, Auckland; William Farrington, Orthopaedic Surgeon, Department of Orthopaedics, Waitemata District Health Board, Auckland; Peter Misur, Orthopaedic Surgeon, Department of Orthopaedics, Waitemata District Health Board, Auckland; Matthew L Walker, Clinical Director of Orthopaedics, Waitemata District Health Board, Auckland; Michal Kluger, Anaesthetist, Department of Anaesthesiology and Perioperative Medicine, Waitemata District Health Board, Auckland; Min Yee Seow, Orthogeriatrician, Department of Orthopaedics, Waitemata District Health Board, Auckland; Penny Andrew, Director of the Institute of Innovation and Improvement, Waitemata District Health Board, Auckland. Corresponding author: Dr Reuben Kirk, Department of Orthopaedics, Waitemata District Health Board, Private Bag 93-503, Takapuna, Auckland 0740. [email protected] URL: http://www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2019/vol-132-no-1490- 22-february-2019/7813

REFERENCES: 1. Tha HS, Armstrong D, New Zealand 2007–2020. 3. Australian & New Zealand Broad J, et al. Hip fracture Wellington: Osteoporosis Hip Fracture Registry. in Auckland: contrasting New Zealand Inc. [Inter- Annual Report for Hip Frac- models of care in two net]. 2007 [cited 2017 ture Care 2015. [Internet]. major hospitals. Intern Nov 24]. Available from: July 2016 [cited 2018 Apr Med J. 2009; 39(2):89–94. http://www.iofbonehealth. 26]. Available from: http:// 2. Brown P, McNeill R, org/sites/default/fi les/ www.hqsc.govt.nz/assets/ Radwan E, Willingale J. The PDFs/white_paper_new_ Falls/PR/ANZHFR-2016-An- burden of osteoporosis in zealand_2007.pdf nual-Report-Sep-2016.pdf

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 23 www.nzma.org.nz/journal ARTICLE

4. Kilci O, Un C, Sacan O, et with increased mortality fractured neck of femur al. Postoperative mortality of hip fracture patients? in New Zealand orthopae- after hip fracture surgery: Systematic review, dic services. N Z Med J. a 3 years follow up. PloS meta-analysis, and meta-re- 2017; 130(1455):77–90. One. 2016; 11:e0162097. gression. Can J Anaesth. 19. Macfi e D, Zadeh R, 5. Fansa A, Huff S, Ebraheim 2008; 55(3):146–54. Andrews M, et al. Periop- N. Prediction of mortality 13. de Luise C, Brimacombe M, erative multimodal in nonagenarians following Pedersen L, Sorensen HT. optimisation in patients the surgical repair of hip Comorbidity and mortality undergoing surgery for fractures. Clin Orthop following hip fracture: a fractured neck of femur. Surg. 2016; 8:140–5. population-based cohort Surgeon. 2012; 10(2):90–4. 6. Vestergaard P, Rejnmark study. Aging Clin Exp 20. Neuman MD, Archan S, L, Mosekilde L. Increased Res. 2008; 20(5):412–18. Karlawish JH, et al. The mortality in patients with 14. NICE clinical guideline: relationship between a hip fracture - effect of The management of hip short-term mortality and pre-morbid conditions fracture in adults [Inter- quality of care for hip and post-fracture compli- net]. London: NICE. [cited fracture: a meta-analysis cations. Osteoporos Int. 2017 Nov 2]. Available of clinical pathways for 2007; 18(12):1583–93. from: http://www.nice. hip fracture. J Am Geriatr 7. Berggren M, Stenvall M, org.uk/guidance/cg124/ Soc. 2009; 57(11):2046–54. Englund U, et al. Co-mor- resources/hip-fracture-man- 21. Elsworthy AM. ICD-10-AM/ bidities, complications and agement-35109449902789 ACHI/ACS – Eighth edition causes of death among 15. American Academy of overview [Internet]. people with femoral neck Orthopaedic Surgeons. 2012 [cited 2018 Apr 30]. fracture – a three-year Management of Hip Available from: http:// follow-up study. BMC fractures in the Elderly: ro.uow.edu.au/cgi/ Geriatr. 2016; 16:120. doi: Evidence-based clinical viewcontent.cgi?arti- 10.1186/s12877-016-0291-5. practice guideline [Inter- cle=1149&context=ahsri 8. Lisk R, Yeong K. Reducing net]. 2014 [cited 2017 Nov 22. Owens WD, Felts JA, mortality from hip frac- 2]. Available from: http:// Spitznagel EL Jr. ASA phys- tures: a systematic quality www.aaos.org/cc_fi les/ ical status classifi cations: improvement programme. aaosorg/research/guide- a study of consistency of BMJ Qual Improv Rep. lines/hipfxguideline.pdf ratings. Anesthesiology. [Internet]. 2014 [cited 2017 16. Australian and New 1978; 49(4):239–43. Nov 24]; 3(1):4. Available Zealand Hip Fracture 23. Wimalasena B, Harris R. from: http://bmjopenqual- Registry. ANZ guideline Auckland City Hospital’s ity.bmj.com/content/3/1/ for hip fracture care Ortho-Geriatric Service: u205006.w2103.short. [Internet]. 2014 [cited 2017 an audit of patients aged 9. Major LJ, North JB. Nov 2]. Available from: over 65 with fractured Predictors of mortality http://anzhfr.org/wp-con- neck of femur. N Z Med J. in patients with femoral tent/uploads/2016/07/ 2016; 129(1437):15–26. ANZ-Guideline-for-Hip- neck fracture. J Orthop 24. Davidson CW, Merrilees Fracture-Care.pdf Surg. 2016; 24(2):150–2. MJ, Wilkinson TJ, et al. 10. Young W, Seigne R, Bright 17. Ministry of Health. Hip fracture mortality S, Gardner M. Audit of Enhanced Recovery and morbidity--can we morbidity and mortality After Surgery [Internet]. do better? N Z Med J. following neck of femur 2015 [cited 2017 Nov 2]. 2001; 114(1136):329–32. Available from: http:// fracture using the POSSUM 25. Walker N, Norton R, www.health.govt. scoring system. N Z Med Vander Hoorn S, et al. nz/our-work/hospi- J. 2006; 119(1234):U1986. Mortality after hip fracture: tals-and-specialist-care/ 11. Chatterton BD, Moores regional variations in enhanced-recov- TS, Ahmad S, et al. Cause New Zealand. N Z Med J. ery-after-surgery of death and factors 1999; 112(1092):269–71. 18. Proudfoot S, Bennett B, associated with early 26. Roche JJ, Wenn RT, Sahota Duff S, Palmer J. Imple- in-hospital mortality after O, Moran CG. Effect of mentation and effects hip fracture. Bone Joint comorbidities and post- of enhanced recovery J. 2015; 97-B:246–51. operative complications after surgery for hip and 12. Shiga T, Wajima Z, Ohe Y. Is on mortality after hip knee replacements and operative delay associated fracture in elderly people:

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 24 www.nzma.org.nz/journal ARTICLE

prospective observa- between wait time and Fragility Fracture Network tional cohort study. BMJ. 30-day mortality in adults Delphi consensus state- 2005; 331(7529):1374. undergoing hip fracture ment on the principles of 27. Kastanis G, Topalidou A, surgery. JAMA. 2017; anaesthesia for patients Alpantaki K, et al. Is the 318(20):1994–2003. with hip fracture. Anaes- ASA score in geriatric hip 31. Forni S, Pieralli F, Sergi thesia. 2018; 73(7):863–74. fractures a predictive factor A, et al. Mortality after 34. Grigoryan KV, Javedan H, for complications and hip fracture in the Rudolph JL. Orthogeri- readmission? Scientifi ca elderly: the role of a atric care models and (Cairo). 2016; 2016:7096245. multidisciplinary approach outcomes in hip fracture 28. Mundi S, Pindiprolu B, and time to surgery in a patients: a systematic Simunovic N, Bhandari M. retrospective observa- review and meta-analysis. Similar mortality rates in tional study on 23,973 J Orthop Trauma. 2014; hip fracture patients over patients. Arch Gerontol 28(3):e49–e55. doi: 10.1097/ the past 31 years. Acta Geriatr. 2016; 66:13–17. BOT.0b013e3182a5a045. Orthop. 2014; 85(1):54–9. 32. Holt G, Smith R, Duncan 35. Förch S, Kretschmer R, 29. Haleem S, Lutchman L, K, McKeown DW. Does Haufe T, et al. Orthogeriat- Mayahi R, et al. Mortality delay to theatre for ric combined management following hip fracture: medical reasons affect the of elderly patients with trends and geographical peri-operative mortality in proximal femoral fracture: variations over the patients with a fracture of results of a 1-year follow- last 40 years. Injury. the hip? J Bone Joint Surg up. Geriatr Orthop Surg 2008; 39(10):1157–63. Br. 2010; 92(6):835–41. Rehabil. 2017; 8(2):109–14. 30. Pincus D, Ravi B, Wasser- 33. White SM, Altermatt F, stein D, et al. Association Barry J, et al. International

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 25 www.nzma.org.nz/journal ARTICLE

Patterns of referral and uptake of BReast CAncer (BRCA) gene testing of eligible women with ovarian cancer in New Zealand Katherine Fraser, Ai Ling Tan, Carrie Innes, Rosalie Stephens, Amanda Tristram, Simone Petrich, Caroline Lintott, Peter H Sykes, Kimberley Gamet, Alice Christian, Bryony Simcock

ABSTRACT AIMS: To determine the proportion of eligible patients with high-grade serous carcinoma of the ovary, fallopian tube or peritoneum discussed at gynaecological oncology multidisciplinary meetings (MDMs) in New Zealand and subsequently referred for genetic counselling and BRCA pathogenic variant testing. METHODS: Eligible cases were identified from Auckland, Wellington, Christchurch and Dunedin gynaecologic oncology MDM databases between 1 January 2015 to 31 December 2016. Patients who met the eligibility criteria for genetics referral were identified, and cross-referenced against genetic services databases to ascertain the rates of referrals received, the numbers attending appointments, genetic testing o ered and range of results. RESULTS: During the two-year period, 205 patients were eligible for referral. Of these, 143 (70%) patients were referred for genetic counselling with 128 (90%) of this group recommended for BRCA pathogenic variant testing. Of the 126 who undertook the test, results were available for 120 (95%). Nineteen patients (16%) tested positive for a germline BRCA pathogenic variant. CONCLUSIONS: The New Zealand rate of referral to genetic counselling for women with high-grade serous cancer, (HGSC), of the ovary, fallopian tube or peritoneum diagnosed between 2015–2016 is encouraging when compared with others internationally. The rate of BRCA positive pathogenic variants is comparable to international data.

varian cancer is the seventh most or primary peritoneal carcinoma is 1.3%;6 common malignancy among New rising to 44% risk for women with a germ- OZealand females and the fourth line BRCA1 pathogenic variant and 17% for leading cause of female cancer death in this those with a BRCA2 pathogenic variant by country.1 The association between ovarian age 80.7,8 cancer risk and germline pathogenic vari- Determining BRCA pathogenic variant ants in BReast CAncer 1 (BRCA1) and BReast status is increasingly advantageous. Genetic CAncer 2 (BRCA2) genes has been recognised testing of patients enables targeted cascade 2,3 for nearly two decades. Several popula- testing of breast and ovarian cancer-free tion level studies have shown that germ- family members. There is currently no line BRCA pathogenic variants account for effective screening for ovarian cancer, but 11.6–16.6% of invasive ovarian carcinomas procedures such as bilateral salpingo-oopho- and are even more likely to be identifi ed rectomy (RRBSO) may reduce risk of HGSC in women with high-grade serous tumour and play a role in the reduction of risk of 4,5 histology. The lifetime population risk of breast cancer. 9 However it should be noted developing an epithelial ovarian, fallopian that recent studies, including a Cochrane

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 26 www.nzma.org.nz/journal ARTICLE

systematic review, question the quality and Dunedin, Wellington and Auckland Gynae- reliability of this evidence due to analysis cologic Oncology Multi-Disciplinary Tumour bias and study design.10,11 These recent Board Meetings (MDM) 2015–2016 databases. studies suggested the role of RRBSO in Patients who met the eligibility criteria were increasing overall survival and lower HGSC cross-referenced against the records of each and breast cancer mortality for BRCA1 and centre’s local genetic services database. BRCA2 pathogenic variant carriers is of very Clinical data including demographics of low certainty.10 the patients (domicile DHB, ethnicity) and In women with ovarian cancer, BRCA tumour information were collected from pathogenic variants are associated with the MDM database. Further information improved progression-free and overall including who made the referral, patient survival compared to pathogenic variant attendance for counselling, time between negative patients.12 In addition, a new each stage of the referral and testing process, therapy – poly-(ADP)-ribose polymerase family and personal history of breast or (PARP) inhibitor – is becoming established ovarian cancer, and current status were internationally as an effective maintenance collected from each participating hospital’s therapy in women with recurrent ovarian clinical database, electronic medical records cancer.13,14 While PARP inhibitor targeted and genetic service database. This data was therapy is not funded in New Zealand, gathered at each locality and de-identifi ed knowledge of a patient’s BRCA status can prior to central collation. infl uence clinical decision-making. Study cohort and eligibility criteria New Zealand testing guidelines Outcomes recommend BRCA pathogenic variant testing The primary outcome for this audit was to for women under 70, or women over 70 with establish the proportion of eligible women a positive family history of breast or ovarian with high-grade serous cancer of the ovary, cancer, or for whom there is limited infor- fallopian tube or peritoneum discussed at 15 mation available from recorded genealogy. the gynaecological oncology MDMs who However, guidelines internationally are were referred for genetic counselling and 16 shifting towards universal referrals. Despite BRCA pathogenic variant testing. this shift, international studies are showing Secondary outcomes were to determine referral rates of only 23.1% to 51.7%.17,18 which specialties were making the referrals, To date there has been no published data the timeliness of the referrals, genetic exploring New Zealand referral patterns. reviews and reporting of results, the The multi-centre retrospective audit proportion of women who underwent BRCA detailed here examines this country’s rate of pathogenic variant testing and of those, referrals to genetic services for counselling how many had a germline BRCA pathogenic and subsequent testing for women with variant. The study also began to explore high-grade serous ovarian cancer diag- reasons why some women were not referred nosed in 2015–2016. This baseline data will to genetics services or a decision was made allow identifi cation of areas where referral not to proceed with testing. rates can be improved and also provide comparison data for future studies. Statistical analysis The results were largely descriptive, however a Chi-square analysis was used to Methods investigate any differences in referral rate Study design by year, by referral centre and between A multi-centre retrospective audit. Māori vs non-Māori. Data source This study was approved by the New Results Zealand Northern Health and Disability Study cohort Ethics Committee (Approval Number 17/ This audit identifi ed 245 women referred NTA/181, 11 October 2017). Eligible cases to the MDMs with high-grade serous cancer were identifi ed from the Christchurch, of the ovary, fallopian tube or peritoneum.

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 27 www.nzma.org.nz/journal ARTICLE

Table 1: Eligible population demographics in numbers and percentages.

