2021: Here We Come Tēnei Te Whakatau!

Te ara tika o te hauora hapori Journal of the New Zealand Medical Association Vol 134 | No 1528 | 15 January 2021

2021: here we come tēnei te whakatau!

Remembering past pandemics to ensure we plan for the future: updated New Zealand survey of memorials to the 1918 influenza pandemic

Counting the cost of major infection and sepsis in New Zealand: an exploratory study using the National Minimum Data Set

Increasing access to contraception Prognostic significance of mid-range Clinician knowledge in New Zealand: assessing the ejection fraction following acute of driving restrictions impact of a new funding initiative coronary syndrome (ANZACS-QI 23) following a stroke event

Te ara tika o te hauora hapori 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 Richard Beer Diana Wolken

Other enquiries to: NZMA To contribute to the NZMJ, first read: PO Box 156 www.nzma.org.nz/journal/contribute The Terrace Wellington 6140 © NZMA 2021 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 six 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 2020 New Zealand subscription rates Overseas subscription rates Individuals* $349 Individual $486 Institutions $604 Institutions $650 Individual article $33 Individual article $33

*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 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 2 www.nzma.org.nz/journal CONTENTS

EDITORIAL 57 7 Prognostic significance 2021: here we come / of mid-range ejection fraction tēnei te whakatau! following Acute Coronary Suzanne Pitama, Tania Huria, Syndrome (ANZACS-QI 23) Cameron Lacey Daniel Chan, Robert N Doughty, Mayanna Lund, Mildred Lee, ARTICLES Katrina Poppe, Andrew J Kerr 10 79 Counting the cost of major infection Radiological imaging of and sepsis in New Zealand: an melanoma: a review to guide exploratory study using the clinical practice in New Zealand National Minimum Data Set Victoria Francis, Tvesa Sehji, Paul J Huggan, Tania A Helms, Mark Barnett, Richard Martin Veronique Gibbons, Katie Reid, Harry Hutchins, Ian Sheerin VIEWPOINTS 26 88 A retrospective analysis Increasing access to of calls to the New Zealand contraception in New Zealand: National Poisons Centre assessing the impact of a new regarding Pacific patients funding initiative Eeva-Katri Kumpula, Rosalina Richards, Orna McGinn, Vicki Mount, Helen Fulcher Pauline Norris, Vanda Symon, Adam C Pomerleau CLINICAL CORRESPONDENCE 35 96 Interviews with health Think about cats in professionals about the National acute vision loss Aaron Yap, Moaz Alshaikhi, Kay Evans Child Protection Alert System Patrick Kelly, Melissa Adam, LETTERS Carmen Basu, Miranda Ritchie, Denise Wilson, Fred Seymour 99 Updated algorithms are 46 required to differentiate type 1 Clinician knowledge from type 2 diabetes using the of driving restrictions Virtual Diabetes Register following a stroke event Lynne Chepulis, Laura Stratton, John Parsons, Christopher Mayo, Ryan Paul Shannon Tisbury, Susan Waterworth, Nicola Starkey

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 3 www.nzma.org.nz/journal CONTENTS

103 100 YEARS AGO Are patients diagnosed 115 with functional symptoms during Cheese Mite Itch and ‘code stroke’ receiving best medical Conjunctivitis—A “Minor Horror” of care? the Great War Jaron Huang, Karim M Mahawish PROCEEDINGS 107 DNA information: access, use and 118 implications for healthcare in Proceedings of the Waikato Aotearoa New Zealand Clinical Campus Research Seminar, Sara Filoche, Jon Cornwall Wednesday 14 October 2020 111 ERRATUM Remembering past 123 pandemics to ensure we plan Erratum for the future: updated New Zealand survey of memorials to the 1918 influenza pandemic Nick Wilson, Geoffrey Rice, George Thomson, Michael G Baker

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 4 www.nzma.org.nz/journal summaries

Counting the cost of major infection and sepsis in New Zealand: an exploratory study using the National Minimum Data Set Paul J Huggan, Tania A Helms, Veronique Gibbons, Katie Reid, Harry Hutchins, Ian Sheerin Infections are a common and growing cause of hospital admission in New Zealand, and a proportion of these admissions are complicated by sepsis. Sepsis (blood poisoning/toto pirau) is what happens when the body’s response to an infection damages its own tissues and organs. Although the exact proportion is unknown, we do know that infections occur more commonly in young children, older adults, people with ‘chronic conditions’ and those who live in areas of economic deprivation. This is the first study in New Zealand to look at the costs of a hospital admission with sepsis. The findings suggest that a quarter of hospital in-patients have an infection that can lead to sepsis. When sepsis develops, it costs $11,000 on average and $13,000 for people who are re-admitted after a sepsis episode, comparable with major trauma. A retrospective analysis of calls to the New Zealand National Poisons Centre regarding Pacific patients Eeva-Katri Kumpula, Rosalina Richards, Pauline Norris, Vanda Symon, Adam C Pomerleau This study aimed to describe contacts to the New Zealand National Poisons Centre (NZNPC) in 2018–2019 relating to Pacific patients. Of all 40,185 patients, Pacific patients comprised 3.4%, non-Pacific 61.9% and people of unknown ethnicities 34.7%. A total of 78.0% of Pacific patients were aged 0–5, 70.4% of exposures were due to child exploratory behaviour, where a child explores their surroundings (eg, putting things in their mouths), and 96.8% occurred in residential settings. Pacific patients were most often exposed to substances commonly found in households. The wellbeing of children and youth is central to the aspirations of Pacific peoples, and the NZNPC service could be a useful resource for families if exposures do occur. Interviews with health professionals about the National Child Protection Alert System Patrick Kelly, Melissa Adam, Carmen Basu, Miranda Ritchie, Denise Wilson, Fred Seymour The National Child Protection Alert System functions in every district health board. It helps health providers share information with colleagues across New Zealand about children where there have been serious concerns about possible maltreatment, so those colleagues may be more likely to provide appropriate help and/or recognise recurrent abuse or neglect. This research describes how the National Child Protection Alert System is regarded by a representative sample of health professionals who work within it and make decisions about whether an alert will be placed. Generally, health professionals regard the system positively, but they identify a number of areas where the system could be enhanced to improve access to information and reduce any risk of misuse. Clinician knowledge of driving restrictions following a stroke event Laura Stratton, John Parsons, Shannon Tisbury, Susan Waterworth, Nicola Starkey A stroke can have a dramatic and lasting impact on someone’s life and how they think, move and interact with their environment. Driving after stroke can be particularly challenging, but it is often seen as vital to successful rehabilitation. Health professionals working in acute and rehabilitation services are often expected to provide information and advice about driving

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 5 www.nzma.org.nz/journal

restrictions to people in the early stages following stroke. We surveyed 49 health professionals about their knowledge of New Zealand Transport Agency guidelines for driving after stroke and found that many health professionals have discussions with people about driving restrictions following a TIA or stroke. However, there appears to be limited knowledge of all the restrictions for each condition as they relate to either private or commercial vehicle use. Insufficient training and education for clinicians might explain this gap. Prognostic significance of mid-range ejection fraction following Acute Coronary Syndrome (ANZACS-QI 23) Daniel Chan, Robert N Doughty, Mayanna Lund, Mildred Lee, Katrina Poppe, Andrew J Kerr Previous studies have shown that patients who have suffered a heart attack and who have reduced heart pump function with an ejection fraction of less than 40% have a high risk of dying. In this study, we followed-up over 6,000 New Zealand patients who had a heart attack in 2015. We found that those with mid-range heart pump function with an ejection fraction of 40–50% had a 1.5-fold higher risk of death compared to those with normal ejection fraction, but they had a lower risk compared to those with reduced ejection fraction. Unlike those with reduced ejection fraction following a heart attack, there are no medications that improve outcomes in patients with mid-range ejection fraction. Hence further study is required to identify medications that can improve outcomes for this group. Radiological imaging of melanoma: a review to guide clinical practice in New Zealand Victoria Francis, Richard Martin, Tvesa Sehji, Mark Barnett The aim of this review is to propose guidelines for initial radiological staging and the follow-up imaging regime for melanoma. This will provide consistency in the access and delivery of quality melanoma care. Radiological imaging plays an important role in assessing the extent of disease, guiding individual treatment and evaluating treatment response. However, there exists limited literature addressing the optimal radiological staging and surveillance imaging regimes for melanoma. The lack of consensus on imaging for melanoma can generate inconsistency in the standard of skin cancer care provided. Increasing access to contraception in New Zealand: assessing the impact of a new funding initiative Orna McGinn, Vicki Mount, Helen Fulcher New Zealand has high rates of unplanned pregnancy and abortion. Many barriers exist to women’s ability to access the most effective forms of contraception such as the IUD and implant—LARCs (long-acting reversible contraception). Funding released by the Ministry of Health last year to improve access to LARC contraception has failed to make an impact, with each area of the country configuring services in their own way and applying variable eligibility criteria. Information accessed via Official Information Act requests shows that most district health boards have failed to deliver the contraception services for which they were contracted. Think about cats in acute vision loss Aaron Yap, Moaz Alshaikhi, Kay Evans Bartonella henselae is a bacteria that can be transferred from cat to humans through cat bites and scratches to cause systemic illness. In a small proportion of cases, it can cause inflammation of the optic nerve and vision loss. In this paper, we describe two such cases that presented to the Palmerston North Eye Department. Research has shown that cases with severe vision loss respond well to a combination of steroid and antibiotic therapy.

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 6 www.nzma.org.nz/journal editorial

2021: here we come / tēnei te whakatau! Suzanne Pitama, Tania Huria, Cameron Lacey

gā mihi o te tau hou ki ngā whanau, were responsive to the Treaty of Waitangi whanui o te NZMJ. and utilised non-racist methodologies. It also N The New Zealand Medical Journal encouraged potential authors of the NZMJ (NZMJ) was first published 47 years after to undertake relevant professional devel- the signing of the Treaty of Waitangi. opment to produce responsive anti-racist The NZMJ honours this legacy with the research that has a focus on health equity. inclusion of a publication from 100 years In another NZMJ editorial, Crampton (2020) ago in each issue. These past articles also highlighted the lack of progress of act as a signpost of changes in scientific Hauora Māori outcomes over the last twenty knowledge and health practices over time. years. He called for a significant cultural A review of these articles documents the change in the whole health sector “to rid our culture of the medical profession and the society of racism, and to demand equity in NZMJ throughout this time. They prioritise the structures, processes, and outcomes for 2 authorship of the mainly white, male work- Māori and Pacific New Zealanders.” force reporting on specific patient cases Therefore the first issue of the NZMJ in and describe an experimental approach to 2021 provides an opportunity to respond further understand or treat the presenting to the need for further evolution with a clear complaints. These articles reflect the kaupapa that not only accepts the critique historical privilege of western science (such of Selak et al (2020),1 but that also acknowl- as anecdotal evidence being presented edges recent changes in health research as fact) in the culture of the New Zealand funding criteria (Māori Health Advancement medical profession. Guidelines),3 health research guidelines,4,5 6,7 By 2002 (when the journal migrated to an the influence of policy on health equity electronic format), the NZMJ had changed to and the extensive body of evidence on the 7–13 embrace an evidence-based framework that impacts of racism on hauora in its consid- was clinically relevant to the New Zealand eration of appropriate publications. health sector. This embracing of evidence- As the new Māori subeditors for the NZMJ, based medicine signposted a cultural we will work alongside the editor and current evolution of the NZMJ. This evolution subeditors to support these cultural changes: includes the publishing of evidence-based • Develop a framework for potential articles from New Zealand health practi- authors so they can work to submit tioners and evidence relevant to the New articles that align with the cultural Zealand health sector. expectations of the NZMJ. The year 2020 marked a further cultural • Utilise a Treaty of Waitangi framework, evolution for the NZMJ, by the identification to advocate for a minimum target of of the journal’s growing stakeholder group articles that report on Hauora Māori that includes tangata Māori, health stake- outcomes. holders (clinicians, scientists, organisational • Ensure the ongoing retention of administrators, researchers and health themed issues in health equity (as was systems) and communities impacted by the 4 September 2020 issue). racism and inequity in health outcomes. This cultural evolution was highlighted by Selak, • Increase the number of Hauora Māori Rahiri, Jackson and Harwood’s editorial in expert peer reviewers. the 4 September 2020 issue.1 They extended • Increase opportunities for commu- a call to the NZMJ to enact further structural nities who are impacted by inequity in changes that would ensure publications health to be published.

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 7 www.nzma.org.nz/journal editorial

• Provide a template to guide peer to stimulate debate that supports leading reviews to align with the cultural our communities to wellbeing. expectations of the NZMJ. Like any cultural evolution it will take • Oversee all publications and provide time, but resources such as the Māori Health further peer review, or have the right Advancement Guidelines3 and the CONSIDER to reject publications that do not align statement5 will provide foundation docu- with the cultural expectations of the ments for these changes. NZMJ stakeholders NZMJ. within their own research and health So what are the aims of the NZMJ going environments are already enacting similar forward in 2021? Perhaps these aims are cultural evolutions. The Treaty of Waitangi, best captured in the new NZMJ byline: as a context, provides a way forward through partnership.6 This current evolution “Te ara tika o te hauora hapora” will now enable us to build on the gains and development of decades of Hauora Māori In concept, these kupu express the intent evidence, innovation and excellence. of the NZMJ to provide a path by which Ngā mihi matou ki a koutou katoa. learning, best practice, clinical practice, health policy and knowledges are presented

Competing interests: Nil. Acknowledgements: We acknowledge Frank Frizelle and David McBride for their active engagement and leadership in this cultural evolution, and for their contribution to the new New Zealand Medical Journal aims. We also acknowledge Lutz Beckert and Maira Patu for their review of this editorial. Author information: Suzanne Pitama: Professor, Māori/Indigenous Health Institute, University of Otago, Christchurch. Tania Huria: Senior Lecturer, Māori/Indigenous Health Institute, University of Otago, Christchurch. Cameron Lacey: Senior Lecturer, Māori/Indigenous Health Institute, University of Otago, Christchurch. Corresponding author: Professor Suzanne Pitama, Māori/Indigenous Health Institute, University of Otago, Christchurch, PO Box 4345, Christchurch [email protected] URL: www.nzma.org.nz/journal-articles/2021-here-we-come-open-access

REFERENCES 1. Selak V, Rahiri J-L, 3. Health Research Council 5. Huria T, Palmer SC, Jackson R, Harwood M. of New Zealand. Māori Pitama S, Beckert L, Acknowledging and acting Health Advancement Lacey C, Ewen S, et al. on racism in the health Guidelines. 2019. Consolidated criteria for sector in Aotearoa New 4. Reid P, Paine S-J, Curtis strengthening report- Zealand. The New Zealand E, Jones R, Anderson A, ing of health research Medical Journal (Online). Willing E, et al. Achieving involving indigenous 2020;133(1521):7-13. health equity in Aotearoa: peoples: the CONSIDER 2. Crampton P. Oh my. strengthening respon- statement. BMC Med Res The New Zealand Medi- siveness to Māori in Methodol. 2019;19(1):173. cal Journal (Online). health research. NZ Med 6. Baker G, Baxter J, 2020;133(1524):8-10. J. 2017;130(1465):96-103. Crampton P. The primary

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 8 www.nzma.org.nz/journal editorial

healthcare claims to 9. Harris RB, Stanley J, experience of Māori and the Waitangi Tribunal. Cormack DM. Racism and Indigenous peoples. Public The primary healthcare health in New Zealand: health. 2019;172:119-24. claims to the Waitangi Prevalence over time 12. Stanley J, Harris R, Tribunal. 2019;132(1505). and associations between Cormack D, Waa A, 7. McLeod M, Gurney J, recent experience of racism Edwards R. The impact Harris R, Cormack D, and health and wellbeing of racism on the future King P. COVID‐19: we measures using national health of adults: proto- must not forget about survey data. PloS one. col for a prospective Indigenous health and 2018;13(5):e0196476. cohort study. BMC public equity. Australian and 10. Paine S-J, Cormack D, health. 2019;19(1):346. New Zealand journal of Stanley J, Harris R. 13. Talamaivao N, Harris public health. 2020. Caregiver experiences of R, Cormack D, Paine 8. Harris RB, Cormack DM, racism are associated with S-J, King P. Racism and Stanley J. Experience of adverse health outcomes health in Aotearoa New racism and associations for their children: a Zealand: a systematic with unmet need and cross-sectional analysis of review of quantitative healthcare satisfaction: data from the New Zealand studies. The New Zealand the 2011/12 adult New Health Survey. Critical Medical Journal (Online). Zealand health survey. Public Health. 2019:1-12. 2020;133(1521):55-5. Australian and New 11. Reid P, Cormack D, Paine Zealand Journal of Public S-J. Colonial histories, Health. 2019;43(1):75-80. racism and health—The

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 9 www.nzma.org.nz/journal article

ABSTRACT AIM: To explore the population-at-risk and potential cost of a sepsis episode in New Zealand. METHOD: Retrospective analysis of the National Minimum Data Set using two code-based algorithms selecting (i) an inclusive cohort of hospitalised patients diagnosed with a ‘major infection’ with the potential to cause sepsis and (ii) a restricted subset of these patients with a high likelihood of clinical sepsis based on the presence of both a primary admission diagnosis of infection and at least one sepsis-associated organ failure. RESULTS: In 2016, 24% of all inpatient episodes were associated with diagnosis of a major infection. The sepsis coding algorithm identified a subset of 1,868 discharges. The median (IQR) reimbursement associated with these episodes was $10,381 ($6,093–$10,964). In both groups, 30-day readmission was common (26.7% and 11% respectively). CONCLUSIONS: Infectious diseases with the potential to cause sepsis are common among hospital inpatients. Direct treatment costs are high for those who present with or progress to sepsis due to these infections.

epsis is defined as “life-threatening Sepsis is a complication of infection. In organ dysfunction due to a dysregu- New Zealand, infection-related public-hos- Slated immune response to infection.”1 pital admissions have increased significantly Sepsis is a major health challenge globally, over time, particularly among Māori and with incidence stratified by geography Pacific people and those facing high levels and national income.2 In high-income coun- of socioeconomic deprivation.7–11 Presen- tries, sepsis-associated mortality remains tations with infectious diseases and sepsis high, with a wide variation based on the are therefore a major barrier to popu- age and underlying health status of the lation health equity, and their prevention, individual.3 A proportion of patients mitigation and treatment are deserving of with sepsis require treatment in an in- investment. Investment requires an under- tensive care unit (ICU), survivors often standing of the scale of the underlying require long stays in hospital and hospital problem and its associated cost. There are readmission is common.4,5 Unsurprising- no studies reporting the cost of infection ly perhaps, sepsis is a leading cause of and sepsis to the New Zealand public health healthcare spending. In the US in 2018, system. We used routine data to estimate USD$22,000,000,000 was charged to the (i) the number of inpatients with infections Medicare and Medicaid budgets for inpa- that can cause sepsis and (ii) the potential tient sepsis management.6 cost of a sepsis episode.

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 10 www.nzma.org.nz/journal article

of Disease study, and which is reported Methods to exhibit 50% sensitivity and 94% speci- This study was registered as a quality ficity against the 2001 consensus definition improvement activity with the Clinical Audit of ‘severe sepsis’.2,8,13 This method was Support Unit at Waikato District Health therefore selected to define sepsis within the Board (WDHB). It was considered a low-risk NMDS and from then on was referred to as observational study and therefore out of the ‘New Zealand Sepsis’ indicator (NZS, see scope for New Zealand Health and Disability Appendix). Ethics Committee review. Funding for an Due to the syndromic nature of sepsis (as independent health-economist (IS) was opposed to the binary presence or absence provided by the Accident Compensation of infections with specific ICD-10-AM codes), Corporation (ACC). clinical validation of the NZS algorithm Defining infection and sepsis was undertaken by reviewing a sample of using routine data clinical records at WDHB. We retrospec- This was a retrospective analysis of tively identified 100 NZS discharges from the National Minimum Data Set (NMDS). WDHB facilities in each of two one-year time The analysis made use of codes derived periods (July to June 2008/09 and 2012/13). from the International Classification of These adult patients were found to have Disease, Tenth Edition, Australian Modifi- confirmed sepsis if their presentations were cation (ICD-10-AM). The a priori design of both consistent with infection and asso- this explorative study addressed several ciated with a new increase of two or more problems known to impact studies of sepsis in the modified-Sequential Organ Failure epidemiology and cost. Assessment (mSOFA).15 Use of the original Sequential Organ Failure Assessment (SOFA) Firstly, we had to identify a source of score is required to satisfy the current data from which to derive estimates of clinical definition of sepsis.1 mSOFA replaces prevalence and cost. Although prospective the cardiovascular and respiratory require- databases are maintained to identify sepsis ments of the original score to make use within intensive care unit admissions, of information typically entered into the limiting studies to ICU-treated populations clinical record. is highly problematic.4 The NMDS is the only resource available to judge the total number Thirdly, we recognised the limited sensi- of infectious disease and sepsis-associated tivity of the NZS algorithm and, therefore, hospital admissions in New Zealand. It has our inability to identify all patients with been the preferred data source for national sepsis from the NMDS. Instead, we sought to reporting of infection-related hospital identify the hospitalised population-at-risk admissions and is linked to hospital reim- of sepsis. This approach is in routine use bursement data.7 The NMDS was therefore in the UK and is used to identify trends in chosen as the data source for this study. the presentation and outcome of specific infectious diseases in NHS hospitals. The Secondly, we needed a method to so-called ‘suspicion of sepsis’ approach was identify sepsis within the NMDS. Signif- first developed by Inada-Kim et al.14 These icant controversy and debate surround authors conducted a consensus review of the contemporary clinical definition of the International Classification of Disease sepsis, and the limitations associated with to extract all infectious disease diagnoses defining it within routine data, are well commonly complicated by sepsis. To these described.4,12–14 Briefly, the clinical defi- codes we added 14 that were part of the nition of sepsis has changed over time, as sepsis coding strategy developed by Huggan have the International Classification of et al.4 From then on we labelled this algo- Disease versions from which sepsis coding rithm as the ‘New Zealand Major Infection’ algorithms are constructed.1,12 Multiple code- (NZMI) indicator. based definitions of sepsis exist, and their accuracy has been reported against different In summary, to estimate the popula- populations in different health systems.3,12 tion-at-risk of sepsis, we identified all The only published study of sepsis incidence patients admitted to New Zealand hospitals in New Zealand was based on an approach with infections known to cause this subsequently adopted by the Global Burden condition (NZMI). From within this cohort,

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 11 www.nzma.org.nz/journal article

we identified a subpopulation with a high Related Groups (DRGs) linked to ICD-10-AM likelihood of having true clinical sepsis (NZS) codes. For cases not covered by the Casemix and validated this assumption by conducting System (namely those paid by Crown a clinical record review. agencies such as ACC), we used the average Data extraction and hospital inlier costs for relevant DRGs. We had no data relating to reimbursements for private reimbursement hospitals or facilities run by community The National Minimum Data Set (NMDS) trusts. To compare case-weighted reim- was used to identify discharges meeting bursement with true inpatient costs at NZS and NZMI criteria for the 2016 calendar Waikato District Health Board, we used year (see Appendix). We extracted 30-day i.Patient Manager (DXC Technology, Tysons readmissions for any reason through to Corner, US) to describe the actual costs of 31 January 2017. The NMDS was accessed care for patients included in the NZS clinical under a pre-existing memorandum of under- validation cohort. standing between the Ministry of Health and ACC. This limited the information provided to the patient’s age, district health board and Results discharge diagnosis codes. Mortality and Regarding validation of the NZS algorithm, ethnicity data were not available. 192 sets of clinical records were available Data were entered into Microsoft Excel for review. Clinical sepsis was identified (2016) and further analysed in SAS Enter- in 165 (86%); 125 (76%) of these satisfied prise Guide (version 7.1). Public-hospital the clinical sepsis definition (mSOFA score reimbursement for each case was derived of two or more) at first presentation to from the New Zealand Casemix System for hospital, 43 (26%) identified as Māori, 36 Publicly Funded Hospitals (WEISNZ16v1.0, (22%) were admitted to ICU and 30 (18%) NCCP Casemix—Cost Weights Project Group, died in hospital. 2016).16 This system uses case-weights to Table 1 shows the number of cases iden- estimate average costs for cases of varying tified using the NZMI and NZS indicators in complexity, as determined by Diagnosis 2016, stratified by age group.

Table 1: Hospital discharges identified by the New Zealand Major Infection (NZMI) and New Zealand Sepsis (NZS) indicators, 2016.

Age group N % N % (years)

NZMI NZS

0–2 13,255 8 18 1

3–19 14,136 8 35 2

20–29 15,053 9 28 2

30–39 11,449 6 41 2

40–49 11,950 7 86 5

50–59 16,337 9 202 11

60–69 22,832 13 330 18

70–79 29,471 17 472 25

80 and over 40,136 23 656 35

Total 174,619 100 1,868 100

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 12 www.nzma.org.nz/journal article

In the 2016 calendar year, we estimated that there were 725,294 non-day-stay Discussion discharges from New Zealand public To our knowledge, this is the first study hospitals (see Appendix). 174,619 that attempts to report hospital resource discharges (24%) were associated with utilisation associated with episodes of a NZMI code. 47% of patients were infection and sepsis in New Zealand. Codes male, 40% were over 70 years of age for ‘major infection’ were associated with and 16% were under 20. NZMI admissions 24% of all hospital discharges, almost absorbed 949,026 hospital bed days, for 1,000,000 hospital bed days and over which $1,191,279,897 was reimbursed. $1,000,000,000 in reimbursement. A high The average length of stay (ALOS) for proportion of patients were readmitted to these admissions was 5.5 days (range hospital within 30 days (27% and 11% of the 1–225 days, median 3.0 days, inter-quartile NZMI and NZS cohorts, respectively). Sepsis ration (IQR) 1–6 days) and the average episodes were high-cost events, and the reimbursement per discharge was actual costs of care for a sepsis cohort iden- $6,822 (range $147–$410,599, median tified at a large district health board were $3,995, IQR $2,231–$6,865). 46,627 26% higher than reimbursement derived NZMI discharges (26.7%) were asso- using the case-weight system. ciated with readmission within 30 days, As an exploratory analysis, our aim was accounting for 341,606 additional bed days to estimate the population-at-risk of sepsis and reimbursement of $373,700,000 (mean and the likely cost of a sepsis episode while $8,014, median $5,167, IQR $2,807–$8,446). recognising the limitations placed on studies We found 3,904 (2.2%) NZMI cases that using routine data. We did this by applying were not reimbursed using the Casemix two entirely different algorithms to a single System. Assigning the casemix average database: one which identified patients to these admissions added $26,300,000 to with the infections that cause sepsis (NZMI), the total. the other which identifies patients with a 1,868 hospital discharges were iden- high likelihood of true clinical sepsis (NZS). tified using NZS codes. Of these patients, Comparison of these cohorts provides two 54% were male and 60% were aged 70 or important observations. Firstly, NZMI codes over. NZS admissions absorbed 15,137 more completely represent the bimodal hospital bed days, for which $21,500,000 distribution of infection-related hospital was reimbursed. The ALOS was 8.1 (range admissions, a pattern observed in the Global 1–86, median 6, IQR 3–10) and the average Burden of Disease study but not by the NZS reimbursement per discharge was $11,552 algorithm.2,8 Secondly, both methods demon- (range $717–$181,988, median $10,381, strate a steep increase in the proportion of IQR $6,177–$10,964). There were 203 NZS cases with age. This is a universal obser- discharges (11%) that were associated vation in studies of infection and sepsis with readmission within 30 days. This incidence, including those reported from accounted for an additional 2,418 bed days New Zealand.7,8 and a further reimbursement of $2,800,000 The NZS algorithm was designed to report (average $13,682, range $717–$179,231, sepsis incidence from hospital coding data. median $10,381, IQR $6,093–$10,964). We Due to concerns about the reliability of found 26 (1.4%) NZS cases that were not coding strategies to identify true clinical reimbursed using the Casemix System. sepsis, it aims to maintain specificity for Assigning the casemix average to these the sepsis syndrome at the expense of admissions added $355,732 to the total sensitivity. This is achieved by requiring reimbursement. an explicit organ failure code while also For the 192 patients in the clinical vali- excluding infection codes other than in the dation cohort at Waikato District Health primary position (see Appendix). Merely Board, 79% of the actual costs of care by including cases with infection codes were identified using national casemix in primary or secondary positions in our methodology (costs of $2,150,209 against database, we would have increased the reimbursement of $1,699,155). number of NZS cases by 64% to 3,073, and

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 13 www.nzma.org.nz/journal article

a further 2,615 cases would have been the reimbursement associated with index identified by combining infection and organ hospitalisation. Large studies in the US failure codes in any position. With 86% of have shown that readmission rates after cases satisfying contemporary sepsis defi- sepsis are similar for heart failure and nitions in our validation work, we conclude myocardial infarction.5 Reasons for hospital that NZS codes can be used to estimate readmission are likely to be heterogenous. the cost of sepsis episodes, although they Possibly for this reason, interventions will underestimate sepsis incidence and focused solely on supporting sepsis prevalence. survivors at discharge have shown little 22–24 This brings earlier findings into question. impact on rates of readmission. Intrigu- In the Waikato region, the NZS algorithm ingly, though, total healthcare utilisation led to an estimate of 107 cases of sepsis per does appear to be reduced by efforts to 100,000 in the year to June 2012.4 This is at identify and treat patients at risk of sepsis the lower limit of sepsis incidence estimated in hospital. A machine-learning algorithm in high-income economies by the Global designed to identify sepsis using electronic Burden of Disease study, which employed medical records reduced 30-day read- code-based methods to estimate 120 to mission rates from a baseline of 46% to 23% 25 200 cases per 100,000 population in high- in one single-centre study. Evaluation of income countries including New Zealand a state-wide sepsis quality improvement and Sweden.2 Swedish studies identifying programme in New South Wales, Australia, the presence of sepsis in patients receiving pointed to a reduction in intensive care 26 intravenous antibiotics report annual sepsis utilisation and the total length of stay. The rates of 800 per 100,000 population.20,21 By hypothesis proposed by these authors and implication, rates of sepsis are much higher others is that early sepsis identification and in New Zealand than previously reported. treatment improves clinical recovery by Better estimates of sepsis incidence are preventing the accumulation of sepsis-as- needed, particularly given the severity of sociated tissue injury. We can’t support this the associated outcomes and the high cost conclusion from the data provided here, but presented to public hospitals. we have shown that quality improvement programmes aimed at preventing, mitigating We also note that critical illnesses and treating infection and sepsis would be requiring complex multidisciplinary relevant to a high proportion of our inpa- care have been associated with deficits tient population. in hospital reimbursement. For example, the average case-weighted inpatient reim- A major weakness of our study is the bursement for major trauma at Whangārei omission of data relating to mortality Hospital from 2015 to 2017 was $17,042, but and ethnicity. The dynamic impacts of the actual costs of care were 36% higher.17 infection are most marked among popu- For both trauma and sepsis, additional lations suffering high rates of chronic costs will extend well beyond hospital care, morbidity and socioeconomic disadvantage, with non-inpatient (‘indirect’) costs adding which unfortunately includes a significant substantially to total spending following proportion of Māori and Pacific people. For critical illness.18 Sepsis has recently been example, compared with non-Māori living shown to cause a durable increase in in the Waikato, Māori are 3.2 times more health spending over at least five years of likely to suffer sepsis and at a much younger 4 follow-up.19 Further work is required to age. Under-reporting rates of infection and establish a better estimate of short- and sepsis at a national level risks obscuring the long-term costs, but a clue to the true extent important contribution of these conditions of resource utilisation associated with to health inequity. infectious disease and sepsis diagnosis is In summary, infection and sepsis are provided by the high readmission rate found costly and previously under-appreciated in this study. sources of direct healthcare spending in The 30-day readmissions in the NZMI New Zealand. Total healthcare spending and NZS cohorts respectively added 31% on sepsis will be significantly higher than ($373,700,000 added to $1,200,000,000) and reported here, due to under-reporting, the 13% ($2,800,000 added to $21,500,000) to ongoing costs of care in the community

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 14 www.nzma.org.nz/journal article

and, potentially, the significant gap between a high probability of sepsis, which can be reimbursement and actual spending. The used to study clinical outcomes and costs. NZMI and NZS approaches have their Both groups would benefit from invest- strengths and weaknesses. The first can ments in infection control, antimicrobial estimate the size of the inpatient population stewardship and sepsis care aimed at at risk of sepsis, and the second can provide preventing or reducing long lengths of stay a representative sample of patients with and readmission.