Number (n=205) Percentage (%) Mean age at diagnosis 64.2

Ethnicity NZ European 131 64%

NZ Māori 20 10%

Pacific Island 12 6%

Other 42 20%

Site of primary Ovary 121 59%

Fallopian tube 41 20%

Primary peritoneal 22 11%

Mullerian (unknown) or 2 origins 21 10%

FIGO stage at diagnosis 1 10 5%

2 14 7%

3 103 50%

4 71 35%

Unknown 7 3%

Personal history of breast or ovarian cancer 14 7%

Family history of breast or ovarian cancer 79 39%

Of this group, 32 were excluded due to not referred or not known did not change failure to meet inclusion criteria, ie, were between 2015 and 2016 (69% vs 71%, over 70 without a family or personal history Pearson X2=0.07 P=0.80). There was also of ovarian or breast cancer. Seven women no difference in proportion of Māori vs were excluded due to having previously non-Māori referred 13/20, (65%) compared undergone BRCA pathogenic variant testing to 130/185 (70%) (Pearson X2(1)=0.04 prior to their diagnosis, and one woman P=0.84). There was no difference between was excluded due to missing date of birth. the referral rates from each centre. Thus, 205 women were eligible for referral These ranged from 68% to 75% (Pearson to genetic counselling and consideration X2(3)=0.32, Pr=0.96). However, when consid- for BRCA testing during the two-year study ering the patient’s domicile district health period. Table 1 shows the demographics of board, referral rates ranged from 27% to the eligible population. 100%. There was no clear explanation for Primary outcome this difference as there was no consistent trend in patient characteristics or number Over a two-year period, 143 of the 205 of cases from low referring DHBs that made (70%) eligible patients were referred for them stand out from other DHBs. genetic counselling. Sixty-two (30%) were not referred. The proportion referred vs Figure 1 demonstrates the results at each stage of the referral process.

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 28 www.nzma.org.nz/journal ARTICLE

Figure 1: Flow chart of results of women eligible for referral for BRCA pathogenic variant testing.

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 29 www.nzma.org.nz/journal ARTICLE

Table 2: Demographic data based on results of test for BRCA pathogenic variant.

Result Number Mean age Family Personal Family or personal history n (%) of women (range) history history BRCA1 6 51 (42–66) 3 1 4 (67%)

BRCA2 13 63 (42–88) 9 2 9 (69%)

VUS 1 or 2 9 64 (49–75) 5 2 5 (56%)

Negative 92 62 (39–83) 52 7 57 (62%) (uninformative)

Total 120 62 (39–88) 69 12 75 (63%)

Characteristics of those not samples for four women. Two underwent referred testing but the results were still pending during the data collection phase of the study. Referrals were not made for 62/205 Thus results were available for 120/126 (30%). Of this group, 24/62 (39%) were over (95%) who underwent testing. 70 years and had no personal or family history of breast or ovarian cancer. Five Results of the test (8%) were over 70 years but had a family Of the 128 women recommended for history (n=3), personal history (breast BRCA testing, 120 received results (94%). cancer) (n=1), or limited information was Of these, 19/120 (16%) tested positive for available from recorded genealogy (n=1). a germline BRCA pathogenic variant. The There was no record of why they were not majority of pathogenic variants (13/19, 68%) referred. The remaining 33/62 (53%) were were in the BRCA2 gene, compared to the 70 years or under. Of these, three declined BRCA1 gene (6/19, 32%). In addition, nine and two were seen privately. There was no women (8%) were found to have a variant of record why a referral was not made for the unknown signifi cance in either BRCA1 or 2. remaining 28. However, the majority tested (92/120 [77%]) Attendance and recommendation had a negative or uninformative result. The majority of women diagnosed with a of review germline pathogenic variant in BRCA1 or Of the 143 women who were referred, 2 were under 70 years of age (16/19, 84%). one is currently on the waitlist for genetic Thirteen had a positive family or personal review while 10 (7%) did not attend their breast or ovarian cancer history (13/19, appointments. Reasons for non-attendance 68%). The three with a germline pathogenic included not having any family members at variant in BRCA1 or BRCA2 who were over risk (n=1), unable to be contacted (n=1) or 70 all had positive family histories. Table 2 not wanting genetic review but choosing to demonstrates the demographics of women store DNA for future use (n=2). There was who have been tested and have results for no information available for the other six BRCA pathogenic variant testing. women. The other 132/143 (92%) attended a Who made referrals? Medical oncologists made the majority of genetics review appointment. A more appro- referrals (91/143 [64%]), followed by gynae- priate relative was chosen to undergo BRCA cological oncologists (35/143 [24%]). General testing for one woman. No recommendation practitioners referred 3/143 (2%) while was found for one woman. One woman 14/143 (10%) came from other healthcare declined testing and one woman chose to professionals such as genetic counsellors have her DNA stored but not tested. Overall and oncology clinical nurse specialists. 128/143 (90%) were recommended to receive BRCA testing. Timeliness of referral and process Two women died before testing was of genetic testing performed and their DNA remains stored. Where dates were reported, the average Therefore, 126/143 (88%) underwent genetic time between diagnosis and referral was testing. The laboratory did not receive DNA 2.6 months (SD 3.4, range 0–22 months). It

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 30 www.nzma.org.nz/journal ARTICLE

took on average 3.4 months (SD 2.7, range The majority of referrals (64%) were 0.2–19.6 months) between referral and made by medical oncologists, followed by review by genetic services and on average gynaecological oncologists (24%). With 7.4 months (SD 3.7, range 1.3–21.5 months) no protocol for referral in place, there is between referral and results being available risk that some women may be referred for the patient. There was an average of multiple times while others are missed.20 3.8 months (SD 2.7, range 0.8–15.4 months) International trials investigating different between review and results, assuming the strategies suggest having a single specialty date of review correlates with the date of in charge of referral, or having a specialised blood sample collection. online referral form. Highlighting the need for referral on histopathology reports may Discussion reduce the numbers of multiple or missed referrals.16,21,22 Cohen et al demonstrated The BRCA status of women with ovarian how having a genetic team member present cancer is becoming increasingly important at weekly MDM meetings increased their in guiding oncological management and referral rate from 26.7% to 51.7%.18 giving unaffected pathogenic variant carriers the opportunity to reduce their risk Our results suggest that the referrals that of developing breast or ovarian cancers in are being made are appropriate, with 128 the future. (90%) of those referred offered BRCA testing. It appears there is signifi cant interest among This study showed the nationwide referral patients for testing as 98% (126/128) of rate for women discussed at the central those recommended undertook the tests. gynaecological oncology MDMs with a high- This is a positive outcome when compared grade serous cancer of the ovaries, fallopian to an uptake of 75% reported by Cohen et tubes or peritoneum diagnosed in 2015–2016 al.18 While not a component of this study, was 70%. This rate is higher than that of an previous research has shown that the main Australian study, which showed an overall motivation for testing is the potential benefi t referral rate of 40.6% over a two-year for family members, while the main concern period, and 51.7% over a one-year period is the effect of testing on insurance policies.23 (during which time there was a genetics In New Zealand, insurers cannot make appli- team member present at the weekly MDM).18 cants take genetic tests, however they do The New Zealand referral rate was also require the results of these tests be disclosed higher than the 23% reported in a Canadian when applying for life or health insurance.24 study of women diagnosed with invasive Pathogenic variants in genes, such as serous ovarian cancer between 2002 and BRCA1 and BRCA2, may increase premiums 2009.17 However, it is of concern that 30% or set exclusions on policies. As genetic of eligible women were not referred to this tests become cheaper and more common, important health service. A recent US study insurance may be further impacted. showed disparities in the referral rate for genetic assessment for women with ovarian Each step in the process of referral was cancer with 61% of Caucasians, compared approximately three months with time from to 40% Asians, 38% Latinos and 33% of diagnosis to referral (2.6 months), referral African-Americans referred.19 However, to review (3.4 months) and review to results our study found no statistical difference (3.8 months). This is signifi cantly shorter in referral rates for indigenous Māori than the >3 years median time between compared to non-Māori (65% vs 70%). patients’ initial gynaecological visit and the genetic counselling referral reported in No statistical difference was observed Meyer et al.20 However, a Norwegian study between the centres in which MDMs were showed a median time of 34 days between held. Referral rates ranged from 68% to 75%. diagnosis and blood sampling (review) and This suggests that the barriers to referral are overall time of 52 days (1.4 months) between not centre-specifi c. There was a signifi cant diagnosis and receiving initial results.25 range of referral rates based on patients’ domicile DHB (27–100%). The reasons for The observed BRCA pathogenic variant this remain unclear at this time. rate in those with known results (16%) is similar to those seen in a range of overseas

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 31 www.nzma.org.nz/journal ARTICLE

studies.4,18,21 Of note, 68% of patients had discussions regarding BRCA testing must be pathogenic variants in their germline done with care and consideration. BRCA2 gene compared to 32% with patho- Recommendations genic variants in their BRCA1 gene. This International studies argue for improving is in contrast to current literature, which rates of referral and testing. Further studies suggests BRCA1 pathogenic variants are are required in order to determine how more likely.7,18 Even when the seven women best to achieve this in New Zealand. This excluded from the current study because of retrospective audit of MDM databases and prior testing are included, the New Zealand medical records gives baseline data but ratios still differ to those reported in the does not give an understanding of how local international studies with 14/25 (56%) practices have evolved. Further exploration BRCA2 pathogenic variants found compared of the reasons behind regional variations, to 11/25 (44%) of BRCA1. including the wide range of referral rates Of those who tested positive, 32% (6/19), noted in this study, would be a useful contri- had no family history of breast or ovarian bution to understanding current practice. cancer. This fi nding is largely consistent Some of this research would need to be done with the Australian Ovarian Cancer study’s with individual DHBs. fi nding of 44%.4 These results thus suggest Once the factors impacting on current testing should be performed regardless practice are understood, a further study is of family history.18,20 However, current recommended to consider New Zealand-ap- resourcing means those over 70 in this situ- propriate strategies to enhance referral ation are unlikely to be tested. and testing rates. This could include an International developments include evaluation of the success of strategies used focusing on increasing the number of overseas, an examination of the potential for women tested. Innovations include having building on existing national initiatives such geneticists attend clinic days for coun- as the New Zealand Familial GastrointestinaI selling sessions22 and training oncologists to Cancer Service, and a report on the impact provide BRCA counselling and pathogenic of New Zealand’s health insurance provision variant testing to shorten the referral and and blend of private and public healthcare review components of testing.26 on any alternatives under consideration. In a ‘Mainstreaming’ model, patients It is recommended that establishing suitable for genetic testing are identifi ed national standards for referrals and testing at MDM meetings, with testing ordered should be examined as one potential by the treating physician after abbre- strategy. Areas of practice identifi ed in viated pre-testing counselling.18,22 A this study that could usefully be addressed number of overseas centres have adopted through national standards include the this approach. Auckland City Hospital is establishment of a systematic approach to reportedly planning to shift to this model in referrals and the setting of standards for the near future.27 timeframes, including those between diag- However, while routine testing would nosis, testing and receiving results. increase the number of women with known Finally, it is recommended that further pathogenic variants, it could create new study be conducted to determine the ways issues with the potential for the system to in which counselling services can be most be overwhelmed with lower-risk women, responsive to changing needs of patients leading to delayed results for those at and physicians. higher risk. While shifting pre-counselling to oncologists or having no pre-test coun- Limitations A signifi cant limitation of this study is selling would relieve the burden on genetic the retrospective design and reliance on services, there are concerns women might databases. This inherently leads to issues not be adequately prepared for the impli- pertaining to data collection and entry cations of the results for either themselves such as ascertainment bias. However, or their families.26 As the diagnostic phase having access to the MDM database, Genetic is already a sensitive and emotional time, Services database and patient electronic

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 32 www.nzma.org.nz/journal ARTICLE

medical records enabled a level of corrobo- ration. Data collection was confi ned to the Conclusion women presented at MDMs. While every This study has shown that New Zealand woman with a gynaecological cancer should has a relatively high rate of referral for be presented to an MDM, it is possible some women with high-grade serous ovarian were not including those who may not have cancer to genetic services for germline BRCA wanted treatment or follow up, or passed testing. However, there remains a signifi cant away in the interim. Data collection was portion of the population who are missing limited by the availability of electronic out on referral. Through identifi cation of medical records for a number of patients. barriers and implementation of strategies These patients were typically from DHBs to improve systems, referral rates may outside of the main centres where data increase further. Germline BRCA pathogenic could not be accessed off-site, or were variant testing is increasingly important women seen in private practice. Despite to the identifi cation of pathogenic variant these limitations, the fi ndings are useful for carriers so that best practice care can be establishing baseline data and determining offered to patients with ovarian cancer and the need for future services. their families.

Competing interests: Nil. Author information: Katherine Fraser, Medical Student, University of Otago, Christchurch; Ai Ling Tan, Gynaecological Oncologist, National Women’s Health, Auckland District Health Board, Auckland; Carrie Innes, Research Fellow, University of Otago, Christchurch; Rosalie Stephens, Medical Oncologist, Auckland District Health Board, Auckland; Amanda Tristram, Gynaecological Oncologist, Obstetrics and Gynaecology, Wellington Hospital, Wellington; Simone Petrich, Gynaecological Oncologist, Dunedin Hospital, Dunedin; Caroline Lintott, Senior Genetic Associate/Team Leader, Genetic Health Service NZ, Southern Hub, Christchurch; Peter H Sykes, Senior Lecturer, Department of Obstetrics & Gynaecology, University of Otago, Christchurch & Gynaecological Oncologist, Christchurch Women’s Hospital, Christchurch; Kimberley Gamet, Team Leader/Senior Genetic Counsellor, Genetic Health Service NZ, Northern Hub, Auckland; Alice Christian, Senior Genetic Counsellor/Team Leader, Genetic Health Service NZ, Central Hub, Wellington; Bryony Simcock, Senior Lecturer, Department of Obstetrics & Gynaecology, University of Otago, Christchurch & Gynaecological Oncologist, Christchurch Women’s Hospital, Christchurch. Corresponding author: Dr Bryony Simcock, Department of Obstetrics & Gynaecology, University of Otago, Christchurch Women’s Hospital, Private Bag 4711, Christchurch 8140. [email protected] URL: http://www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2019/vol-132-no-1490- 22-february-2019/7814

REFERENCES: 1. Cancer: New registrations for the breast and ovarian cancer susceptibility gene and deaths 2013. In: Health cancer susceptibility gene BRCA2. Nature. 1995; Mo, editor. Wellington: BRCA1. Science. 1994; 378(6559):789–92. Ministry of Health; 2016. 266(5182):66–71. 4. Alsop K, Fereday S, 2. Miki Y, Swensen J, Shat- 3. Wooster R, Bignell G, Meldrum C, deFazio A, tuck-Eidens D, Futreal PA, Lancaster J, Swift S, Seal Emmanuel C, George Harshman K, Tavtigian S, S, Mangion J, et al. Iden- J, et al. BRCA Mutation et al. A strong candidate tifi cation of the breast Frequency and Patterns

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 33 www.nzma.org.nz/journal ARTICLE

of Treatment Response in salpingo-oophorectomy 16. Bednar EM, Oakley HD, BRCA Mutation–Positive in women with BRCA1 Sun CC, Burke CC, Munsell Women with Ovarian or BRCA2 mutations. MF, Westin SN, et al. A Cancer: A Report from the Cochrane Database Syst universal genetic testing Australian Ovarian Cancer Rev. 2018; 8:Cd012464. initiative for patients with Study Group. Journal 11. Heemskerk-Gerritsen BA, high-grade, non-mucinous of Clinical Oncology. Seynaeve C, van Asperen epithelial ovarian cancer 2012; 30(21):2654–63. CJ, Ausems MG, Collee and the implications 5. Nelson HD, Huffman LH, JM, van Doorn HC, et al. for cancer treatment. Fu R, Harris EL, Force Breast cancer risk after Gynecologic Oncology. USPST. Genetic risk salpingo-oophorectomy 2017; 146(2):399–404. assessment and BRCA in healthy BRCA1/2 17. Demsky R, McCuaig J, mutation testing for mutation carriers: Maganti M, Murphy breast and ovarian cancer revisiting the evidence KJ, Rosen B, Armel SR. susceptibility: systematic for risk reduction. J Natl Keeping it simple: Genetics evidence review for the Cancer Inst. 2015; 107(5). referrals for all invasive U.S. Preventive Services 12. Hyman DM, Zhou Q, serous ovarian cancers. Task Force. Ann Intern Iasonos A, Grisham RN, Gynecologic Oncology. Med. 2005; 143(5):362–79. Arnold AG, Phillips MF, 2013; 130(2):329–33. 6. SEER Cancer Statistics et al. Improved survival 18. Cohen PA, Nichols CB, Review, 1975–2011 for BRCA2-associated Schofi eld L, Van Der Werf [Internet]. National serous ovarian cancer S, Pachter N. Impact of Cancer Institute. Available compared with both Clinical Genetics Atten- from: http://seer.cancer. BRCA-negative and dance at a Gynecologic gov/csr/1975_2011/ BRCA1-associated serous Oncology Tumor Board 7. Arts-de Jong M, de Bock GH, ovarian cancer. Cancer. on Referrals for Genetic van Asperen CJ, Mourits 2012; 118(15):3703–9. Counseling and BRCA Muta- MJ, de Hullu JA, Kets CM. 13. Coleman RL, Oza AM, tion Testing. Int J Gynecol Germline BRCA1/2 muta- Lorusso D, Aghajanian C, Cancer. 2016; 26(5):892–7. tion testing is indicated Oaknin A, Dean A, et al. 19. Manrriquez E, Chapman in every patient with Rucaparib maintenance JS, Mak J, Blanco AM, Chen epithelial ovarian cancer: treatment for recurrent L-m. Disparities in genetics A systematic review. Eur J ovarian carcinoma after assessment for women with Cancer. 2016; 61:137–45. response to platinum ovarian cancer: Can we do 8. Kuchenbaecker KB, therapy (ARIEL3): a better? Gynecologic Oncolo- Hopper JL, Barnes DR, randomised, double-blind, gy. 2018; 149(1):84–8. Phillips KA, Mooij TM, placebo-controlled, phase 20. Meyer LA, Anderson Roos-Blom MJ, et al. Risks 3 trial. Lancet. 2017; ME, Lacour RA, Suri A, of Breast, Ovarian, and 390(10106):1949–61. Daniels MS, Urbauer DL, Contralateral Breast Cancer 14. Mirza MR, Monk BJ, Herrst- et al. Evaluating women for BRCA1 and BRCA2 edt J, Oza AM, Mahner S, with ovarian cancer Mutation Carriers. JAMA. Redondo A, et al. Niraparib for BRCA1 and BRCA2 2017; 317(23):2402–16. Maintenance Therapy in mutations: missed oppor- 9. Kauff ND, Domchek SM, Platinum-Sensitive, Recur- tunities. Obstet Gynecol. Friebel TM, Robson ME, rent Ovarian Cancer. New 2010; 115(5):945–52. Lee J, Garber JE, et al. England Journal of Medi- 21. Petzel SV, Vogel RI, Risk-reducing salpin- cine. 2016; 375(22):2154–64. Bensend T, Leininger A, go-oophorectomy for the 15. NSW CI. Genetic testing Argenta PA, Geller MA. prevention of BRCA1- and for heritable mutations in Genetic Risk Assessment BRCA2-associated breast BRCA1 and BRCA2 genes for Women with Epithelial and gynecologic cancer: NSW: Cancer Institute Ovarian Cancer: Referral a multicenter, prospec- NSW; 2010 [updated Patterns and Outcomes in tive study. J Clin Oncol. 14/6/2018. Available a University Gynecologic 2008; 26(8):1331–7. from: http://www.eviq. Oncology Clinic. Journal 10. Eleje GU, Eke AC, Ezebi- org.au/cancer-genetics/ of Genetic Counseling. alu IU, Ikechebelu JI, genetic-testing-for-her- 2013; 22(5):662–73. Ugwu EO, Okonkwo OO. itable-mutations/620-ge- 22. Kentwell M, Dow E, Antill Risk-reducing bilateral netic-testing-for-herita- Y, Wrede CD, McNally O, ble-mutations-in-the

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 34 www.nzma.org.nz/journal ARTICLE

Higgs E, et al. Mainstream- Human Rights Commis- in women with ovarian ing cancer genetics: A sion; 2007. Report No.: cancer: results of the Genet- model integrating germline 978-478-28641-0. ic Testing in Epithelial BRCA testing into routine 25. Hoberg-Vetti H, Bjorvatn Ovarian Cancer (GTEOC) ovarian cancer clinics. C, Fiane BE, Aas T, Woie K, study. J Med Genet. Gynecologic Oncology. Espelid H, et al. BRCA1/2 2016; 53(10):655–61. 2017; 145(1):130–6. testing in newly diagnosed 27. Wong E, Wilson M, Chrystal 23. Metcalfe K, Liede A, Hood- breast and ovarian cancer K, Wilson L, Stephens R. far E, Scott A, Foulkes W, patients without prior A retrospective audit of Narod S. An evaluation of genetic counselling: the compliance with BRCA needs of female BRCA1 and DNA-BONus study. Euro- testing guidelines in the BRCA2 carriers undergoing pean Journal of Human Medical Oncology clinic genetic counselling. Journal Genetics. 2016; 24(6):881–8. in Auckland City Hospital. of Medical Genetics. 26. Plaskocinska I, Shipman H, Poster presented at: 2000; 37(11):866–74. Drummond J, Thompson New Zealand Society of 24. Guidelines: Insurance E, Buchanan V, Newcombe Oncology 49th Annual and the Human Rights B, et al. New paradigms Meeting; Auckland, New Act 1993. New Zealand: for BRCA1/BRCA2 testing Zealand Oct 14–15 2017.

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 35 www.nzma.org.nz/journal ARTICLE

The cost of teri unomide in the treatment of relapsing- remitting multiple sclerosis J Alasdair Millar

ABSTRACT AIMS: Teriflunomide, used globally to treat multiple sclerosis (MS) and widely subsidised for this indication including in Australia and New Zealand, is the main metabolite of leflunomide, an older immune- modulating drug. Leflunomide therefore represents a potential alternative therapy for MS. Teriflunomide is about 50–500 times more expensive than leflunomide, depending on prices in each jurisdiction. I wished to study how this situation arose. METHODS: Web search to obtain the publicly available minutes of eight international regulatory bodies that have approved teriflunomide for the governments of the US, Canada, Europe, England, Scotland, Australia (TGA and PBS) and New Zealand, and examination of the processes and minuted discussions concerning the metabolic, e icacy, toxicity and cost relationship between teriflunomide and leflunomide. RESULTS: The relationship between the two drugs and their relative e icacy or toxicity in MS was considered by three of eight agencies (Food and Drug Administration (FDA), European Medicines Agency (EMA) and the Canadian Agency for Drugs and Technology in Health (CADTH)). The remaining agencies accepted teriflunomide applications at face value, assessed cost-e ectiveness against contemporaneous drugs used for treating MS, and did not discuss the potential role of leflunomide as a therapy for MS. No agency minuted the implications of the cost di erence. CONCLUSIONS: E icacy for leflunomide in MS is likely but unproven. The sponsor presented a case for teriflunomide that was within the established procedures for drug agencies in establishing cost-e ectiveness, and agencies did not stray from their normal procedures. As a result, an opportunity to decrease the cost of treating MS has been missed. Though o -label use of leflunomide is possible, this is unlikely without a publicly-funded trial to demonstrate non-inferiority with regard to e icacy and safety.

n several jurisdictions worldwide, in- and safety in the treatment of disease. cluding New Zealand and Australia, new Sponsor applications for registration or Idrugs are subjected to an examination of subsidy deploy the best possible arguments, cost-effectiveness as a condition of govern- including use of economic models based on ment subsidy. This process is subject to sev- clinical trials, to argue that the drug is as eral imponderables that can be infl uenced or more cost-effective compared to existing by patient or specialist lobby groups, or drug treatments. The sponsors often design and manufacturers, to maximise the chances of fund the clinical trials intended to demon- success in the subsidy application. The usual strate equivalence or superiority when applicant for subsidy (the ‘sponsor’) is the compared against current standard therapy. pharmaceutical company that developed the By so doing they defi ne a ‘comparator’ agent, drug, in the case of terifl unomide, Sano- the cost of which is included in the economic fi -Aventis or its subsidiary Genzyme. model and which is thus used to justify a The development process is expensive. It minimum or higher price. Naturally, the includes devising the basic chemistry of the sponsor has an interest in achieving the drug, synthesising it in volume production, highest possible price in order to obtain an testing its physiological effects in animal acceptable return on its sizeable investment models and volunteers, registering it for a and to create profi t for shareholders. Teri- clinical indication, and defi ning the effi cacy fl unomide (Aubagio) is a good example of these processes.

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 36 www.nzma.org.nz/journal ARTICLE

Terifl unomide was granted subsidy in example, in New Zealand the cost of lefl un- both New Zealand and Australia following omide is NZ$2.90 for 30 tablets (9.7 c/tab!) assessment by the respective advisory versus NZ$1,582.62 for 28 tablets of teri- committees: the Pharmacology and Thera- fl unomide (NZ$56.52/tab),10 a price ratio peutics Advisory Committee (PTAC) (which per tablet of 584. Corresponding fi gures advises PHARMAC and in turn the New in Australia are AU$1.47 and $65.62 per Zealand Government)1 and the Pharmaceu- tablet, a ratio of 44.6.11 The low New Zealand tical Benefi ts Advisory Committee (PBAC),2 price for lefl unomide presumably refl ects which advises the Australian Government the PHARMAC drug tendering process as via the Department of Health. As licensing applied to generic drugs. A respectable and subsidy approvals are separate logical but unproven case can be made processes in Australia, the drug was also that terifl unomide, which has been deter- examined by the Therapeutic Goods mined in each of the above jurisdictions Administration (TGA).3 Terifl unomide was to be cost-effective compared to current also registered in the US (Food and Drugs treatments for multiple sclerosis, would be Administration, FDA)4 and subsidised across dominated by lefl unomide in an economic Europe following the advice of the European model that used it as a comparator and Medicines Agency (EMA),5 and in England assumed therapeutic and safety equiva- (National Institute of Health and Care Excel- lence. Or to put it another way, it could be lence, NICE),6 Scotland (Scottish Medicines argued that having demonstrated cost-ef- Consortium, SMC)7 and Canada (Canadian fectiveness of terifl unomide, prescribing Agency for Drugs and Technologies in lefl unomide to patients with multiple Health, CADTH).8 Subsidised availability in sclerosis would be reasonable because of these jurisdictions is for the treatment of its pro-drug status, would probably obtain relapsing-remitting multiple sclerosis, with identical clinical benefi ts, be more cost-ef- minor variations among jurisdictions such fective and avoid substantial costs. as restriction, or not, to single drug use. It is legitimate to ask why this logic was Terifl unomide is the active metabolite of not used by any of the international drug lefl unomide, an older immunosuppressant regulatory bodies that considered the regis- drug also developed by Sanofi -Aventis. tration and subsidy of terifl unomide. Lefl unomide was registered in 1998 for use as a biological DMARD (disease modifying Methods anti-rheumatic drug) in rheumatoid and This discussion is based on internet psoriatic arthritis. Seventy percent of lefl un- searches entitled ‘terifl unomide minutes in omide is metabolised to terifl unomide.4,9 [name of jurisdiction]’ or specifi c searches Lefl unomide has four minor metabolites with the search term ‘terifl unomide’ within whose biological activity is uncertain. websites for specifi c regulatory bodies such Thus lefl unomide is a pro-drug for teri- as the TGA, PBAC and PTAC. Countries such fl unomide, and terifl unomide is the active as Switzerland where the relevant docu- metabolite. The unit tablet sizes are 20 and ments are not in English were excluded. 14mg daily respectively. A patient taking a The minutes obtained were scrutinised to daily dose of 20mg of lefl unomide generates, determine the extent to which a connection by metabolism at the above rate, the same between lefl unomide and terifl unomide quantity of terifl unomide per day that is in was acknowledged and further discussed, 4 one tablet of that drug. It therefore appears and the nature of these discussions. Some logical that, other things being equal, a available minutes are of a summary nature benefi t of terifl unomide 14mg daily in any and the full minutes are confi dential. For registered indication for either would be example, in Australia only a Public Summary obtained if the patient was instead given Document of PBAC decisions is published. lefl unomide 20mg. However, lefl unomide Subsidy data for PBS (Australian) use of has not been tested in multiple sclerosis, nor terifl unomide for the 2017–2018 fi nancial has terifl unomide been trialled in rheu- year were obtained from Australian matoid or psoriatic arthritis. Government Department of Human In all jurisdictions, there is a substantial Services website.12 cost difference between the two drugs. For

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 37 www.nzma.org.nz/journal ARTICLE

Table 1: Drug regulatory agency minute contents relating to comparisons between lefl unomide and terifl unomide. The ‘IMPLIED’ entry for CADTH discussions of safety matters is explained in the text.