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 15 www.nzma.org.nz/journal article

Appendix Figure A1: New Zealand Major Infection and New Zealand Sepsis indicator methodologies Hospital discharge episodes in 2016 (eg, from 1 January 2016 to 31 December 2016) were identified using two separate algorithms applied to the National Minimum Data Set (NMDS). The resulting cohorts were analysed separately. For each episode, readmission within 30 days was identified. In both cohorts, admission more than 30 days after the index discharge was counted as a separate episode. To estimate total in-patient discharges in calendar year 2016, we first subtracted day-case admissions from total reported hospital episodes provided by the New Zealand Ministry of Health (tables available at https://www.health.govt.nz/nz-health-statistics/ health-statistics-and-data-sets/publicly-funded-hospital-discharges-series/publicly-funded-hospital-discharges-series/publicly-funded-hospital-discharges-series. [Accessed October 2020]). We then calculated an average based on numbers derived for the period July–June 2016/2017 and 2015/2016002E. New Zealand Major Infection indicator The ‘New Zealand Major Infection’ (NZMI) indicator is comprised of the ICD-10-AM codes identified by Inada-Kim et al,14 with the addition of 14 ICD-10-AM codes used in a Waikato-based study conducted by Huggan et al4 These ICD-10 codes are applied to the first 30 diagnosis codes entered into the NMDS. Codes are listed under ICD-10-AM chapter headings.

I. Certain infectious and parasitic diseases 1. A01 Typhoid and paratyphoid fevers (incl. A01.0, A01.1, A01.2, A01.3, A01.4) 2. A02 Other salmonella infections (incl. A02.0, A02.1, A02.2, A02.8, A02.9) 3. A03 Shigellosis (incl. A03.0, A03.1, A03.2, A03.3, A03.8, A03.9) 4. A04 Other bacterial intestinal infections (incl. A04.0, A04.1, A04.2, A04.3, A04.4, A04.5, A04.6, A04.7, A04.8, A04.9) 5. A06 Amoebiasis (incl. A06.0, A06.1, A06.2, A06.3, A06.4, A06.5, A06.6, A06.7, A06.8, A06.9) 6. A15 Respiratory tuberculosis (incl. A15.0, A15.2, A15.3, A15.4, A15.5, A15.6, A15.7, A15.8, A15.9) 7. A16 Respiratory tuberculosis, not confirmed bacteriologically or histologically (incl. A16.0, A16.1, A16.2, A16.3, A16.4, A16.5, A16.7, A16.8, A16.9) 8. A17 Tuberculosis of nervous system (incl. A17.0, A17.1, A17.8, A17.9) 9. A18 Tuberculosis of other organs (incl. A18.0, A18.1, A18.2, A18.3, A18.4, A18.5, A18.6, A18.7, A18.8) 10. A19 Miliary tuberculosis (incl. A19.0, A19.1, A19.2, A19.8, A19.9) 11. A27 Leptospirosis (incl. A27.0, A27.8, A27.9) 12. A32 Listerosis (incl. A32.0, A32.1, A32.7, A32.8, A32.9) 13. A37 Whopping cough (all subcategories) 14. A38 Scarlet fever 15. A39 Meningococcal infection (incl. A39.0, A39.1, A39.2, A39.4, A39.5, A39.8, A39.9) 16. A40 Streptococcal sepsis (incl. A40.0, A40.1, A40.2, A40.3, A40.8, A40.9) 17. A41 Other Sepsis (incl. A41.0, A41.1, A41.2, A41.3, A41.4, A41.5, A41.8, A41.9) 18. A42 Actinomycosis (all subcategories) 19. A43 Nocardiosis (all subcategories)

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 16 www.nzma.org.nz/journal article

20. A44 Bartenollosis (all subcategories) 21. A46 Erysipelas 22. A48 Other Bacterial diseases, not elsewhere classified (incl. A48.0, A48.1, A48.2, A48.3, A48.4, A48.8) 23. A49 Bacterial infection of unspecified site (incl. A49.0, A49.1, A49.2, A49.3, A49.8, A49.9) 24. A51 Early syphilis (all subcategories) 25. A54 Gonococcal infection (incl. A54.1, A54.2, A54.3, A54.4, A54.5, A54.6, A54.8, A54.9) 26. A55 Chlamydial lymhogranuloma (venereum) 27. A56 Other sexually transmitted chlamydial diseases (incl. A56.0, A56.1, A56.2, A56.3, A56.4, A56.8) 28. A68 Relapsing fevers (all subcategories) 29. A69.2 Lyme disease 30. A70 Chlamydia psittaci infection 31. A75 Typhus fever (all subcategories) 32. A77 Spotted fever (all subcategories) 33. A78 Q fever 34. A79 Other rickettsioses (all subcategories) 35. B59 Pneumocystosis

VI. Diseases of the nervous system 36. G00 Bacterial meningitis, not elsewhere classified (incl. G00.0, G00.1, G00.2, G00.3, G00.8, G00.9) 37. G01 Meningitis in bacterial diseases classified elsewhere 38. G04.2 Bacterial meningoencephalitis and meningomyelitis, not elsewhere classified 39. G06 Intracranial and intraspinal abscess and granuloma (incl. G06.0, G06.1, G06.2)

VIII. Diseases of the ear and mastoid process 40. H60 Otitis externa (incl. H60.0, H60.1, H60.2, H60.3) 41. H66 Suppurative and unspecified otitis media (incl. H66.0, H66.4, H66.9) 42. H67.0 Otitis media in bacterial diseases classified elsewhere 43. H68.0 Eustachian salpingitis 44. H70 Mastoiditis and related conditions (incl. H70.0, H70.9) 45. H73.0 Acute myringitis

IX. Diseases of the circulatory system 46. I00 Rheumatic fever without mention of heart involvement 47. I01 Rheumatic fever with heart involvement (incl. I01.0, I01.1, I01.2, I01.8, I01.9) 48. I02 Rheumatic chorea (incl. I02.0, I02.9) 49. I33 Acute and subacute endocarditis (incl. I33.0, I33.9) 50. I38 Endocarditis, valve unspecified

X. Diseases of the respiratory system 51. J01 Acute sinusitis (incl. J01.0, J01.1, J01.2, J01.3, J01.4, J01.8, J01.9) 52. J02 Acute pharyngitis (incl. J02.0, J02.9) 53. J03 Acute tonsillitis (incl. J03.0, J03.9)

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 17 www.nzma.org.nz/journal article

54. J05.1 Acute epiglottitis 55. J06.9 Acute upper respiratory infection, unspecified 56. J13 Pneumonia due to Streptococcus pneumoniae, 57. J14 Pneumonia due to Haemophilus influenza, 58. J15 Bacterial pneumonia, not elsewhere classified (J15.0, J15.1, J15.2, J15.3, J15.4, J15.5, J15.6, J15.7, J15.8, J15.9) 59. J16 Pneumonia due to other infectious organisms, not elsewhere classified (incl. J16.0, J16.8)] 60. J17.0 Pneumonia in bacterial diseases classified elsewhere (incl. J17.0, J17.8) 61. J18 Pneumonia, organism unspecified (including J18.0, J18.1, J18.2, J18.8 and J18.9) 62. J20 Acute bronchitis (incl. J20.0, J20.1, J20.2, J20.8, J20.9) 63. J22 Unspecified acute lower respiratory infection 64. J36 Peritonsillar abscess 65. J39 Other diseases of upper respiratory tract (incl. J39.0, J39.1) 66. J44.0 Chronic obstructive pulmonary disease with acute lower respiratory infection 67. J69 Pneumonitis due to solids and liquids (incl. J69.0, J69.8) 68. J84.9 Interstitial pulmonary disease unspecified (interstitial pneumonia NOS) 69. J85 Abscess of lung and mediastinum (incl. J85.1, J85.2, J85.3) 70. J86 Pyothorax (incl. J86.0, J86.9) 71. J95.0 Sepsis of tracheostomy stoma 72. J98.5 Diseases of mediastinum, not elsewhere classified- Mediastinitis

XI. Diseases of the digestives system (dental disorders omitted) 73. K22.3 Perforation of oesophagus 74. K35 Acute appendicitis (incl. K35.2, K35.3, K35.8) 75. K37 Unspecified appendicitis 76. K57 Diverticular disease of intestine (incl. K57.0, K57.2, K57.4, K57.8,) 77. K61 Abscess of anal and rectal regions (incl. K61.0, K61.1, K61.2, K61.3, 61.4) 78. K63.0 Abscess of intestine 79. K63.1 Perforation of intestine (nontraumatic) 80. K65.0 Acute peritonitis (incl. K65.0, K65.8, K65.9) 81. K67 Disorders of peritoneum in infectious diseases classified elsewhere (all subcategories) 82. K75.0 Abscess of liver 83. K80.0 Calculus of gallbladder with acute cholecystitis/cholangitis (incl. K80.0, K80.1, K80.3, K80.4) 84. K81 Cholecystitis (incl. K81.0, K81.1, K81.8, K81.9) 85. K82.2 Perforation of gallbladder 86. K83.0 Cholangitis 87. K83.2 Perforation of bile duct

XII. Diseases of skin and subcutaneous tissue 88. L00 Staphylococcal scalded skin syndrome 89. L01 Impetigo (L01.0, L01.1)

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 18 www.nzma.org.nz/journal article

90. L02 Cutaneous abscess, furuncle and carbuncle (incl. L02.0, L02.1, L02.2, L02.3, L02.4, L02.8, L02.9) 91. L03 Cellulitis (including L03.0, L03.1, L03.2, L03.3, L03.8 and L03.9) 92. L05.0 Pilonidal cyst with abscess 93. L08 Other local infections of skin and subcutaneous tissue (incl. L08.0, L08.8, L08.9) 94. L30.3 Infective dermatitis 95. L53.3 Erythema marginatum 96. L98.0 Pyogenic granuloma

XIII. Diseases of the musculoskeletal system and connective tissue 97. M00 Pyogenic arthritis (incl. M00.0, M00.1, M00.2, M00.8, M00.9) 98. M01 Direct infections of joint in infectious and parasitic diseases classified elsewhere (incl. M01.0, M01.1, M01.2, M01.3) 99. M46.2 Osteomyelitis of vertebra 100. M46.4 Discitis, unspecified 101. M65 Synovitis and tenosynovitis (incl. M65.0, M65.1) 102. M71.0 Abscess of bursa 103. M72.6 Necrotizing fasciitis 104. M86 Osteomyelitis

XIV. Diseases of genitourinary system 105. N10 Acute tubulo-interstitial nephritis 106. N11 Chronic tubulo-interstitial nephritis (incl. N11.0, N11.1, N11.8, N11.9) 107. N12 Tubulo-intestitial nephritis, not specified as acute or chronic 108. N13.6 Pyonephrosis 109. N15.1 Renal and perinephric abscess 110. N15.9 Renal tubulo-interstitial disease, unspecified 111. N30 Cystitis, unspecified (including N30.0, N30.8, N30.9) 112. N34.0 Urethral abscess 113. N39.0 Urinary tract infection, site not specified 114. N41.0 Acute prostatitis 115. N43.1 Infected hydrocele 116. N45 Orchitis and epididymitis (incl. N45.0, N45.9) 117. N48.2 Other disorders of penis (incl. N48.1, N48.2) 118. N49.9 Inflammatory disorder of unspecified male genital organ 119. N61 Inflammatory disorders of breast 120. N70 Salpingitis and oophoritis (incl. N70.0, N70.9) 121. N71 Inflammatory disease of uterus, except cervix (incl. N71.0, N71.9) 122. N73 Other female pelvic inflammatory diseases (incl. N73.0, N73.1, N73.2, N73.4, N73.9) 123. N75.1 Abscess of Bartholin gland 124. N76 Other inflammation of vagina and vulva (incl. N76.0, N76.1, N76.3, N76.4, N76.8)

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 19 www.nzma.org.nz/journal article

XV. Pregnancy, childbirth and the puerperium 125. O08.0 Genital tract and pelvic infection following abortion and ectopic and molar pregnancy 126. O23 Infections of genitourinary tract in pregnancy (incl. O23.0, O23.1, O23.2, O23.3, O23.4, O23.5, O23.9) 127. O41.1 Infection of amniotic sac and membranes 128. O85 Puerperal sepsis 129. O86 Other puerperal infections (incl. O86.0, O86.1, O86.2, O86.3, O86.4, O86.8) 130. O88.3 Obstetric pyaemic and septic embolism 131. O91 Infections of breast associated with childbirth (incl. O91.0, O91.1)

XVI. Certain conditions originating in the perinatal period 132. P36 Bacterial sepsis of newborn (incl. P36.0, P36.1, P36.2, P36.3, P36.4, P36.5, P36.8, P36.9) 133. P39 Other infections specific to the perinatal period (incl. P39.0, P39.2, P39.3, P39.4, P39.8, P39.9) 134. P78 Other perinatal digestives system disorders (P78.0, P78.1,) 135. T814 Infection following a procedure, not elsewhere classified 136. T845 Infection and inflammatory reaction due to internal joint prosthesis

XVIII. Symptoms, signs and abnormal clinical and laboratory findings, not elsewhere classified 137. R57.2 Septic shock 138. R65 Systemic Inflammatory Response syndrome (incl. R65.0, R65.1, R65.9)

The following 14 ICD-10-AM codes were added to the NZMI indicator as they are included in the NZS indicator and part of the study conducted by Huggan et al.

A241 Acute and fulminating melioidosis B377 Candidal sepsis B387 Disseminated coccidioidomycosis B393 Disseminated histoplasmosis capsulati B407 Disseminated blastomycosis B417 Disseminated paracoccidioidomycosis B427 Disseminated sporotrichosis B447 Disseminated aspergillosis B457 Disseminated cryptococcosis B464 Disseminated mucormycosis A4150 Sepsis due to unspecified Gram-negative organisms A4151 Sepsis due to Escherichia coli [E Coli] A4152 Sepsis due to Pseudomonas A4158 Sepsis due to other Gram-negative organisms

New Zealand Sepsis indicator The New Zealand Sepsis (NZS) indicator is present when a ‘Primary Infection’ code is found together with an ‘Organ Failure’ code.

Definition of ‘Primary Infection’: Where a pre-specified ICD10 code defining infectious

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 20 www.nzma.org.nz/journal article

disease was present in the first (primary) diagnosis position the indicator ‘Primary_ infection’ was assigned. Where the primary position was occupied by an ICD10 Z-code and an indicator code (as defined below) was in the second position, the ‘Primary_infection’ indicator was also assigned. Note that in the original study by Huggan et al4 identified only Infection Codes in the first (primary) position.

1 A010 Typhoid fever 2 A021 Salmonella sepsis 3 A190 Acute miliary tuberculosis of a single specified site 4 A191 Acute miliary tuberculosis of multiple sites 5 A192 Acute miliary tuberculosis, unspecified 6 A198 Other miliary tuberculosis 7 A199 Miliary tuberculosis, unspecified 8 A241 Acute and fulminating melioidosis 9 A327 Listerial sepsis 10 A394 Meningococcaemia, unspecified 11 A400 Sepsis due to streptococcus, group A 12 A401 Sepsis due to streptococcus, group B 13 A402 Sepsis due to streptococcus, group D 14 A403 Sepsis due to Streptococcus pneumoniae 15 A408 Other streptococcal sepsis 16 A409 Streptococcal sepsis, unspecified 17 A410 Sepsis due to Staphylococcus aureus 18 A411 Sepsis due to other specified staphylococcus 19 A412 Sepsis due to unspecified staphylococcus 20 A413 Sepsis due to Haemophilus influenzae 21 A414 Sepsis due to anaerobes 22 A4150 Sepsis due to unspecified Gram-negative organisms 23 A4151 Sepsis due to Escherichia coli [E Coli] 24 A4152 Sepsis due to Pseudomonas 25 A4158 Sepsis due to other Gram-negative organisms 26 A418 Other specified sepsis 27 A419 Sepsis, unspecified 28 A427 Actinomycotic sepsis 29 A430 Pulmonary nocardiosis 30 A481 Legionnaires' disease 31 A483 Toxic shock syndrome 32 A499 Bacterial infection, unspecified 33 A548 Other gonococcal infections 34 B377 Candidal sepsis 35 A78 Q fever (coded in logic as A780) 36 B387 Disseminated coccidioidomycosis 37 B393 Disseminated histoplasmosis capsulati 38 B407 Disseminated blastomycosis 39 B417 Disseminated paracoccidioidomycosis 40 B427 Disseminated sporotrichosis

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 21 www.nzma.org.nz/journal article

41 B447 Disseminated aspergillosis 42 B457 Disseminated cryptococcosis 43 B464 Disseminated mucormycosis 44 P360 Sepsis of newborn due to streptococcus, group B 45 P361 Sepsis of newborn due to other and unspecified streptococci 46 P362 Sepsis of newborn due to Staphylococcus aureus 47 P363 Sepsis of newborn due to other and unspecified staphylococci 48 P364 Sepsis of newborn due to Escherichia coli 49 P365 Sepsis of newborn due to anaerobes 50 P368 Other bacterial sepsis of newborn 51 P369 Bacterial sepsis of newborn, unspecified d. Organ failure: These ICD-10 codes, applied to the first 30 diagnosis codes, were used to identify organ failure. In addition, the ‘Organ_failure’ indicator was also applied when one of the three operation/procedure codes appeared within the first 30 operation/procedure codes.

1 I950 Idiopathic hypotension 2 I951 Orthostatic hypotension 3 I959 Hypotension, unspecified 4 R031 Nonspecific low blood-pressure reading 5 R572 Septic shock 6 R570 Cardiogenic shock (missing) 7 R571 Hypovolaemic shock 8 R578 Other shock 9 R579 Shock, unspecified 10 D65 Disseminated intravascular coagulation [defibrination syndrome] 11 D688 Other specified coagulation defects 12 D689 Coagulation defect, unspecified 13 D695 Secondary thrombocytopenia 14 D696 Thrombocytopenia, unspecified 15 K720 Acute and subacute hepatic failure 16 E872 Acidosis 17 F050 Delirium not superimposed on dementia, so described 18 F051 Delirium superimposed on dementia 19 F058 Other delirium 20 F059 Delirium, unspecified 21 G934 Encephalopathy, unspecified 22 R400 Somnolence 23 R401 Stupor 24 R402 Coma, unspecified 25 N000 Acute nephritic syndrome, minor glomerular abnormality 26 N001 Acute nephritic syndrome, focal and segmental glomerular lesions 27 N002 Acute nephritic syndrome, diffuse membranous glomerulonephritis 28 N003 Acute nephritic syndrome, diffuse mesangial proliferative glomerulonephritis 29 N004 Acute nephritic syndrome, diffuse endocapillary proliferative

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 22 www.nzma.org.nz/journal article

glomerulonephritis 30 N005 Acute nephritic syndrome, diffuse mesangiocapillary glomerulonephritis 31 N006 Acute nephritic syndrome, dense deposit disease 32 N007 Acute nephritic syndrome, diffuse crescentic glomerulonephritis 33 N008 Acute nephritic syndrome, other 34 N009 Acute nephritic syndrome, unspecified 35 N010 Rapidly progressive nephritic syndrome, minor glomerular abnormality 36 N011 Rapidly progressive nephritic syndrome, focal and segmental glomerular lesions 37 N012 Rapidly progressive nephritic syndrome, diffuse membranous glomerulonephritis 38 N013 Rapidly progressive nephritic syndrome, diffuse mesangial proliferative glomerulonephritis 39 N014 Rapidly progressive nephritic syndrome, diffuse endocapillary proliferative glomerulonephritis 40 N015 Rapidly progressive nephritic syndrome, diffuse mesangiocapillary glomerulonephritis 41 N016 Rapidly progressive nephritic syndrome, dense deposit disease 42 N017 Rapidly progressive nephritic syndrome, diffuse crescentic glomerulonephritis 43 N018 Rapidly progressive nephritic syndrome, other 44 N019 Rapidly progressive nephritic syndrome, unspecified 45 N170 Acute kidney failure with tubular necrosis 46 N172 Acute kidney failure with medullary necrosis 47 N178 Other acute kidney failure 48 N179 Acute kidney failure, unspecified 49 N171 Acute kidney failure with acute cortical necrosis 50 J80 Adult respiratory distress syndrome 51 J951 Acute pulmonary insufficiency following thoracic surgery 52 J952 Acute pulmonary insufficiency following nonthoracic surgery 53 J9600 Acute respiratory failure, type I 54 J9601 Acute respiratory failure, type II 55 J9609 Acute respiratory failure, type unspecified (J6909) 56 J9690 Respiratory failure unspecified, type I 57 J9691 Respiratory failure unspecified, type II 58 J9699 Respiratory failure unspecified, type unspecified 59 J960 Acute respiratory failure 60 J969 Respiratory failure, unspecified 61 R092 Respiratory arrest

Procedure codes:

66 1388200 Management of continuous ventilatory support, <= 24 hours 67 1388201 Management of continuous ventilatory support, more than 24 hours and less than 96 hours 68 1388202 Management of continuous ventilatory support, 96 hours or more

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 23 www.nzma.org.nz/journal article

Competing interests: Dr Paul Huggan is trustee of the New Zealand Sepsis Trust. Dr Ian Sheerin’s contribution was funded by an ACC project grant. Acknowledgements: The authors acknowledge the help of Dr K Walland and Dr G Mills in reviewing this manuscript, as well as the efforts of Ms Linda Shepherd and Mr Nick Kendall, ACC, who sponsored this work and ensured its completion. The authors also thank Mr Roger Williams, who provided management accounting information at Waikato District Health Board. Author information: Paul J Huggan: Senior Medical Officer, Waikato District Health Board; Honorary Lecturer, University of Auckland. Tania A Helms: Senior Prevention Intelligence Advisor, Accident Compensation Corporation. Veronique Gibbons: Clinical Effectiveness Manager, Quality and Patient Safety, Waikato District Health Board. Katie Reid: Medical Officer, Department of Medicine, Waikato Hospital. Harry Hutchins: Clinical Research Fellow, London School of Hygiene and Tropical Medicine, UK. Ian Sheerin: Consultant and Health Economist, recently retired from University of Otago Medical School. Corresponding author: Dr Paul Huggan, Department of Infectious Disease, Waikato Hospital and District Health Board, Waikato 3240 [email protected] URL: www.nzma.org.nz/journal-articles/counting-the-cost-of-major-infection-and-sepsis-in-new- zealand-an-exploratory-study-using-the-national-minimum-data-set

REFERENCES 1. Singer M, Deutschman CS, Regional Sample of the tion-based epidemiology Seymour CW, et al. The New Zealand Population. of Staphylococcus aureus Third International Consen- Open Forum Infect bloodstream infection sus Definitions for Sepsis Dis. 2017;4:ofx106. in Canterbury, New and Septic Shock (Sepsis-3). 5. Chang DW, Tseng CH, Zealand. Intern Med JAMA. 2016 23;315:801-10. Shapiro MF. Rehospital- J. 2010;40:117-125. 2. Rudd KE, Johnson SC, izations following sepsis: 9. Safar A, Lennon D, Stewart Agesa KM, et al. Global, Common and costly. J, et al. Invasive Group A regional, and national Critical Care Medicine. Streptococcal Infection and sepsis incidence and 2015;43:2085-2093. Vaccine Implications, Auck- mortality, 1990–2017: 6. Buchman TG, Simpson SQ, land, New Zealand. Emerg analysis for the Global Sciarretta KL, et al. Sepsis Infect Dis. 2011;17:983-989. Burden of Disease Study. Among Medicare Benefi- 10. Milne RJ, van der Hoorn S. Lancet. 2020;395:200-211. ciaries: 1. The Burdens of Burden and cost of hospital 3. Angus DC, Linde-Zwirble Sepsis, 2012-2018. Critical admissions for vaccine-pre- WT, Lidicker J, et al. Care Med. 2020;48:276-288. ventable paediatric Epidemiology of severe 7. Baker MG, Barnard pneumococcal disease and sepsis in the United States: LT, Kvalsvig A, et al. non-typable Haemophilus analysis of incidence, Increasing incidence of influenzae otitis media outcome, and associated serious infectious diseases in New Zealand. Appl costs of care. Critical Care and inequalities in New Health Econ Health Med. 2001;29:1303-1310. Zealand: A national Policy. 2010;8:281-300. 4. Huggan PJ, Bell A, Waetford epidemiological study. 11. Bremner C, Lennon D, J, et al. Evidence of High Lancet. 2012;379:1112-1119. Martin D, et al. Epidemic Mortality and Increasing 8. Huggan PJ, Wells JE, Meningococcal Disease Burden of Sepsis in a Browne M, et al. Popula- in New Zealand: Epide-

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 24 www.nzma.org.nz/journal article

miology and Potential for 16. The New Zealand Casemix Open Forum Infect Prevention by Vaccine. N System – An Overview Dis. 2016;3:ofw207. Z Med J. 1999:112:257-9. | Ministry of Health 22. Goodwin AJ, Ford DW. 12. Wilhelms SB, Huss FR, NZ. https://www.health. Readmissions among sepsis Granath G, Sjöberg F. govt.nz/publication/ survivors: risk factors and Assessment of incidence new-zealand-casemix-sys- prevention. Clin Pulm of severe sepsis in Sweden tem-overview-0. Accessed Med. 2018;25:79-83. April 27, 2020. using different ways of 23. Cuthbertson BH, Rattray abstracting International 17. Lee H, Croft R, Monos O and J, Campbell MK, et al. Classification of Diseases Harmston C. Counting the The PRaCTICaL study of codes: difficulties with costs of major trauma in a nurse led, intensive care methods and interpretation provincial trauma centre. follow-up programmes of results. Critical Care N Z Med J. 2018;131:57-63. for improving long term Med. 2010;38:1442-1449. 18. Report on New Zealand outcomes from critical 13. Iwashyna T, Odden A, Cost-of-Illness Studies on illness: a pragmatic Rohde J, et al. Identifying Long-Term Conditions. randomised controlled patients with severe sepsis Ministry of Health NZ. trial. BMJ 2009; 339:b3723. using administrative https://www.health.govt. 24. Schmidt K, Worrack S, von claims: patient-level nz/publication/report-new- Korff M, et al. Effect of a validation of the Angus zealand-cost-illness-stud- primary care management implementation of the ies-long-term-conditions. intervention on mental international consensus Accessed April 27, 2020. health-related quality of life conference definition 19. Jukarainen S, Mildh H, among survivors of sepsis of severe sepsis. Med Pettilä V, et al. Costs and a randomized clinical trial. Care. 2014:52;e39-43. cost-utility of critical care JAMA 2016;315:2703-2711. 14. Inada-Kim M, Page B, and subsequent health 25. McCoy A, Das R. Reducing Maqsood I, Vincent C. care. Critical Care Med. patient mortality, length Defining and measuring 2020:48;e345-e355. of stay and readmis- suspicion of sepsis: an 20. Ljungström L, Andersson sions through machine analysis of routine data. R, Jacobsson G. Incidences learning-based sepsis BMJ Open. 2017;7:e014885. of community onset severe prediction in the emergen- 15. Raymond NJ, Nguyen M, sepsis, sepsis-3 sepsis, and cy department, intensive Allmark S, et al. Modified bacteremia in Sweden care unit and hospital sequential organ failure – A prospective popula- floor units. BMJ Open assessment sepsis score in tion-based study. PLoS Qual. 2017;6:e000158. an emergency department One. 2019;14: e0225700. 26. Burrell AR, McLaws ML, setting: retrospective 21. Mellhammar L, Wullt Fullick M, et al. SEPSIS assessment of prognostic S, Lindberg Å, et al. KILLS: early interven- value. Emerg Med Austral- Sepsis incidence: a tion saves lives. Med as. 2019;31:339-346. population-based study. J Aust 2016; 204:73.