Drug Metabolic link Possible use of Discussion of Discussion Discussion agency acknowledged leflunomide in therapeutic of safety of price MS mentioned equivalence equivalence implications FDA YES NO YES YES NO

TGA YES NO YES (BRIEFLY) NO NO

PBAC YES NO NO NO NO

PTAC YES NO NO NO NO

CADTH YES NO NO IMPLIED NO

NICE YES NO NO NO NO

SMC YES NO NO NO NO

EMA YES NO YES* YES NO

*In discussion over whether teriflunamide should be assigned ‘new drug’ status. For agency acronyms, see text.

arose around the question of whether teri- Results with comment fl unomide should be allocated ‘new drug’ The results are summarised in Table 1. status, as discussed below. All agency minutes acknowledged, usually The remaining agencies adopted a tradi- in a single sentence, that terifl unomide is tional approach in which the comparator the predominant metabolite of lefl unomide. drug or drugs (mainly interferon β 1a, Only the FDA and EMA minutes4,5 state that interferon β 1b or glatiramer) were appar- the 20mg tablet of lefl unomide will generate ently accepted at face value and the the unit dose (14mg) of terifl unomide, and economic models were scrutinised in the only these two agencies explicitly discuss the usual way, taking into account the trials implications for this relationship for either against placebo and active drug that had effi cacy or safety (particularly the latter). been organized by the sponsor. Each agency One FDA delegate was critical of the sponsor concluded that terifl unomide had modest for allegedly seeking to minimise or obscure effi cacy advantages over these drugs, and the toxicity of terifl unomide given the known its ability to be administered by mouth was toxicity of lefl unomide, especially hepato- given due weight. Thus it was accepted toxicity and toxicity related to pregnancy.12 as being economically dominant over These were aspects that in the past nearly led contemporaneous treatments, and subsi- to the withdrawal of lefl unomide. Suggested dised accordingly. The relative clinical and brief organ-specifi c toxicity proposed by economic performance over fi ngolimod the sponsor for the US terifl unomide Drug (also given orally) was not tested. Product Information sheet were rejected in most cases by the FDA in favour of the Extracts from agency reports descriptions already listed for lefl unomide.12 According to the EMA report,5 “Teri- Thus in effect, the FDA suggested that the two fl unomide is the active main metabolite of drugs must have the same toxicity. However, lefl unomide. The non-clinical development the FDA did not discuss any effect of drug of terifl unomide was originally guided by similarity on whether terifl unomide should experience gained with the parent compound be given marketing approval. lefl unomide in terms of study designs and anticipated target organ toxicities. Never- CADTH provided a summary of lefl un- theless, an independent self- standing omide toxicity for each organ system before development of terifl unomide was later discussing the corresponding trial evidence pursued non-clinically [refers to pre-clinical for terifl unomide, strongly implying by that development: Auth], because of the intended device that similarity of toxicity was antic- use of terifl unomide as treatment in multiple ipated.8 In the case of EMA, the discussion sclerosis as opposed to the different indication

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 38 www.nzma.org.nz/journal ARTICLE

of lefl unomide (Report, p 14)…Terifl unomide as an investigational drug for the treatment of exposure after a single 20mg dose of lefl un- multiple sclerosis (MS) is based largely on the omide was ~70% of that after a single 20mg ability to administer the active drug directly. dose of terifl unomide. Lefl unomide was Direct ingestion of only the active metabolite shown to be effective in patients with rheu- negates the need for formation of the active matoid arthritis at doses of 10 and 20mg. In metabolite and ensures a more consistent Phase 2 of development of lefl unomide, higher delivery and higher bioavailability inde- incidences of adverse events were observed pendent of enzymic conversion which may be with the dose of 25mg and consequently, the impaired in the presence of concomitant drugs decision was made to test the doses of 10mg or under certain disease conditions.” These and 20mg of lefl unomide in Phase 3 studies. are hardly substantial advantages. The FDA These results were also taken into account report also voiced strong criticisms related for the development of terifl unomide and the to toxicity. The FDA required rewriting of decision was made not to test doses of teri- detailed toxicity by organ to ensure that the fl unomide higher than 14mg. The applicant descriptions matched the corresponding made a presumption of a 1:1 transmission of entries for lefl unomide. One contributor effective doses developed in RA to MS referring was scathing over a six-year delay in notifi - to the hypothesis that in both indications, the cation of a case of toxicity, and in describing effect of terifl unomide on the DHO-DH enzyme alleged behaviour of the sponsor in mini- would result in a clinical effect” (Report, p 33). mising terifl unomide toxicity. The delegate These entries suggest that the EMA accepted stated “It is understandable that the sponsor that the actions of lefl unomide and terifl un- might like to extend the useful life of an old omide would be identical. drug lefl unomide, now off patent, by seeking The EMA also assessed whether terifl un- new patients and new uses for terifl un- omide is a ‘new drug’ because of its policy omide, but not by concealing adverse events that “The different salts, esters, ethers, to physicians who may not be fully aware of isomers, mixtures of isomers, complexes or them”. It is not clear whether the charge of derivatives of an active substance shall be “concealing” has any evidential basis. considered to be the same active substance, unless they differ signifi cantly in prop- Discussion erties with regard to safety and/or effi cacy”. The usual comparator drug accepted Whether terifl unomide is a new drug thus in pharmacoeconomic analysis is the relied on the interpretation of “derivatives”. drug most likely to be replaced in clinical The complex argument on this point makes practice. In Australia, this may be the interesting reading. One difference cited by PBS-listed drug for the same indication, the sponsor was the potential for greater but unlisted drugs or even “existing phar- hepatotoxicity with lefl unomide because its macological analogues” may serve as metabolism generates 4 minor metabolites comparators.13 This makes sense because not present after terifl unomide. However, the economic effect of a new drug can the agency felt that “…the potentially lower only be judged according to a comparison hepatotoxic potential of terifl unomide is of with current treatment and how it will be minor clinical relevance with no signifi cant improved by substitution of its effects on benefi ts for patients treated with terifl un- disease outcomes and toxic reactions by omide”. The conclusion was that though the new agent. There is no provision in any terifl unomide satisfi ed other criteria for jurisdiction for considering a metabolic marketing, it was not a new active drug, and precursor as a comparator drug and indeed further, that the effects of lefl unomide were this does not make sense, and is not possible wholly mediated via terifl unomide. The according to current procedures, if the sponsor appealed this decision using even parent drug is not registered and therefore more complex arguments and succeeded in not used in the proposed indication. overturning it on the basis of possible toxicity The questions arising from this work are differences. A dissenting opinion to the whether the benefi t of terifl unomide might effect that the evidence of such a difference be achieved at a fraction of the cost by using was unacceptably weak was signed by nine lefl unomide, whether a predominant metab- members of the Committee and published as olite of a pro-drug should be given new an Appendix to the EMA report. drug status, and why these questions and 12 The FDA documents state that “The the matter of relative cost were apparently advantage of terifl unomide versus lefl unomide

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 39 www.nzma.org.nz/journal ARTICLE

not considered by any of eight international assume that the two drugs differ in either drug regulatory agencies. their mode of action or clinical effects, and With the possible qualifi cation related to hence no scientifi c rationale to suggest that the comments in FDA minutes over minimi- terifl unomide might be more effective than sation of potential toxicity of terifl unomide, lefl unomide in MS, and scant evidence of a the sponsor of terifl unomide made legal difference in toxicity. The opportunity for use of the systems created internationally a drastic reduction in the cost of treating for marketing and subsidy of new drugs, multiple sclerosis by seeking a registration as it was entitled to do. It is likely that for lefl unomide and testing that drug in Sanofi -Aventis/Genzyme recognised the patients with MS was not in the sponsor’s possibility of marketing as a new drug the commercial interest. The agencies may principal metabolite of the out-of-patent have discussed but were unable to apply drug lefl unomide, in a high-cost indication the presumptive argument that lefl unomide which was not a registered indication for the and terifl unomide were likely to be equiv- parent drug. If the ‘new’ drug was shown alent, because of absence of trial data to in trials to be effective when tested against confi rm it. They were obliged to deal with existing costly therapies for MS, the sponsor the applications before them using prede- could seek a high price while satisfying termined procedures in good faith, and the usual criteria of cost-effectiveness. In did so. On the positive side, it can be said spite of reservations regarding new drug that the global success of the terifl unomide status and criticisms over relative toxicity, applications expanded the therapeutic it appears that under prevailing regulatory options available to neurologists in the systems, success for the sponsor was likely treatment of MS and did not lead to a signif- from the outset, as the applications satisfi ed icant increase in the cost of treating MS the defi ned scientifi c and economic criteria compared to other treatments. for granting subsidy that exist similarly worldwide. Nevertheless, it was a high-risk Conclusion strategy, as no prior human evidence of Governments in states that have subsi- possible effi cacy for terifl unomide existed dised terifl unomide pay about 50–500 (studies in animal models had suggested times the cost (depending on the local a benefi t). The company also faced the prices including secretly negotiated prices tangible risk that one or more of the regu- for terifl unomide) arising if the patients latory agencies it had to deal with would were given lefl unomide 20mg daily. This not recognise terifl unomide as a new drug would generate by metabolism the same or would ask why it should not be adminis- dose of terifl unomide that is now used in tered as lefl unomide, leading to application clinical practice. In Australia, the cost to failure. In the event, most agencies (and government for terifl unomide amounted to the governments they report to) treated AU$29.7m in the year to 30 June 2018.14 The the terifl unomide application in a routine estimated corresponding cost of lefl unomide manner, and probing questions about the is about $666,000, subject to variation in the relationship with lefl unomide were not mix of patient category under the scheme, asked or, if discussed, not minuted. Though implying a possible saving of around $29m. absence of a formal minute does not neces- Thus a potential opportunity to decrease sarily mean that an issue was not discussed, substantially the cost of treating MS may in this case it is possible that discussions have been missed. The only routes to obtain ensued but remained unminuted because the cheaper option now lie in (a) exec- no decision in respect of lefl unomide was utive action by a drug regulatory body or possible in the absence of supporting government to register lefl unomide in MS; trial data. This possibility and this reason (b) off-label use of lefl unomide by neurolo- would not, however, have prevented any gists; or (c) by the design and execution of committee from minuting the detail of the a randomised comparative trial of lefl un- metabolic relationship between the drugs, omide versus terifl unomide in MS, funded their common mechanism of action, the by a non-commercial organisation such as dosing equivalence and the substantial cost NIH, MRC or NH&MRC. Given that equiv- differences, to allow consideration by higher alence has not been demonstrated and government authority. possible differences in toxicity, the fi rst two These events happened in circumstances options are not attractive, and only the third where there was no prima facie reason to option appears practicable.

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 40 www.nzma.org.nz/journal ARTICLE

Competing interests: Nil. Author information: J Alasdair Millar, Physician and Clinical Pharmacologist, Adjunct Clinical Professor, Curtin University Medical School, Bentley, Western Australia 6102, Australia. Corresponding author: J Alasdair Millar, Physician and Clinical Pharmacologist, Adjunct Clinical Professor, Curtin University Medical School, Bentley, Western Australia 6102, Australia. [email protected] URL: http://www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2019/vol-132-no-1490- 22-february-2019/7815

REFERENCES: 1. Pharmaceutical Manage- sessment-report_en.pdf 10. New Zealand Pharmaceu- ment Agency, New Zealand Accessed 7/11/18. tical Schedule. Section H (PHARMAC). Minutes of 6. UK Department of Health. (Hospital Pharmaceuticals). a meeting of the Pharma- National Institute for Pharmaceutical Manage- cology and Therapeutics Health and Care Excellence. ment Agency (PHARMAC) Advisory Committee. http:// http://www.nice.org.uk/ New Zealand. 1 July 2018. www.pharmac.govt. guidance/ta303/documents/ 11. Australian Government, nz/assets/ptac-minutes- multiple-sclerosis-relaps- Pharmaceutical Benefi ts 2014-11-updated.pdf ing-terifl unomide-fi nal-ap- Scheme. http://www. Accessed 13/11/2018. praisal-determination3 pbs.gov.au/pbs/home 2. Australian Government, Accessed 13/11/2018. Accessed 13/11/2018. Pharmaceutical Benefi ts 7. National health Service, 12. Food and Drug Adminis- Scheme. Public Summary Scotland. Scottish tration, USA government. Document for terifl uno- Medicines Consortium Food and Drug Adminis- mide. http://www.pbs.gov. Advice. Terifl unomide tration, USA government. au/info/industry/listing/ (Aubagio). http://www. http://www.accessdata. elements/pbac-meetings/ scottishmedicines.org. fda.gov/drugsatfda_docs/ psd/2013-07/terifl unomide. uk/medicines-advice/ nda/2012/202992Orig- Accessed 13/11/2018. terifl unomide-aubagio-full- 1s000OtherR.pdf Accessed 3. Australian Government, submission-94014/ 13/11/2018. (Other reviews: Department of Health and Accessed 13/11/2018. noting safety data). Ageing. Therapeutic Goods 8. Canadian Agency for Drugs 13. Australian Government, Administration. Extract and Technologies in Health Department of Health. from the clinical evaluation (CADTH) http://www.cadth. Guidelines for preparing a report for terifl unomide. ca/sites/default/fi les/cdr/ submission to the Pharma- http://www.tga.gov.au/ clinical/SR0350_Auba- ceutical Benefi ts Advisory fi le/1630/download. gio_CL_Report_e.pdf Committee (Version 5.0). Accessed 13/11/2018. Accessed 13/11/2018. http://pbac.pbs.gov.au/ 4. Food and Drug Adminis- 9. Kalgutkar AS, Nguyen HT, content/information/fi les/ tration, USA government. Vaz ADN, et al. In vitro pbac-guidelines-version-5. Center for Drug Evaluation metabolism studies on pdf Accessed 7/11/20. and Research http:// the isoxazole ring scission 14. Australian Government, www.accessdata.fda.gov/ in the anti-infl ammatory Department of Human drugsatfda_docs/nda/2012/ agent lefl unomide to Services. Pharmaceutical 202992Orig1s000MedR. its active α-cyanoenol Benefi ts Schedule Item pdf Accessed 13/11/2018. metabolite A771726: Reports. http://medicar- 5. European Medicines mechanistic similarities estatistics.humanservices. Agency. Assessment Report with the cytochrome gov.au/statistics/pbs_item. – Aubagio. http://www. P450-catalyzed dehydra- jsp (report generated using ema.europa.eu/docu- tion of aldoximes. Drug Item No 2898M) Accessed ments/assessment-report/ Metabolism and Disposi- 8/11/201 8. aubagio-epar-public-as- tion 2003; 31:1240–50.

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 41 www.nzma.org.nz/journal ARTICLE

Multimorbidity and multiple social disadvantage in a New Zealand high-needs free primary healthcare clinic population: a cross- sectional study Sharmaine Sreedhar, Lauralie Richard, Tim Stokes

ABSTRACT AIMS: Multimorbidity is a major issue in primary healthcare. The study aim was to determine the prevalence of multimorbidity and multiple social disadvantage in relation to age, gender, deprivation and ethnicity in a New Zealand high-needs primary healthcare clinic population. METHODS: A cross-sectional study using data manually extracted from electronic medical records was conducted on all patients registered with a Dunedin free ‘third sector’ primary healthcare clinic. The data were analysed in terms of the number and type of morbidities, and prevalence of multimorbidity in relation to age, sex and multiple social disadvantage. RESULTS: Most patients had multimorbidity (76.5%, 95% CI 72.0–85.5%) and half (49.9%, 95% CI 44.8–54.9%) had long-term physical and mental health comorbidities. The prevalence of multimorbidity in patients in all ethnic groups increased with age and was high across all ethnic groups and deprivation quintiles. Seven of the 10 most prevalent long-term conditions were mental health conditions. A majority of patients (54.7%, 95% CI 49.4–59.8%) had at least one multiple social disadvantage domain recorded. CONCLUSIONS: The high prevalence of multimorbidity in a high-needs population served by a ‘third sector’ clinic raises important issues relating to equity in the New Zealand health system.

ultimorbidity (the presence of two outcomes: it is associated with high mortal- or more chronic conditions in a sin- ity, reduced functional status and quality of gle patient)1 is a major issue for pri- life, increased use of inpatient and ambula- M 2 10,11 mary healthcare both in New Zealand and tory healthcare and polypharmacy. 3,4 internationally. There is a limited evidence To date, there has been no research base on the epidemiology of multimorbidity reporting on the prevalence of multimor- 5 in New Zealand, particularly in relation bidity and its relationship with multiple to its prevalence in a primary healthcare social disadvantage (MSD) in a New Zealand 6 population. Nonetheless, in line with the high-needs primary healthcare population. 7–9 international literature, the New Zealand The term MSD12,13 is helpful in under- prevalence increases with age, is more standing the pathways by which social common and occurs earlier in those living in inequalities may lead to individuals expe- areas with high socioeconomic deprivation riencing disadvantage in multiple areas and disproportionately affects indigenous of life concurrently14 and how these indi- 5,6 people. In addition, cohort studies show viduals engage with a range of health and that multimorbidity leads to poorer health social services aimed at addressing these

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 42 www.nzma.org.nz/journal ARTICLE

disadvantages (eg, primary healthcare, of 1 November 2017 (n=375) were included housing, criminal justice system).13 MSD in the study. can be framed as a patient level measure Data collection which encompasses a range of social disad- The patient-level data was manually vantage domains: health, income, housing, extracted from electronic medical records education, employment, material well- onto an Excel spreadsheet by SS under the being, safety and social connectedness.14 supervision of TS. The dataset comprised Homelessness, previous imprisonment and of administratively and clinically recorded problem drug use are some domains that data. This included age, sex, ethnicity, depri- have been highlighted as being important vation quintile, chronic medical conditions, markers for MSD in a high-needs (specialist multimorbidity, multiple social disadvantage homeless service) UK primary healthcare (MSD) and service utilisation (defi ned population.15 The extent of MSD was also as engagement with selected secondary found to be positively associated with multi- care and community services; emergency morbidity in this study.15 department attendances and primary In New Zealand, patients experiencing healthcare consultations over the previous MSD will have diffi culty accessing main- calendar year). Deprivation was deter- stream primary healthcare services, due at mined by the area in which the patient lived least in part to the fi nancial barrier to access (NZ Dep2013).25,26 We selected 38 common caused by the co-payment model.16,17 In the morbidities seen in a high-needs primary 2016/2017 New Zealand Health Survey, 28% healthcare population8,15 derived from the of respondents reported an unmet need list of 40 long-term conditions presented in for general practice services, with 20% of a widely cited Scottish primary care multi- those living in the most socioeconomically morbidity cross-sectional study.8 These 38 deprived areas indicating cost as a reason.18 morbidities (See Supplementary Table 1) There exist, however, a number of ‘third were selected through discussion between sector’ (non-government, non-profi t)19,20 TS, SS and LR. The inclusion criteria were New Zealand primary healthcare clinics, the condition being common in a high- which provide care to vulnerable popula- needs population and being able to be tions including those experiencing MSD.21 recorded through use of general practice One such clinic is Dunedin’s Servants disease classifi cation (READ codes) alone. Health Centre (SHC), which provides free Multimorbidity was defi ned as two or more healthcare through the voluntary services long-term health conditions (LTCs).8 Social of doctors, nurses and counsellors to a disadvantage domains included previous ‘high needs’ population22,23 (defi ned as the imprisonment, in receipt of health and general practice having >50% Māori, Pacifi c disability benefi t, homelessness, domestic and New Zealand deprivation quintile 5 violence or other reports of violence and patients).24 lack of food. These social disadvantage The aim of this study was to determine domains were based on those used in a high- the prevalence of multimorbidity and MSD needs UK primary healthcare population.15 A in relation to age, gender, deprivation and detailed description of data collected can be ethnicity in a New Zealand high-needs found in Table 1. primary healthcare clinic population (SHC). Data analysis A descriptive statistical analysis was Methods conducted by SS and TS using frequencies, Study design, setting and percentages and cross tabulations with participants reported exact 95% confi dence intervals (CI) as appropriate. SPSS Statistics 24 was used A cross-sectional study of manually for the descriptive analysis. 95% CI were extracted patient-level data from calculated using Epi Info 7. 2. 2. 2. computerised medical records (practice management system in use: MedTech This study was granted ethical approval software) was conducted in January 2018. by the University of Otago ethics committee All patients permanently enrolled at SHC as (H17/122).

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 43 www.nzma.org.nz/journal ARTICLE

Table 1: Collected data variables and data source.

Variable Data collected Data source—relevant section of MedTech clinical record Demographics Age Date of birth Sex Sex

Ethnicity Self-reported ethnicity

Deprivation Deprivation quintile (1–5)

Smoking status Self-reported smoking status

Chronic medical Common morbidities seen in general practice Clinical problems recorded in the conditions identified (38 ‘long-term conditions’; multimor- classification section. The appro- bidity defined as the presence of two or more of priate general practice disease these morbidities in one patient).8 classification (READ codes) is used.

Engagement Patient is currently attending: Correspondence in document with selected • Community Mental Health Services (CMHS) inbox section of MedTech of secondary care • Community Alcohol and Drug Service (CADS) engagement with service over the and community previous year services (1/11/16–31/10/17).

Multiple social Multiple social disadvantage domains of rele- Homelessness: READ code ever disadvantage vance to this population were identified:15 recorded • Homelessness Previous imprisonment: letter from • Previous imprisonment prison service filed in document In addition we included: inbox section • Domestic violence or other reports of Reports of violence: READ code violence ever recorded • Health and disability benefit Health and disability benefit: • Lack of food electronic WINZ* medical certifi- cate filed in document outbox in previous year Lack of food: READ code ever recorded

Healthcare Number of GP consultations Recorded in consultation section utilisation Number of practice nurse consultations of records over the previous year Number of presentations to the emergency ED attendances: letter from ED department (ED) filed in document inbox section over the previous year.

*WINZ = Work and Income New Zealand.

Results Multimorbidity and comorbidity Most patients had multimorbidity (76.5%, The medical records of all permanently 95% CI 72.0–85.5%) and half (49.9%, 95% registered patients (n=375) as of 1 November CI 44.8–54.9%) had long-term physical and 2017 were reviewed. The practice demo- mental health comorbidities. These fi ndings graphics are presented in Table 2. The mean were consistent across all ethnic groups age of patients was 41.9 years ((median age (New Zealand European; Māori; Pacifi c and 42, Interquartile range (IQR) 30-53)). other). A majority of patients (58.7%, 95%

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 44 www.nzma.org.nz/journal ARTICLE

Table 2: Prevalence of multimorbidity by age, sex, ethnicity and New Zealand deprivation quintile.

N (%) Mean Median Percentage Percentage number of number of (95% CI) with (95% CI) with morbidities morbidities multimorbidity physical and (SD) (interquartile mental health range) comorbidity All patients 375 (100%) 3.20 (2.22) 3 (2–4) 76.5 (72.0–80.5) 49.9 (44.8–54.9)

Sex

Female 155 (41.3%) 3.67 (2.58) 3 (2–5) 78.7 (71.4–84.9) 57.4 (49.2–65.3)

Male 220 (58.7%) 2.87 (1.87) 3 (1.5–4) 75.0 (68.7–80.6) 44.5 (37.9–51.4)

Age, years

0–14 3 (0.8%) 0.33 (0.58) 0 (0–1) 0.0 0.0

15–24 35 (9.3%) 1.69 (1.75) 2 (0–3) 51.4 (34.0–68.6) 22.9 (10.4–40.1)

25–34 86 (22.9%) 2.77 (1.93) 3 (1–4) 69.8 (58.9–79.2) 41.9 (31.3–53.0)

35–44 96 (25.6%) 3.70 (2.33) 3 (2–5) 83.3 (74.4–90.2) 57.3 (46.8–67.3)

45–54 76 (20.3%) 3.58 (2.21) 3 (2–5) 82.9 (72.5–90.6) 52.6 (40.8–64.2)

55–64 62 (16.5%) 3.65 (2.39) 3 (2–5) 82.3 (70.5–90.8) 66.1 (53.0–77.7)

65–74 13 (3.5%) 2.92 (1.38) 3 (2–3) 92.3 (64.0–99.8) 46.2 (19.2–74.9)

>75 4 (1.1%) 2.75 (1.26) 3 (2–3.5) 75.0 (19.4–99.4) 25.0 (0.6–80.6)

Deprivation quintile*

1 (a luent) 14 (3.7%) 3.92 (3.36) 3 (2–5) 78.6 (49.2–95.3) 28.5 (8.4–58.1)

2 33 (8.8%) 2.64 (1.95) 3 (1–4) 69.7 (51.3–84.4) 39.4 (22.9–57.9)

3 55 (14.7%) 3.18 (2.03) 3 (2–4) 81.8 (69.1–90.9) 60.0 (45.9–73.0)

4 113 (30.1%) 3.12 (2.16) 3 (2–4) 77.0 (68.1–84.4) 46.9 (37.5–56.5)

5 (deprived) 160 (42.7%) 3.32 (2.26) 3 (2–5) 75.6 (68.2–82.1) 52.5 (44.5–60.4)

Ethnicity

NZ European 226 (60.3%) 3.21 (2.10) 3 (2–4) 77.9 (71.9–83.1) 50.0 (43.3–56.7)

Māori 100 (26.7%) 3.34 (2.46) 3 (1.5–5) 75.0 (65.3–83.1) 49.0 (38.9–59.2)

Pacific 15 (4.0%) 2.67 (1.84) 2 (1–4) 73.3 (44.9–92.2) 60.0 (32.3–83.7)

Other** 33 (8.8%) 2.85 (2.39) 2 (1–4) 72.7 (54.5–86.7) 45.5 (28.1–63.7)

Number of long-term conditions

0 40 (10.7%) - - - -

1 48 (12.8%) - - - -

2 67 (17.9%) - - - -

3 72 (19.2%) - - - -

4 55 (14.7%) - - - -

>5 93 (24.8%) - - - -

*NZDep2013; ** Other: Asian, Other European, African and Middle Eastern.

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 45 www.nzma.org.nz/journal ARTICLE

CI 53.5–63.7%) had three or more long-term 3.7, median 2, IQR 1–4, range 0–46), GP conditions and a quarter (24.8%, 95% CI or practice nurse (mean 8.2, median 5, 20.5–29.5%) had fi ve or more long-term IQR 3–12, range 0–84). A third of patients conditions. The prevalence of multimorbidity (34.4%) had attended the local emer- in patients in all ethnic groups increased gency department (ED) at least once in the with age and was high across all deprivation previous year. Of those patients with a quintiles. Table 2 presents the prevalence long-term mental health problem (n=281), of multimorbidity and physical and mental a fi fth (20.3%) had at least one recorded health comorbidity by age, sex, ethnicity and encounter with community mental health New Zealand deprivation quintile. services (CMHS) in the previous year. Long-term conditions (LTCs) Of those patients with substance misuse (alcohol, opioids or other psychoactive Both long-term physical and mental health substance misuse) (n=153), 16.3% had problems were common: most patients had at least one recorded encounter with a long-term mental health condition (74.9%, community alcohol and drug services 95% CI 70.3–79.1%) and a majority had a (CADS) in the previous year. long-term physical health condition (64.3%, 95% CI 59.3–69.0%). The 10 most prevalent long-term conditions were: depression Discussion (48.5%, 95% CI 43.4–53.7%), substance misuse This is the fi rst New Zealand study to (40.8%, 95% CI 35.9–45.8%), anxiety (34.9%, report the epidemiology of multimorbidity 95% CI 30.1–40.0%), asthma (26.9%, 95% CI in a high-needs primary healthcare popu- 22.5–31.7%), hypertension (13.3%, 95 % CI lation. Most patients had multimorbidity 10.3–17.2%), severe mental illness (schizo- (76.5%) and half (49.9%) had long-term phrenia, bipolar disorder, non-organic physical and mental health comorbidities. psychosis) (13.1%, 95% CI 10.0–16.7%), hepa- These fi ndings were consistent across all titis C (11.7%, 95% CI 8.9–15.4%), learning ethnic groups (New Zealand European; disabilities (10.7%, 95% CI 7.9–14.2%), Māori; Pacifi c and Other). The prevalence personality disorder (8.8%, 95% CI 6.3–12.1%) of multimorbidity in patients in all ethnic and deliberate self-harm (8.5%, 95% groups increased with age and was high 6.1–11.8%). Of these 10 conditions, seven across all deprivation quintiles. Multimor- were long-term mental health conditions. bidity was the norm for patients aged over Multiple social disadvantage (MSD) 25. Seven of the 10 most prevalent long-term A majority of patients had at least one conditions were mental health conditions MSD domain recorded (54.7%, 95% CI 49.4– (depression, substance misuse, anxiety, 59.8%). More Māori had at least one MSD severe mental illness (schizophrenia, bipolar domain recorded than New Zealand Euro- disorder, non-organic psychosis), learning peans (61.0%, 95% CI 55.9–66.0 and 53.5%, disabilities, personality disorder and delib- 95% CI 48.1–58.5% respectively). The prev- erate self-harm). A majority of patients had alence for the MSD domains was: in receipt at least one MSD domain recorded (54.7%), of Work and Income New Zealand (WINZ) the most common being in receipt of health health and disability benefi t in the previous and disability benefi t in the previous year year (28.5%, 95% CI 24.0–33.4%), previous and previous imprisonment. imprisonment (16.3%, 95% CI 12.7–20.0%), A key strength of this study is that we domestic violence or other violence (4.8%, conducted a manual review of the clinic 95% CI 2.9–7.5%), homelessness (2.7%, 95% electronic medical records, which hold the CI 1.3–4.9%) and lack of food (1.3%, 95% CI most complete health data for patients, of 0.4–3.0%). all permanently registered patients. This is in contrast to an earlier study of the clinic’s Service utilisation patients, which relied on an opportunistic Most patients had seen either a GP or survey of a sample of clinic attenders.21 A a practice nurse in the previous year (31 limitation of using data from the electronic October 2016–1 November 2017) (80.8% medical record is that its completeness relies and 78.4% respectively). The consultation on information being recorded. We consider rates were: GP (mean 4.5, median 4, IQR we have captured all data relating to clinic 1–7, range 0–45), practice nurse (mean healthcare utilisation and electronic fi ling

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 46 www.nzma.org.nz/journal ARTICLE

of WINZ medical certifi cates (health and deprivation and deprivation is associated disability benefi ts) in the previous year. with increased multimorbidity.8 Our fi ndings The clinic has high levels of recording of of a high multimorbidity prevalence, high active and past medical problems through physical and mental health comorbidity and READ codes (the standard clinical termi- the most prevalent conditions being mental nology system used in general practice in health conditions are, however, consistent New Zealand). At the time of conducting with a UK study of multimorbidity in a the study, however, discussion with the specialist homeless health service.15 We also clinic team revealed that there was limited found that while the mean age of patients use of READ social codes for MSD: there was 41.9 years, the level of multimorbidity was no standardised method or process for was comparable to those aged 65 and recording MSD in operation. It is therefore over in a general population of primary likely that the recording of homelessness, healthcare attenders.8 In contrast to the violence and lack of food is incomplete and UK homeless study,15 however, we found a thus the true prevalence of these domains lower prevalence of MSD. This is likely to at the clinic is higher than that reported be explained by the fact that SHC offers free here. We were also not able to explore any care to a broader range of patients than a association between extent of MSD and specialist homeless service and is also due to degree of multimorbidity due to these data the incomplete recording of MSD data noted recording limitations. A further limitation of above. In line with previous studies the the study is that it is set in a single free clinic health practitioner consultation rates were study in one New Zealand city, nonetheless higher than those in ‘standard’ New Zealand it is considered that the fi ndings are likely to primary healthcare19,25 and consistent with be generalisable to other New Zealand free the UK homeless study.15 It is also notable, clinics serving similar high-needs popula- when the results are compared with a tions.22 A fi nal limitation of the study is its key Scottish primary care multimorbidity cross-sectional design. This does not allow cross-sectional study,8 that we did not fi nd us to explore the relationship between increasing levels of multimorbidity with social inequalities, MSD and engagement increasing levels of area deprivation. In with health and social services. Recent UK contrast, we found that the prevalence of research has explored the role of social multimorbidity in patients was high across deprivation and individual lifestyle risk all deprivation quintiles. This is likely to factors (eg, smoking, alcohol consumption, be explained by the high-needs nature of body mass index) in developing multi- the SHC population, and we would add that morbidity across the lifecourse.27 While there is considerable uncertainty around the lifestyle risk factors partially mediate the multimorbidity estimates in the lower depri- relationship between deprivation and multi- vation quintiles due to small numbers. morbidity, the majority of the relationship This study has a number of implications between deprivation and multimorbidity for healthcare policy and practice. The high 27 remains unexplained. prevalence of mental health conditions, There is limited New Zealand6 and inter- physical and mental health comorbidity and national literature15 on the prevalence of use of specialist mental health and addiction multimorbidity in high-needs or vulnerable services indicates the need for better inte- populations attending primary healthcare gration of mental health and social services28 clinics. We found, however, that the preva- with primary healthcare, particularly in lence of multimorbidity in Māori and Pacifi c those serving a high-needs population. It is patients attending SHC was higher than well recognised that geographical, fi nancial, that in those enrolled with a large tradi- organisational and cultural barriers limit tional general practice in the same city (53% such integration internationally.26 There of Māori and 64% of Pacifi c patients aged is scope for increased use of READ social over 35 were multimorbid)6 and there were codes for MSD in primary healthcare to also higher rates of physical and mental better establish the prevalence of MSD health comorbidity in the SHC population. in this population. A further issue is the One explanation for this fi nding is that the need to ensure that primary healthcare patients at SHC had higher socioeconomic funders and providers explicitly consider