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 25 www.nzma.org.nz/journal article

A retrospective analysis of calls to the New Zealand National Poisons Centre regarding Pacific patients Eeva-Katri Kumpula, Rosalina Richards, Pauline Norris, Vanda Symon, Adam C Pomerleau

ABSTRACT AIMS: Poisoning is a common type of injury in New Zealand. The New Zealand National Poisons Centre (NZNPC) offers a free 24/7 specialist assessment service for enquiries about substance exposures for all New Zealanders. This study aimed to characterise calls to the NZNPC relating to Pasifika patients to explore the potential for unmet need or health disparity in this area. METHODS: A retrospective analysis of 2018–2019 human exposure call data was performed. Patients were stratified into three groups: those with at least one Pacific ethnicity listed (Pasifika); those with known ethnicities but no Pacific ethnicity listed (non-Pasifika); those of unknown ethnicity (unknown). Demographic variables and substance groups were described. RESULTS: Of the 40,185 human exposure patients, 1,367 (3.4%) were Pasifika, 24,892 (61.9%) were non-Pasifika and 13,926 (34.7%) were of unknown ethnicity. The median age of Pasifika patients was 2.0 years, with 78.0% aged 0–5, and the exposure most commonly involved a liquid product (46.6%) and a simple analgesic (8.3%). CONCLUSIONS: The NZNPC receives a relatively small number of calls about exposures to Pasifika patients, especially given the youthful population demographic. It is unclear whether there is unmet need for this service, and this study suggests the need for further research.

oisons information centres (PICs) offer with clinical toxicologists providing clinical advice on chemical safety, perform supervision and support. Depending on the Prisk assessments over the phone in scenario and risk assessment, the NZNPC cases where a person has been exposed to a may recommend that a patient present to a substance and provide clinical advice about medical facility for further management, or management of the exposure. The term that the exposure can be safely managed at ‘exposure’, in the context of poison centres, home with instructions on what to look out refers to any route of contact (eg, ingestion, for, saving the patient and family an un- inhalation, dermal, etc) between a person necessary trip to the doctor or hospital. The and a substance (which can be a pharma- NZNPC does not routinely follow up cases ceutical, chemical, commercial product, etc). and verify whether advice was followed; The New Zealand National Poisons Centre however, a previous study found a high rate (NZNPC) offers a 24/7 specialist service for of adherence.1 exposure enquiries from all New Zealand- Pasifika communities in New Zealand ers, including both the general public and have a youthful demographic, which healthcare professionals. The service is is relevant here because a substantial available through a toll-free number (0800 proportion of NZNPC calls relate to young POISON (0800 764 766)) staffed by specially children.2 While Pasifika families are trained poison information officers (PIOs), diverse in their collective health-literacy

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 26 www.nzma.org.nz/journal article

resources and experiences in accessing lauan, Tongan, Other Pacific Peoples, Pacific health services,3,4 for many families there Peoples not further defined));10 non-Pas- are challenges in accessing health services,5 ifika (at least one other known ethnicity which may be heightened in an urgent situ- and no Pacific ethnicity listed); unknown ation such as suspected poisoning. While ethnicity (missing information, or caller there is little specific literature available refused to provide it). For the main analysis, on Pasifika peoples and poisoning in New call clusters where multiple calls are made Zealand, the risks of poisoning events likely about the same patient in the same exposure vary by age group in the Pasifika population, event (‘linked calls’) were excluded to not as in the general population.2 Compared over count these patients. to other ethnicities, the rates of death and Ministry of Health (MOH) prioritised hospital presentations due to unintentional ethnicity level 1 (Māori, Pacific Peoples, poisoning were highest for Pasifika New Asian, MELAA (Middle Eastern, Latin 6 Zealanders in the age group 65 and older. American, African), Other, European, Not On the other hand, Pasifika women and stated/Refused to say)10 was used to give men had lower rates of presenting to New a general description of the total patient Zealand hospitals due to intentional self-poi- population by ethnicity, but the three main soning (self-harm) compared to Māori and groups described above were used for the people of other, non-Asian ethnicities, but at main analysis to identify as many Pasifika 7 similar rates to people of Asian ethnicities. patients from the data as possible. Patient Telehealth is a means of service delivery demographics, reason for the exposure that has the potential to minimise the impact event, crude rates of exposure, caller of some barriers to access, such as trans- relationship to the patient, caller location, portation. Significant questions remain, site of exposure, most frequent substance however, about how telehealth can deliver categories and management advice were services equitably to Pasifika populations, analysed in RStudio (1.1.456; RStudio Inc.) with a recent evaluation reporting that use and Excel (2016; Microsoft Corp.). of the Healthline and Quitline services (two The study used previously collected other national toll-free telehealth services) de-identified health data and was approved was lower among Pasifika communities by the Human Ethics Committee of the 8 than for other New Zealanders. This study University of Otago (ref# HD20/002). The expands this knowledge base by exploring study was conducted in accordance with the Pasifika engagement with the NZNPC. The Declaration of Helsinki.11 aim of this study was to describe Pasifika access to the NZNPC service and unpack the characteristics of this engagement through Results routinely collected health data from the During the two-year study period, the service. NZNPC provided advice during 42,631 calls to a total of 44,234 human patients who had been exposed to a substance. After linked Methods calls were excluded (3,781 calls with 4,049 This retrospective analysis investigated patients), there were 40,185 patients in de-identified call data from NZNPC relating 38,760 calls. Their MOH level 1 prioritised to human patients exposed to any substance ethnicities are described in Table 1, along from 1 January 2018 to 31 December 2019. with a breakdown by the three ethnicity Ethnicity information is collected in the call groups used in the study. records of the NZNPC, either by directly When the three pre-planned ethnicity collecting the information from the caller, groups were created from the data, there or by extracting ethnicity information from were 1,367 Pasifika patients, 24,892 non-Pa- the National Health Index (NHI)9 number sifika, and 13,926 people of unknown database when the patient’s NHI is identified ethnicity. The characteristics of these three at the time of the call. Exposure patients groups of patients are summarised in Table were stratified into three ethnicity groups 2. The median age of Pasifika patients was for analysis: Pasifika (any Pacific ethnicity 2.0 years (interquartile range (IQR) 1–4 listed in patient ‘ethnicity’ data field (Cook years; 0.2% unknown age), while the median Island Māori, Fijian, Niuean, Samoan, Toke-

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 27 www.nzma.org.nz/journal article

Table 1: Ethnicity of exposure patients managed by the New Zealand National Poisons Centre, 2018–2019.

Patient ethnicity n (% of all n (% of patients of patients)* known ethnicity)* MOH prioritised ethnicity, level 1

Māori 4,360 (10.8%) 4,360 (16.6%)

Pacific Peoples 986 (2.5%) 986 (3.8%)

Asian 1,327 (3.3%) 1,327 (5.1%)

MELAA 326 (0.8%) 326 (1.2%)

Other ethnicity 911 (2.3%) 911 (3.5%)

European 18,349 (45.7%) 18,349 (70.0%)

Unknown (Refused to answer/Not stated) 13,926 (34.7%) -

Grand total 40,185 (100%) 26,259 (100%)

Ethnicity groups created in study

Pasifika (any Pacific ethnicities recorded) 1,367 (3.4%) 1,367 (5.2%)

Non-Pasifika (no Pacific but other ethnicities recorded) 24,892 (61.9%) 24,892 (94.8%)

Unknown (Refused to answer/Not stated) 13,926 (34.7%) -

Grand total 40,185 (100%) 26,259 (100%)

*Linked calls excluded. MOH=Ministry of Health; MELAA=Middle Eastern/Latin American/African. was 2.0 (IQR 1–20 years; 0.3% unknown (44.4%) had a therapeutic agent involved, age) for non-Pasifika and 14.0 (IQR 2–32; 18,393 (30.2%) had a household chemical 34.0% unknown age) for people of unknown involved and 16,831 (27.6%) had a miscel- ethnicity. The properties and circumstances laneous chemical involved in the exposure. of the calls are described in Table 3. All three groups had most frequently been Substances involved in the exposed to a simple analgesic (eg, paracetamol (Table 4)). The NZNPC mostly advised exposures management at home (either no treatment The exposure event involved a median or self-treatment (Table 5)). of 1.0 substance per patient for Pasifika, non-Pasifika and patients of unknown Discussion ethnicity (IQR 1–1 for all three groups). The There is very little published literature on most common substance types were liquid Pasifika peoples and poisoning, and espe- products (46.6% for Pasifika, 45.9% for cially about calls to a poisons information non-Pasifika, 37.5% for people of unknown centre (PIC) regarding Pasifika patients. ethnicity) and capsules and tablets (36.1% This study begins to address this gap in for Pasifika, 30.9% for non-Pasifika, 40.2% knowledge. Over the two years covered for people of unknown ethnicity). The route in this study, most Pasifika patients were of exposure was ingestion for 78.3% of under the age of six, and the people calling Pasifika, 73.1% of non-Pasifika and 65.7% of about them were most often members patients of unknown ethnicity. of the public calling about an exposure occurring in a residential environment. A total of 720 Pasifika patients (52.7%) This is important information for Pasifika had a therapeutic agent involved in their parents and communities, in terms of high- exposure event, while 391 (28.6%) had lighting the need to limit children’s access a household chemical involved and 269 to chemicals and substances within home (19.7%) a miscellaneous chemical. On the environments. The most common substance other hand, 27,061 non-Pasifika patients

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 28 www.nzma.org.nz/journal article

Table 2: Patient characteristics from exposure calls to the New Zealand National Poisons Centre, 2018–2019.

Patient characteristics* Pasifika Non-Pasifika Unknown ethnicity

Gender n (%) n (%) n (%) Female 675 (49.4%) 12,922 (51.9%) 6,232 (44.8%)

Male 691 (50.5%) 11,957 (48.0%) 6,301 (45.2%)

Other 0 1 (0.0%) 7 (0.1%)

Unspecified or unknown 1 (0.1%) 12 (0.0%) 1,386 (10.0%)

Total patients 1,367 (100.0%) 24,892 (100.0%) 13,926 (100.0%)

Age group n (%) n (%) n (%) 0–5 1,066 (78.0%) 16,342 (65.7%) 3,575 (25.7%)

6–12 74 (5.4%) 1,457 (5.9%) 687 (4.9%)

13–19 34 (2.5%) 783 (3.1%) 1,289 (9.3%)

20–64 165 (12.1%) 4,871 (19.6%) 2,926 (21.0%)

65+ 25 (1.8%) 1,362 (5.5%) 708 (5.1%)

Unknown 3 (0.2%) 77 (0.3%) 4,741 (34.0%)

Total patients 1,367 (100.0%) 24,892 (100.0%) 13,926 (100.0%)

Reason for exposure event n (%) n (%) n (%) Abuse 7 (0.5%) 82 (0.3%) 377 (2.7%)

Child exploratory 963 (70.4%) 14,826 (59.6%) 4,026 (28.9%)

Intentional 20 (1.5%) 353 (1.4%) 1,873 (13.4%)

Other 0 11 (0.0%) 21 (0.2%)

Therapeutic error 190 (13.9%) 4,115 (16.5%) 1,693 (12.2%)

Unintentional 182 (13.3%) 5,416 (21.8%) 5,510 (39.6%)

Unknown 5 (0.4%) 89 (0.4%) 426 (3.1%)

Total patients 1,367 (100.0%) 24,892 (100.0%) 13,926 (100.0%)

Crude rates of exposures n (%) n (%) n (%) Patients 1,367 24,892 13,926

Reference population 386,616 4,406,742** N/A

Crude rate per 100,000 353.6 564.9 N/A population

*Linked calls excluded. **Total New Zealand population (4,793,358) minus Pacific population (Census 2018)12 equals non-Pacific population. N/A=not applicable.

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 29 www.nzma.org.nz/journal article

Table 3: Characteristics of exposure calls to the New Zealand National Poisons Centre, 2018–2019.

Characteristic* Pasifika Non-Pasifika Unknown ethnicity

Caller relationship to patient n (%) n (%) n (%) Healthcare professionals 55 (4.0%) 787 (3.2%) 5,473 (39.3%)

Workplace or school related 2 (0.1%) 15 (0.1%) 893 (6.4%)

General public 1,310 (95.8%) 24,086 (96.8%) 7,318 (52.5%)

Other 0 3 (0.0%) 165 (1.2%)

Unknown 0 1 (0.0%) 77 (0.6%)

Total patients 1,367 (100.0%) 24,892 (100.0%) 13,926 (100.0%)

Caller location n (%) n (%) n (%) Healthcare facility 40 (2.9%) 569 (2.3%) 3,296 (23.7%)

Workplace or school related** 24 (1.8%) 569 (2.3%) 2,827 (20.3%)

Residential 1,281 (93.7%) 23,235 (93.3%) 7,267 (52.2%)

Other 3 (0.2%) 67 (0.3%) 75 (0.5%)

Public area 15 (1.1%) 352 (1.4%) 242 (1.7%)

Environment 0 21 (0.1%) 37 (0.3%)

Unknown 4 (0.3%) 79 (0.3%) 182 (1.3%)

Total patients 1,367 (100.0%) 24,892 (100.0%) 13,926 (100.0%)

Site of exposure event n (%) n (%) n (%) Healthcare facility 3 (0.2%) 15 (0.1%) 69 (0.5%)

Workplace or school related 19 (1.4%) 705 (2.8%) 1,893 (13.6%)

Residential 1,323 (96.8%) 23,417 (94.1%) 10,629 (76.3%)

Other 5 (0.4%) 80 (0.3%) 143 (1.0%)

Public area 10 (0.7%) 408 (1.6%) 356 (2.6%)

Environment 0 99 (0.4%) 107 (0.8%)

Unknown 7 (0.5%) 168 (0.7%) 729 (5.2%)

Total patients 1,367 (100%) 24,892 (100%) 13,926 (100%)

*Linked calls excluded. **This category contains calls originating from the ambulance service setting.

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 30 www.nzma.org.nz/journal article

Table 4: Ten most-frequently encountered substance classes in the three ethnicity groups; % of all substances within patient group.

Pasifika n (%)* Non-Pasifika n (%)* Unknown ethnicity n (%)*

Simple analgesic** 126 (8.3%) Simple analgesic** 2,338 (8.5%) Simple analgesic** 1,220 (7.2%)

Cosmetic 106 (7.0%) Household 2,186 (7.9%) Industrial 1,213 (7.2%) miscellaneous chemicals

Household 105 (7.0%) Household 1,548 (5.6%) Household 988 (5.8%) miscellaneous cleaner miscellaneous

Topical medication 80 (5.3%) Cosmetic 1,479 (5.4%) Miscellaneous 837 (5.0%) chemicals

Household cleaner 79 (5.2%) Plant 1,464 (5.3%) Plant 834 (4.9%)

Medications for 71 (4.7%) Miscellaneous 1,263 (4.6%) Household cleaner 798 (4.7%) infections/infestations chemicals

Antihypertensives 68 (4.5%) Topical medication 1,121 (4.1%) Antidepressants 692 (4.1%)

NSAID 68 (4.5%) NSAID 1,102 (4.0%) NSAID 581 (3.4%)

Miscellaneous 59 (3.9%) Vitamins/minerals 1,100 (4.0%) Opioid analgesic 514 (3.0%) chemicals

Vitamins/minerals 55 (3.6%) Industrial chemicals 803 (2.9%) Antihypertensive 492 (2.9%)

NSAID = non-steroidal anti-inflammatory drug. *Linked calls excluded. **Eg, paracetamol.

Table 5: Management advice given by the New Zealand National Poisons Centre, stratified by ethnicity group.

Management advice* Pasifika Non-Pasifika Unknown n (%) n (%) ethnicity n (%) Medical referral 204 (14.9%) 3,202 (12.9%) 5,836 (41.9%)

Referral to other service 4 (0.3%) 70 (0.3%) 126 (0.9%)

No referral needed 1,143 (83.6%) 21,299 (85.6%) 7,479 (53.7%)

Further information required** 16 (1.2%) 321 (1.3%) 485 (3.5%)

Total patients 1,367 (100.0%) 24,892 (100.0%) 13,926 (100.0%)

*Linked calls excluded. **Not enough information available at the time of the call to make a definite recommendation yet.

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 31 www.nzma.org.nz/journal article

classifications in these exposures included because they may look to other resources to simple analgesics (eg, paracetamol), manage exposures. It is impossible to define cosmetics and miscellaneous household an ‘optimal’ rate of using a PIC service. chemicals (eg, baking supplies, batteries, The rate of healthcare professional calls cigarettes, etc). All of these products are relating to Pasifika patients was similar for commonly found in New Zealand house- non-Pasifika in this study, which indicates holds. The wellbeing of children and youth that at least some people from both groups are central to the aspirations of Pasifika did seek medical care for exposures. The peoples, with parents wanting to give their reasons suggested for their lower uptake of children the best start in life.13 Ideally, expo- two other free national telehealth services, sures/poisonings are prevented. However, Healthline and Quitline, included reluctance as an urgent source of support, services to make contact, as they had no previous like NZNPC are a valuable resource, as they relationship with the services, concern that can be immediately accessed via phone for they were not being understood and that the expert advice. In this study, the majority advice was being tailored to their ethnicity of exposures were able to be managed at rather than their acute needs.8,15 People may home, saving families potentially expensive also fear being judged or accused of being trips to a medical centre or hospital. careless and ‘allowing’ a child to get injured Only 3.4% of all NZNPC patients, and in this way, and experiences of racism 5.2% of patients with known ethnicity, were within the health system may exacerbate 16,17 Pasifika. These observed rates are relatively these concerns. While NZNPC staff focus low, as the Pasifika population comprises on providing non-judgmental advice to all 8.1% of the total New Zealand population.12 callers, potential callers who have not previ- Further, as the Pasifika population aged 0–4 ously used the service may not be aware of comprises 14.3% of the total New Zealand this. This is an area where further research population of that age,12 and over two-thirds would be useful. of the patients with known ethnicity who Another interesting line of inquiry is to were managed by the NZNPC were aged explore concepts of ‘poisoning’ within the under six, more calls relating to Pasifika diverse Pasifika cultures in New Zealand patients could be expected. Further research and traditional models of managing the is needed to directly explore possible risks of poisoning across the Pacific region. barriers to service access. Health promotion approaches that align Possibilities for exploration include that with these models could then be developed Pacific peoples have a lower prevalence of and evaluated to see whether they resonate poisoning compared to other ethnicities, or and engage families more than current that they experience very mild exposures approaches. that they do not feel require formal care. Implications This seems unlikely, as rates of poisoning As most exposures reported to the NZNPC generally increase with increasing levels could be managed at home, the service of deprivation,6,7 and Pasifika people are has the potential to save people time and over-represented in these areas.14 Addi- resources, as they do not need to obtain the tionally, it might be expected that larger same assessment and advice elsewhere. On and/or multigenerational Pasifika house- the other hand, in cases requiring medical holds could create additional challenges for care, the NZNPC can assist by advising storing medications and other chemicals in people to seek such care in a timely fashion. spaces that are inaccessible to children. Such There are valuable opportunities here to multigenerational household compositions promote the safe storage of chemicals and could function as a protective factor as well, substances in Pasifika homes and to build as other, non-parental relatives may be able awareness and relationships between to provide additional supervision of children the NZNPC, the Pasifika health sector and in the household. Pasifika communities. The facilitators and Alternately, it may be that there is a barriers to Pasifika families contacting degree of unmet need in accessing poisons the NZNPC for advice regarding substance information. This could be because Pasifika exposures need to be explored further, families are not aware of the service, or while taking into account the rich variety of

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 32 www.nzma.org.nz/journal article

different Pasifika cultures represented in the an underestimation of rates of medical New Zealand population. referrals in Pasifika patients. Further, the Limitations relatively high proportion of patients of unknown age (34.0%) in the unknown-eth- Some Pasifika people were likely misclas- nicity subgroup of patients limits the sified due to being in the unknown ethnicity reliability of the median age reported for category. Even when an NHI number is that group. But, as this study did not seek to available, ethnicity details may not correctly perform statistical comparisons between the identify a person’s Pasifika ethnicity,18 and groups, no corrections such as imputations19 indications of Pasifika ethnicity may be were performed. lost over time as information is updated, effectively ‘changing’ a person’s ethnicity in the NHI database.5 Although the caller Conclusion being a healthcare professional and the The New Zealand National Poisons Centre exposure being intentional do not neces- receives what appears to be a relatively sarily mean that the exposure was more small number of calls about exposures to toxicologically significant, it should be noted Pasifika patients—disproportionately small that the unknown-ethnicity category had compared to the population prevalence higher rates of these characteristics than the of Pasifika people in New Zealand. It is Pasifika and non-Pasifika categories. Such unclear whether this finding represents a exposures may be more likely to result in a low prevalence of exposures in the Pasifika recommendation to be medically assessed. communities, a low utilisation of the NZNPC Any Pasifika patients ‘lost’ in this unknown by the Pasifika communities or some other ethnicity group may therefore lead to combination of factors.

Competing interests: Nil. Acknowledgements: We would like to acknowledge and thank the poisons information officers of the New Zealand National Poisons Centre for serving callers and collecting the data used in this study as part of their routine documentation practice. Author information: Eeva-Katri Kumpula: Postdoctoral Research Fellow, National Poisons Centre, University of Otago, Dunedin. Rosalina Richards: Associate Professor, Health Science Divisional Office, Centre for Pacific Health, Va’a o Tautai, University of Otago, Dunedin. Pauline Norris: Professor, Health Science Divisional Office, Centre for Pacific Health, Va’a o Tautai, University of Otago, Dunedin. Vanda Symon: Postdoctoral Fellow, Health Science Divisional Office, Centre for Pacific Health, Va’a o Tautai, University of Otago, Dunedin. Adam C Pomerleau: Director, Clinical Toxicologist, Emergency Physician (FACEM), National Poisons Centre, University of Otago, Dunedin. Corresponding author: Dr Eeva-Katri Kumpula, National Poisons Centre, PO Box 56, Dunedin 9054, 03-479 5168 (no fax available) [email protected] URL: www.nzma.org.nz/journal-articles/a-retrospective-analysis-of-calls-to-the-new-zealand-national-poisons-centre-regarding-pacific-patients

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 33 www.nzma.org.nz/journal article

REFERENCES 1. Watts M, Fountain JS, Reith national registry data: 14. Craig E, Taufa S, Jackson D, Schep L. Compliance exploring the challenges. C, Han DY. The Health with poisons center refer- Aust N Z J Public Health. of Pacific Children and ral advice and implications 2017;41(4):535-40. Young People in New for toxicovigilance. Clin 8. Sapere Research Group. Zealand Dunedin: Minis- Toxicol. 2004;42(5):603-10. FINAL Post-Implementa- try of Health; 2008. 2. Kumpula E-K, Shieffel- tion Review Report of the 15. Sapere Research Group. bien LM, Pomerleau AC. National Telehealth Service Phase 2 Report on the Enquiries to the New Wellington: Ministry of National Telehealth Service Zealand National Poisons Health – Manatū Hauora; Evaluation Wellington: Centre in 2018. Emerg 2017. Cited 03/03/2020. Ministry of Health – Med Australas. 2020; doi: Available from: https:// Manatū Hauora; 2019. Cited 10.1111/1742-6723.13563. www.health.govt.nz/ 03/03/2020. Available from: 3. Sa’uLilo L, Tautolo E-S, system/files/documents/ https://www.health.govt. Smith M. Health litera- pages/post-implementa- nz/system/files/documents/ cy, culture and Pacific tion-review-national-tele- pages/phase_2_report_ peoples in Aotearoa, New health-service.pdf. on_the_national_tele- Zealand: A review. 9. Ministry of Health – health_service_evalua- Pacific Health. 2020;3. Manatū Hauora. National tion_28_02_19_redacted. pdf. 4. Pio FH, Nosa V. Health Health Index Data literacy of Samoan mothers Dictionary version 5.3. 16. Harris RB, Stanley J, and their experiences Wellington: Ministry Cormack DM. Racism and with health profession- of Health – Manatū health in New Zealand: als. J Prim Health Care. Hauora; 2009. Prevalence over time 2020;12(1):57-63. 10. Ministry of Health and associations between recent experience of racism 5. Southwick M, Kenealy – Manatū Hauora. and health and wellbeing T, Ryan D. Primary Care Ethnicity Data Protocols measures using national for Pacific People: A for the Health and Disabil- survey data. PLoS One. Pacific and Health Systems ity Sector. Wellington: 2018;13(5):e0196476. approach. Report to the Ministry of Health – Health Research Council Manatū Hauora; 2004. 17. Kapeli SA, Manuela S, and the Ministry of 11. World Medical Associ- Sibley CG. Perceived Health Wellington: Pacific ation. World Medical discrimination is associ- Perspectives; 2012. Cited Association Declaration of ated with poorer health 03/03/2020. Available Helsinki: ethical principles and well-being outcomes from: https://www.health. for medical research among Pacific peoples in govt.nz/system/files/ involving human subjects. New Zealand. J Commu- documents/publications/ JAMA. 2013;310:2191-4. nity Appl Soc Psychol. 2020;30(2):132-50. primary-care-pacific-peo- 12. Statistics New Zealand. Age ple-pacific-health-sys- and sex by ethnic group 18. Lepa T, Norris P, Hors- tems-approach.pdf. (grouped total responses), burgh S, Taungapeau 6. Peiris-John R, Kool B, for census night popula- F. Accuracy of National Ameratunga S. Fatalities tion counts, 2018 Census Health Index numbers and hospitalisations due 2020. Cited 23/01/2020. for Pacific people in New to acute poisoning among Available from: http:// Zealand. Aust N Z J Public New Zealand adults. Intern nzdotstat.stats.govt.nz/ Health. 2013;37(2):189-90. Med J. 2014;44(3):273-81. wbos/Index.aspx?DataSet- 19. Kwak SK, Kim JH. Statistical 7. Kumpula E-K, Nada-Raja Code=TABLECODE8317 data preparation: manage- S, Norris P, Quigley P. 13. Ministry for Pacific Peoples. ment of missing values and A descriptive study of Pacific Aotearoa: Lalanga outliers. Korean J Anesthe- intentional self-poison- Fou. Wellington: Ministry siol. 2017;70(4):407-11. ing from New Zealand of Pacific Peoples; 2018.

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 34 www.nzma.org.nz/journal article

Interviews with health professionals about the National Child Protection Alert System Patrick Kelly, Melissa Adam, Carmen Basu, Miranda Ritchie, Denise Wilson, Fred Seymour

ABSTRACT AIM: The New Zealand National Child Protection Alert System is administered by multidisciplinary teams in every district health board. The aim of this study was to investigate the factors that influence multidisciplinary child protection teams’ (MDTs’) decisions about whether to place a child protection alert. METHOD: Members of the Child Protection Alert System teams were invited to participate in semi-structured interviews. Interview data were coded and grouped into themes using inductive thematic analysis. RESULTS: Six themes were identified: the system works well; a wide range of factors are considered in multidisciplinary team decision-making; there are some difficulties with multidisciplinary team meetings; there are problems with the administration of the system across district health boards; there is concern about the potential for the Child Protection Alert System to stigmatise families or cause unjustified responses; improvements can be made to the system. CONCLUSION: There is overall support for the National Child Protection Alert System and a consensus that the benefits outweigh any potential risks. There is a need for further improvements to the system, including consistent training, further standardisation and increased accessibility of the information to health professionals, including making information on the system available to primary healthcare.

ew Zealand has a publicly funded appointments with well-child nurses in healthcare system with services the community and 30 visits to general provided through 20 semi-auton- practitioners at four practices. “Collec- N 1 omous district health boards (DHBs). All tively the health sector had available a DHBs have policies requiring staff to notify telling picture… [which] was never put child protection concerns to statutory child together because of poor communication. protective services (CPS). However, there is Information was not passed on or was no nationally shared patient record system, incomplete. Previous records within the and many families are highly mobile.2 As a same hospital or practice were not viewed result, a past history of child protection con- and… social and medical histories were not cerns is often invisible to health providers. sought or provided.”3 The failure of other The potential consequences of this were agencies to share information with the tragically illustrated in the case of a child health system was also important and was murdered in the Hawkes Bay in 1999 despite highlighted again in the death of two sisters 4 40 contacts with health providers: four in 2001. Similar issues occur repeatedly hospital emergency department visits, two in reports into child-abuse deaths in other 5,6 admissions, one clinic, three face-to-face countries.