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 47 www.nzma.org.nz/journal ARTICLE

health equity when planning and deliv- charges, is a legacy of the compromises ering services for a high-needs population struck between the medical profession and (Māori, Pacifi c and New Zealand deprivation the government during the implementation quintile 5 patients), for example, by using of the 1938 Social Security Act.31 Now, 80 the New Zealand Health Equity Assessment years on, the time is surely right for a review Tool (HEAT).29 The study also has impli- of direct patient charges and their impact on cations for improving access to primary access to primary healthcare for vulnerable healthcare for high-needs populations. One groups as part of the ongoing review of New widely cited conceptualisation of access to Zealand primary healthcare.32 healthcare sees access as the product of the interaction between accessibility of services Conclusions (approachability, acceptability, availability The high prevalence of multimorbidity and accommodation, affordability, appropri- and physical and mental health comorbid- ateness) and the individual’s abilities (ability ities in a free primary healthcare population to perceive, ability to seek, ability to reach, serving a high-needs population experi- ability to pay and ability to engage).30 This encing MSD raises important issues relating conceptualisation emphasises the impor- to health equity in the New Zealand health tance of considering multilevel interventions system.33 There is a need to re-orient New to improve access that take into account the Zealand primary healthcare delivery around range of determinants of access. Clearly, multimorbidity.4,34,35 There is also a need to improving access involves much more than further the integration of health and social simply removing a fi nancial barrier to care. services.28,36 In addressing these two key We acknowledge, however, that in New issues the ‘third sector’ is likely to continue Zealand cost is a major barrier to equitable to be an important provider of healthcare to access to primary healthcare for high-needs those high-needs groups who cannot access and vulnerable populations18 and that ‘third mainstream primary healthcare. Further sector’ organisations remain an important New Zealand research is also required to provider of care for this population.19,22 The explore the relationship between MSD, current New Zealand business model of multimorbidity and social inequalities. general practice, with its use of direct patient

Supplementary Table 1: List, data source defi nitions and prevalence of the 38 long-term conditions included in the multimorbidity analysis.

Condition and disease group Data source definition— Prevalence relevant section of n (%, 95% CI) MedTech clinical record Cardiovascular

1. Hypertension READ code ever recorded 50 (13.3, 10.3–17.2)

2. CVD (CHD, stroke/TIA, PVD, heart failure, READ code ever recorded 21 (5.6, 3.7–8.4) atrial fibrillation)

Respiratory

3. Asthma READ code ever recorded 101 (26.9, 22.5–31.7)

4. COPD READ code ever recorded 27 (7.2, 4.8–10.4)

5. Bronchiectasis READ code ever recorded 4 (1.1, 0.4–2.7)

Gastrointestinal

6. Hepatitis C (Hep C) READ code ever recorded 44 (11.7, 8.9–15.4)

7. Irritable bowel syndrome (IBS) READ code ever recorded 14 (3.7, 2.2–6.2)

8. Chronic liver disease READ code ever recorded 7 (1.9, 0.9–3.8)

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 48 www.nzma.org.nz/journal ARTICLE

Supplementary Table 1: List, data source defi nitions and prevalence of the 38 long-term conditions included in the multimorbidity analysis (continued).

9. Diverticular disease of intestine READ code ever recorded 7 ((1.9, 0.9–3.8)

10. Hepatitis B (Hep B) READ code ever recorded 3 (0.8, 0.3–2.3)

11. Inflammatory bowel disease (IBD) READ code ever recorded 3 (0.8, 0.3–2.3)

Renal

12. Chronic kidney disease READ code ever recorded 2 (0.5, 0.2–1.9)

Endocrine

13. Diabetes READ code ever recorded 24 (6.4, 4.3–9.4)

14. Thyroid disorder READ code ever recorded 14 (3.7, 2.2–6.2)

Neuro

14. Migraine READ code ever recorded 30 (8.0, 5.7–11.2)

16. Epilepsy READ code ever recorded 22 (5.9, 3.9–8.7)

17. Multiple sclerosis READ code ever recorded 2 (0.5, 0.2–1.9)

18. Dementia READ code ever recorded -

19. Parkinson’s disease READ code ever recorded -

Musculoskeletal

20. Gout READ code ever recorded 17 (4.5, 2.9–7.1)

21. Rheumatoid arthritis READ code ever recorded 2 (0.5, 0.2–1.9)

22. Other inflammatory polyarthropathies & READ code ever recorded - systemic connective tissue disorders

Mental health

23. Depression READ code ever recorded 182 (48.5, 43.4–53.7)

24. Substance misuse READ code ever recorded 153 (40.8, 35.9–45.8)

25. Anxiety READ code ever recorded 131 (34.9, 30.1–40.0)

26. Severe mental illness (schizophrenia, bipo- READ code ever recorded 49 (13.1, 10.0–16.7) lar disorder, non-organic psychosis)

27. Learning disability READ code ever recorded 40 (10.7, 7.9–14.2)

28. Personality disorder READ code ever recorded 33 (8.8, 6.3–12.1)

29. Self-harm (deliberate)/suicide READ code ever recorded 32 (8.5, 6.1–11.8)

30. Physical or sexual abuse READ code ever recorded 24 (6.4, 4.3–9.4)

31. Anorexia/bulimia READ code ever recorded 10 (2.7, 1.5–4.8)

Other

32. Eczema or psoriasis READ code ever recorded 32 (8.5, 5.9–11.8)

33. Blindness or low vision READ code ever recorded 20 (5.3, 1.3–4.5)

34. New diagnosis of cancer in last five years READ code ever recorded 20 (5.3, 1.3–4.5)

35. Hearing loss READ code ever recorded 15 (4.0, 2.4–6.5)

36. Chronic sinusitis READ code ever recorded 0

37. Glaucoma READ code ever recorded 0

38. Prostate disorders READ code ever recorded 0

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 49 www.nzma.org.nz/journal ARTICLE

Competing interests: Nil. Acknowledgements: We thank the trustees and clinic staff of Servants Health Centre for their participation and support. The study was conducted while TS worked as a volunteer GP at the Health Centre. SS held a University of Otago summer studentship scholarship. SS was awarded the 2018 Gil Barbezat Summer studentship prize by the Dunedin School of Medicine and the Southern District Health Board for this work. Author information: Sharmaine Sreedhar, Medical Student, Dunedin School of Medicine, University of Otago, Dunedin; Lauralie Richard, Research Fellow, Department of General Practice & Rural Health, Dunedin School of Medicine, University of Otago, Dunedin; Tim Stokes, Elaine Gurr Professor of General Practice, Department of General Practice & Rural Health, Dunedin School of Medicine, University of Otago, Dunedin. Corresponding author: Professor Tim Stokes, Department of General Practice and Rural Health, Dunedin School of Medicine, University of Otago, 55 Hanover Street, Dunedin 9016. [email protected] URL: http://www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2019/vol-132-no-1490- 22-february-2019/7816

REFERENCES: 1. Valderas JM, Starfi eld national-level hospital 10. Salisbury C, Johnson L, B, Sibbald B, Salisbury and pharmaceutical data. Purdy S, Valderas JM, C, Roland M. Defi ning BMJ Open. 2018; 8(5). Montgomery AA. Epide- comorbidity: implications 6. Stokes T, Azam M, miology and impact of for understanding health Doolan-Noble F. Multi- multimorbidity in primary and health services. morbidity in Māori care: a retrospective cohort Annals of Family Medicine. and Pacifi c patients: study. The British Journal 2009; 7(4):357–363. cross-sectional study in a of General Practice : the 2. Mercer SW, Smith SM, Dunedin general practice. Journal of the Royal College Wyke S, O'Dowd T, Watt GC. Journal of Primary Health of General Practitioners. Multimorbidity in primary Care. 2018; 10(1):39–43. 2011; 61(582):e12–21. care: developing the 7. Violan C, Foguet-Boreu 11. France EF, Wyke S, Gunn research agenda. Family Q, Flores-Mateo G, et al. JM, Mair FS, McLean G, Practice. 2009; 26(2):79–80. Prevalence, determinants Mercer SW. Multimor- 3. Mangin D, Heath I, and patterns of multimor- bidity in primary care: Jamoulle M. Beyond bidity in primary care: a systematic review of diagnosis: rising to the a systematic review of prospective cohort studies. multimorbidity challenge. observational studies. PloS The British Journal of BMJ (Clinical research one. 2014; 9(7):e102149. General Practice : the Journal of the Royal College ed). 2012; 344:e3526. 8. Barnett K, Mercer SW, of General Practitioners. 4. Goodwin N, Dixon A, Norbury M, Watt G, Wyke 2012; 62(597):e297–307. Anderson GF, Wodchis S, Guthrie B. Epidemiology W. Providing integrated of multimorbidity and 12. Manderbacka K, Arffman care for older people with implications for health M, Sund R, Karvonen S. complex needs Lessons care, research, and medical Multiple social disadvan- from seven international education: a cross-sec- tage does it have an effect case studies. London: tional study. Lancet. on amenable mortality: King's Fund; 2014. 2012; 380(9836):37–43. a brief report. Interna- tional Journal for Equity 5. Stanley J, Semper K, Millar 9. Gracey M, King M. in Health. 2014; 13:67. E, Sarfati D. Epidemiology Indigenous health part of multimorbidity in 1: determinants and 13. Valentine K. Complex needs New Zealand: a cross-sec- disease patterns. Lancet. and Wicked Problems: tional study using 2009; 374(9683):65–75. How Social Disadvan-

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 50 www.nzma.org.nz/journal ARTICLE

tage Became Multiple 22. Loh L, Trevallyan S, Main 30. Levesque JF, Harris MF, Social Policy & Society. SJ, Revell L, Patton V, Ojo A. Russell G. Patient-cen- 2016; 15(2):237–249. The case for a systematic tred access to health 14. Bramley G, Fitzpatrick policy approach to free care: conceptualising S, Edwards J, Ford D, primary health care for access at the interface Johnsen S, Sosenko F, vulnerable groups in of health systems and Watkins D. Hard Edges: New Zealand. The New populations. International Mapping Severe and Zealand Medical Journal. Journal for Equity in Multiple Disadvantage in 2015; 128(1424):45–53. Health. 2013; 12:18. England. London: Lankelly 23. Loh L, Jaye C, Dovey S, 31. Gauld R. Questions about Chase Foundation; 2015. Lloyd H, Rowe J. Dunedin's New Zealand's health 15. Queen AB, Lowrie R, free clinic: an exploration system in 2013, its 75th Richardson J, Williamson of its model of care using anniversary year. The New AE. Multimorbidity, case study methodology. Zealand Medical Journal. disadvantage, and patient Journal of Primary Health 2013; 126(1380):68–74. engagement within Care. 2015; 7(2):145–152. 32. Ministry of Health. Draft a specialist homeless 24. Ministry of Health. Very Terms of Reference health service in the Low Cost Access Scheme Review of New Zealand UK: an in-depth study (VLCA) 2016; http://www. Health and Disability of general practice data. health.govt.nz/our-work/ Sector. Wellington: BJGP Open. 2017; 1(3). primary-health-care/ Ministry of Health, 2018. 16. Jatrana S, Crampton P. primary-health-care- 33. Sheridan NF, Kenealy TW, Primary health care in subsidies-and-services/ Connolly MJ, et al. Health New Zealand: who has very-low-cost-access- equity in the New Zealand access? Health policy scheme health care system: a (Amsterdam, Nether- 25. Salmond CE, Crampton national survey. Interna- lands). 2009; 93(1):1–10. P. Development of New tional Journal for Equity 17. Osborn R, Squires D, Doty Zealand's deprivation index in Health. 2011; 10:45–45. MM, Sarnak DO, Schnei- (NZDep) and its uptake 34. Stokes T, Tumilty E, der EC. In New Survey as a national policy tool. Doolan-Noble F, Gauld R. Of Eleven Countries, US Canadian journal of public Multimorbidity, clinical Adults Still Struggle With health = Revue canadienne decision making and Access To And Afford- de sante publique. 2012; health care delivery in ability Of Health Care. 103(8 Suppl 2):S7–11. New Zealand Primary Health Affairs. 2016. 26. Atkinson J, Salmond C, care: a qualitative study. 18. Ministry of Health. Crampton P. NZDep2013 BMC Family Practice. Annual Data Explorer Index of Deprivation. 2017; 18(1):51. 2016/17: New Zealand University of Otago: 35. Mays N. Reorienting Health Survey Wellington: Division of Health the New Zealand health Ministry of Health, 2017. Sciences; 2014. care system to meet the 19. Crampton P, Dowell AC, 27. Katikireddi SV, Skivington challenge of long-term Bowers S. Third sector K, Leyland AH, Hunt K, conditions in a fi scally primary health care in Mercer SW. The contri- constrained environment. New Zealand. The New bution of risk factors to Paper prepared for New Zealand Medical Journal. socioeconomic inequalities Zealand Treasury Long- 2000; 113(1106):92–96. in multimorbidity across term Fiscal External Panel, the lifecourse: a longi- November 2012, and Chair 20. Crampton P, Dowell A, tudinal analysis of the of Public Finance, Victoria Woodward A. Third sector Twenty-07 cohort. BMC University of Wellington primary care for vulner- medicine. 2017; 15(1):152. and New Zealand Treasury able populations. Social conference, Affording Science & Medicine (1982). 28. Timmins N, Ham C. The our Future, Wellington, 2001; 53(11):1491–1502. quest for integrated health and social care A 10–11 December 2013. 21. Loh L, Dovey S. Who case study in Canterbury, 36. Ham C. Successful integrat- attends Dunedin's free New Zealand. London: ed care. A Canterbury tale: clinic? A study of patients King's Fund; 2013. an invaluable education. facing cost barriers to The Health Service Journal. primary health care access. 29. Signal L, Martin J, Cram 2013; 123(6367):18–19. Journal of Primary Health F, Robson B. The Health Care. 2015; 7(1):16–23. Equity Assessment Tool: A user's guide. Wellington: Ministry of Health, 2008.