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 35 www.nzma.org.nz/journal article

There is one feature of the New Zealand substance abuse, parental criminality or healthcare system that makes it possible to parental mental health. Many have multiple share health information nationwide: the concerns.13 There are now 48,530 child National Health Index number (NHI), “a protection alerts on the NCPAS.14 unique identifier assigned to every person When an individual presents to a DHB who uses health and disability support and their NHI is registered, the CPA is 7 services.” Linked to the NHI is the National automatically downloaded from the NMWS Medical Warning System (NMWS), a simple and appears as follows: “Child protection database first designed in the 1960s to share concerns, contact X DHB,” where X is the information about risks such as anaphy- name of the DHB that placed the CPA. 8 lactic medication reactions. The NHI and Policies and training emphasise that the the NMWS are the only databases shared by NCPAS is a flag to past concerns, not a all 20 DHBs, which otherwise use multiple child protection register (that is, it is not a information and clinical record systems, database containing up-to-date information most of which cannot communicate with about a subset of children at a known level 9 each other. of risk). The health provider is expected In response to these issues, the National to access those past concerns to inform a Child Protection Alert System (NCPAS) was thorough assessment of the new presen- developed to use the NMWS to share infor- tation. A CPA does not necessarily indicate mation about child protection concerns. It ongoing risk, and no CPA does not neces- is a collaboration between the Paediatric sarily mean no risk. Society of New Zealand, the Ministry of Although the NCPAS is designed to enable Health and DHBs. Hawkes Bay DHB placed information sharing, audits show a wide the first national child protection alert (CPA) variation in the ratio between the number in 2003 and was joined by the Auckland of CPS notifications submitted to each MDT DHB in 2009. By 2011, a comprehensive for consideration of a CPA and the number framework was in place, incorporating of CPAs that each MDT agrees to place. Some national policy, a privacy impact assessment, MDTs place CPAs on 99% of children whom governance, DHB policy templates, terms their DHB has notified to CPS, whereas of reference for multidisciplinary child others place CPAs on only 60%.13 Variation protection teams (MDTs) in each DHB, in this ratio (the ‘conversion rate’) was proformas to document factors considered not eliminated in a study of inter-rater when placing CPAs, guidelines, training agreement where MDTs were presented 10,11 packages and regular auditing. The with the same cases, so it appears there system was implemented on an opt-in basis are factors other than case information as each DHB achieved compliance with this alone that influence decisions to share child framework. All 20 DHBs were participating protection concerns.13 by December 2016. The aim of this study was therefore A CPA can only be placed where a DHB to further investigate factors that may has notified CPS (or where CPS are already influence decision-making within NCPAS involved). It must be approved by the MDT MDTs. and supported by comprehensive information. It remains until the eighteenth birthday but may be removed earlier by Method application to the MDT that approved Research approval was obtained from the it.10,11 CPAs can also be placed on siblings Auckland DHB. at risk12 and pregnant women who have Every MDT member in all 20 DHBs was been notified to CPS. Antenatal CPAs can identified by direct contact with team coor- be transferred to the baby after birth if dinators, and all agreed to provide their concerns persist. Concerns leading to CPAs email addresses to the study investigators. cover a wide range of issues, including In December 2016 and January 2017, all emotional abuse, neglect, physical abuse, were invited to participate in an online sexual abuse, mental-health or behavioural survey. The size, composition and frequency issues, poor engagement with services, of MDT meetings nationwide, and the family violence, parenting capacity, parental years of experience in child health of MDT

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 36 www.nzma.org.nz/journal article

members, is described in detail in the results codes and the full dataset. Codes and themes of that survey, published elsewhere.15 At the were then reviewed by a third member end of the survey, participants were asked of the research team (FS).19 Once data if they would be willing to be interviewed. saturation had been reached, no further We aimed to recruit up to 25 interviewees, interviews took place. Because the inter- a number often sufficient to reach data views were semi-structured (ie, questions saturation in qualitative studies.16 Data did not follow the same precise order and saturation occurs when there is no new were not necessarily expressed in identical information in the data.17 Using an online manner), the number of respondents for research randomiser, interviewees were each theme are not reported, but descriptors randomly selected from those who said yes. such as ‘few’, ‘some’ and ‘many’ are used.20 The sample was checked to confirm that randomisation had achieved representation Results of multiple professional roles, multiple Ninety-one of 160 MDT members (57%) DHBs (eg, both small rural DHBs and large completed the online survey, representing urban DHBs) and multiple perceptions all 20 DHBs and a diverse range of profes- of the NCPAS (ie, those selected included sional groups and levels of experience.15 Of both those who rated the NCPAS highly in these, 61 (67%) volunteered for an interview, the online survey and those who rated it and 18 (29.5%) were interviewed. Thirteen poorly).15 were female and five were male. Sixteen In April and May 2017, telephone inter- were of European ethnicity, one was Māori views were conducted by an Honours and the ethnicity of one was unknown. student in clinical psychology (MA) and Interviewees came from nine DHBs, with no the Starship Foundation Child Protection more than two from any one DHB, and they Research Fellow (CB). Interviews followed a represented the full range of perceptions of semi-structured format (Figure 1). Prompts the NCPAS, from disapproval to approval, (in brackets) were used to seek clarifi- as expressed in the online survey.15 There cation of initial responses. Interviews were were six social workers, five doctors, three recorded, transcribed and de-identified. violence intervention programme coordi- Thematic analysis followed the guidelines nators, two nurses and two midwives. The proposed by Braun and Clarke.18 Transcripts professional mix reflected the national were read and re-read to gain familiarity make-up of MDTs, as described in the 2016 with the data, then initial codes (features study of inter-rater agreement13 and the evident across the entire dataset) were online survey,15 but Māori health profes- generated and collated into themes. Each sionals (who are present in 44% of MDT theme was reviewed in an iterative process meetings)13 were under-represented. to ensure that it worked in relation to the Six themes were identified (Figure 2).

Figure 1: Interview schedule.

Tell me how you feel the NCPAS is functioning at a national and local level? (Negative and positive)

What is the process you use when deciding whether to place an alert? (Factors that influence decision-making, personal threshold)

Do you have any concerns about NCPAS?

What else does the NCPAS need?

Is there anything else you would like to say about the NCPAS that we have not already talked about?

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 37 www.nzma.org.nz/journal article

Theme 1: the system works well Several described other advantages, Many supported the system and believed including antenatal alerts and benefits in it was working well. providing care for transient families. “…my feeling is that locally here for Theme 2: a wide range of factors us it is functioning very well and are considered effective. We’ve got a good strong A wide range of factors are considered group in that we’ve got very good when MDTs are deciding whether to place a buy in from a multi-professional CPA. In addition to physical or sexual abuse group” (Participant 1) or neglect, many mentioned the child’s age, “I think the process of identifying multiple presentations, parental mental children and putting the alerts on is health, family violence, substance abuse working pretty well” (Participant 5) and gang affiliations. Some mentioned “I think it is making quite a signif- transience, truancy, concerns from icant difference to the safety of multiple professionals and factors asso- children and young people” ciated with parenting stress, such as being (Participant 12) a solo mother or children from multiple fathers. Many said they took a common- Many commented on the strength and sense approach and that some cases were utility of MDT meetings. obvious. “We do have some healthy discus- “We basically, we look at the sions where, at times, maybe just one parents and siblings, we look at person in the group doesn’t agree, all their presentations… I guess and so, we’ll continue to talk about the more risk factors the more that, to try and reach a consensus” likelihood of harm. So, you know, (Participant 13) engagement, mobility, mental health, Several stated that the NCPAS supported maternal mental health, what the continuity of care. child presents with… alcohol, drug “It’s good for professionals and and the CPS history would probably for families because it… it can be be the biggest triple, the top a continuation, rather than a trio… the other thing we look at repeating, repeating over and over would be the age of the children… again of potential missing important intimate partner violence…” information” (Participant 17) (Participant 11)

Figure 2: Themes.

The system works well

A wide range of factors are considered when multidisciplinary teams are deciding whether to place a Child Protection Alert

There are difficulties with multidisciplinary team meetings

There are concerns about potential negative impact of Child Protection Alerts

There are problems with system administration

Improvements can be made

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 38 www.nzma.org.nz/journal article

Several mentioned factors that influ- Theme 4: concerns about potential enced them not to place an alert, such as negative impact engagement with services, presence of a A few were concerned that alerts might safety plan, imminent plans for adoption or cause unjustified or judgemental responses, other changes in living circumstances and/ or that they are not regularly reviewed. or the child being an older adolescent. “It has the potential to… skew the… “…if they were engaged with a approach that a subsequent clinician number of services not just one… might have to that child towards if mum or the partner were seeking child protection. And that’s not neces- help, seeking advice” (Participant 10) sarily a bad thing, but it… could lead A few placed great reliance on whether to further investigations which are information had been substantiated by CPS, unnecessary and unwarranted, and but others disagreed. also potentially removing a child, or “I just think you should be basing another notification to CPS, which it on… the information that you may or may not be necessary” have in front of you. And if there’s (Participant 5) enough concern on that to warrant “…my biggest concern would be that an alert… anything else you’re people may get judged by clinicians… actually gathering from other I just worry that people might get agencies is something that is not treated differently” (Participant 11) available in the health record” “…the fact that it hasn’t been (Participant 14) reviewed in 10 or 12 years. And so, In summary, a wide range of factors were you know this family or this child… considered, but there was variation in has been treated in a way that understanding of how much information actually is no longer relevant for was needed before an alert was placed. them” (Participant 1) Theme 3: difficulties with MDT Participants were specifically asked about meetings stigmatisation. A few expressed concerns Although many commented positively on about the possibility but felt that it was MDT meetings, some had concerns including outweighed by the need to ensure child the adequacy of the information presented, safety. A few noted the information was time required, staff turnover and challenges helpful even when abuse was no longer a engaging doctors. risk (eg, in adolescents presenting with later mental health concerns). “There’s some reluctance around paediatricians wanting to pick up “Isn’t it worthwhile us… having child protection as part of a port- this information? And sharing it so folio” (Participant 11) that actually we can put better stuff around this mum… I don’t think A few expressed concerns about the stigma is an issue at all, because if impact of the work. you’re being professional. That’s like “I am worried about burnout… some- saying… because she smokes there’s times you just got to de-frag and try a stigma around her, do you know and leave that stuff here at work” what I mean, we’re anti-smoke? And (Participant 13) we know the damage smoking does” One commented that it was difficult to (Participant 6) get staff to attend meetings, as there was A few felt that there is misunderstanding a feeling that the paediatrician overruled about what an alert is (eg, a register of them. children at high risk, rather than simply a “I know that has caused a few flag to past concerns that may or may not issues, for some… I’ve (heard) a remain relevant). couple of comments, ‘well, why do “I think people get it into their head we bother if it’s just the paediatrician that… it’s somehow marking a child making all the decisions’” and discriminating against them, and (Participant 8)

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 39 www.nzma.org.nz/journal article

a family. And they don’t feel that’s “…on our paediatric ward… this okay, it’s certainly not what’s meant child’s been through a couple of to happen, and shouldn’t be used that times, and no one’s bothered to get way because it is just alert infor- this information” (Participant 8) mation that’s held on the medical “what we’ve noticed a lot lately is record that can be used for people to that people are emailing each other make decisions about the particular instead of going through medical child” (Participant 14) records” (Participant 11) Another felt the name causes confusion. A few raised the issue of cost versus benefit. “I think the word ‘alert’ sometimes “…we don’t know whether it works freaks people out, I don’t know what or not, that would be my biggest a better word would be, but the concern. That we are spending heaps actual term alert, is in some ways of money, and time, and energy, and already indicative of big concerns” doing all of this stuff, and we don’t (Participant 18) really know whether it is making Theme 5: problems with system a difference to children and their administration families. And whether it’s making Electronic record systems are generally a positive difference, or a negative not shared between one DHB and another. difference, and what the balance of Therefore, staff in one DHB who see a CPA outcomes is. And it’s really critical that was entered on the NMWS by another that we know that because otherwise DHB must contact the medical records it’s going to be hard to maintain its department of that other DHB to access continued presence” (Participant 5) the information behind it. When the infor- Theme 6: improvements can be mation does arrive, it may be of varying made quality and may be presented in an unfa- Many recommended a more cohesive miliar way. Many interviewees struggle with system with greater standardisation of these challenges. computer systems, documentation and “[The system] is clunky… every DHB MDT composition and more comprehensive has a different computing system… guidelines. Many felt it should be acces- the only system… that every hospital sible to primary health, including birthing has access to, is the national alert centres. system. And it’s, you know built in “What I’d love to have is one health the 70s… so it’s not the easiest way record… the national alert sitting of sharing information. It would be on a health record… from birth to nicer if you could have that infor- death. And that’s your spine, and mation automatically come up” then along the way we have all of (Participant 14) these little branches that come of It is especially difficult after hours. A few that are maternity services, well child felt that, even when accessed, the infor- providers, national immunisation mation was often not helpful register… nothing sitting in silos” “…after hours, the alert system (Participant 18) doesn’t work very well anyway, full A few discussed the content of notifica- stop… whether it’s a child protection tions to CPS. issue or… penicillin allergy… the “…if you’ve got a really poor- paediatrician I talked to had no idea quality notification it’s difficult to where the alert was… there’s a lot do anything with it. So, I think a lot of information that is irrelevant and of the education would be useful there’s a lot of relevant information around what you put in… teaching that isn’t there” (Participant 9) people how to… add the right, A few commented that, because of these correct information, may actually issues, staff were either not accessing infor- improve the alert system signifi- mation or using workarounds. cantly. Because then your basis, your

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 40 www.nzma.org.nz/journal article

information is better, so then your since 2015 to lodge a ‘Child-At-Risk’ alert in decision-making… is better” the electronic medical record.23 Those clini- (Participant 15) cians also regard their system positively,23 A few highlighted the importance of although there are no studies to date that training front-line staff. investigate the ultimate question: are such systems effective in reducing the recurrence “…there’s a child protection alert… and/or morbidity of abuse? can you send a social worker. And so, the staff are not always trained to go However, despite the positive views, oh, there’s a child protection alert, our data suggest that the information I need to look at that information” behind alerts is not always accessed and (Participant 17) considered—consistent with both the online survey15 and implementation research into “I don’t think there’s been a lot other clinical decision support systems.24 of education about how the Issues include alert visibility, the under- system works” (Participant 9) standing of front-line staff of an alert, ease A few mentioned the need to improve of access to the information (especially after alert visibility. hours) and more consistent and embedded “…it should pop up and you should procedures for response to alerts. have to acknowledge the alert… Also, there was evidence of other issues I have to scroll down to see the that may contribute to inconsistency in alert… So technically you could go implementation of the NCPAS and may go into this woman’s file and not see it” some way towards explaining the variation (Participant 7) in conversion rate seen both in NCPAS audits A few noted the importance of cultural and in the inter-rater agreement study.13 representation at MDT meetings. Many participants viewed the NCPAS “We don’t have any representative simply as a flag to information that may from… our Pasifika or āM ori health or may not remain relevant (the concept team… We make those decisions, to which the system is designed), but a few with the team that we have, and so, it regarded it as a register for children at could be better” (Participant 18) a certain level of risk,25 which highlights the issue of variation and gaps in multi- Discussion disciplinary child protection team (MDT) members’ understanding of all aspects of This study found a general perception the NCPAS. This issue was also identified that the National Child Protection Alert in the online survey.15 The perception System (NCPAS) was functioning according that the NCPAS is a register of ‘high-risk’ to the purpose for which it was designed: children leads naturally to a view that alerts to support information sharing. Our data should only be placed for high-risk cases are therefore consistent with the generally (a notoriously complex issue to predict)26 positive findings of the online survey15 but to be reviewed regularly as circumstances enable more intensive exploration of system change. However, when the NCPAS is used strengths and weaknesses. as intended, that view is less cogent. Even The fact that the NCPAS was regarded if the risk of recurrent abuse diminishes, positively suggests that the information the health impact of past abuse27 means is genuinely useful to clinicians. This is that an alert may have enduring value for consistent with international research, informing appropriate care while ensuring where computer-based alert systems (as one participant put it) that children and have been shown to be useful in reducing families are not required endlessly to repeat 21 prescription errors and (in one US emer- accounts of their past experiences. gency department) for identifying children A view that alerts should be placed only who merit further evaluation for physical in substantiated cases arises from the same abuse.22 The closest parallel to the NCPAS perception. While it is tempting to view is found in northern New South Wales, substantiated cases as ‘true’ abuse and Australia, where clinicians who notify child unsubstantiated ones as ‘false,’28 the reality protection services (CPS) have been required

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 41 www.nzma.org.nz/journal article

is far more complex. The risk profiles and Limitations 29,30 long-term outcomes are almost identical Varying familiarity with the NCPAS may and some suggest that it is “time to leave have had some effect on responses. In 31 substantiation behind.” some DHBs, the NCPAS had been in oper- MDTs are incorporated in the NCPAS ation for more than a decade, and in others to support high-quality decision-making for less than a year.15 However, the most by drawing on a range of experience and important limitation is that non-European expertise.32,33 Data in this study suggest that MDT members were under-represented the factors they consider align with NCPAS in our sample. This is a crucial limitation policy and guidelines. However, quality is in an area of practice where Māori and not determined simply by the information Pasifika children are so greatly over-rep- on the table. The dynamics may be complex resented, and it raises the possibility that and cohesion and shared goals are key to issues of equity and unconscious bias may good decisions.34 This study confirms that not have been adequately investigated. We many MDTs, but not all, function well. are attempting to address this deficit in Further research into the functioning of our data in a current prospective multi-site MDTs is needed, including better assessment mixed-methods study, where investi- of the role of team dynamics.35 gators observe MDT decision-making in MDTs are also there to reduce the risk of multiple DHB meetings across New Zealand bias in decisions to place an alert.13 Inter- and interview individual team members national literature is clear that bias plays a afterwards. part in the notification of minorities and the poor to child protective services,36 possibly Conclusion including notifications from child protection The NCPAS is well-regarded by the MDT 33,37 MDTs. In New Zealand, Māori and members that were involved in this study, Pasifika children are far more likely to be but there are challenges to overcome. The 38 notified to CPS. Such evidence supports the technical challenges could be addressed by concern of some participants in this study technical solutions, which would fit well that bias might affect decisions to place a with the recent proposals for DHB reform CPA or actions taken in response to a CPA. and a greater engagement with primary That concern was also evident in the online care.9 In addition, such solutions must be survey, where a statement that “the NCPAS accompanied by thorough training in the stigmatises families” received the agreement conceptual framework of the NCPAS, how to of 16% of MDT respondents and another complete a high-quality notification to CPS, 15 30% neither agreed nor disagreed. It is how to access a CPA and what to do when therefore important for MDTs to ensure that one is seen. minority voices are well-represented and that all voices represented are heard.

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 42 www.nzma.org.nz/journal article

Competing interests: Miranda Ritchie is employed by contract to the Ministry of Health to audit implementation of the National Child Protection Alert System. Patrick Kelly is a member of the Clinical Reference Group of the Child Protection Special Interest Group of the Paediatric Society of New Zealand, which participates in the governance structure and quality assurance mechanisms of the National Child Protection Alert System. This position is unpaid. Acknowledgements: Dr Carol Chan for her assistance in identifying all MDT team members and in the design phase of this study. The Starship Foundation for funding Dr Carmen Basu as the Child Protection Research Fellow. The health professionals from across New Zealand who gave their time to participate in interviews. Author information: Patrick Kelly: Paediatrician, Service Clinical Director, Te Puaruruhau (Child Protection Team), Starship Children’s Health, Auckland; Department of Paediatrics, Child and Youth Health, Faculty of Medical and Health Sciences, University of Auckland. Melissa Adam: Honours Student, School of Psychology, University of Auckland. Carmen Basu: Child Protection Research Fellow, Te Puaruruhau (Child Protection Team), Starship Children’s Health, Auckland. Miranda Ritchie: National Violence Intervention Programme Manager, Health Networks Limited. Denise Wilson: Professor of Māori Health, Co-Director, Taupua Waiora Centre of Māori Health Research, AUT University, Auckland. Fred Seymour: Emeritus Professor, School of Psychology, University of Auckland. Corresponding author: Dr Patrick Kelly, Te Puaruruhau, Starship Children’s Health, Park Road, Private Bag 92024, Auckland 1, New Zealand, +64 9 307 2860 [email protected] URL: www.nzma.org.nz/journal-articles/interviews-with-health-professionals-about-the-national-child-protection-alert-system

REFERENCES 1. New Zealand Ministry of 3. Office of the Commissioner Commissioner for Children; Health. District Health for Children. Final report 2003 [Available from: Boards Wellington: on the investigation into https://library.nzfvc.org.nz/ Ministry of Health; 2018 the death of Riri-o-te-Rangi cgi-bin/koha/opac-detail. [Available from: https:// (James) Whakaruru, born pl?biblionumber=2944. www.health.govt.nz/ 13 June 1994, died 04 5. Laming H. The Victoria new-zealand-health-system/ April 1999 Wellington, Climbie Inquiry Norwich: key-health-sector-or- New Zealand: Office of the The Stationery Office; 2003 ganisations-and-people/ Commissioner for Children; [Available from: https:// district-health-boards. 2000 [Available from: www.gov.uk/government/ 2. Jiang N, Pacheco G, Dasgup- https://www.worldcat.org/ publications/the-victoria- ta K. Residential movement title/final-report-on-the-in- climbie-inquiry-report-of- within New Zealand: Quan- vestigation-into-the-death- an-inquiry-by-lord-laming. of-riri-o-te-rangi-james- tifying and characterising 6. Campbell D, Jones S, whakaruru-born-13-june- the transient population Brindle D. 50 injuries, 60 1994-died-04-april-1999/ Auckland, New Zealand: visits - failures that led to oclc/154314237. New Zealand Work the death of Baby P London Research Institute, AUT 4. Office of the Commissioner [updated 12/11/2008. University; 2017 [Available for Children. Report of Available from: https:// from: https://workresearch. the Investigation into the www.theguardian.com/ aut.ac.nz/__data/assets/ Deaths of Saliel Jalessa society/2008/nov/12/ pdf_file/0020/350606/ Aplin & Olympia Marisa baby-p-child-protec- Transient-population-re- Aplin Wellington, New tion-haringey. port-FINAL.pdf. Zealand: Office of the

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 43 www.nzma.org.nz/journal article

7. New Zealand Ministry of 13. Kelly P, Chan C, Reed P, opment of an electronic Health. National Health Ritchie M. The national medical record-based child Index Wellington: New child protection alert physical abuse alert Zealand Ministry of Health; system in New Zealand: A system. J Am Med Inform 2017 [Available from: prospective multi-centre Assoc. 2018;25:142-9. https://www.health.govt.nz/ study of inter-rater agree- 23. Flaherty R, Meiksans J, our-work/health-identity/ ment. Children and Youth McDougall S, Arney F. national-health-index. Services Review. 2020;116. Examining the power of 8. National Medical Warn- 14. Ritchie M. [Personal Child-At-Risk electronic ing System. Wellington: communication]. 1/05/2020. medical record (eMR) New Zealand Ministry of 15. Kelly P, Basu C, Graham VT, alerts to share interper- Health; 2012 [Available et al. Health professionals’ sonal violence, abuse from: https://www. perception of the New and neglect concerns: Do health.govt.nz/our-work/ Zealand National Child child protection alerts health-identity/nation- Protection Alert System: An help? Sydney, New South al-medical-warning-system. online survey. J Paediatr Wales: Australia’s National 9. Health and Disability Child Health. 2020. Research Organisation for Women’s Safety; 2018 System Review. Health 16. Guest G, Namey E, Chen [Available from: https:// and Disability System M. A simple method www.anrows.org.au/ Review - Interim Report. to assess and report publication/examining-the- Hauora Manaaki ki thematic saturation in power-of-child-at-risk-elec- Aotearoa Whānui – qualitative research. PLoS tronic-medical-record-emr- Pūrongo mō Tēnei Wā. One. 2020;15:e0232076. Wellington: Health alerts-to-share-interper- 17. Saunders B, Sim J, and Disability System sonal-violence-abuse-and- Kingstone T, et al. Satu- Review; 2019 [Available neglect-concerns-do-child- ration in qualitative from: https://system- protection-alerts-help/. research: exploring its review.health.govt. 24. Ash JA, Sittig DF, Campbell conceptualization and nz/interim-report/ EM, et al. Some Unin- operationalization. Qual download-the-report/. tended Consequences of Quant. 2018;52:1893-907. 10. Kelly P, Ritchie M, Belt K. Clinical Decision Support 18. Braun V, Clarke V. Using Privacy Impact Assess- Systems. AMIA Sympo- thematic analysis in ment. National Child sium Annual Symposium psychology. Qualitative Protection Alert System Proceedings American Research in Psycholo- Wellington: New Zealand Medical Informatics Asso- gy. 2006;3:77-101. Ministry of Health; 2011 ciation 2007:26-30. [Available from: https:// 19. Morrow SL. Quality 25. Low D. The Child Protec- media.starship.org.nz/ and trustworthiness tion Information Sharing child-protection-alert-sys- in qualitative research Project (CP-IS): electronic tem-privacy-impact/ in counseling psychol- information sharing to Child_Protection_Alert_ ogy. J Couns Psychol. improve the assessment System_Privacy_Impact_ 2005;52:250-60. of known vulnerable or Assessment.pdf. 20. Neale J, Miller P, West R. at-risk children. Adoption 11. Kelly P, Ritchie M, Wills R, Reporting quantitative & Fostering. 2016;40:293-6. McLaren Z. Child Protection information in qualitative 26. van der Put CE, Assink M, Alert System within Health research: guidance for Boekhout van Solinge NF. Wellington: New Zealand authors and reviewers. Predicting child maltreat- Ministry of Health; 2011 Addiction. 2014;109:175-6. ment: A meta-analysis [Available from: https:// 21. Sethuraman U, Kanni- of the predictive valid- www.starship.org.nz/ keswaran N, Murray ity of risk assessment health-professionals/clini- KP, et al. Prescription instruments. Child Abuse cal-network-for-child-pro- errors before and after Negl. 2017;73:71-88. tection. introduction of electronic 27. Bellis MA, Hughes K, Ford 12. Hamilton-Giachritsis CE, medication alert system K, et al. Life course health Browne KD. A retrospective in a pediatric emergency consequences and associat- study of risk to siblings in department. Acad Emerg ed annual costs of adverse abusing families. J Fam Med. 2015;22:714-9. childhood experiences Psychol. 2005;19:619-24. 22. Berger RP, Saladino RA, across Europe and North Fromkin J, et al. Devel- America: a systematic

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 44 www.nzma.org.nz/journal article

review and meta-analysis. 31. Kohl PL, Jonson-Reid M, 36. Wood JN, Hall M, Schilling Lancet Public Health. Drake B. Time to leave S, et al. Disparities in the 2019;4:e517-e28. substantiation behind: evaluation and diagnosis of 28. Benbenishty R, David- findings from a nation- abuse among infants with son-Arad B, Lopez M, al probability study. traumatic brain injury. et al. Decision making Child maltreatment. Pediatrics. 2010;126:408-14. in child protection: An 2009;14:17-26. 37. Benbenishty R, Jedwab M, international comparative 32. Kolbo JR, Strong E. Chen W, et al. Predicting study on maltreatment Multidisciplinary Team the decisions of hospital substantiation, risk assess- Approaches to the Inves- based child protection ment and interventions tigation and Resolution of teams to report to child recommendations, and the Child Abuse and Neglect: protective services, police role of professionals’ child A National Survey. Child and community welfare welfare attitudes. Child maltreatment. 1997;2:61-72. services. Child Abuse Abuse Negl. 2015;49:63-75. 33. Benbenishty R, Chen W. Negl. 2014;38:11-24. 29. Hussey JM, Marshall JM, Decision making by the 38. Rouland B, Vaithianathan English DJ, et al. Defining child protection team of R, Wilson D, Putnam-Horn- maltreatment according to a medical center. Health stein E. Ethnic Disparities substantiation: distinction Soc Work. 2003;28:284-92. in Childhood Prevalence without a difference? Child 34. Robinson M, Cottrell D. of Maltreatment: Evidence Abuse Negl. 2005;29:479-92. Health professionals in From a New Zealand 30. Modernising Child, Youth multi-disciplinary and Birth Cohort. Am J Public and Family. Expert Panel: multi-agency teams: Health. 2019;109:1255-7. Interim Report Wellington: changing professional Ministry of Social Devel- practice. J Interprof opment; 2015 [Available Care. 2005;19:547-60. from: https://www.msd. 35. Harr C, Souza L, Fairchild govt.nz/documents/ S. International models of about-msd-and-our-work/ hospital interdisciplinary publications-resources/eval- teams for the identification, uation/modernising-cyf/ assessment, and treatment interim-report-ex- of child abuse. Soc Work pert-panel.pdf. Health Care. 2008;46:1-16.