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 51 www.nzma.org.nz/journal LETTER

A response to: The ‘elephants in the room’ for New Zealand’s health system in its 80th anniversary year: general practice charges and ownership model Kate Baddock, Jan White, Lesley Clarke

rofessor Gauld et al’s editorial calls comments including their view that private for debate on GP fees and ownership general practice providers are less likely to Pmodels without acknowledging the cater to diversity.1 As well as being wrong, good-quality care GPs provide; we see this as this is insulting to all doctors and espe- an inherent failure in the paper. cially GPs working in public and/or private In New Zealand GPs work hard, providing practice, as GPs cater to a wide range of the public with an excellent standard and ethnicity, geography, sexual orientation, age, quality of care. They have shown themselves religious beliefs, different experiences and to be adaptable and have evolved smoothly backgrounds and want better health for all from the solo practitioner often working their patients. from the family home to many different The two published papers referenced in iterations. They have also shown themselves Table 1 were an editorial by Les Toop and a to be willing early adopters of new tech- viewpoint paper by Peter Crampton in 2005 nology and ideas. Any debate on the health based on 2001 information that was pre-PHO of New Zealanders involving changes to the data. Almost 20 years later the primary care GP ownership model must not disregard the environment is very different to that prior value currently provided by GPs under the to the development of PHOs. Since 2005 current models. we have seen the rollout of capitation, the The authors of the editorial state that advent of Very Low Access practices, the further research is needed but imply that evolution of Independent Practitioner Asso- GP fees and current ownership models need ciations (IPAs), Primary Health Organisations to be discarded. This would be a mistake as, (PHOs) and Primary Care Networks. Very although we acknowledge there are issues different to fee for service, General Medical that need fi xing such as equitable access for Services (GMS) payments, and no subsidies all our patients and a funding model that for a substantial part of the population. works to address this (and other) inequities, Other points of concern include the general practice does work. statement that DHBs are funders of public We note that the editorial stemmed from hospitals, ignoring the fact that they are a panel discussion that the authors partic- expected to fund the health of popula- ipated in hosted by the Centre for Health tions. There is certainly not equity in the Systems and Technology at the University public health system at the moment, so one of Otago on 30 July 2018. This may go wonders why would GPs believe they would some way to explain how the authors have get adequate funding if the DHB was respon- failed to back up some of their subjective sible for it?

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 52 www.nzma.org.nz/journal LETTER

Furthermore, the statements: “patient Other unanswered questions include: fees have arguably played a part in this would GPs retain autonomy over how a ranking” and “there is growing, anecdotal practice is run; would the government buy agreement within the GP community that practices from owners who have invested cost has become an unacceptable barrier hundreds of thousands of dollars in estab- for many people accessing general practice lishing them; would the pro-rata basis service in New Zealand” are unsupported guideline of 1,250 patients per GP continue? assumptions with no place in any mean- To fully answer these and other ques- ingful debate as the implication is that tions we welcome evidence-based dialogue patient fees are the barrier to general taking account of all that general practice practice healthcare. We need only look currently does. at the data presented in the editorial that However, any new model must provide suggests that in the NHS some 7% of patients confi dence to current and future GPs that do not access healthcare, citing cost as a it will enable the best primary healthcare barrier—and this is in a system which has available, and that it will be both funded no patient fees. satisfactorily, maintained over time and In terms of business ownership models indexed to the Health Infl ation Index. there are changes afoot internationally, most NZMA wants to ensure that its GP recently in the UK where a new GP contract members’ voice is heard in this debate and deal is being developed that may be a game to that end will be asking its members for changer for primary care.2 There is also new both their thoughts and comments, which research in the US that receipt of primary will inform future conversations with care was associated with signifi cantly more policy makers. high-value care, slightly more low-value care, and better healthcare experience. Poli- Further change must be clearly thought cymakers and health system leaders seeking through and it is vital that GPs and GP to improve value should consider increasing leaders are involved from the start in both investments in primary care.3 While we the discussion and the decision making, may agree that “most GPs have benefi tted otherwise a profession that is already from increased capitation payments but facing workforce challenges may be left in a remain(ing) reliant on patient charges for perilous state. a signifi cant portion of their income”; the It is with some concern that we note that questions remain, should general practice be Peter Crampton is both a co-author of the completely funded by third parties so that it original editorial and a member of the is free to patients at the point of care? What Health and Disability Sector Review panel would this mean for patients in the future? and this interest was not declared. We What would it mean to new GPs? And what would hope that his thoughts as expressed would it mean to GPs who own their own in the editorial do not predicate his stance general practices? Should funding belong to on the panel. the patient?

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 53 www.nzma.org.nz/journal LETTER

Competing interests: Nil. Author information: Kate Baddock, Chair, New Zealand Medical Association, Wellington; Jan White, General Practitioner, Mt Eden Medical Centre, Auckland; Lesley Clarke, Chief Executive, New Zealand Medical Association, Wellington. Corresponding author: Kate Baddock, Chair, New Zealand Medical Association, Wellington. [email protected] URL: http://www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2019/vol-132-no-1490- 22-february-2019/7817

REFERENCES: 1. http://www.nzma.org.nz/ 2. http://www.bmj.com/ jamainternalmedicine/arti- journal/read-the-journal/ content/364/bmj.l531 cle-abstract/2721037?guest- all-issues/2010-2019/2019/ 3. http://jamanetwork. AccessKey=0983f57a-7a4e- vol-132-no-1489-1-febru- com/journals/ 480a-b0b4-7bbcc2c6649b ary-2019/7795 : Table 1.

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 54 www.nzma.org.nz/journal LETTER

Getting surgical antibiotic prophylaxis right, lessons from the National Orthopaedic Surgical Site Infection Improvement Programme: a call for action! Arthur J Morris, Sally A Roberts, Nikki Grae, Deborah Jowitt

oint replacement greatly improves qual- fl ucloxacillin (two 1g doses) with genta- ity of life and with an ageing population micin (single dose at 4mg/kg) was associated Jthe number of procedures is increasing. with a 94% increase in acute renal injury.9 A feared complication is subsequent surgical The authors concluded that “gentamicin site infection (SSI), especially deep peripros- should be avoided in orthopaedic patients thetic joint infection (PJI), which results in in the perioperative period”.9 In early 2015 signifi cant patient morbidity and healthcare the Programme wrote to the clinical heads costs.1,2 In New Zealand each PJI adds an of all DHB orthopaedic surgery services excess mean treatment cost of $40k and an with several recommendations concerning additional 42 days in hospital.3 Similar costs surgical prophylaxis, including; that the and excess hospital stays occur following use of gentamicin for prophylaxis should deep spinal SSIs and Staphylococcus aureus stop, a second dose of prophylaxis be given mediastinitis after sternotomy.4,5 before the second side of bilateral arthro- The National Orthopaedic Surgical Site plasty procedures, to use clindamycin Infection Improvement Programme (the in preference to vancomycin in cases of Programme) was started in 2013 to promote cephalosporin allergy because of better adherence with evidence-based practices compliance with timing, and to stop prophy- known to reduce SSI.6,7 The Programme has laxis within 24 hours of surgery. resulted in signifi cantly improved use of This report examines in more detail the alcohol-based skin preparations and the relationship between the timing of surgical timing and dosing of surgical antibiotic antibiotic prophylaxis and the SSI rate prophylaxis.7,8 The national surveillance, and the changes in practice following the intervention and quality improve- Programme’s March 2015 recommendations. ment-based Programme has been associated Full details of the Programme’s methods, with a reduction in the median SSI rate data collection and reporting have been following hip and knee arthroplasties from published previously.6–8 The US National 1.36% to 0.91% (p<0.01).8 Healthcare Safety Network SSI defi nitions 6 In 2014 early observations of the were used. Programme and the publication of a report from Scotland on the use of gentamicin Results and discussion in orthopaedic procedures led to further From 1 July 2013 to 31 December 2017, recommendations on surgical prophylaxis. 46,360 publicly funded hip and knee arthro- A change to Scotland prophylaxis in 2009, in plasties were performed. There were 498 an effort to reduce the burden of Clostridium SSIs, a rate of 1.07% (95% CI 0.98–1.17%). diffi cile disease which was an epidemic at Prophylaxis was defi ned as “on time” if the time, from a cephalosporin alone to given ≤60 minutes of knife to skin (KTS),

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 55 www.nzma.org.nz/journal LETTER

Table 1: Surgical site infection rates by timing, dose and type of prophylaxis: July 2013 to December 2017.

Prophylaxis SSI rates Odds ratio p value Comment On time prophylaxis vs 1.05% (95% CI 0.96–1.15) 0.52 (95% CI 0.34–0.78) <0.003 On-time prophylaxis late/early vs 2.0% (95% CI 1.3–3.0) reduces SSI rate.

Gentamicin and cepha- 0.85% (95% CI 0.51–1.2) 0.78 (95% CI 0.51–1.2) p=0.29 Combination pro- losporin prophylaxis vs vs 1.08% (95% CI 0.99–1.18) phylaxis does not cephalosporin alone reduce SSI rate.

early if given >60 minutes before KTS and Timing for bilateral procedures improved, late when given after KTS.6,10 use of clindamycin increased and excess There was no difference in the SSI rate duration of prophylaxis reduced (Table 2). depending on the antibiotic used, ie, The Progamme has recently started an cefazolin, cefuroxime, clindamycin, vanco- identical process for cardiac surgery.11 mycin or other agents (data not shown). While the adherence with recommended On-time prophylaxis halved the SSI rate interventions is higher than that at the (Table 1). start of the orthopaedic Programme there Following the 2015 recommendations, are ~3–5% of cardiac surgery patients who signifi cant improvements in prophylaxis do not receive either the correct antibiotic were achieved (Table 2). Combination dose on time or have an alcohol-based skin 12 prophylaxis continued to decline signifi - preparation used. It is probable that the cantly (Table 2 and Figure 1). Combination practice gaps identifi ed in both orthopaedic prophylaxis did not reduce the SSI rate and cardiac surgery exist in other surgical (Table 1) another reason to stop its use. disciplines as well.

Table 2: Change in antimicrobial surgical prophylaxis practice following Programme recommendations in March 2015.

2015 recommendations Change in practice Comments The routine/common use Gentamicin, in combination In 2013 five DHBs used it >20%. In 2017 five of gentamicin for prophy- with a cephalosporin, dropped DHBs never used combination prophylaxis laxis in orthopaedic surgery from >6% to <1.5%, p=0.0002, and only one used it >10% (14.5%). should stop. see Figure 1. For 2017 and 2018 this means >3,000 fewer patients were exposed to gentamicin and its toxicity.

A second dose of prophylax- Prophylaxis timing has 10% of bilateral procedures still not getting is is given before the second improved from 76% to 90%, prophylaxis on time. side of a bilateral procedure p=0.002, due to a second dose A second dose is not required if vancomy- is commenced. being given more frequently. cin is used because of its long half-life.

Clindamycin to be used in Clindamycin is given on time In cases of cephalosporin allergy there is preference to vancomycin 96% vs 70% for vancomycin. still scope to increase clindamycin use to when there is cephalosporin When either is used the pro- achieve higher on time prophylaxis. allergy because it is asso- portion of clindamycin has ciated with better on time increased from <50% to 75%, compliance. p<0.0001.

Prophylaxis should stop Stopping prophylaxis within Standing orders can assist in ensuring within 24hrs of a procedure. 24hrs has increased from <80% prophylaxis is stopped within 24hrs. to >97%, p<0.0001. Continuing prophylaxis until drains or catheters have been removed is unproven and not recommended.

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 56 www.nzma.org.nz/journal LETTER

Figure 1: Proportion of procedures receiving combination prophylaxis of gentamicin and a cephalosporin: July 2013–December 2017.

The Programme’s experience to date eg, cardiac surgery.10 Consider the use suggests the following steps be undertaken of pharmacy stop orders by all surgical service providers, public and • Ensuring alcohol-based skin prepa- private, to generalise the lessons from the rations are used for procedures orthopaedic and cardiac Programmes to all involving skin incisions surgical specialties to reduce their SSI rate. • Asking those undergoing elective These should include: procedures about possible previous • Ensuring an agreed up-to-date S. aureus infections and, if a history is evidence based surgical prophylaxis confi rmed, implement an “anti-staph” 10 guideline is in place protocol • Performing snap audits of an • Ensuring adequate infection adequate number of procedures in prevention and control resources are a sub-specialty/theatre to measure in place and utilised to reduce SSI compliance with the prophylaxis • That the relevant infection control guideline, ie, choice, dose, timing, and committee has input into, and audit duration of, the surgical prophylaxis guideline • When used, recommending 2g of and regularly reviews local SSI data. cefazolin for all adults as our expe- The signifi cant strides made in the rience shows that weight-based dosing improvement in orthopaedic prophylaxis is inadequately adhered to compliance, with its attendant reduction • Not using gentamicin in prophylactic in SSIs, are transferrable to other surgical regimes unless its use is included in specialities. If the lessons learned from authoritative prophylaxis guidelines, the national orthopaedic Programme were eg, for certain abdominal or pelvic applied across all surgical sub-specialties a 10 procedures reduction in SSI rates will occur. Surgeons • Surgeons utilise the surgical safety and anaesthetists should take the lead to checklist to ensure prophylaxis has ensure indicated surgical prophylaxis is been given on time administered correctly to maximise its • Ensuring that prophylaxis is not ability to reduce SSIs and their often signif- continued after surgery unless recom- icant human and fi nancial costs. mended in authoritative guidelines,

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 57 www.nzma.org.nz/journal LETTER

Competing interests: Arthur Morris states he is the Clinical Lead for the NZ Surgical Site Infection Improvement Programme. Sally Roberts is the National Clinical Lead for the Health Quality and Safety Commission Infection Prevention and Control Programme. Acknowledgements: We recognise the essential role DHB IPC staff have had in the implementation of the SSII Programme. We also acknowledge the important contributions of New Zealand surgeons and anaesthetists, and their teams, who have responded positively to the Programme’s interventions and recommendations. Author information: Arthur J Morris, Clinical Microbiologist, Auckland City Hospital, Auckland; Clinical Lead NZ SSIIP; Sally A Roberts, Clinical Microbiologist, Auckland City Hospital, Auckland; Clinical Lead Health Quality and Safety Commission Infection Prevention and Control Programmes; Nikki Grae, Infection Prevention and Control Specialist, Infection Prevention and Control Programme, Health Quality and Safety Commission, Wellington; Deborah Jowitt, Infection Prevention and Control Advisor, Infection Prevention and Control Programme, Health Quality and Safety Commission, Wellington. Corresponding author: Arthur J Morris, Clinical Microbiology Laboratory, Auckland City Hospital, Park Road, Grafton, Auckland 1010. [email protected] URL: http://www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2019/vol-132-no-1490- 22-february-2019/7818

REFERENCES: 1. Zimmerli W, Trampuz A, nitis. NZ Med J. 2005; 118: gentamicin as surgical Ochsner P. Prosthetic-joint URL:http://www.nzma.org. prophylaxis. J Am Soc infections. N Engl J Med. nz.journal/118-1210/1316/ Nephrol. 2014; 25:2625–32. 2004; 351:1645–54. 6. Health Quality and 10. Bratzler DW, Dellinger EP, 2. Tande A, Patel R. Prosthetic Safety Commission New Olsen KM, et al. Clinical joint infection. Clin Micro- Zealand. Surgical Site practice guidelines for biol Rev. 2014; 27:302–345. Infection Improvement antimicrobial prophylaxis 3. Gow N, McGuinness C, Programme Orthopaedics in surgery. Am J Health-Syst Morris A, et al. Excess Surgery Implementation Pharm. 2013; 70:195–283. cost associated with Manual v 1.4. 2015. 11. Health Quality and primary hip and knee 7. Morris A, Panting A, safety Commission. joint arthroplasty surgical Roberts S, Shuker C. A National cardiac surgery site infections: a driver new surgical site infection report, 1 October - 31 to support investment improvement programme December 2017. https:// in quality improvement for New Zealand: early www.hqsc.govt.nz/ strategies to reduce progress. NZ Med J. our-programmes/infec- infection rates. NZ Med 2015; 128(1414):51–9. tion-prevention-and-control/ J. 2016; 129(1432):51–8. 8. Morris AJ, Roberts SA, publications-and-resources/ 4. Barnacle J, Wilson D, Little Hamblin R, et al. The New publication/3401/ C, Hoffman C, Raymond Zealand Surgical Site 12. Health Quality and Safety N. Excess cost and Infection Improvement Commission. SSIIP Cardiac inpatient stay of treating Programme: a national Surgery. Summary fi ndings deep spinal surgical site quality improvement for current quarter - April infections. NZ Med J. programme reducing to June 2018. https:// 2018; 131(1475):27–34. orthopaedic surgical site wwwpublic.tableau.com/ 5. Upton A, Smith P, Roberts infections. NZ Med J. 2018; profi le/hqi2803#!/vizhome/ S. Excess cost associated 131(1479):45–56. SSIcardiacdashboard/ with Staphylococcus aureus 9. Bell S, Davey P, Nathwani SSIIPcardiacsurgery?- poststernotomy mediasti- D, et al. Risk of AKI with publish=yes