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 45 www.nzma.org.nz/journal article

Clinician knowledge of driving restrictions following a stroke event Laura Stratton, John Parsons, Shannon Tisbury, Susan Waterworth, Nicola Starkey

ABSTRACT AIM: To investigate the knowledge and practice of health professionals when advising persons on driving restrictions after a transient ischaemic attack (TIA) or stroke in a tertiary hospital in New Zealand. METHODS: Health professionals working in the area of stroke care across the acute and rehabili tation services in a large tertiary hospital were invited to complete an electronic survey around knowledge of driving restrictions based on the New Zealand Transport Agency (NZTA) guidelines. Knowledge was assessed for both private and commercial vehicle use. The other information gathered included participant profession, level of seniority and experience working in stroke care, previous education around medical-related driving restrictions and how and what driving recommendations were discussed with patients. Knowledge of driving restrictions was established by the number and percentage of correct responses for each condition (single TIA, multiple TIA and stroke with full recovery) relating to the recommended restrictions in both private and commercial vehicle use. RESULTS: Forty-nine participants’ surveys were analysed with representation from across all the health professions (24.5% (12/49) doctors, 38.8% (19/49) nurses and 36.7% (18/49) allied health). Only 38.8% reported having had received training around post-stroke driving restrictions. Knowledge around driving restrictions was highest for a single episode TIA for private vehicle use (73.5% (36/49)). For all other categories, fewer than 50% of participants answered correctly, with knowledge of commercial vehicle restrictions being the least accurate. CONCLUSIONS: Many health professionals have discussions with people about driving restrictions following a TIA or stroke. However, there appears to be limited knowledge of all the restrictions for each condition as they relate to either private or commercial vehicle use. Insufficient training and education for clinicians might explain this gap.

hen a person experiences a titioners regarding the medical aspects of medical event or condition that fitness to drive.1 These are safety guide- affects their ability to drive safely, lines for clinicians to advise patients about W 1 they pose a risk to themselves or to others. returning to driving after suffering from A stroke presents with a variety of impair- certain medical conditions. Such guidance ments that could impact on a person’s ability is in place to ensure that drivers on New to safely drive a car, such as altered visuo- Zealand roads are fit and competent to spatial perception, reaction time, attention, drive, with the aim of reducing harm caused sensation and muscle power.2,3 A safe return by road accidents.1 The New Zealand Clinical to driving is an important part of recovery, Guidelines for Stroke Management6 high- and the inability to do so has a profound light driving post-stroke as a chapter in its impact on community participation.4,5 recommendations. These consensus-based The New Zealand Transport Agency recommendations include: asking all (NZTA) provides a guide for health prac- admitted patients whether they intend to

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 46 www.nzma.org.nz/journal article

drive again; providing information about Many were of the view that further training assessments for fitness to drive to those around FTD was required, with some profes- who do intend to drive again; advising sional groups reporting that they obtained people with a diagnosis of stroke or tran- their knowledge of where to find return- sient ischaemic attack (TIA) not to drive to-driving (RTD) information through their for a minimum of one month; visiting their colleagues.13 general practitioner (GP) or specialist prior It has been estimated that a quarter of to resuming driving; and an occupational New Zealand’s population will be aged 65 therapy assessment for driving, which years and over by 2030.14 Therefore, the should be undertaken (if required), with its number of older drivers on the roads is outcomes documented and provided to the going to see a dramatic rise.15 Of all the 6 patient and the GP. In New Zealand, the people in New Zealand who have a current legal obligation of the health practitioner learner, restricted or full driver’s licence, is to assess the fitness of an individual to 551,754 are over the age of 65. Of this group, drive and to report those persons who 7,569 people are aged 90 years and older.16,17 continue to drive against medical recom- Driving is one of the main transport options mendation. However, the NZTA guidance for our older population in New Zealand, booklet, although part of legislation, is a especially for those outside of urban centres, guideline for best practice only and not where public transport is limited.16,18 With 1 legally enforceable. Therefore, advice current trends demonstrating that stroke about returning to driving can be confusing, mortality is falling quicker than the stroke resulting in variance in health professionals’ incidence, and along with New Zealand’s knowledge and their provision of return-to- population growth and ageing demo- 5,7–9 driving information to patients. graphic, a rise in stroke related disability To support health professionals when is inevitable.19 With this information, it is giving advice, the Stroke Foundation of reasonable to assume that there will be New Zealand has produced a free printable more drivers on the road who may have resource titled Fact sheet: driving and experienced a stroke. transport after stroke. This is located on the Being able to return to driving is a priority Stroke Foundation website and provides a for many patients after a stroke or TIA. summary of the NZTA driving recommen- However, driving too soon after an acute dations following a TIA or stroke for both medical event, such as a stroke or TIA, 10 private and commercial drivers. The fact has safety implications. Given the lack of sheet provides clear rationale for the driving research examining people who return to restrictions as well as information for driving following an acute medical event on-going driving rehabilitation and alter- in New Zealand, this study aimed to assess native transport options. health professionals’ knowledge of driving Internationally, regulations related to restrictions following stroke and TIA in a post-stroke driving vary, and it has been tertiary hospital setting in New Zealand demonstrated that knowledge of post- and how they deliver this information to stroke driving restrictions is low among patients. healthcare professionals.11 In the UK, a study by Batool et al found that only 29% of Methods physicians and 36% of allied health profes- Ethical approval for this study was sionals knew the restrictions for domestic granted from the University of Waikato driver’s licence holders following a stroke (UoW HREC (Health) #2017-26). or a TIA.7 In addition, a Swedish review of medical records indicated that a report of a Participants discussion relating to driving cessation post- An advertisement for the survey was stroke was missing in 81% of the medical posted on the intranet of a large tertiary records.12 Despite health professionals district health board in New Zealand, reporting they were aware of fitness-to- inviting medical, nursing and allied health drive (FTD) guidelines, only 33.9% of the professionals working across acute and respondents had utilised these guidelines rehabilitation services involved with the within stroke care in the past two years.13 care of stroke patients to participate. In

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 47 www.nzma.org.nz/journal article

addition, information about the survey was provided. At the end of the survey, partici- circulated via email by the service leaders of pants were given the option to provide their relevant disciplines/departments. email address to receive a summary of the Procedure and questionnaire findings, and they were provided with a list of relevant resources/additional information design about fitness to drive and return to driving. The questionnaire was developed on the Qualtrics online survey software. As described above, an advertisement Results containing a link to the survey was circu- A total of 66 participants began the survey, lated via email, and information about but 17 were excluded from the analyses due the study was posted on the intranet of to missing data, leaving a final sample of a large tertiary district health board in 49. Twelve of these 49 (24.5%) were doctors, New Zealand. Upon clicking a link to a 19/49 (38.8%) were nurses and 18/49 (36.7%) website, the respondents were provided were allied health professionals. Fifteen with a short overview of the purpose of the of the 49 participants (30.6%) had been research, reassured that their responses practising between 0–5 years, 18/49 (36.8%) were anonymous and that they had the between 6–15 years and 16/49 (32.7%) had ability and right to withdraw at any time. been practising for over 15 years. Just over The respondents were asked to complete a third of participants (15/49) worked in the the questionnaire without accessing any medical area, 6/49 (12.2%) in cardiology, reference materials. Respondents were 8/49 (16.3%) in community/rural health, then presented with the survey questions, 8/49 (16.3%) in geriatrics/stroke and 12/49 which took approximately 10–15 minutes to (24.5%) in other clinical areas. complete (a copy of the survey can be found Training and knowledge of driving in Appendix). The questionnaire was based restrictions on the version used by Batool et al7 and With regard to training received about modified to reflect the New Zealand advice driving restrictions, 20/49 (40.8%) reported and restrictions as per the NZTA guide- having received no training, 19/49 (38.8%) lines.1 Participants were asked to choose the had received training about driving post- appropriate private and commercial driving stroke and 10/49 (20.4%) had received restrictions from a pre-populated list for a training about driving restrictions in number of conditions, including single and relation to other medical conditions. multiple transient ischaemic attack (TIA) and stroke. The restrictions recommended The percentage of respondents choosing by NZTA are as follows: the correct response for the driving restrictions and TIA or stroke is shown in Table 1. • Single TIA with full recovery: one As can be seen in the table, knowledge of month for a private vehicle and six driving restrictions was best for a single TIA months for commercial driving. for private vehicles. However, fewer than • Multiple TIA: three months after last 50% of participants knew the driving restric- incident for a private vehicle and 12 tions for the other categories. months for commercial driving. • Stroke with no residual deficits: one Table 1: Percentage of correct responses among month for a private vehicle and three health professionals for driving restrictions for years (with medical approval) for TIA and stroke for private and commercial licence commercial driving. holders. Participants were also asked to provide Private Commercial information on their profession, level of (n, %) (n, %) seniority, experience working in the field Single TIA 36, 73.5 9, 18.4 of stroke and, if any, previous fitness- to-drive education they had received. Multiple TIA’s 11, 22.4 11, 22.4 Finally, they were asked about when and how they discuss driving restrictions with Stroke with 22, 44.9 9, 18.4 their patients and whether they thought no residual that patients adhered to the advice they deficits

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 48 www.nzma.org.nz/journal article

A high percentage of respondents indi- permanent driving restrictions, with only cated that they did not know the restrictions 14.3% reporting that they would notify the related to operating a commercial vehicle GP instead. However, 22.5% (11/49) stated (38.8% (19/49) for single TIA, 44.9% (22/49) that they did not know who to report to. for multiple TIA, 42.9% (21/49) for stroke). Furthermore, 79.5% (39/49) of respondents For each scenario, some respondents identified that medical professionals (eg, reported the driving restrictions as being doctors, GPs, specialists) were responsible shorter than the NZTA guidelines: For for reporting these restrictions, with a driving a private vehicle, 3/49 respondents further 8.1% of those respondents (4/49) (6.1%) reported there was no restriction also stating other health professionals after a single TIA (NZTA recommend one should also notify the appropriate agency. month), 7/49 (14.3%) reported a restriction Seven respondents (14.2%) thought it was of one month following multiple TIA (NZTA the patients’ responsibility, one (2.1%) was recommend three months) and 3/49 (6.1%) unsure and one (2.1%) identified everyone reported no restrictions after stroke with no as responsible for reporting. residual deficits (NZTA recommended one Nearly two-thirds of respondents (62.5% month). For driving a commercial vehicle, (31/49)) reported that they knew where to 8/49 participants (16.3%) selected restric- refer patients for post-discharge driving tions of less than six months for a single TIA, assessments. The preferences for onward 3/49 respondents (6.1%) selected periods of referral were 36.8% (18/49) to occupational less than 12 months after multiple TIA and therapy (OT), 14.3% (7/49) to medical or 15/49 participants (30.6%) selected periods nursing professionals and 10.4% (5/49) of less than three years after a stroke with reported other referrals as appropriate no residual deficits. post-discharge driving follow-up. Action and advice provided by Over half of the respondents (55.1% clinicians (27/49)) reported that the advice provided to Overall, 83.3% (41/49) of respondents patients about driving restrictions is inad- reported that they discussed driving restric- equate, and 83.7% (41/49) indicated that tions with their patients. The majority of the information should be improved. In these discussions followed an assessment or the general, respondents noted that many were provided prior to discharge home (50% patients do not understand the need for the (25/49)). Half of the respondents reported driving restrictions and that, as they are that the advice they provided included when only guidelines, they are hard to enforce. to resume driving and where to obtain There was also some suggestion that certain further information if needed. Thirty-five demographic groups may be more resistant percent (17/49) of respondents discussed only to the restrictions, particularly older men. one of these options with the patient. The remainder of respondents reported that they Discussion would advise the patient to seek GP advice This local New Zealand study showed and approval before resuming driving. that health professionals working in a large Seventy-five percent of clinicians reported tertiary hospital have poor knowledge that patients had questions regarding around driving restrictions following a TIA driving, yet only 35.4% (17/49) of clinicians or minor stroke. Fewer than 50% of the staff reported providing written information surveyed indicated that they knew what regarding driving restrictions and RTD advice to provide around driving to this to patients. This was despite 60.4% of the patient population, with the least knowledge respondents (30/49) reporting that they being demonstrated around restrictions for thought that patients did not adhere to returning to commercial driving. driving recommendations and that only The majority of staff (83%) surveyed in two-thirds of patients (66.6% (33/49)) under- this study reported they were providing stood the driving restrictions. driving recommendations to patients, The majority of respondents (63.2% which is a lot higher than what was found (31/49)) were aware that reporting to NZTA amongst allied health staff (33%) across was required for conditions that led to Australia.5 However, it is concerning that

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 49 www.nzma.org.nz/journal article

half of these respondents (48%) reported strategies should be considered and utilised having had no formal training around the depending on the needs and characteristics return-to-driving issues, and they did not of the patient. know the correct recommendations for Limitations each condition, which calls into question The main limitation to this study is the the accuracy of the information being small sample size that is local to one tertiary provided to patients. Also, three-quarters hospital in New Zealand and therefore of clinicians reported that patients often may not be representative of the rest of the had questions relating to RTD, but only a country. Future research in New Zealand third would actually provide written infor- regarding safe return-to-driving advice mation regarding these restrictions and following stroke should be multi-centre and recommendations. It therefore cannot be also include GPs to better understand clini- expected that persons’ driving behaviours cians’ knowledge and delivery of driving following a TIA or stroke align with what restrictions and support. Furthermore, is outlined by the NZTA and Stroke Foun- investigating patients’ understanding and dation of New Zealand, given that the adherence to return-to-driving advice information provided in the first instance following stroke would add to this body is inaccurate and not supported by written of knowledge and aid the development of documentation. clinician (and/or NZTA) education and infor- Evidence suggests that, within New mation regarding safe return to driving. Zealand, 27–35% of persons return to driving within one month following a Recommendations Formal training and education regarding stroke, often against medical or allied health the driving restrictions around common recommendations.20,21 The characteristics of medical condition and safe RTD education the persons who had resumed driving were should be provided within undergraduate most likely to be male or the main income programmes and also to clinicians as part of earner, and they were often independent ongoing professional development. with activities of daily living (ADL).21 Over half of these subjects interviewed—of Provision of accurate, standardised whom 15% were found to have cognitive written information regarding driving impairment—reported that they could not restrictions and safe RTD (eg, Fact sheet: remember being given any specific advice driving and transport after stroke)10 to all around RTD. In mild stroke, cognitive persons diagnosed with a TIA or mild stroke deficits can persist for several months after prior to discharge from hospital. injury,22 and this may predispose individuals Patient education should include recom- to an increased risk of having a motor-ve- mended standdown period appropriate for hicle accident. The education and advice each licence type and where to obtain advice provided by physicians during hospital and driving assessments if required. stays has been shown to significantly reduce Acute hospital services could benefit from the annual rate of road crashes in people having clear processes for identifying which considered to be medically unfit to drive members of the team will provide advice 23 by 45% per 1,000 persons. Therefore, on driving restrictions and RTD and written the provision of correct driving infor- information alongside NZTA reporting mation during acute hospital admission procedures. is important, and a range of educational

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 50 www.nzma.org.nz/journal article

Appendix Figure A1: Driving after a stroke/cardiovascular event. This study aims to explore the driving-related advice health professionals provide to patients following a stroke/cardiovascular event, and how they deliver this information to patients. Your participation is entirely voluntary and your responses to the survey are anonymous and cannot be linked with you personally. Completion of the questionnaire and the submission of the responses will be taken as your consent to participate. You are free to withdraw without penalty from the study at any time, by closing the browser window. If you have any questions about the study please contact a member of the study team (Nicola Starkey, [email protected]; Victoria [email protected]; John Parsons, [email protected]; Samuel Charlton, [email protected] and Jess Leov, [email protected]). The study has received approval from the University of Waikato Human Research Ethics Committee (Health) #2017-26. We would like to encourage you to answer the questions based on you current knowledge and how you have dealt with patients over the last few months (rather than trying to find the correct answers).

Q1 What professional group are you from? a. SMO b. RMO c. HO d. Nurse e. Social Worker f. Physiotherapist g. Occupational therapist h. Other Q2 How many years have you been practicing? a. 0 - 5 years b. 6 - 10 years c. 10 - 15 years d. 15 + years Q3 What clinical area do you practice?

Q4 Have you worked with stroke patients? a. yes b. no Q5 Have you received teaching about driving restrictions surrounding medical conditions? (tick all that apply) a. No Training b. Epilepsy c. Cardiac Diseases d. Stroke e. TIA f. Other (6) Q6 Where should medical conditions that permanently restrict driving be reported? a. Insurance company b. NZTA c. GP practice d. Don't know

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 51 www.nzma.org.nz/journal article

Q7 Who is responsible for reporting such conditions?

Q8 What should you do if patients continue to drive when they are not fit to do so?

Q9 What are the restrictions for the following conditions? Q10 Single TIA - Full recovery - Vehicle private driving? a. No restriction b. 1 week c. 1 month d. 6 months e. 1 year f. Don't know g. Outside my area of practice h. Other (8) Q11 Single TIA - Full recovery - Commercial driving: a. No restrictions b. 1 week c. 1 month d. 6 months e. 1 year f. Don't know g. Outside my area of practice h. Other Q12 Multiple TIA - Private driving: a. No restrictions b. 1 week c. 1 month d. 3 months e. 6 months f. Don't know g. Outside my area of practice h. Other Q13 Multiple TIA - Commercial driving: a. No restrictions b. 1 month c. 3 months d. 6 months e. 12 months f. Don't know g. Outside my area of practice h. Other Q14 Stroke - No residual deficits - Private driving: a. No restrictions b. 1 week c. 1 month d. 6 months e. 12 months f. Don't know g. Outside my area of practice h. Other (8)

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 52 www.nzma.org.nz/journal article

Q15 Stroke - No residual deficits - Commercial driving: a. No restrictions b. 3 Months c. 6 months d. 12 months e. 3 years f. Don't know g. Outside my area of practice h. Other Q16 Seizure (unprovoked): a. No restrictions b. 1 week c. 1 month d. 6 months e. 12 months f. Don't know g. Outside my area of practice h. Other Q17 Myocardial infarction: a. No restrictions b. 1 week c. 2 weeks d. 1 month e. 2 months f. 3 months g. Don't know h. Outside my area of practice i. Other Q18 Pacemaker implant: a. No restrictions b. 1 week c. 2 weeks d. 4 weeks e. 6 weeks f. Don't know g. Outside my area of practice h. Other Q19 Coronary artery bypass graft (CABG): a. No restrictions b. 1 week c. 1 month d. 3 months e. 6 months f. Don't know g. Outside my area of practice h. Other Q30 Name two residual impairments that may require discontinuation of driving following a stroke. a. (1) ______b. (2) ______Q21 Do you discuss driving restrictions with your patients? Yes (1) No (2)

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 53 www.nzma.org.nz/journal article

Q22 When do you typically tell the patient about the driving restriction? a. After I have assessed them b. Before they discharge home c. When I remember d. Rely on someone else telling them e. Other (5) Q23 Where do you typically tell the patient about their driving restrictions? a. When they are in the hospital bed b. When I give them their discharge papers c. Over the phone d. During assessment e. Other Q24 What information do you tell the patient about driving restrictions? a. What the restrictions are b. Where to get further information c. What to do to resume driving d. all of the above e. Other Q25 Do you give the patient any written information about driving restrictions? Yes / No Q26 Do patients ask questions about driving restrictions? Yes / No Q27 What kind of questions do patients have about driving restrictions Q28 Do you think that patients understand the driving restrictions? Yes / No Q29 Do you think that patients adhere to driving restrictions? Yes / No Q30 Do you know where to refer patients to for driving assessments post discharge? Yes / No Q31 Name where you refer patients to for driving assessment post discharge

Q32 Do you think that the information you provide to patients about driving restrictions is adequate? Yes / No Q33 Do you think there needs to be any improvements about patient information regarding driving restrictions? a. Yes (1) b. No (2) c. Add comment (3) ______Q47 Are there any other comments you would like to make about driving restrictions?

Q34 Thank you for completing the survey. If you would like to receive a summary of the findings please enter your email address below (this will not be liked with your answers) If you would like more information about current recommendations for driving after stroke, the following resources may be helpful: NZTA- Medical Aspects of Fitness to Drive NZTA- Information for Medical Practitioners, NZTA- Occupational Therapy Assessments for Driving Fitness, Stroke Foundation Resources for Health Professionals.

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 54 www.nzma.org.nz/journal article

Competing interests: Nil. Author information: Laura Stratton: Bupa Research Associate, Institute of Healthy Ageing, Waikato District Health Board, Hamilton. John Parsons: Bupa Fellow in Allied Health Research, Institute of Healthy Ageing, Waikato District Health Board, Hamilton; Associate Professor, School of Nursing, The University of Auckland, Auckland. Shannon Tisbury: Bupa Research Associate, Institute of Healthy Ageing, Waikato District Health Board, Hamilton. Susan Waterworth: Senior Lecturer, School of Nursing, The University of Auckland, Auckland. Nicola Starkey: Professor, School of Psychology, Waikato University, Hamilton. Corresponding author: John Parsons, Room 231A, Level 2, Building 505, 85 Park Road, Grafton, Auckland 1142, New Zealand, 09 923 3935 [email protected] URL: www.nzma.org.nz/journal-articles/clinician-knowledge-of-driving-restrictions-following-a-stroke-event

REFERENCES 1. New Zealand Transport 6. Stroke Foundation of New 10th February ); Available Agency, Medical aspects of Zealand and New Zealand from: https://www.stroke. fitness to drive: A guide for Guidelines Group, Clinical org.nz/sites/default/files/ health practitioners. 2014, Guidelines for Stroke inline-files/Driving%20 New Zealand Government: Management 2010. 2010, and%20Transport%20 Palmerston North. Stroke Foundation of New after%20Stroke%20 2. Yu, A.Y., et al., Increased Zealand: Wellington. Print%204%20pages.pdf. risk of traffic injury 7. Batool, S., et al., Health 11. Ng, K.Y.B., et al., Knowledge after a cerebrovascular professionals’ knowledge of Driving Vehicle Licens- event. Stroke, 2018. of driving restrictions ing Agency guidelines 49(12): p. 3006-3011. following stroke and among NHS doctors: a 3. Hird, M.A., et al., Is it safe TIA: experience from multicentre observational to drive after acute mild a hyperacute stroke study. JRSM open, 2015. stroke? A preliminary centre. Postgraduate 6(10): p. 2054270415601586. report. Journal of the medical journal, 2014. 12. Mårdh, S., P. Mårdh, neurological sciences, 90(1065): p. 370-376. and A. Anund, Driving 2015. 354(1-2): p. 46-50. 8. McNamara, A., et al., The restrictions post-stroke: 4. Frith, J., et al., Returning need for consistency and Physicians’ compliance to driving after stroke: A equity in driver education with regulations. Traffic systematic review of adher- and assessment post-stroke. injury prevention, 2017. ence to guidelines and Journal of Transport & 18(5): p. 477-480. legislation. British Journal Health, 2014. 1(2): p. 95-99. 13. Hawley, C., Road Safety of Occupational Therapy, 9. Ranchet, M., et al., Fitness- Research Report No. 91 2015. 78(6): p. 349-355. to-drive agreements The Attitudes of Health 5. Frith, J., et al., Shifting after stroke: medical Professionals to Giving gears: An inpatient versus practical recom- Advice on Fitness to medical record audit and mendations. European Drive 2010, Department post-discharge survey of journal of neurology, 2016. for Transport: London. return-to-driving following 23(9): p. 1408-1414. 14. Statistics New Zealand, stroke/transient ischaemic 10. Stroke Foundation of 2013 Census QuickStats attack. Australian occu- New Zealand. Driving About People Aged 65 and pational therapy journal, and transport after a Over. 2016, Statistics New 2017. 64(3): p. 264-272. stroke. 2016 (cited 2020 Zealand: Wellington.

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 55 www.nzma.org.nz/journal article

15. New Zealand Transport qualitative studies on 21. Yu, S., et al., Driving in Agency. Senior driver health and well-being, stroke survivors aged 18–65 statistics and resources. 2018. 13(1): p. 1503908. years: the Psychosocial 2019 (cited 2020 21st 18. Liddle, J., K. McKenna, Outcomes In strokE (POISE) February); Available from: and K. Broome, Older cohort study. Interna- https://www.nzta.govt. road users: from driv- tional Journal of Stroke, nz/safety/driving-safely/ ing cessation to safe 2016. 11(7): p. 799-806. senior-drivers/senior-driv- transportation. Sleep, 22. Wolf, T.J. and M.C. er-statistics-and-resources/. 2003. 57(59.6): p. 59.1. Rognstad, Changes in 16. Hempel, M.E., et al., 19. Feigin, V.L., et al., 30-Year cognition following mild Scared behind the wheel: trends in stroke rates and stroke. Neuropsycho- what impact does driving outcome in Auckland, logical rehabilitation, anxiety have on the health New Zealand (1981-2012): 2013. 23(2): p. 256-266. and well-being of young a multi-ethnic popula- 23. Redelmeier, D.A., et al., older adults? International tion-based series of studies. Physicians’ warnings psychogeriatrics, 2017. PloS one, 2015. 10(8). for unfit drivers and the 29(6): p. 1027-1034. 20. McCarron, M., A. Loftus, risk of trauma from road 17. Neville, S., et al., “Engaging and P. McCarron, Driving crashes. New England in my rural community”: after a transient ischaemic Journal of Medicine, 2012. perceptions of people attack or minor stroke. 367(13): p. 1228-1236. aged 85 years and over. Emergency medicine jour- International journal of nal, 2008. 25(6): p. 358-359.

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 56 www.nzma.org.nz/journal article

Prognostic significance of mid-range ejection fraction following acute coronary syndrome (ANZACS-QI 23) Daniel Chan, Robert N Doughty, Mayanna Lund, Mildred Lee, Katrina Poppe, Andrew J Kerr

ABSTRACT AIM: Recommendations regarding medication use after acute coronary syndrome (ACS) are dichotomised according to whether left ventricular ejection fraction (LVEF) is <40% or≥ 40%. In the context of heart failure (HF), a mid-range EF (mrEF, 40–49%) confers an intermediate prognosis between reduced EF (rEF, <40%) and preserved EF (pEF, ≥50%). The aim of this study was to describe, in the context of ACS, the frequency of each EF subgroup and their associated outcomes. METHODS: Consecutive patients presenting with ACS who underwent coronary angiography during 2015 were enrolled in the ANZACS-QI (All New Zealand Acute Coronary Syndrome—Quality Improvement) registry. Outcomes were obtained using anonymised linkage to national datasets. Cox proportional hazards models were used to adjust for confounding variables. RESULTS: Of the cohort of 6,216 patients, 31% did not have an LVEF assessment. Of those with a recorded LVEF, 63% had pEF, 21% had mrEF and 16% had rEF. Mean follow-up was 1.5 years. After adjusting for age, sex, clinical risk factors and post-ACS management, those with mrEF and rEF had a higher adjusted risk of all-cause mortality compared to pEF (HR 1.55, 95% CI 1.12–2.15 and HR 2.57, 95% CI 1.89–3.48, respectively). After adjustment, rEF was associated with an increased risk of subsequent HF hospitalisation (HR 2.32, 95% CI 1.75–3.08). CONCLUSIONS: One in five patients post-ACS have mrEF, which is associated with an intermediate risk of morbidity and mortality compared to those with pEF and rEF. Further study is warranted to determine the optimal management for these patients.

educed left ventricular ejection dations for any subgroups of LVEF greater fraction (LVEF) is associated with than 40% following ACS. adverse clinical outcomes after acute For patients presenting with heart failure R 1 coronary syndrome (ACS), and clinical trials (HF), the European Society of Cardiology 2 have established the role of beta-blockers, (ESC) heart failure guidelines have drawn angiotensin converting enzyme inhibitors attention to the patients with LVEF inter- 3 (ACEi), angiotensin receptor blockers (ARB) mediate between those with reduced 4 and aldosterone antagonists in patients LVEF (HFrEF; LVEF<40%) and those with with LVEF<40% post-ACS. Use of these preserved LVEF (HFpEF; LVEF≥50%), agents post-ACS is a Class I indication in so-called HF with mid-range LVEF (HFmrEF; international guidelines for LVEF<40%, but LVEF 40% to 49%).7 Studies have shown that recommendations are less consistent for the patients with HFmrEF have an inter- 5,6 LVEF greater than 40%. Current guidelines mediate risk of all-cause mortality between make no distinction in treatment recommen- those with HFrEF and HFpEF.8-10 However,

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 57 www.nzma.org.nz/journal article

there are little data available regarding the Coronary Events (GRACE) admission-to-six prognostic significance of mid-range LVEF in month score14 was calculated using standard patients following ACS. If the same interme- variables: age, admission heart rate and diate prognosis occurs for mid-range LVEF systolic blood pressure, admission plasma after ACS, it may lead to a re-evaluation of creatinine level, cardiac arrest, presence of the role of beta-blockers and ACEi in these ST deviation on electrocardiogram (ECG), patients. elevated cardiac enzymes on admission Since 2015, all New Zealand hospitals and initial Killip Class. History of HF prior admitting patients with ACS have partic- to the index ACS event was captured on ipated in the All New Zealand Acute the ANZACS-QI registry. The Charlson 15 Coronary Syndrome—Quality Improvement Comorbidity Index, modified to exclude (ANZACS-QI) registry,11 which records a cardiac conditions and diabetes, was calcu- mandatory dataset in all patients with lated for each patient using the linked ACS referred for coronary angiography, national datasets. The New Zealand Index including LVEF measurement using the of Socioeconomic Deprivation (NZDep) LVEF bands embraced by the 2016 ESC heart is a socioeconomic-deprivation score failure guideline. Linkage of individual derived from national census data (1–2 patient data to national hospitalisation and least deprived quintile, 9–10 most deprived 16 mortality data allows assessment of quintile). post-discharge outcomes. Left ventricular ejection fraction Our aim was to evaluate, in a contem- measured on echocardiography, ventric- porary and comprehensive national ACS ulogram or other modality during the cohort, the characteristics, management index admission was classified as either and outcomes of patients with preserved, preserved LVEF (pEF, LVEF≥50%), mid-range and reduced LVEF. mid-range LVEF (mrEF, LVEF 40–49%) or reduced LVEF (rEF, LVEF<40%), consistent with the 2016 ESC heart failure guide- Methods lines.7 Medication dispensing data was Patients enrolled in the ANZACS-QI obtained by linkage to national pharma- registry with confirmed ACS, who ceutical dispensing dataset, as previously underwent invasive coronary angiog- described.17 Medications investigated in raphy and were discharged during the this study were aspirin, second antiplaletet 2015 calendar year, were included in (clopidogrel, ticagrelor or prasugrel), statin, the study. Patients who were not New beta-blocker, ACEi, ARB, spironolactone Zealand residents were excluded. For and anticoagulants (warfarin, dabigatran patients with multiple admissions with and rivaroxaban). Medication dispensing ACS during the study period, the first on discharge was defined as medication admission with complete data was used as dispensed between zero and three months the index admission. Subsequent admis- following discharge, as medications are sions with ACS were counted as outcomes. typically prescribed in New Zealand in a All patients had follow-up available through three-month supply. linkage of national datasets to at least one year post-discharge or death (whichever Outcomes happened first). All New Zealanders have a unique National Health Identifier (NHI) number. Data and definitions We used an encrypted version of the NHI Only patients with a confirmed diag- to anonymously link in-hospital ANZACS-QI nosis of ST-segment elevation myocardial patient records to subsequent outcomes infarction (STEMI), non-ST segment captured in the New Zealand Ministry of elevation myocardial infarction (NSTEMI) or Health mortality and public hospitalisation unstable angina were included. Myocardial datasets. The encryption and linkage meth- infarction (MI) was defined according to odology has been previously described.11 12 the third universal definition. Unstable In-hospital and post-discharge deaths were angina was defined according to guideline identified using the national mortality 13 definitions. The Global Registry of Acute dataset. Deaths were categorised as cardio-

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 58 www.nzma.org.nz/journal article

vascular disease (CVD) or non-CVD-related non-fatal cardiovascular disease rehos- from International Classification of pitalisation. Covariates were selected a Diseases, Tenth Revision, Australian Modi- priori based on their clinical relevance for fication (ICD-10-AM) coded hospital and patients post-ACS and included age, gender, mortality datasets, with CVD death defined modified Charlson Comorbidity Index, prior as deaths associated with atherosclerotic, MI, prior HF, diabetes, GRACE score, type cardiac, cerebral or peripheral vascular of ACS, coronary angiography findings, disease.18 Subsequent hospital admission revascularisation (percutaneous coronary with cardiovascular disease was defined as intervention (PCI) or coronary artery the composite of MI, stroke or HF, with HF bypass grafting (CABG)) and medications defined as an admission with a primary or dispensed at discharge. Sensitivity analyses secondary diagnosis ICD-10-AM code of I110, were performed to assess the association I130, I132, I50, I501 or I509. between EF band and mortality for STEMI Ethics and NSTE-ACS subtypes separately and for the subgroup of patients who survived the The ANZACS-QI registry is part of the first 30 days post-ACS. The association of EF Vascular Informatics using Epidemi- band with cardiac and non-cardiac mortality ology and Web (VIEW) programme at the separately was also assessed. University of Auckland, which oversees the implementation, quality and academic All p-values were two tailed and p<0.05 use of national datasets. It is funded by was considered statistically significant. Data the Health Research Council and the were analysed using SAS version 9.4 (SAS National Heart Foundation of New Zealand. Institute, Cary, NC), and survival plots were The VIEW programme was approved by created using RStudio version 1.0.143. Northern Region Ethics Committee Y in 2003 (AKY/03/12/314) and updated to ANZACS-QI Results registries with annual approvals by the Study cohort national Multi-Region Ethics Committee From 1 January 2015 to 31 December since 2007 (MEC07/19/EXP). 2015, 6,216 patients with confirmed ACS Statistics who underwent coronary angiography were Continuous variables were summarised enrolled in the ANZACS-QI registry. A record as mean with standard deviation (SD) of LVEF was available for 4,290 patients and compared using Student’s t-test. Cate- (69%) and not available for 1,926 (31%) gorical data are reported by frequency and during their index ACS admission. Of the percentage and compared using Pearson’s 4,290 patients with a recorded LVEF, 2,704 chi-squared test or Fisher exact test where (63%) had pEF, 897 (21%) had mrEF and 689 appropriate. (16%) had rEF. All-cause death is reported as the unad- Baseline characteristics (Table 1) justed percentage of patients who had died Compared to patients with pEF, patients by one year and as deaths per 1,000-patient with rEF were slightly older (67±12 vs 64±12 years with 95% confidence interval (CI) years) and more likely to be male (77% vs calculated using an exact mid-p method 68%), and New Zealand Māori and Pacific (www.openepi.com). Survival functions patients were relatively over-represented. were estimated by a Kaplan–Meier estimator Patients with mrEF were more likely to be and Kaplan–Meier plots were truncated at male compared to patients with pEF (74% vs 1.5 years. The significance of the difference 68%). between pEF, mrEF, rEF and no EF groups Patients with rEF and mrEF were more were tested using the log-rank test. The likely to have had a prior MI compared proportional hazard assumption was tested to those with pEF (27%, 21% vs 17% respec- using SAS ASSESS statement in PROC PHREG tively) and to have prior documented HF and was met. Multivariable Cox regression (11%, 4% vs 2% respectively). Charlson models were used to estimate the adjusted scores and the proportion of patients hazard ratios of mrEF and rEF compared with diabetes were higher in those with to pEF for all-cause mortality, heart failure rEF and similar between pEF and mrEF readmission and all-cause mortality and groups.

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 59 www.nzma.org.nz/journal article

Table 1: Baseline characteristics.

Preserved EF Mid-range EF Reduced EF No EF pEF vs mrEF mrEF vs rEF pEF vs no EF (n=2,704) (n=897) (n=689) (n=1,926) P-value P-value P-value Demographics

Age, years (mean±SD) 64.4±11.7 64.7±12.3 67.3±11.5 66.1±11.5 0.474 <.001 <.001

Male, n (%) 1,849 (68.4) 665 (74.1) 530 (76.9) 1,288 (66.9) 0.001 0.202 0.280

Ethnic group, n (%) 0.477 <.001 0.001 NZ Māori 261 (9.7) 98 (10.9) 100 (14.5) 188 (9.8) Pacific 127 (4.7) 32 (3.6) 54 (7.8) 56 (2.9) Indian 131 (4.8) 40 (4.5) 26 (3.8) 69 (3.6) Other Asian 84 (3.1) 25 (2.8) 23 (3.3) 43 (2.2) European/other 2,101 (77.7) 702 (78.3) 486 (70.5) 1,570 (81.5)

NZDep, n (%) * 0.461 0.121 0.878 1–2 452 (16.7) 171 (19.1) 107 (15.5) 311 (16.1) 3–4 458 (16.9) 146 (16.3) 98 (14.2) 341 (17.7) 5–6 525 (19.4) 172 (19.2) 135 (19.6) 390 (20.2) 7–8 627 (23.2) 213 (23.7) 170 (24.7) 438 (22.7) 9–10 633 (23.4) 192 (21.4) 178 (25.8) 442 (22.9)

Comorbidities

Prior CVD, n (%) 775 (28.7) 256 (28.5) 258 (37.5) 851 (44.2) 0.944 <.001 <.001

Prior MI, n (%) 450 (16.6) 191 (21.3) 189 (27.4) 513 (26.6) <.001 0.727 0.364

COPD, n (%) 216 (8.0) 74 (8.2) 71 (10.3) 179 (26.6) 0.803 0.159 0.117

Current smoker, n (%) 602 (22.3) 241 (26.9) 174 (25.3) 417 (21.7) 0.005 0.469 0.620

Diabetes, n (%) 559 (20.7) 191 (21.3) 204 (29.6) 445 (23.1) 0.692 <.001 0.048

BMI (kg/m2), n (%)* 0.089 0.364 0.567 18.5–<25 462 (17.1) 185 (20.6) 148 (21.5) 352 (18.3) 25–<30 841 (31.1) 298 (33.2) 199 (28.9) 621 (32.2) 30–<35 531 (19.6) 168 (18.7) 120 (17.4) 361 (18.7) ≥35 324 (12.0) 93 (10.4) 82 (11.9) 252 (13.1)

eGFR (ml/min/1.73m2), 0.023 <.001 0.003 n (%)* <30 60 (2.2) 29 (3.2) 42 (6.1) 60 (3.1) 30–<60 573 (21.2) 219 (24.4) 243 (35.3) 473 (24.6) ≥60 2,070 (76.6) 649 (72.4) 404 (58.6) 1,393 (72.3)

Dialysis prior to 6 (0.7) 22 (3.2) 31 (1.6) 0.082 <.001 0.572 38 (1.4) admission, n (%)

Prior CHF, n (%) 47 (1.7) 34 (3.8) 74 (10.7) 72 (3.7) <.001 <.001 <.001

Modified Charlson 0.586 <.001 <.001 Comorbidity Index, n (%) 0 2,049 (75.8) 694 (77.4) 455 (66.0) 1,348 (70.0) 1–2 486 (18.0) 148 (16.5) 156 (22.6) 401 (20.8) ≥3 169 (6.3) 55 (6.1) 78 (11.3) 177 (9.2)

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 60 www.nzma.org.nz/journal article

Table 1: Baseline characteristics (continued).

Preserved EF Mid-range EF Reduced EF No EF pEF vs mrEF mrEF vs rEF pEF vs no EF (n=2,704) (n=897) (n=689) (n=1,926) P-value P-value P-value Admission data

Admission HR (bpm), 74.3±17.8 77.8±19.7 85.9±23.0 74.2±17.9 <.001 <.001 0.841 mean±SD

Admission SBP (mmHg), 143.9±26.2 139.2±26.2 133.9±26.6 142.5±26.9 <.001 <.001 0.071 mean±SD

Initial Killip Class, n (%) <.001 <.001 0.016 I 2,573 (95.2) 790 (88.1) 520 (75.5) 1,801 (93.5) II–IV 131 (4.8) 107 (11.9) 169 (24.5) 125 (6.5)

Cardiac arrest on 68 (2.5) 49 (5.5) 63 (9.1) 51 (2.6) <.001 0.005 0.778 admission, n (%)

GRACE score, n (%)* <.001 <.001 0.396 <1% 860 (31.8) 174 (19.4) 67 (9.7) 577 (30.0) 1–<3% 1,148 (42.5) 343 (38.2) 207 (30.0) 846 (43.9) ≥3% 694 (25.7) 380 (42.4) 415 (60.2) 501 (26.0)

Type of ACS, n (%) <.001 0.522 <0.001 Unstable angina 394 (14.6) 75 (8.4) 47 (6.8) 533 (27.7) NSTEMI 1,680 (62.1) 440 (49.1) 344 (49.9) 1,044 (54.2) STEMI 630 (23.3) 382 (42.6) 298 (43.3) 349 (18.1)

LDL (mmol/L), mean±SD* 2.59±1.28 2.68±1.37 2.36±1.30 2.46±1.38 0.066 <.001 0.002

HDL (mmol/L), mean±SD* 1.14±0.51 1.14±0.56 1.06±0.55 1.11±0.57 0.933 0.007 0.127

Coronary angiogram <.001 <.001 <.001 findings, n (%) No significant CAD 414 (15.3) 79 (8.8) 51 (7.4) 282 (14.6) 1–2 vessel disease 1460 (54.0) 541 (60.3) 359 (52.1) 1223 (63.5) 3 vessel disease and/or 830 (30.7) 277 (30.9) 279 (40.5) 421 (21.9) LM>50%

EF=ejection fraction; pEF=preserved ejection fraction; mrEF=mid-range ejection fraction; rEF=reduced ejection fraction; NZDep=New Zealand Index of Deprivation 2013; CVD=cardiovascular disease; MI=myocardial infarction; COPD=chronic obstructive pulmonary disease; BMI=body mass index; eGFR=glomerular filtration rate estimated by the Chronic Kidney Disease Epidemiology Collaboration equation; CHF=congestive heart failure; HR=heart rate; SBP=systolic blood pressure; STEMI=ST elevation myocardial infarction; NSTEMI=non-ST elevation myocardial infarction; LDL=low-density lipoprotein; HDL=high-density lipoprotein; CAD=coronary artery disease; LM=left main; SD=standard deviation. *Contains missing data, reported in Appendix, Table A1.

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 61 www.nzma.org.nz/journal article

A higher Killip Class, GRACE score and group and 105 (5.5%) deaths in the no EF proportion of patients having a cardiac group. There were 23.3 (95% CI 18.9–28.4) arrest on admission were found in the rEF deaths/1,000-patient years in those with pEF, group compared to the pEF group. The 50.6 (95% CI 39.3–64.2) deaths/1,000-patient mrEF group were intermediate in regard years in those with mrEF and 114.1 (95% to these variables, with higher propor- CI 93.9–137.4) deaths/1,000-patient years in tions compared to the pEF group but lower those with rEF. compared to the rEF group. The proportion The unadjusted hazard ratio for death of patients with STEMI were similar was 2.1 in those with mrEF and 4.8 among between the mrEF and rEF groups (43% vs those with rEF, compared to those with pEF. 43%) but were significantly lower in the pEF After adjusting for covariates (age, gender, group (23%). Obstructive coronary artery modified Charlson score, prior MI, prior disease on coronary angiography was more HF, diabetes, GRACE score, coronary angi- prevalent among those with mrEF compared ography findings, revascularisation and to pEF (91% vs 85%). Three vessel and/or left evidence-based medications on discharge), main disease was more common in the rEF mrEF and rEF were still associated with a group (41%) compared to those with mrEF respective 1.6 and 2.6 times higher all-cause (31%) and pEF (31%). mortality rate. The independent predictors In-hospital management and of all-cause mortality included age, modified outcomes (Table 2) Charlson score, diabetes, GRACE score and The proportion of patients who received lack of dispensing of aspirin, second anti- revascularisation was highest in the mrEF platelet, statin, beta-blocker, ACEi/ARB and group (78%) and the same in the pEF and anticoagulation on discharge (see Appendix, rEF groups (70%). This difference was Table A2)—however, the model with all largely driven by higher proportion of predictors was underpowered to show all patients treated with PCI in the mrEF group. statistically significant effects. Length of stay was significantly longer for Mid-range EF and rEF were associated those with rEF. with a higher risk of all-cause mortality Beta-blocker dispensing on discharge in both ACS subtypes: NSTE-ACS (adjusted were similar among those with mrEF and HR 1.5 and 2.6 respectively) and STEMI rEF (90% and 92% respectively) and lower (adjusted HR 2.7 and 3.6 respectively). among those with pEF (83%). ACEi/ARB Increased mortality in those with no dispensing increased with declining EF (pEF EF assessment was only seen in those 68%, mrEF 81%, rEF 88%). The proportion presenting with STEMI (see Appendix, of patients who were dispensed aspirin Table A5). The association between mrEF was similar among those who had LVEF and rEF and adverse outcome was found measured. for both cardiac and non-cardiac mortality, and when only 30-day survivors were A reduction in LVEF was associated with included in the analysis (see Appendix, progressively higher in-hospital mortality; Table A6–A7). 0.1% of patients with pEF, 1.2% of patients with mrEF and 2.9% of patients with rEF Both mrEF and rEF were associated with died in hospital. Those who did not have EF an increased risk of subsequent hospital- assessed during their index admission had a isation with heart failure compared to those relatively high in-hospital mortality of 2.2%. with pEF with unadjusted hazard ratios of 1.5 and 4.9 respectively. After adjusting Longer-term outcomes (Table 3 and for covariates, rEF was still associated Figures 1–3) with a 2.3 times higher HF hospitalisation Mean follow-up was 1.5 years. In the rate, whereas mrEF was associated with entire cohort, 93 (3%) patients with pEF a higher, although not statistically signif- died, 64 (7%) patients with mrEF died, 106 icant, risk of HF hospitalisation (HR 1.30, (15%) with rEF died and 127 (7%) patients 95% CI 0.94–1.80). Similar findings were with no EF assessed died. At one-year seen with the composite outcome of death follow-up, there were 70 (2.6%) deaths or non-fatal CVD rehospitalisation with an in the pEF group, 51 (5.7%) deaths in the adjusted HR of 1.24 and 1.75 for mrEF and mrEF group, 87 (12.6%) deaths in the rEF rEF respectively.

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 62 www.nzma.org.nz/journal article

Table 2: In-hospital management and outcomes.

Preserved EF Mid-range EF Reduced EF No EF pEF vs mrEF mrEF vs rEF pEF vs no EF (n=2,704) (n=897) (n=689) (n=1926) P-value P-value P-value Management

Total revascularisation, 1,896 (70.1) 696 (77.6) 483 (70.1) 1,323 (68.7) <.001 <.001 0.299 n (%)

PCI, n (%) 1,394 (51.6) 553 (61.6) 379 (55.0) 1,261 (65.5) <.001 0.008 <.001

Referred for CABG, n (%) 528 (19.5) 153 (17.1) 110 (16.0) 73 (3.8) 0.102 0.562 <.001

Reperfusion therapy 0.529 0.026 <.001 for STEMI, n (%) Primary PCI 355 (56.4) 220 (57.6) 171 (57.4) 146 (41.8) Thrombolysis 140 (22.2) 91 (23.8) 51 (17.1) 139 (39.8) None 135 (21.4) 71 (18.6) 76 (25.5) 64 (18.3)

Length of stay (days), 4 (3–8) 4 (4–8) 6 (4–13) 4 (2–5) 0.640 <.001 <.001 median (IQR)

In-hospital events

In-hospital death, n (%) 3 (0.1) 11 (1.2) 20 (2.9) 43 (2.2) <.001 0.017 <.001

Recurrent MI, n (%) 20 (0.7) 7 (0.8) 14 (2.0) 15 (0.8) 0.903 0.031 0.879

Worst Killip Class, n (%) <.001 <.001 0.003 I 2,527 (93.5) 750 (83.6) 428 (62.1) 1,755 (91.1) II–IV 177 (6.5) 147 (16.4) 261 (37.9) 171 (8.9)

Medication dispensed on discharge

Aspirin, n (%) 2,533 (93.8) 840 (94.8) 625 (93.4) 1,717 (91.2) 0.262 0.247 0.001

Second antiplatelet, n (%) 2,106 (78.0) 726 (81.9) 525 (78.5) 1,575 (83.6) 0.012 0.088 <.001

Statin, n (%) 2,508 (92.9) 833 (94.0) 618 (92.4) 1,722 (91.5) 0.234 0.200 0.080

Beta-blocker, n (%) 2,239 (82.9) 797 (90.0) 618 (92.4) 1,555 (82.6) <.001 0.099 0.782

ACEi/ARB, n (%) 1,833 (67.9) 713 (80.5) 589 (88.0) 1,300 (69.0) <.001 <.001 0.400

Spironolactone, n (%) 73 (2.7) 44 (5.0) 144 (21.5) 55 (2.9) 0.001 <.001 0.659

Anticoagulant, n (%) 236 (8.7) 108 (12.2) 116 (17.3) 174 (9.2) 0.003 0.004 0.557

EF=ejection fraction; pEF=preserved ejection fraction; mrEF=mid-range ejection fraction; rEF=reduced ejection fraction; PCI=percutaneous coronary intervention; CABG=coronary artery bypass grafting; STEMI=ST elevation myocardial infarction; MI=myocardial infarction; ACEi=angiotensin converting enzyme inhibitor; ARB=angiotensin receptor blocker; IQR=interquartile range. *Contains missing data, reported in Appendix, Table A1.

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 63 www.nzma.org.nz/journal article

other literature is difficult as older ACS Discussion studies did not report on echocardiography In a contemporary nationwide cohort findings and used Killip Class as a marker of patients presenting with ACS, both rEF for heart failure.19 More recent studies and mrEF were associated with a worse reporting on LVEF in patients with ACS prognosis compared to those with pEF. have largely focused on STEMI patients.20,21 Patients with mrEF had intermediate Other studies have reported different LVEF outcomes between rEF and pEF. One in cut-offs, which also makes comparisons five patients had mrEF, and mrEF was difficult. In a single centre cohort of STEMI associated with a 1.6-fold increased risk patients who underwent both primary PCI of death compared to those with pEF and echocardiography during their index after adjustment for covariates. One in admission, 48% had pEF, 41% had mrEF and six patients had rEF, which was asso- 10% had rEF.20 Among patients with existing ciated with a three-fold increased risk of HF in cross-sectional registries or clinical death and repeat hospitalisation for heart trials, the prevalence of heart failure with failure compared to those with pEF after mrEF is 13–24%.22 adjustment for covariates. Outcomes Prevalence of mrEF following ACS This study adds to the growing liter- This study applied the LVEF categories ature base confirming patients with mrEF introduced by the recent ESC heart failure post-ACS are a prognostically distinct group guidelines to patients with all types of ACS. from those with pEF or rEF. One previous In this cohort of patients presenting with study has reported outcomes in patients ACS who had an assessment of LVEF, 63% presenting with STEMI and mrEF.20 This had preserved EF, 21% had mid-range EF study reported that patients with mrEF and 16% had reduced EF. Comparison to and STEMI had a similar 30-day mortality

Table 3: Unadjusted and adjusted risk of outcomes related to ejection fraction.

N All-cause mortality Heart failure hospitalisation Death or non-fatal CVD hospitalisation

# events HR (95% CI) P-value # events HR (95% CI) P-value # events HR (95% CI) P-value Unadjusted Preserved EF 2704 93 1.00 - 115 1.00 - 383 1.00 - Mid-range EF 897 64 2.14 (1.56–2.94) <.001 55 1.50 (1.09–2.07) 0.013 170 1.40 (1.17–1.68) <.001 Reduced EF 689 106 4.79 (3.62–6.32) <.001 124 4.86 (3.77–6.27) <.001 239 2.78 (2.37–3.27) <.001 No EF 1,926 127 1.96 (1.50–2.56) <.001 97 1.22 (0.93–1.60) 0.147 349 1.31 (1.14–1.52) <.001

Model A Preserved EF 1.00 - 1.00 - 1.00 Mid-range EF 2.04 (1.49–2.81) <.001 1.47 (1.07–2.03) 0.018 1.37 (1.15–1.65) 0.001 Reduced EF 4.09 (3.09–5.42) <.001 4.37 (3.39–5.65) <.001 2.55 (2.17–3.00) <.001 No EF 1.83 (1.40–2.39) <.001 1.14 (0.87–1.49) 0.353 1.25 (1.08–1.45) 0.003

Model B Preserved EF 1.00 - 1.00 - 1.00 - Mid-range EF 1.55 (1.12–2.15) 0.009 1.30 (0.94–1.80) 0.115 1.24 (1.03–1.49) 0.024 Reduced EF 2.57 (1.89–3.48) <.001 2.32 (1.75–3.08) <.001 1.75 (1.46–2.09) <.001 No EF 1.63 (1.23–2.15) 0.001 1.01 (0.77–1.34) 0.928 1.17 (1.01–1.36) 0.041

HR=hazard ratio; EF=ejection fraction; CVD=cardiovascular disease. Model A: Adjusted for age and gender. Model B: Model A + prior myocardial infarction, prior congestive heart failure, diabetes (ANZACS-QI), modified Charlson Comorbidity score, GRACE score, type of acute coronary syndrome, coronary angiogram findings, percutaneous coronary intervention, coronary artery bypass grafting, dispensed meds (aspirin, second antiplatelet, statin, angiotensin converting enzyme inhibitor/angiotensin receptor blocker, beta-blocker, spironolactone, anticoagulant).

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 64 www.nzma.org.nz/journal article

Figure 1: Kaplan–Meier curve for all-cause mortality.

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 65 www.nzma.org.nz/journal article

Figure 2: Kaplan–Meier curve for heart failure readmission.

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 66 www.nzma.org.nz/journal article

Figure 3: Kaplan–Meier curve for composite outcome of all-cause mortality and non-fatal cardiovascular disease rehospitalisation.

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 67 www.nzma.org.nz/journal article

(2% vs 1%) but higher long-term mortality Medical therapy for mrEF (10% vs 7%) compared to those with pEF. The patients with mrEF comprise 21% of In an analysis of patient-level data from the cohort with known LVEF and contribute 23 five randomised controlled trials, LVEF to 24% (39/161) of the overall cardiac of 40–49% in the context of PCI for ACS mortality in that cohort. Cardiovascular was associated with increased all-cause death accounts for 61% of total deaths, mortality over five years (HR 1.54). Other and these events are potentially modi- studies have also demonstrated increased fiable with medical therapy. In this study mortality and morbidity with declining of ACS patients, beta-blocker dispensing LVEF post-ACS, but they have not specifi- was similarly high in the mrEF and rEF cally reported outcomes for an EF 40–49% groups at 90% and 92% respectively. ACEi/ group. In patients 65 years or older who ARB dispensing was slightly lower in the presented with an MI, those who had an mrEF group compared to rEF: 81% and 88% EF≤35% had an increased risk of one-year respectively. These dispensing rates are mortality, all-cause readmission and heart similar to dispensing data reported by other failure readmission (adjusted HR 1.58, 1.20 contemporary ACS and MI cohorts.25,26 and 2.43 respectively) compared to the The benefit of beta-blockers,2 ACEi/ reference group with an EF≥55%.1 Those ARB3 and mineralocorticoid antagonists4 with EF of 35–45% and 45–55% had inter- following an ACS has previously been mediate outcomes in that study. demonstrated in those with rEF following In contrast, the prognostic significance of ACS. In a meta-analysis of patients with mid-range ejection fraction in patients with coronary artery disease and without HF heart failure is less clear with conflicting or LV systolic dysfunction, ACEi have results from registry studies. In the ESC been shown to reduce all-cause mortality, Heart Failure Long-Term Registry of ambu- non-fatal MI, stroke and heart failure.27 latory HF patients, one-year mortality in The role of beta-blockers in patients with patients with HFrEF, HFmrEF and HFpEF LVEF>40% after ACS is less clear, given a 8 was 8.8%, 7.6% and 6.3% respectively. lack of recent randomised controlled trials However, in the Get With The Guide- and conflicting observational studies.28–30 lines—Heart Failure registry of patients Prescription of ACEi/ARB, particularly in hospitalised with HF, mortality at five years the mrEF group, could be improved in 24 was similar across EF categories. A recent this cohort, based on the above published international, multi-ethnic prospective evidence. cohort study of patients with HF has shown There are no studies to our knowledge that patients with HFmrEF had a risk of demonstrating improvement in clinical all-cause mortality similar to those with outcomes with medical therapy in patients HFpEF and lower than that for patients specifically with mrEF following ACS. with HFrEF, while the risk of the composite Recent data from heart failure studies have outcome of death or HF readmission was suggested that ARBs and beta-blockers are intermediate between the HFpEF and HFrEF beneficial for patients with HFmrEF. The groups.10 CHARM-Preserve trial found candesartan Observed outcomes in patients with reduced the primary outcome of death and HFmrEF and HFrEF do not necessarily HF hospitalisation in patients with HF and translate to patients with mrEF and rEF LVEF 40–50%.31 In a recent meta-analysis post-ACS. Although coronary artery disease of individual patient data from major heart is the most common underlying cause for failure RCTs comparing beta-blockers to heart failure, the pathophysiology, clinical placebo, there was a reduction in all-cause factors, management and outcomes can and cardiovascular mortality in patients vary significantly between those with acute with LVEF 40–49% who were in sinus coronary syndromes and heart failure exac- rhythm. This benefit was not seen in those erbations. Furthermore, only a minority in atrial fibrillation.32 It is unclear whether of the ACS patients with mrEF or rEF in the benefit from these treatments will our study had clinical heart failure as translate to patients with mrEF following defined by a Killip Class II–IV (16% and 38% ACS. Given the poorer clinical outcomes respectively). seen in ACS patients with mrEF compared

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 68 www.nzma.org.nz/journal article

to pEF, there should be strong consideration the ANZACS-QI registry. The timing of LVEF for commencing these medications in those measurement in relation to admission and with mrEF post-ACS. revascularisation was not standardised, as Study limitations this was an observational registry study. However, this was unlikely to influence Three out of ten patients did not have the results of this study, given the median an assessment of LVEF during their index hospital length of stay was short (four days). admission with ACS, which may have intro- Retrospective chart review was not possible, duced a selection bias, however, it may as data linkage is anonymised, hence data also be representative of real-life practice. analysis is limited only to variables collected Guidelines have emphasised the need for as part of the ANZACS-QI registry during the LVEF measurement in the investigation and index hospital admission. We were unable management of patients with myocardial to identify whether patients had a repeat LV infarction.5,6 The proportion of patients who assessment following their index admission had a LV assessment in this study is compa- with ACS, and whether a change in LVEF rable to other ACS registries internationally. influenced subsequent outcomes. Doses of In the Myocardial Ischaemia National Audit medications such as ACEi/ARB and beta- Project (MINAP) registry between 2016 blockers prescribed on discharge where not and 2017, only 73% of patients with a final able to be identified from the ANZACS-QI diagnosis of STEMI had an echocardiogram registry. during their index admission.33 Patients with no LVEF measurement were likely a mixed group. This group had a rela- Conclusion tively high in-hospital mortality, suggesting One in five patients post-ACS have mrEF that some of these patients were ‘high-risk’ and they are a clinically distinct group. They ACS presentations and died prior to an LVEF have an intermediate risk of morbidity and assessment being able to be performed. The mortality compared to those with pEF and post-discharge event rates were fairly low in rEF. Compared with those with pEF, they this group, suggesting that overall they were have higher in-hospital mortality, all-cause a low-risk group where it is possible that mortality after discharge and subsequent HF the responsible clinicians felt that they did hospitalisation. There is a lack of evidence not require an inpatient LVEF assessment. for medical therapy specifically for those There was also a relatively high prevalence with mrEF following ACS, unlike for those of prior cardiovascular disease, MI and HF, with rEF. Recent studies have suggested suggesting that these patients may have had the clinical benefit of ACEi/ARB and beta- a recent LVEF assessment. blockers extends to those with HF with a LVEF up to 50%. Further study is required to Only patients with confirmed ACS who determine the optimal management regimes underwent coronary angiography were that can improve outcomes for patients with included in this study, because data from mrEF post-ACS. ACS patients who did not undergo coronary angiography are not routinely entered into

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 69 www.nzma.org.nz/journal article

Appendix

Table A1: Missing data.