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 58 www.nzma.org.nz/journal LETTER

Smokefree signage at railway stations: a survey of 54 stations in 11 local government areas Nick Wilson, George Thomson

s with a growing number of nations, measured. From the collected photographs New Zealand has adopted a national we examined qualitative issues (eg, evidence Asmokefree goal. Achieving this goal is of vandalism), along with selected aspects of likely to require multiple strategies, includ- sign design and content. The latter built on ing the expansion of outdoor smokefree ar- what is known about size and issues around eas. There has been some progress with ex- clutter from research on tobacco and panding such areas,1 but gaps in coverage2–6 alcohol warning labels (for further details, and processes7 still exist. For example, we see online Report10). identifi ed only two reports on smoking in New Zealand public transportation settings. Results Data from a 2008 survey on non-smokers’ exposure at bus stops and railway stations Station descriptions The 54 surveyed railway stations provided (inside or out) found 11.9% of respondents passenger services on seven different reported secondhand smoke exposure in the train lines.10 These stations were run by last month.8 Other research on health and two different train companies and the nuisance concerns by train users reported stations were located in 11 different local that smoking at transport waiting areas was government jurisdictions (territorial author- a problem for breathing, example to youth, ities) in the southern half of the North and the smell on clothes.9 Indeed, some wait- Island. ing passengers felt they were forced to move from seating because of the smoking. Given Station use of smokefree signage this background and international evidence At the studied stations there were a total around smoking at railway stations (see of 301 outdoor smokefree signs, and 77.8% online Report10), we aimed to identify the (42/54) of the stations had some smokefree extent and nature of outdoor smokefree sig- signage (although 18.5% [10/54] had only nage at a sample of such stations in a region one sign for the whole station). The median of New Zealand. number of signs per station was 2.0, but the range was up to a very large total of 192 Methods (Wellington station, which had nine plat- forms). The Capital Connection Line had the We surveyed all 54 passenger railway most such stations with no signs within the stations in the lower half of the North surveyed area for fi ve stations, followed by Island of New Zealand (from National the Northern Explorer Line at four stations). Park in the middle of the island, south to Each of the seven train lines had at least one Wellington). This was 47% (54/114) of all station with no smokefree signs. such stations in the country. Field observa- tions were conducted between November When considering separate platforms, 2017 and January 2018 by both authors. Any 76.4% (55/72) of these had some smokefree smokefree signage was photographed and signage. The median number of signs per platform was also 2.0 (range: 0 to 21.3).

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 59 www.nzma.org.nz/journal LETTER

There were no smokefree signs at the 186 pedestrian entrances to all the platforms (ie, Discussion 0/159 for well-defi ned entrances and 0/27 The results for this survey of railway for more general areas where pedestrians stations indicate highly diverse, but could access the platform). Most signage was generally poor levels of smokefree signage attached to the outside of platform struc- at both stations and at the platform level. tures at 88.1% (37/42), the rest being on just Particularly problematic was that there poles (4.8%, 2/42) or a mix of both (7.1%, were no such signs at any of the 186 pedes- 3/42) (eg, see fi gures in the online Report10). trian entrances to the platforms—which means that train users could easily miss Sign characteristics seeing those smokefree signs that did exist, There were 10 different types of typically on the platform buildings. Added smokefree sign observed.10 The median to this were the problems that some of the number of sign types per station was one, signs were placed high up on buildings or but it ranged up to eight different types poles above eye level, and were sometimes (Wellington station). Smokefree signage very small (ie, only 5x10cm). was often part of other signage (eg, prohib- iting skateboarding and cycling) and was So overall this situation suggests also located near to other types of signs. substantial scope for improvements to The smokefree signs were also sometimes smokefree signage by relevant agencies positioned quite high off the ground (ie, (transport authorities, local or central requiring a person to look upward to see government) in these particular settings. it).10 There was no smokefree signage Such improvements could specifi cally painted on the actual platform or entrance include: (i) Ensuring that all railway station surface, in contrast to a sign banning platforms have at least some smokefree alcohol.10 signage; (ii) Ensuring that the signage is frequent enough to be readily visible for Sign size was relatively modest overall people arriving via the main entrances to (median = 300cm2 similar to “A4” size) the platform (ie, improving on the current and ranged from 50cm2 to a very large level of zero entrances with signage found in 18,000cm2. Vandalism levels were rela- this study); (iii) Considering ways to reduce tively high for the signs that were just small the visual clutter of signage; (iv) Consid- stickers (at 63% [10/16]) but were extremely ering ways to improve signage visibility low for all the other signs (0.4%, 1/285). by lowering sign height; and (v) A national None of the signs found had wording that standard for the minimum size and legibility indicated that railway staff could or would of smokefree signs in all outdoor settings. enforce the outdoor smokefree policy. Policy options to achieve more complete There were also no references to any legis- smokefree status of such settings include an lation that would empower staff to require upgraded central government law, or use of non-smoking behaviour in outdoor areas. bylaws by local governments.

Competing interests: Nil. Author information: Nick Wilson, Department of Public Health, University of Otago, Wellington; George Thomson, Department of Public Health, University of Otago, Wellington. Corresponding author: Professor Nick Wilson, Department of Public Health, University of Otago, Wellington. [email protected] URL: http://www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2019/vol-132-no-1490- 22-february-2019/7819

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 60 www.nzma.org.nz/journal LETTER

REFERENCES: 1. Thomson G, Wilson N. 5. Wilson N, Thomson G. air sampling. N Z Med J Local and regional smoke- Suboptimal smokefree 2011; 124(1330):34–47. free and tobacco-free action signage at some hospi- 9. Russell M, Wilson N, in New Zealand: highlights tals: Field observations Thomson G. Health and and directions. N Z Med J and the use of Google nuisance impacts from 2017; 130(1462):89–101. Street View. N Z Med J outdoor smoking on 2. Thomson G, Wilson N. 2015; 128(1415):56–9. public transport users: Smokefree signage at 6. Wilson N, Thomson G. data from Auckland and children’s playgrounds: Surveying all outdoor Wellington. N Z Med J Field observations smokefree signage in 2012; 125(1360):88–91. and comparison with contrasting suburbs: 10. Wilson N, Thomson G. Google Street View. Tob methods and results. Full Report on Smokefree Induc Dis 2017; 15:37. Health Promot J Austr Signage at Railway Stations: 3. Thomson G, Wilson N. 2017; 28:264–65. A Survey of 54 Stations Smokefree signage at 7. Murad M, Marsh L, McGee in 11 Local Government New Zealand racecourses R. The contribution of Areas. Wellington: Univer- and sports facilities with smokefree outdoor areas sity of Otago Wellington, outdoor stands. N Z Med in achieving New Zealand’s 2019. http://aspire2025.fi les. J 2017; 130(1464):80–86. Smokefree 2025 goal. wordpress.com/2019/02/ 4. Wilson N, Thomson G. Health Promot J Austr report-on-smokefree-rail- Survey of smokefree 2018:(Online 28 February). way-stations-feb-2019-nw. signage at playgrounds: 8. Wilson N, Edwards R, Parry pdf The potential value of R. A persisting secondhand comparisons with dog smoke hazard in urban control signage. Aust N Z J public places: results from Public Health 2016; 40:395. fi ne particulate (PM2.5)

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 61 www.nzma.org.nz/journal METHUSELAH

Effect of aspirin on cardiovascular events and bleeding in the healthy elderly Aspirin is a well-established therapy for the secondary prevention of cardiovascular events. However, its role in the primary prevention of cardiovascular disease is unclear, especially in older persons, who have an increased risk. To evaluate this issue, these researchers enrolled 19,114 community dwelling healthy men and women in Australia and the US. Half were assigned to receive aspirin and the other half placebo. At a median follow-up of 4.7 years the rate of cardiovascular disease was 10.7 events per 1,000 person years in the aspirin group and 11.3 events in the placebo cohort. The rate of major haemorrhage was signifi cantly worse in the aspirin group (hazard ratio 1.38). It was concluded that the use of low-dose aspirin as a primary prevention strategy in older adults resulted in a signifi cantly higher risk of major haemorrhage and did not result in a signifi cantly lower risk of cardiovascular disease than placebo. An editorial reviewer noted these results and opined that beyond diet maintenance, exercise and smoking cessation, the best strategy for the use of aspirin in the primary prevention of cardiovascular disease may simply be to prescribe a statin instead. N Engl J Med 2018; 379:1509–18 & 2018; 379:1572–74 Intramuscular versus intravenous oxytocin to prevent postpartum haemorrhage at vaginal delivery Postpartum haemorrhage (PPH) is one of the leading causes of maternal mortality worldwide, accounting for nearly a quarter of maternal deaths. This report is of a study to determine whether intra- venous oxytocin is more effective than intramuscular oxytocin at preventing postpartum haemorrhage at vaginal delivery. One thousand and thirty-fi ve pregnant women were included in the study. Half received intramuscular oxytocin and the other half received intravenous oxytocin at vaginal delivery. The incidence of PPH was not signifi cantly lower in the intravenous group. However, the incidence of severe PPH and the need for blood transfusion was signifi cantly lower in the patients receiving the intravenous treatment. Intravenous oxytocin for the third stage of labour results in less frequent severe PPH, blood transfusion and admission to a high dependency unit than intramuscular oxytocin, and without excess side effects. BMJ 2018; 362:k3546 Association of dairy intake with cardiovascular disease and mortality in 21 countries fom fi ve continents Dietary guidelines recommend minimising consumption of whole-fat dairy products, as they are a source of saturated fats and presumed to adversely affect blood lipids and increase cardiovascular disease and mortality. The authors of this paper note that evidence for this contention is sparse. They enrolled 136,384 individuals aged between 35–70 years from 21 countries in fi ve continents. Dietary intakes of dairy products (milk, yoghurt and cheese) were recorded using validated question- naires. They further grouped these foods into whole-fat and low-fat dairy. The primary outcome was the composite of mortality or major cardiovascular events (non-fatal myocardial infarction, stroke or heart failure). The results of this study were that dairy consumption was associated with lower risk of mortality and major cardiovascular disease events in a diverse multinational cohort. Lancet 2018; 392:2288–97

URL: http://www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2019/vol-132-no-1490-22- february-2019/7820

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 62 www.nzma.org.nz/journal 100 YEARS AGO

Etiology and Treatment of In uenza February 1919

Ambulances in Wellington during the 1918 flu epidemic. Ref: PAColl-7489-69. Alexander Turnbull Library, Wellington, New Zealand. /records/22511121

he infl uenza bacillus was described by The disease is said to be endemic in North- Pfeiffi er and others in 1892, not long ern Central Asia. There is no doubt that the Tafter the severe European epidemic infection spreads with the quickest means of in 1889–90, and for some considerable time transit at the speed of an express train. The this germ was thought to be the sole origi- germs found in cases of infl uenza include nating cause of the disease, producing con- Pfeiffer’s bacillus, Grampositive diplo-strep- ditions in the respiratory tract favourable to tococci, Gram-negative diplococci, pneumo- the development of the micro-organisms of cocci, Friedlander’s bacilli, and micrococcus pneumonia. With the development of bac- catarrhalis. Nicolle and Lebailly on 14th teriological research it later became evident October, 1918, showed that the disease could that symptoms usually described as those be produced in monkeys by the fi ltrate of of infl uenza could proceed from organisms the infl uenza expectoration when injected other than Pfeiffer’s bacillus. In the last by the subconjunctival and nasal routes, and edition of the Manual of Medicine in the Uni- they also reported that the inoculation of versity series it is stated that “this organism the fi ltered bronchial secretion of infl uenza appears to have become much less common patients produced the disease in two men in Britain in recent years than it was in the inoculated by the subcutaneous route. A nineties, and to have been in great measure week later, Dujarrie de la Rivière successful- replaced by the micrococcus catarrhalis ly inoculated himself with infl uenza from a and other organisms even in cases where fi ltrate made from the blood of cases infect- clinical features would suggest infl uenza.” ed with infl uenza, showing that the fi ltrable

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 63 www.nzma.org.nz/journal 100 YEARS AGO

virus is present in the blood as well as in considerable hesitation must be felt in the bronchial secretion. In December, 1918, advising vaccine treatment as a curative Gibson, Bowman, and Connor confi rmed measure.” We have read many reports from these results. It appears certain, therefore, all over the world of the results of vaccine that the virus of infl uenza can pass through treatment as a prophylactic and curative a Chamberland fi lter, and the weight of evi- agency in infl uenza, and they are very dence favours the conclusion that there is a confl icting, and the only verdict at present is yet undiscovered fi ltrable organism which is “not proven.” At the same time the prospect the cause of the disease. This is the opinion is very bright that in the near future, in view of the majority of the members of the War of the recent advances in the investigation Offi ce Conference. There has lately been of the etiology of the disease, a vaccine will issued an unconfi rmed cablegram announc- be available which will revolutionise the ing that a defi nite organism of this kind has treatment of a scourge that causes little less been discovered. On the other hand, it may devastation than war. be that Pfeiffer’s bacillus may be able to pass A Royal Commission has been established through a fi lter in one phase of the cycle of in New Zealand to investigate the recent its development, and Hort’s recent investiga- epidemic of infl uenza which caused great tions lend some colour to this view. death and disablement in this eountry. It As regards the vexed question of may be that the masses still hope for some treatment, all are agreed that at present great result from Royal Commissions, but we know of no drug which has any specifi c most medical men, being well-informed, action in infl uenza. In the epidemic in the will cherish no delusions of that kind. A nineties quinine was vaunted, but it was body composed entirely of laymen is as found of little use recently, when salic- competent to investigate a serious and ylate of soda received general favour. The baffl ing disease and its consequences as a supplies of this drug in New Zealand are committee of doctors would be to report of inferior therapeutic quality, and salicin upon the Confession of Faith or hydro- is preferable, and is highly recommended electric power. We think it was Palmerston by Mr. E. B. Turner, who has had much who said that he prayed fervently that the experience in its use. The Royal College of people would never realise with how little Physicians states in a memorandum: “In wisdom they are governed. the uncertainty of our present knowledge

URL: http://www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2019/vol-132-no-1490- 22-february-2019/7821

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 64 www.nzma.org.nz/journal published by the New Zealand Medical Association

NZMJDigest http://www.nzma.org.nz/publications/nzmjdigest

The NZMA publishes the e-magazine NZMJDigest 10 times a year. It contains news and views from the profession and the NZMA, including the NZMA Chair's editorial, along with highlights from and links to the New Zealand Medical Journal. Click on the image above to view the latest issue. We welcome contributions from members and readers. To contribute to the NZMJDigest, please email [email protected]

NZMJ 22 February 2019, Vol 132 No 1490 ISSN 1175-8716 © NZMA 65 www.nzma.org.nz/journal