Preserved EF Mid-range EF Reduced EF No EF (n=2704) (n=897) (n=689) (n=1926) Demographics

NZDep 13, n (%) 9 (0.3) 3 (0.3) 1 (0.1) 4 (0.2)

Comorbidities

BMI, n (%) 546 (20.2) 153 (17.1) 140 (20.3) 340 (17.7)

Estimated GFR, n (%) 1 (0.04) 0 (0) 0 (0) 0 (0)

Admission data

GRACE score, n (%) 2 (0.1) 0 (0) 0 (0) 2 (0.1)

LDL, n (%) 10 (0.4) 2 (0.2) 1 (0.1) 6 (0.3)

HDL, n (%) 8 (0.3) 2 (0.2) 1 (0.1) 5 (0.3)

EF=ejection fraction; NZDep=New Zealand Index of Deprivation 2013; BMI=body mass index; eGFR=estimated glomerular filtration rate; LDL=low density lipoprotein; HDL=high density lipoprotein

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 70 www.nzma.org.nz/journal article

Table A2: Multivariable associations with all-cause mortality outcome.

Outcome Model A Model B Variable HR (95% CI) HR (95% CI)

All-cause mortality Left ventricular ejection fraction Preserved EF (ref) 1.00 1.00 Mid-range EF 2.04 (1.49–2.81) 1.55 (1.12–2.15) Reduced EF 4.09 (3.09–5.42) 2.57 (1.89–3.48) No EF 1.83 (1.40–2.39) 1.63 (1.23–2.15)

Demographics Age (years) 1.06 (1.05–1.07) 1.02 (1.01–1.04) Male sex 1.26 (1.01–1.57) 1.26 (1.00–1.58)

Comorbidities Prior MI 1.23 (0.98–1.55) History of HF 1.27 (0.90–1.79) Diabetes 1.87 (1.50–2.34)

Modified Charlson Comorbidity Index 0 (ref) 1.00 1–2 1.60 (1.26–2.03) ≥3 1.84 (1.39–2.43)

GRACE score <1% (ref) 1.00 1–<3% 2.47 (1.43–4.27) ≥3% 5.65 (3.20–9.96)

Type of acute coronary syndrome USA (ref) 1.00 NSTEMI 2.46 (1.64–3.68) STEMI 2.51 (1.61–3.92)

Coronary angiogram findings No significant CAD (ref) 1.00 1–2 vessel disease 2.30 (1.49–3.54) 3 vessel disease or Left main stenosis 3.41 (2.22–5.25)

Revascularisation PCI 1.12 (0.87–1.46) CABG 0.47 (0.31–0.72)

Medications dispensed Aspirin 0.35 (0.26–0.46) Second antiplatelet 0.44 (0.33–0.60) Statin 0.39 (0.30–0.52) Beta-blocker 0.63 (0.48–0.81) ACEi/ARB 0.52 (0.41–0.66) Spironolactone 1.14 (0.76–1.71) Anticoagulant 0.69 (0.51–0.93)

HR=hazard ratio; EF=ejection fraction; MI=myocardial infarction; HF=heart failure; USA=unstable angina, NSTEMI=non-ST elevation myocardial infarction; STEMI=ST elevation myocardial infarction; PCI=percutaneous coronary intervention; CABG=coronary artery bypass grafting; ACEi=angiotensin converting enzyme inhibitor; ARB=angiotensin receptor blocker.

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 71 www.nzma.org.nz/journal article

Table A3: Multivariable associations with heart failure hospitalisation outcomes.

Outcome Model A Model B Variable HR (95% CI) HR (95% CI)

Heart failure hospitalisation Left ventricular ejection fraction Preserved EF (ref) 1.00 1.00 Mid-range EF 1.47 (1.07–2.03) 1.30 (0.94–1.80) Reduced EF 4.37 (3.39–5.65) 2.32 (1.75–3.08) No EF 1.14 (0.87–1.49) 1.01 (0.77–1.34)

Demographics Age (years) 1.06 (1.05–1.07) 1.03 (1.01–1.04) Male sex 0.87 (0.71–1.08) 0.85 (0.69–1.06)

Comorbidities Prior MI 1.27 (1.02–1.59) History of HF 2.31 (1.73–3.08) Diabetes 1.69 (1.36–2.10)

Modified Charlson Comorbidity Index 0 (ref) 1.00 1–2 1.79 (1.42–2.26) ≥3 2.15 (1.62–2.86)

GRACE score <1% (ref) 1.00 1–<3% 1.74 (1.14–2.64) ≥3% 2.99 (1.91–4.68)

Type of Acute Coronary Syndrome USA (ref) 1.00 NSTEMI 1.34 (0.98–1.85) STEMI 0.91 (0.60–1.36)

Coronary angiogram findings No significant CAD (ref) 1.00 1–2 vessel disease 0.97 (0.67–1.41) 3 vessel disease or ;eft main stenosis 1.32 (0.92–1.90)

Revascularisation PCI 0.79 (0.61–1.02) CABG 0.64 (0.43–0.94)

Medications dispensed Aspirin 0.77 (0.54–1.09) Second antiplatelet 0.93 (0.69–1.27) Statin 0.93 (0.67–1.30) Beta-blocker 1.12 (0.82–1.52) ACEi/ARB 1.11 (0.86–1.44) Spironolactone 1.96 (1.47–2.63) Anticoagulant 1.57 (1.21–2.03)

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 72 www.nzma.org.nz/journal article

Table A4: Multivariable associations with death or non-fatal hospitalisation with cardiovascular disease outcome.

Outcome Model A Model B Variable HR (95% CI) HR (95% CI)

Death or non-fatal hospitalisation with cardiovascular disease Left ventricular ejection fraction Preserved EF (ref) 1.00 1.00 Mid-range EF 1.37 (1.15–1.65) 1.24 (1.03–1.49) Reduced EF 2.55 (2.17–3.00) 1.75 (1.46–2.09) No EF 1.25 (1.08–1.45) 1.17 (1.01–1.36)

Demographics Age (years) 1.04 (1.03–1.04) 1.02 (1.01–1.02) Male sex 1.02 (0.90–1.16) 1.03 (0.90–1.17)

Comorbidities Prior MI 1.08 (0.94–1.24) History of HF 1.56 (1.26–1.93) Diabetes 1.50 (1.31–1.71)

Modified Charlson Comorbidity Index 0 (ref) 1.00 1–2 1.55 (1.35–1.78) ≥3 1.73 (1.45–2.07)

GRACE score <1% (ref) 1.00 1–<3% 1.17 (0.96–1.42) ≥3% 1.55 (1.24–1.95)

Type of Acute Coronary Syndrome USA (ref) 1.00 NSTEMI 1.67 (1.37–2.02) STEMI 1.65 (1.31–2.09)

Coronary angiogram findings No Significant CAD (ref) 1.00 1–2 vessel disease 1.89 (1.48–2.41) 3 vessel disease or left main stenosis 2.73 (2.13–3.49)

Revascularisation PCI 0.90 (0.77–1.05) CABG 0.51 (0.40–0.65)

Medications dispensed Aspirin 0.55 (0.46–0.67) Second antiplatelet 0.83 (0.70–1.00) Statin 0.57 (0.48–0.68) Beta-blocker 0.72 (0.62–0.84) ACEi/ARB 0.80 (0.70–0.92) Spironolactone 1.47 (1.19–1.83) Anticoagulant 1.14 (0.96–1.36)

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 73 www.nzma.org.nz/journal article

Table A5: Unadjusted and adjusted risk of all-cause mortality outcome, related to ejection fraction and type of acute coronary syndrome.

NSTE-ACS STEMI

HR (95% CI) P-value HR (95% CI) P-value Unadjusted Preserved EF 1.00 - 1.00 - Mid-range EF 1.86 (1.25–2.78) 0.002 3.26 (1.75–6.08) <0.001 Reduced EF 4.74 (4.31–6.60) <0.001 6.11 (3.39–11.01) <0.001 No EF 1.25 (0.91–1.71) 0.178 7.00 (3.95–12.40) <0.001

Model A Preserved EF 1.00 - 1.00 - Mid-range EF 1.66 (1.12–2.48) 0.013 3.24 (1.74–6.05) <0.001 Reduced EF 4.11 (2.95–5.73) <0.001 4.88 (2.70–8.83) <0.001 No EF 1.16 (0.84–1.60) 0.152 6.57 (3/71–11.65) <0.001

Model B Preserved EF 1.00 - 1.00 - Mid-range EF 1.54 (1.02–2.32) 0.038 2.72 (1.43–5.18) <.001 Reduced EF 2.63 (1.81–3.80) <.001 3.62 (1.97–6.64) <.001 No EF 1.01 (0.73–1.41) 0.939 4.16 (2.29–7.57) <.001

HR=hazard ratio; EF=ejection fraction, NSTE-ACS=non-ST elevation acute coronary syndrome, STEMI=ST-elevation myocardial infarction. Model A: Adjusted for age and gender. Model B - Model A + prior myocardial infarction, prior congestive heart failure, diabetes (ANZACS-QI), modified Charlson Comorbidity score, GRACE score, type of acute coronary syndrome, coronary angiogram findings, percutaneous coronary intervention, coronary artery bypass grafting, dispensed meds (aspirin, second antiplatelet, statin, angiotensin converting enzyme inhibitor/angiotensin receptor blocker, beta-blocker, spironolactone, anticoagulant).

Table A6: Unadjusted and adjusted risk of cardiac and non-cardiac mortality, related to ejection fraction.

Cardiac mortality Non-cardiac mortality

# events HR (95% CI) P-value # events HR (95% CI) P-value Unadjusted Preserved EF 52 1.00 - 41 100 - Mid-range EF 39 2.32 (1.53–3.51) <0.001 25 1.92 (1.17–3.16) 0.010 Reduced EF 70 5.60 (3.91–8.02) <0.001 36 3.76 (2.40 - 5.88) <0.001 No EF 98 2.70 (1.93–3.79) <0.001 29 1.02 (0.63–1.64) 0.945

Model A Preserved EF 1.00 - 1.00 - Mid-range EF 2.20 (1.45–3.33) <0.001 1.85 (1.13–3.05) 0.015 Reduced EF 4.74 (3.30–6.80) <0.001 3.28 (2.09–5.15) <0.001 No EF 2.51 (1.79–3.52) <0.001 0.95 (0.59–1.53) 0.834

Model B Preserved EF 1.00 - 1.00 - Mid-range EF 1.55 (1.01–2.37) 0.045 1.64 (0.99–2.73) 0.057 Reduced EF 2.75 (1.86–4.05) <.001 2.43 (1.48–3.97) <.001 No EF 2.01 (1.42–2.84) <.001 0.80 (0.49–1.31) 0.368

HR=hazard ratio; EF=ejection fraction. Model A: Adjusted for age and gender. Model B - Model A + prior myocardial infarction, prior congestive heart failure, diabetes (ANZACS-QI), modified Charlson Comorbidity score, GRACE score, type of acute coronary syndrome, coronary angiogram findings, percutaneous coronary intervention, coronary artery bypass grafting, dispensed meds (aspirin, second antiplatelet, statin, angiotensin converting enzyme inhibitor/angiotensin receptor blocker, beta-blocker, spironolactone, anticoagulant).

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 74 www.nzma.org.nz/journal article

Table A7: Unadjusted and adjusted risk of all-cause mortality in 30-day survivors, related to ejection fraction.

N All-cause mortality of those alive at least 30 days

# events HR (95% CI) P-value Unadjusted Preserved EF 2,686 77 1.00 - Mid-range EF 883 50 2.04 (1.43–2.92) <0.001 Reduced EF 655 72 4.01 (2.91–5.53) <0.001 No EF 1,871 74 1.38 (1.01–1.90) 0.046

Model A Preserved EF 1.00 - Mid-range EF 1.94 (1.36–2.77) <0.001 Reduced EF 3.38 (2.45–4.67) <0.001 No EF 1.28 (0.93–1.76) 0.133

Model B Preserved EF 1.00 - Mid-range EF 1.67 (1.16–2.40) 0.006 Reduced EF 2.19 (1.54–3.12) <.001 No EF 1.14 (0.82–1.58) 0.446

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 75 www.nzma.org.nz/journal article

Competing interests: Nil. Acknowledgements: ANZACS-QI programme implementation, coordination and analysis: The ANZACS-QI software was developed and supported by Enigma Solutions. Programme implementation is coordinated by the National Institute for Health Innovation (NIHI) at the University of Auck- land. The ANZACS-QI programme is funded by the New Zealand Ministry of Health. ANZACS-QI Governance group: Andrew Kerr (chair), Chris Nunn, David Smyth, Gary Sutcliffe, Gerry Devlin, Harvey White, Jonathon Tisch, Sue Riddell, Kim Marshall, Michael Williams, Nick Fisher, Paul Bridgeman, Peter Larsen, Tony Scott, Mayanna Lund. ANZACS-QI Project management: Kristin Sutherland (Project Manager), Charmaine Flynn (Northern coordinator), Maxine Rhodes (Southern coordinator), Anna-Marie Rattray (Research Assistant). Data management and analysis: Mildred Lee, Michelle Jenkins, John Faatui. We acknowledge all the New Zealand cardiologists, physicians, nursing staff and radiographers who have supported and contributed to ANZACS-QI. ANZACS-QI hospital coordinators: Ascot Angiography: Summerscales, I. Money, J. Ashburton: Wilson, S. Auckland Hospital: Belz, L. Stewart, R. Marshall, K. Bay of Islands: Cochran, G. Christchurch Hospital: Jackson, M. Clutha Hospital: Reed, G. Campbell, B. Dargaville: Cripps, J. Katipa, K. Dunedin Hospital: Foote, C. Glenie, T. Dunstan: Nixon, G. Shaw, M. Gisborne: Low, T. Gore: Lindley, G. Grey Base Hospital: Smith, L. Hawke’s Bay Hospital Soldiers Memorial: Brown, G. Grant, P. Hutt Hospital: Pinfold, S. Ferrier, K. Kitchen, R. Kaikoura Hospital: McNabb, A. Kaitaia Hospital: Thompson, R. Smith, N. Lakes District Hospital: Burt, J. Mercy Angiography: Shah, A. Ubod, B. Mercy Heart Centre: Hall, S. Middlemore: McLachlan, A. Midland Cardiovascular Services: Phillips, K. Nelson Hospital: Besley, J. Abernethy, H. North Shore Hospital: Gray, L. Oamaru: Gonzales, R. Palmerston North Hospital: Kinloch, D. Rawene Hospital: Dorsay, C. Rotorua Hospital: Colby, C. Invercargill Hospital: Byers, R. St Georges Hospital: Lissette, J. Taranaki Base Hospital: Ternouth, I. Spurway, M. Taumaranui: Pointon, L. Taupo Hospital: McAnanay, J. Tauranga Hospital: Goodson, J. Te Kuiti: Te Wano, T. Thames: Stutchbury, D. Timaru Hospital: Addidle, D. Tokorao: Huitema, V. Waikato Hospital: Emerson, C. Pilay, R. Wairarapa Hospital: Matthews, T. Wairau Hospital: Langford, S. Waitakere Hospital: Long, L. Waitemata Hospital: Newcombe, R. Wakefield Private Hospital: Murphy, S. Wellington Hospital: Scott, B. Whaktane Hospital: Bentley-Smith, M. Whanganui Hospital: Thompson, T. Whangarei Hospital: Vallancey, S. Author information: Daniel Chan: Department of Cardiology, Counties Manukau District Health Board, Auckland, New Zealand. Robert N Doughty: Department of Medicine, University of Auckland, Auckland, New Zealand; Greenlane Cardiovascular Service, Auckland City Hospital, Auckland, New Zealand. Mayanna Lund: Department of Cardiology, Counties Manukau District Health Board, Auckland, New Zealand. Mildred Lee: Department of Cardiology, Counties Manukau District Health Board, Auckland, New Zealand. Katrina Poppe: School of Population Health, University of Auckland, Auckland, New Zealand. Andrew J Kerr: Department of Cardiology, Counties Manukau District Health Board, Auckland, New Zealand; School of Population Health, University of Auckland, Auckland, New Zealand. Corresponding author: Daniel Chan, c/o Department of Cardiology, Middlemore Hospital, Otahuhu, Auckland 93311, New Zealand [email protected] URL: www.nzma.org.nz/journal-articles/prognostic-significance-of-mid-range-ejection-fraction-following-acute-coronary-syndrome-anzacs-qi-23

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 76 www.nzma.org.nz/journal article

REFERENCES 1. Sutton NR, Li S, Thomas L, for the management of study. Eur Heart J 2018; 39: et al. The association of left acute coronary syndromes 1770-1780. 2018/02/02. DOI: ventricular ejection frac- in patients presenting 10.1093/eurheartj/ehy005. tion with clinical outcomes without persistent ST-seg- 11. Kerr A, Williams MJ, after myocardial infarction: ment elevation: Task Force White H, et al. The All Findings from the Acute for the Management of New Zealand Acute Coronary Treatment and Acute Coronary Syndromes Coronary Syndrome Intervention Outcomes in Patients Presenting Quality Improvement Network (ACTION) Regis- without Persistent Programme: Implemen- try-Get With the Guidelines ST-Segment Elevation of tation, Methodology and (GWTG) Medicare-linked the European Society of Cohorts (ANZACS-QI 9). N database. Am Heart J Cardiology (ESC). Eur Heart Z Med J 2016; 129: 23-36. 2016; 178: 65-73. DOI: J 2016; 37: 267-315. DOI: 12. Thygesen K, Alpert JS, 10.1016/j.ahj.2016.05.003. 10.1093/eurheartj/ehv320. Jaffe AS, et al. Third 2. Dargie HJ. Effect of 7. Ponikowski P, Voors AA, universal definition of carvedilol on outcome Anker SD, et al. 2016 ESC myocardial infarction. after myocardial infarc- Guidelines for the diagnosis Circulation 2012; 126: tion in patients with and treatment of acute 2020-2035. DOI: 10.1161/ left-ventricular dysfunc- and chronic heart failure: CIR.0b013e31826e1058. tion: the CAPRICORN The Task Force for the 13. Anderson JL, Adams CD, randomised trial. Lancet diagnosis and treatment Antman EM, et al. ACC/AHA 2001; 357: 1385-1390. of acute and chronic heart 2007 guidelines for the failure of the European 3. Pfeffer MA, McMurray management of patients Society of Cardiology (ESC) JJ, Velazquez EJ, et al. with unstable angina/non Developed with the special Valsartan, captopril, ST-elevation myocardial contribution of the Heart or both in myocardial infarction: a report of Failure Association (HFA) infarction complicated by the American College of of the ESC. Eur Heart J heart failure, left ventric- Cardiology/American Heart 2016; 37: 2129-2200. DOI: ular dysfunction, or Association Task Force 10.1093/eurheartj/ehw128. both. N Engl J Med 2003; on Practice Guidelines 349: 1893-1906. DOI: 8. Chioncel O, Lainscak (Writing Committee to 10.1056/NEJMoa032292. M, Seferovic PM, et al. Revise the 2002 Guidelines 4. Pitt B, Remme W, Zannad Epidemiology and one-year for the Management of F, et al. Eplerenone, a outcomes in patients with Patients With Unstable selective aldosterone chronic heart failure and Angina/Non ST-Elevation blocker, in patients with preserved, mid-range and Myocardial Infarction): left ventricular dysfunc- reduced ejection fraction: developed in collaboration tion after myocardial an analysis of the ESC with the American College infarction. N Engl J Med Heart Failure Long-Term of Emergency Physicians, 2003; 348: 1309-1321. DOI: Registry. Eur J Heart the Society for Cardio- 10.1056/NEJMoa030207. Fail 2017; 19: 1574-1585. vascular Angiography DOI: 10.1002/ejhf.813. 5. Ibanez B, James S, Agewall and Interventions, and S, et al. 2017 ESC Guidelines 9. Koh AS, Tay WT, Teng the Society of Thoracic for the management of THK, et al. A comprehen- Surgeons: endorsed by the acute myocardial infarction sive population-based American Association of in patients presenting characterization of heart Cardiovascular and Pulmo- with ST-segment elevation: failure with mid-range nary Rehabilitation and the The Task Force for the ejection fraction. Eur Society for Academic Emer- management of acute J Heart Fail 2017; 19: gency Medicine. Circulation myocardial infarction in 1624-1634. 2017/09/28. 2007; 116: e148-304. DOI: patients presenting with DOI: 10.1002/ejhf.945. 10.1161/CIRCULATIO- ST-segment elevation 10. Lam CSP, Gamble GD, NAHA.107.181940. of the European Society Ling LH, et al. Mortality 14. Granger CB, Goldberg RJ, of Cardiology (ESC). Eur associated with heart Dabbous O, et al. Predic- Heart J 2017. DOI: 10.1093/ failure with preserved vs. tors of hospital mortality eurheartj/ehx393. reduced ejection fraction in the global registry of 6. Roffi M, Patrono C, Collet JP, in a prospective interna- acute coronary events. et al. 2015 ESC Guidelines tional multi-ethnic cohort Arch Intern Med 2003; 163:

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 77 www.nzma.org.nz/journal article

2345-2353. DOI: 10.1001/ coronary intervention in 28. Dondo TB, Hall M, West archinte.163.19.2345. patients with left ventricu- RM, et al. beta-Blockers 15. Quan H, Sundarara- lar ejection fraction <40% and Mortality After Acute jan V, Halfon P, et al. versus >/= 40% (from the Myocardial Infarction in Coding algorithms for HORIZONS-AMI trial). Patients Without Heart defining comorbidities Am J Cardiol 2013; 111: Failure or Ventricular in ICD-9-CM and ICD-10 12-20. DOI: 10.1016/j. Dysfunction. J Am Coll administrative data. Med amjcard.2012.08.040. Cardiol 2017; 69: 2710- Care 2005; 43: 1130-1139. 22. Bhambhani V, Kizer JR, 2720. 2017/06/03. DOI: 10.1016/j.jacc.2017.03.578. 16. Atkinson J. SC, and Cramp- Lima JAC, et al. Predictors ton P. NZDep2013 Index and outcomes of heart 29. Puymirat E, Riant E, Aiss- of Deprivation. Dunedin: failure with mid-range aoui N, et al. beta blockers University of Otago ejection fraction. Eur and mortality after myocar- J Heart Fail 2017. DOI: dial infarction in patients 17. Kerr A, Exeter D, Hanham 10.1002/ejhf.1091. without heart failure: G, et al. Effect of age, multicentre prospective gender, ethnicity, socioeco- 23. Siontis GC, Branca M, cohort study. BMJ 2016; nomic status and region on Serruys P, et al. Impact of 354: i4801. 2016/09/22. dispensing of CVD second- left ventricular function on DOI: 10.1136/bmj.i4801. ary prevention medication clinical outcomes among in New Zealand: the Atlas patients with coronary 30. Montenegro Sa F, of Health Care Variation artery disease. Eur J Prev Carvalho R, Ruivo C, et al. CVD cohort (VIEW-1). N Z Cardiol 2019; 26: 1273- Beta-blockers for post- Med J 2014; 127: 39-69. 1284. 2019/04/11. DOI: acute coronary syndrome 10.1177/2047487319841939. mid-range ejection fraction: 18. Poppe KK, Doughty RN, a nationwide retrospective Wells S, et al. Devel- 24. Shah KS, Xu H, Matsouaka study. Eur Heart J Acute oping and validating RA, et al. Heart Failure Cardiovasc Care 2019; 8: a cardiovascular risk With Preserved, Border- 599-605. 2019/02/05. DOI: score for patients in the line, and Reduced Ejection 10.1177/2048872619827476. community with prior Fraction: 5-Year Outcomes. cardiovascular disease. J Am Coll Cardiol 2017. DOI: 31. Lund LH, Claggett B, Liu Heart 2017; 103: 891-892. 10.1016/j.jacc.2017.08.074. J, et al. Heart failure with 2017/02/25. DOI: 10.1136/ 25. Kueh SH, Devlin G, Lee M, mid-range ejection fraction heartjnl-2016-310668. et al. Management and in CHARM: characteristics, outcomes and effect of 19. Steg PG, Dabbous OH, Feld- Long-Term Outcome of candesartan across the man LJ, et al. Determinants Acute Coronary Syndrome entire ejection fraction and prognostic impact of Patients Presenting with spectrum. Eur J Heart heart failure complicating Heart Failure in a Contem- Fail 2018 2018/02/13. acute coronary syndromes: porary New Zealand Cohort DOI: 10.1002/ejhf.1149. observations from the (ANZACS-QI 4). Heart Lung Global Registry of Acute Circ 2016; 25: 837-846. DOI: 32. Cleland JGF, Bunting Coronary Events (GRACE). 10.1016/j.hlc.2015.10.007. KV, Flather MD, et al. Circulation 2004; 109: 26. Shah ND, Dunlay SM, Beta-blockers for heart 494-499. DOI: 10.1161/01. Ting HH, et al. Long-term failure with reduced, CIR.0000109691.16944.DA. medication adherence after mid-range, and preserved ejection fraction: an 20. Margolis G, Khoury S, myocardial infarction: individual patient-level Ben-Shoshan J, et al. experience of a community. analysis of double-blind Prognostic Implications of Am J Med 2009; 122: 961 randomized trials. Eur Mid-Range Left Ventric- e967-913. DOI: 10.1016/j. Heart J 2018; 39: 26-35. DOI: ular Ejection Fraction on amjmed.2008.12.021. 10.1093/eurheartj/ehx564. Patients Presenting With 27. Dagenais GR, Pogue J, Fox ST-Segment Elevation K, et al. Angiotensin-con- 33. Myocardial Ischaemia Myocardial Infarction. verting-enzyme inhibitors National Audit Project Am J Cardiol 2017; 120: in stable vascular disease 2016/2017 Summary 186-190. DOI: 10.1016/j. without left ventricular Report, https://www. amjcard.2017.04.005. systolic dysfunction or nicor.org.uk/wp-content/uploads/2018/11/ 21. Daneault B, Genereux P, heart failure: a combined MINAP-Summary-Re- Kirtane AJ, et al. Compari- analysis of three trials. port-2016-17.pdf (2017). son of Three-year outcomes Lancet 2006; 368: 581-588. after primary percutaneous 2006/08/15. DOI: 10.1016/ S0140-6736(06)69201-5.

NZMJ 15 January 2021, Vol 134 No 1528 ISSN 1175-8716 © NZMA 78 www.nzma.org.nz/journal article

Radiological imaging of melanoma: a review to guide clinical practice in New Zealand Victoria Francis, Tvesa Sehji, Mark Barnett, Richard CW Martin

ABSTRACT The aim of this review is to propose guidelines for initial radiological staging and the follow-up imaging regime for melanoma. This will provide consistency in the access and delivery of quality melanoma care. Radiological imaging plays an important role in assessing the extent of disease, guiding individual treatment and evaluating treatment response. However, there exists limited literature addressing the optimal radiological staging and surveillance imaging regimes for melanoma. The lack of consensus on imaging for melanoma can generate inconsistency in the standard of skin cancer care provided. This review considers the appropriate imaging techniques for both initial melanoma staging and follow-up specifically in the New Zealand clinical environment. The recommendations in this article are based on evaluation of the currently available literature and consensus of feedback from consultation with a working group of New Zealand clinicians involved in providing care to patients with melanoma. The proposed guidelines are considered the standard of care, but regional practice may differ based on access to imaging technology, cost limitations and the clinical experience of healthcare professionals.

n recent decades, there has been a sub- The most widely used histopathological stantial worldwide rise in reported rates staging system of melanoma is the American Iof skin cancers, including melanoma.1 Joint Committee on Cancer (AJCC) TNM New Zealand has one of the highest rates system. The most recent AJCC staging update of melanoma in the world.2 In 2016, 2,571 is the eighth edition,5 which is used in patients with melanoma (including melano- this review. Radiological imaging plays an ma in situ (MIS)) were recorded by the New important role in the evaluation of disease, Zealand Cancer Registry (age standardised guiding treatment and assessing response. rate 36.4 per 100,000).3 Although melanoma However, the optimal imaging technique is less common than other forms of skin for staging and the appropriate regime cancer (accounting for less than 5% of skin for surveillance imaging for melanoma cancers), the mortality rate is significantly remains controversial, as the literature greater, causing the majority of skin can- assessing this is limited. The application of cer related deaths.3 New Zealand has the positron emission tomography-computed highest rate of melanoma mortality (age tomography (PET-CT) in the management of standardised rate 4.5 per 100,000) with 362 patients with melanoma is rapidly evolving. deaths in 2016.3 The high mortality rate is re- PET-CT has superior diagnostic accuracy lated to the aggressive nature of the disease, over computed tomography (CT) and is more which results in metastasis and various accurate for staging of distant metastasis.6 barriers in the effective early detection of Surveillance imaging is used to detect local melanoma.4 recurrence/metastasis of melanoma and

assess treatment response. Previously, the In those with thick melanomas (eg, T4, role of surveillance imaging was considered stage IIB and IIC disease) there are conflicting unjustified due to the limited treatment views in the literature, with currently little options available. However, with the emer- evidence to support a significant benefit gence of new and improved therapies, of initial staging with PET-CT or CT, due to surveillance imaging may have a benefit on low yield and high false positive rates.15,16 overall survival. However, in certain high-risk clinical situ- This review evaluates the current ations, such as thick T4 tumours, baseline literature and proposes guidelines for radio- PET-CT may add value with regard to logical staging and surveillance imaging for altering the proposed treatment or therapy. melanoma based on the AJCC TNM stage Therefore, in patients with high-risk stage of disease. The guidelines proposed in this II disease with thick melanomas (eg, T4, review are considered to be the standard stage IIB and IIC disease), initial staging of care. However, it is acknowledged that with PET-CT can be considered and should regional practice may vary according to be discussed at a multidisciplinary meeting the availability of imaging technology, cost (MDM). Also consider PET-CT staging if 17 limitations and clinical experience. sentinel node biopsy failed or was declined. Stage IIIA Initial staging imaging For patients with stage IIIA disease Cross-sectional body imaging (positive sentinel lymph nodes with All staging imaging investigations should low nodal tumour volume), there is ideally be completed within two weeks of little evidence to support the value of referral to radiology. Indication for radio- baseline cross-sectional imaging. In logical staging is dependent on TMN status particular, staging imaging in this group has and intended treatment. The available a high false positive rate, which may lead to literature assessing various imaging tech- inappropriate further investigation and/or 18 niques for radiological staging of melanoma interventions. However, the rate is limited. Most studies are of retrospective of relapse in this group is not negligible, design and are difficult to compare due but it may be that the volume of locore- to variability in both methodology and gional or distant metastatic disease is the patient groups assessed.7 PET-CT has below the threshold for imaging detection at 19 improved diagnostic accuracy over CT, partic- the initial diagnosis. Follow-up surveillance ularly for detection of extracerebral distant imaging should be considered at an appro- metastatic disease.6 A small retrospective priate time interval based on the study compared staging PET-CT with CT alone risk of recurrence. However, for patients and found major therapy changes based on with positive sentinel lymph nodes where the PET-CT findings, particularly with regard therapeutic lymphadenectomy is planned to surgical management.8 (which is not considered the standard of care), baseline PET-CT is recommended Stages 0, I and II instead. Routine radiological staging for asymp- Therefore, for patients with stage IIIA tomatic patients with stage 0 (MIS), I and II disease under clinical or ultrasound obser- disease is generally not recommended, due vation, baseline cross-sectional staging to low rates of true positive findings and imaging such as PET-CT is not routinely comparatively high rates of false positive indicated in asymptomatic patients. PET-CT findings.9–13 A reasonably large percentage is recommended for patients with a positive of the recurrence seen in patients with sentinel node where therapeutic lymph- stage 0 (MIS), I and II disease is local (nodal, adenectomy is planned. satellite or in-transit) and is often detected by the patient or clinician.14 Therefore, for Stages IIIB, IIIC and IIID patients with stage 0 (MIS), I or II disease, In patients with high-risk stage III disease and excluding sentinel node biopsy (where (stage IIIB, IIIC and IIID disease), baseline indicated), baseline cross-sectional imaging PET-CT detection of occult metastasis may is not routinely recommended in asymp- upstage the patient, which can have signif- tomatic patients. icant implications for further management.

A small retrospective study by Groen The risk of cerebral metastasis in stage I et al reported that 18% of patients with and II disease is low, and routine staging is stage III disease were upstaged to stage generally not recommended. For patients IV. 20 Therefore, baseline PET-CT is recom- with stage III disease, macroscopic nodal mended for patients with stage IIIB, IIIC and in-transit disease has been associated or IIID disease, as potential upstaging may with an increased risk of brain metastases.21 influence a change in treatment.13 In stage IV disease, the risk of concurrent cerebral and extracerebral metastasis at Stage IV diagnosis has been described as being Patients with stage IV disease may present present in up to a third of patients, and clinically or as an unexpected finding there is also a small group where the brain on imaging with or without a history of is the only site of metastatic disease. melanoma. For patients with stage IV disease, PET-CT is recommended if the result will It is widely accepted that contrast-en- change management. Baseline PET-CT for hanced magnetic resonance imaging (MRI) stage IV disease should be guided by the is superior to contrast-enhanced CT for MDM and recommended in certain clinical the detection of cerebral metastases and is circumstances, such as if there is oligometa- therefore preferable due to improved diag- 6 static disease demonstrated on conventional nostic accuracy. If low-dose CT is performed CT that would be amenable to surgery or as part of the PET-CT examination, it may radiotherapy, as a baseline for systemic not be of diagnostic quality for the detection therapy or if there are equivocal findings on of brain metastases, and additional diagnos- conventional CT that could potentially change tic-quality brain imaging may therefore be treatment decisions. Otherwise, contrast-en- required. hanced staging CT of the chest, abdomen and Given the prognostic implications and pelvis should be performed. Neck CT should treatment options now available, staging be added if the primary malignancy is in the high-resolution brain imaging is recom- head, neck or upper trunk. mended for patients with stage IIIB, IIIC, IIID 22 Brain imaging and IV disease. The AJCC recognises patients with central Lymph node ultrasound nervous system metastases as having the An ultrasound of the lymph node basins worst prognosis of all melanoma patients draining the primary site may be considered with distant metastatic disease (M1d in selected clinical situations. Examples category).5 The incidence of developing include high-risk stage II patients with brain metastases increases with TNM stage. equivocal clinical examination or when

Table 1: Initial staging imaging recommendations for melanoma by AJCC stage.

AJCC stage Staging imaging

Not routinely indicated in asymptomatic patients. 0, I, II Regional lymph node ultrasound in selected clinical situations.

IIB, IIC Consider baseline PET-CT for high-risk T4 tumours, following MDM discussion.

Not routinely indicated in asymptomatic patients. IIIA PET-CT is recommended if therapeutic lymphadenectomy is planned.

PET-CT and high-resolution brain imaging recommended (MRI preferable over IIIB, IIIC, IID contrast-enhanced CT).

PET-CT, if the result will change clinical management or therapy, and MDM discussion. Otherwise, baseline CT chest, abdomen and pelvis +/- neck. IV High-resolution brain imaging recommended. MRI preferable over contrast-enhanced CT.

clinical examination is limited by patient Stage IIC body habitus or when sentinel node For patients with thick melanomas (eg, T4 biopsy has failed or was declined. There tumours), baseline staging with PET-CT is is evidence that ultrasound can detect controversial due to low yield and high false lymph node metastasis with a reasonable positive rate. However, there are signif- degree of accuracy, with literature to icant relapse rates, particularly in patients support increased sensitivity of ultrasound with stage IIC disease. In a retrospective compared to clinical examination.23,24 study of pathologic stage II patients by Lee Although the sensitivity of ultrasound is et al, 46% of stage IIC patients relapsed, higher than clinical examination, it is no and of those, 52% of first relapses were substitute to sentinel node biopsy. systemic. Imaging detected relapse in 31% of these patients.27 Stage IIC patients notably Follow-up and relapsed earlier with a higher proportion of systemic metastases (especially lung and surveillance imaging brain) when compared to other stage II Cross-sectional body imaging subgroups.27 Therefore, for patients with In conjunction with intensive clinical stage IIC disease, CT chest, abdomen and follow-up, the addition of routine cross-sec- pelvis +/- neck and brain MRI or CT head tional imaging does allow earlier detection at six months and then at twelve months is of recurrent disease, but the impact on recommended. Consider annual surveillance overall survival is still unclear.25 Cross-sec- imaging in years 3–5 following diagnosis. If tional imaging follow-up should be guided there are equivocal findings on routine CT by the probability of recurrence at any surveillance, PET-CT should be considered if stage and determined by stage, symptoms, it would influence a treatment change. clinical findings and suitability for therapy. With emerging systemic therapies, Stage IIIA From the limited data available, baseline routine follow-up cross-sectional imaging staging cross-sectional imaging in patients also provides assessment of therapeutic with a positive sentinel lymph node (stage response. In particular, the apparently high IIIA with low nodal tumour volume) appears negative predictive value of PET-CT seems to be of limited benefit, with low yield and to be reasonably consistent and notably high rates of false positive tests.18,28,29 This reassuring.26 can lead to further unnecessary investi- The optimal cross-sectional imaging gations, some of which may be invasive/ (PET-CT or CT) surveillance regime for morbid. However, the rate of recurrence in high-risk melanoma remains controversial, this group is not insignificant. Although a and there is currently no international high percentage of first relapses are locore- consensus. Even in high-risk melanoma gional and often detected by the patient patients, there is no high-quality data or clinician, less frequent surveillance to support improved survival outcomes cross-sectional imaging in this group has following routine follow-up cross-sectional been shown to detect asymptomatic recur- imaging. It is generally agreed that PET-CT rence/progression.28 Therefore, for patients has superior diagnostic accuracy over with stage IIIA disease, CT chest, abdomen conventional CT.6 In those clinical settings and pelvis +/- neck at six months and then at where CT findings are equivocal or there twelve months, and then annually there- are clinical findings highly suspicious for after until three years following diagnosis, is recurrence, PET-CT results may alter the recommended. treatment course, particularly when surgery is being considered.8 There is, however, no Stages IIIB, IIIC, IIID and IV prospective data that directly compares the The approach to cross-sectional imaging two modalities with regard to the magnitude surveillance of patients with higher stage of differences in survival outcomes. III disease and stage IV disease varies widely. For example, the National Compre- Stages 0, I, IIA and IIB hensive Cancer Network suggests follow For patients with stage 0, I, IIA and IIB up PET-CT or CT every 3–12 months.13 disease, routine surveillance imaging is not Regarding salvage curative surgery, radio- recommended if the patient is asymptomatic. therapy or emerging systemic therapies,

there is an impression that treatments are and in-transit disease has been associated more effective in the setting of low tumour with an increased risk of brain metastases.21 volume, making early detection of recur- There has also been an association between rence and/or distant metastatic disease primary tumour ulceration and devel- relevant.26,30 For patients with asymp- opment of brain metastasis.32 tomatic stage IIIB, IIIC, IIID or IV disease, Previously, the poor prognosis of those more frequent cross-sectional imaging (eg, with brain metastases may have precluded 3–6 monthly in the first three years) should routine surveillance for those at risk. be considered, when the rates of recurrence However, given the prognostic implications are highest. Particularly in stage III disease, and recent advances in surgery, stereo- a substage approach to follow-up regimes tactic radiotherapy and systemic therapy, 28 may indeed be beneficial. This approach there are improved treatment outcomes acknowledges that, although the actual (particularly in the setting of smaller benefit of earlier imaging detection on tumour volume and asymptomatic lesions). survival outcomes is not yet known, there This would suggest that earlier detection are now more treatment options available. increases the treatment options available Therefore, in asymptomatic patients with to patients, although there remains little stage IIIB, IIIC, IIID and IV disease, routine evidence to directly support this. Therefore, follow-up with contrast-enhanced CT of surveillance high-resolution brain imaging chest, abdomen and pelvis +/- neck can be (brain MRI or contrast-enhanced CT head) considered at 3–6 monthly intervals for the should also be considered in high-risk first three years. Annual follow-up imaging patients (stages IIC, IIIB, IIIC, IIID or IV) in years 3–5 following diagnosis is recom- at 3–6 monthly intervals in the first three mended. If there are equivocal findings on years, with less frequent surveillance routine CT surveillance, PET-CT should be following this. considered if it would influence a treatment Lymph node ultrasound change. If there is biopsy proven local (nodal, satellite or in-transit) recurrence or surveillance Ultrasound of the draining regional oligometastatic disease, PET-CT should be lymph node basins may provide a useful considered if the patient is a candidate for adjunct to clinical examination, particularly surgery, radiotherapy or systemic therapy. when clinical examination is limited (such The PET-CT imaging request should be as obese patients), when sentinel node discussed at the MDM. biopsy (SNB) has failed or not performed For patients with stage III and IV disease when indicated or as surveillance of SNB on active treatment (systemic therapy positive nodal stations when therapeutic or radiotherapy), the follow-up imaging lymphadenectomy is not performed. schedule will be determined by the Following the results of the MSLT-II trial, oncology team, who will likely base their local surveillance with ultrasound is likely decision on individual patient symptoms to increase.33 There is evidence that ultra- and/or the need to assess treatment sound can detect lymph node metastasis response. The above schedule may be a with a reasonable degree of accuracy, with useful guide to the minimum frequency of literature to support an increased sensi- imaging. tivity of ultrasound compared to clinical Brain imaging examination.34,35 MRI is superior to contrast-enhanced CT Sonographic features suspicious for for the detection of cerebral metastases and nodal malignancy, as defined by Vassallo is preferred (particularly if there is prior et al, remain as consistent criteria in the documented metastatic brain disease). As literature for lymph node malignancy with relapse at other sites, development of and include longitudinal to transverse brain metastases generally occurs in the diameter ratio <2, echogenic central first three years.21,31 Patients with stage IIC hilum narrowed or absent (suggesting disease have been found to relapse earlier, diffuse hypoechogenicity) and concentric with a higher proportion of systemic metas- or eccentric widening of the peripheral tases compared to other stage II subgroups.27 cortex.36 Nodal size alone is not a good For stage III patients, macroscopic nodal discriminator, as small nodes may have

malignant features and benign reactive knowledge. Equivocal sonographic findings nodes may be notably enlarged. Other may need short-interval follow-up ultra- suspicious features include peripheral sound or fine needle aspiration (FNA). vascularity on colour Doppler sonography Therefore, ultrasound imaging of the 37 and intranodal necrosis. A combination draining nodal basins should be performed of more than one suspicious finding has in a select group of patients, in conjunction been shown to increase the sensitivity of with routine clinical follow-up and appro- 38 detection. The success of sonographic priate cross-sectional imaging as per nodal assessment therefore relies on the TNM stage. Ultrasound imaging should be expertise of the sonographer, who requires performed in patients with stage IB, IIA, IIB a certain level of technical skill and or IIC disease where SNB is not performed

Table 2: Surveillance imaging recommendations for melanoma by AJCC stage.

AJCC stage Surveillance imaging

Not routinely indicated in asymptomatic patients.

0, IB, IIA, IIB For patients with stage IB, IIA, IIB or IIC disease where SNB is not performed when clinically indicated, or where SNB has failed or clinical examination is limited, regional lymph node ultrasound is recommended 6 monthly for 2 years.

CT chest, abdomen and pelvis +/- neck and brain MRI or CT head 6 monthly for 3 years.

Consider annual surveillance CT imaging in years 3–5.

IIC Consider PET-CT if there are equivocal findings on CT that would influence treatment change.

If SNB was not performed, regional lymph node ultrasound is recommended 6 monthly for 2 years.

CT chest, abdomen and pelvis +/- neck at 6 months and then at 12 months and annually thereafter until 3 years from the initial diagnosis. IIIA Regional lymph node ultrasound 6 monthly for 2 years for patients with SNB positive stage III disease where therapeutic lymphadenectomy is not performed.

CT chest, abdomen and pelvis +/- neck and brain MRI or CT head 3–6 monthly for 3 years and annually in years 3–5 following diagnosis.

Consider PET-CT if there are equivocal findings on CT surveillance and if it would IIIB, IIIC, IID, IV influence a treatment change.

Consider PET-CT if there is biopsy proven local recurrence or oligometastatic disease and the patient is a candidate for surgery, radiotherapy or systemic therapy.

when clinically indicated, and in patients with SNB positive stage III disease where Conclusion therapeutic lymphadenectomy is not Although the available literature is limited performed and where SNB has failed. with regard to the effect on overall survival, Ultrasound imaging should also be radiological staging and surveillance imaging considered for patients where clinical plays an integral role in guiding decision examination is limited. The recommended making with respect to management of frequency of ultrasound imaging is six melanoma. With the emergence of new and monthly for two years. For those patients targeted therapies for melanoma, there is undergoing ultrasound surveillance and now even more need to accurately stage and who have not had SNB, baseline ultrasound restage disease. As of yet, there is no inter- is also advised. national consensus to guide the mode or interval of imaging and there is significant There may be more than one draining variability in clinical practice. The imaging nodal basin. For example, a primary recommendations outlined above are based tumour in the central torso may drain to on a review of the current literature and either axillary or inguinal stations, and clinical consultation with the aim to provide ultrasound assessment should include a pragmatic and standardised approach all relevant nodal stations. For primary to radiological imaging for patients with tumours in the head and neck, bilateral melanoma in New Zealand. neck ultrasound is advised.

Competing interests: Nil. Acknowledgements: We thank the Melanoma Working Group for their contribution. Author information: Victoria Francis: Radiologist, Department of Radiology, Waitematā District Health Board, Auckland. Tvesa Sehji: Medical student, Department of Cutaneous Oncology, Waitematā District Health Board, Auckland. Mark Barnett: Radiologist, Department of Radiology, Waitematā District Health Board, Auckland. Richard CW Martin: Cutaneous Surgical Oncologist, Department of Cutaneous Oncology, Waitematā District Health Board, Auckland. Corresponding author: Richard CW Martin, North Shore Hospital, Private Bag, Takapuna, Auckland 0740, 021671777 [email protected] URL: www.nzma.org.nz/journal-articles/radiological-imaging-of-melanoma-a-review-to-guide-clinical-practice-in-new-zealand

REFERENCES 1. Schadendorf D, Van 2016. Wellington: Ministry Cancer Staging Manual Akkooi ACJ, Berking C, et of Health; 2016. [cited 30 (8th edition). Springer al. Melanoma. The Lancet. July, 2020]. Available from: International Publishing: 2018; 15;392(10151):971-84. www.health.govt.nz American Joint Commis- 2. Liang JJ, Robinson E, 4. Tas F. Metastatic behavior sion on Cancer. 2017. Martin RC. Cutaneous in melanoma: timing, 6. Xing Y, Bronstein Y, Ross melanoma in New Zealand: pattern, survival, and MI, et al. Contemporary 2000–2004. ANZ J Surg influencing factors. J diagnostic imaging modal- 2010; 80: 312–316. Oncol. 2012;2012:647684. ities for the staging and 3. New Zealand Ministry doi:10.1155/2012/647684. surveillance of melanoma of Health. Cancer: new 5. Amin MB, Edge S, patients: a meta-analysis. registrations and deaths Greene F, et al. AJCC Journal of the National

Cancer Institute. 2011 Imaging and Radionuclide to detect distant metas- 19;103(2):129-42. Therapy. 2018;27(2):66. tases at initial staging of 7. Cancer Council Australia 13. National Comprehensive melanoma patients with Melanoma Guidelines Cancer Network. Clinical metastatic involvement Working Party. Clinical Practice Guidelines of sentinel lymph node. practice guidelines for the in Oncology (NCCN British Journal of Derma- diagnosis and management Guidelines). 2019. [cited tology. 2011;164(6):1235-40. of melanoma. Sydney: 30 July 2020]. Available 20. Groen LC, Lazarenko Cancer Council Australia. from: https://www.nccn. SV, Schreurs HW, Richir 2019. [cited 30 July 2020]. org/professionals/physi- MC. Evaluation of PET/ Available from: https://wiki. cian_gls/default.aspx CT in patients with stage cancer.org.au/australia/ 14. Swetter SM, Tsao H, III malignant cutaneous Guidelines:Melanoma Bichakjian CK, et al. melanoma. American 8. Schüle SC, Eigentler TK, Guidelines of care for the journal of nuclear medicine Garbe C, et al. Influence of management of primary and molecular imag- 18 F-FDG PET/CT on thera- cutaneous melanoma. ing. 2019;9(2):168. py management in patients Journal of the American 21. Samlowski WE, Moon with stage III/IV malignant Academy of Dermatology. J, Witter M, et al. High melanoma. European 2019;1;80(1):208-50. frequency of brain journal of nuclear medicine 15. Arrangoiz R, Papavasiliou metastases after adjuvant and molecular imaging. P, Stransky CA, et al. therapy for high‐risk 2016; 1;43(3):482-8. Preoperative FDG-PET/CT melanoma. Cancer medi- 9. Barsky M, Cherkassky L, is an important tool in the cine. 2017;6(11):2576-85 Vezeridis M, Miner TJ. management of patients 22. Vosoughi E, Lee JM, Miller The role of preoperative with thick (T4) melanoma. JR, et al. Survival and clin- positron emission tomogra- Dermatology research and ical outcomes of patients phy/computed tomography practice. 2012;1;2012. with melanoma brain (PET/CT) in patients with 16. Danielsen M, Kjaer A, metastasis in the era of high‐risk melanoma. Wu M, et al. Prediction of checkpoint inhibitors and Journal of surgical positron emission tomogra- targeted therapies. BMC oncology. 2014;109(7):726-9. phy/computed tomography cancer. 2018;1;18(1):490. 10. Bikhchandani J, Wood J, (PET/CT) positivity in 23. Bafounta ML, Beauchet Richards AT, Smith RB. No patients with high-risk A, Chagnon S, Saiag P. benefit in staging fluoro- primary melanoma. Amer- Ultrasonography or deoxyglucose‐positron ican journal of nuclear palpation for detection emission tomography in medicine and molecular of melanoma nodal clinically node‐negative imaging. 2016;6(5):277. invasion: a meta-analysis. head and neck cutaneous 17. NICE guidelines. National The lancet oncology. melanoma. Head & neck. Collaborating Centre for 2004;1;5(11):673-80. 2014;36(9):1313-6. Cancer – Melanoma: assess- 24. Machet L, Nemeth‐ 11. Orfaniotis G, Mennie JC, ment and management. Normand F, Giraudeau B, Fairbairn N, Butterworth 2015. [cited 30 July 2020]. et al. Is ultrasound lymph M. Findings of computed Available from: https:// node examination superior tomography in stage www.nice.org.uk/guidance/ to clinical examination in IIB and IIC melanoma: ng14/evidence/full-guide- melanoma follow‐up? A a six-year retrospective line-pdf-250314589 monocentre cohort study study in the South-East of 18. Holtkamp LH, Read RL, of 373 patients. British Scotland. Journal of plastic, Emmett L, et al. Futility of journal of dermatology. reconstructive & aesthetic imaging to stage melanoma 2005;152(1):66-70. surgery. 2012;1;65(9):1216-9. patients with a positive 25. Podlipnik S, Carrera C, 12. Topuz ÖV, Görtan FA, Döner sentinel lymph node. Sánchez M, et al. Perfor- ZR, et al. Usefulness of Melanoma research. mance of diagnostic tests 18F-FDG PET/CT in Cuta- 2017;1;27(5):457-62. in an intensive follow-up neous Melanoma Patients 19. Wagner T, Meyer N, protocol for patients with with Negative Sentinel Zerdoud S, et al. Fluoro- American Joint Committee Lymph Nodes and High deoxyglucose positron on Cancer (AJCC) stage Clark Levels. Molecular emission tomography fails IIB, IIC, and III localized

primary melanoma: a 30. Freeman M, Laks S. invasion: a meta-analysis. prospective cohort study. Surveillance imaging for The lancet oncology. Journal of the American metastasis in high-risk 2004;1;5(11):673-80. Academy of Dermatology. melanoma: importance 35. Machet L, Nemeth‐ 2016;1;75(3):516-24. in individualized patient Normand F, Giraudeau B, 26. Leon-Ferre RA, Kottschade care and survivorship. et al. Is ultrasound lymph LA, Block MS, et al. Asso- Melanoma manage- node examination superior ciation between the use of ment. 2019;18;6(1). to clinical examination in surveillance PET/CT and 31. Fife KM, Colman MH, melanoma follow‐up? A the detection of potentially Stevens GN, et al. Deter- monocentre cohort study salvageable occult recur- minants of outcome in of 373 patients. British rences among patients with melanoma patients with journal of dermatology. resected high-risk melano- cerebral metastases. 2005;152(1):66-70. ma. Melanoma research. Journal of Clinical Oncolo- 36. Vassallo P, Wernecke 2017;1;27(4):335-41. gy. 2004;1;22(7):1293-300. K, Roos N, Peters PE. 27. Lee AY, Droppelmann N, 32. Zakrzewski J, Geraghty LN, Differentiation of Panageas KS, et al. Patterns Rose AE, et al. Clinical vari- benign from malignant and timing of initial ables and primary tumor superficial lymphade- relapse in pathologic stage characteristics predictive of nopathy: the role of II melanoma patients. the development of mela- high-resolution US. Radiolo- Annals of surgical oncol- noma brain metastases gy. 1992;183(1):215-20. ogy. 2017;1;24(4):939-46. and post‐brain metas- 37. Ahuja AT, Ying M, Ho 28. Lewin J, Sayers L, Kee tases survival. Cancer. SY, et al. Ultrasound D, Walpole I, et al. 2011;15;117(8):1711-20. of malignant cervical Surveillance imaging 33. Faries MB, Thompson lymph nodes. Cancer with FDG-PET/CT in JF, Cochran AJ, et al. Imaging. 2008;8(1):48. the post-operative Completion dissection 38. Moehrle M, Blum A, follow-up of stage 3 mela- or observation for senti- Rassner G, Juenger M. noma. Annals of Oncology. nel-node metastasis in Lymph node metastases 2018;1;29(7):1569-74. melanoma. New England of cutaneous melanoma: 29. Scheier B, Lao CD, Kidwell Journal of Medicine. diagnosis by B-scan and KM, Redman BG. Utility 2017;8;376(23):2211-22. color Doppler sonography. of pre-operative PET/CT 34. Bafounta ML, Beauchet Journal of the American staging in sentinel lymph A, Chagnon S, Saiag P. Academy of Dermatology. node-positive melanoma. Ultrasonography or 1999;1;41(5):703-9. Journal of Clinical Oncol- palpation for detection ogy. 2015;20; 33(15). of melanoma nodal

Increasing access to contraception in New Zealand: assessing the impact of a new funding initiative Orna McGinn, Vicki Mount, Helen Fulcher

ABSTRACT AIM: This paper offers a grassroots view of the impact of a recent government initiative designed to increase access to contraception and improve health and social outcomes for women in New Zealand. METHOD: District health board and primary health organisation project leads were contacted to request information on how each region had chosen to configure contraception services under the new contract in August 2019, a month after the rollout of the initiative, and again in August 2020. In addition, feedback from individual general practitioners was sought via social media groups. RESULTS: There is significant variation in regional funding and provision of contraception services. Further, complex eligibility criteria can create unnecessary barriers to access for women. CONCLUSION: Variation in funding and access to contraception continues to be a feature of service provision in New Zealand and may have been exacerbated by the recent Ministry of Health funding initiative. This perpetuates inequity, particularly for vulnerable women. Urgent consideration should be given to a whole-of-system approach with contraception being free at the point of access for all women in New Zealand.

develop a ‘sexual and reproductive health Background: the work programme document’ of unspec- $6m ‘increasing ified actions with ‘a focus on equity and equity of access government priorities’.1 Contraception provision was seen as to contraception’ one of these priority areas, due to New initiative Zealand’s relatively high rates of abortion and unplanned pregnancy2 and variable In November 2019, New Zealand’s long- access to the most effective methods of awaited draft cross-sector Reproductive contraception, such as long-acting reversible Health Action Plan was withdrawn. This contraception (LARC).3 Cost and a dearth of plan had been in development since 2017. trained providers have contributed to this However, email correspondence from the lack of access.3 Ministry of Health (MoH) to stakeholders explained that ‘given the complexity and Funding of $4,500,000 was announced breadth of the plan, and the divergence in April 2019 and was distributed between of advice received, an agreement [was] New Zealand’s twenty district health boards not able to be reached with the sector (DHBs); the amount to each individual advisory group on the text of a final plan’. DHB varied according to the number of The Ministry of Health instead aimed to low-income women aged 15 to 44 years

residing within their area. Additionally, $1,500,000 was allocated to Family Planning Implementation of New Zealand, a non-governmental organ- the contract: where isation (NGO) with 30 clinics around the country. Further discussion of Family Plan- did the money go? ning’s approach is outside the scope of this In August 2019, the authors contacted article. DHB and PHO project leads to request information on how each region had The intention of the funding, as described chosen to configure contraception services in a press release at the time,4 was to under the new contract. The providers provide ‘better access to free or very were then contacted again in August 2020. low-cost contraception’ and increase the Eighteen DHBs supplied information ability of women ‘to manage their fertility including service specifications and details and reduce unwanted pregnancies’. The of payment schedules. Two DHBs (Auckland objectives described in the service specifi- and Waitematā) declined to supply cation from the Ministry of Health were: information. • Increase equity of access for In addition, in July 2020 feedback was low-income women and those living sought from individual general practi- in deprivation. tioners (GPs) via a private New Zealand • Reduce poor health and social GP social media group with over 1,000 outcomes for women and infants members. associated with an unplanned preg- This research has revealed variable and nancy and birth. inconsistent service provision and funding • Provide more women with support, between and within regions, raising so that they can make a decision concerns about equitable access and the about their fertility and when to have effectiveness of the initiative. Specific issues 5 children. have been identified relating to eligibility, Specifically, the MoH contract service service configuration and funding. Primary description included provision of free LARC care practitioners continue to report a lack insertion and removal for the target popu- of information on the available training lation, low-cost contraceptive consultations opportunities, and as yet there has been no and facilitation of a training programme for indication on when the national training LARC providers. scheme (to be administered by Family The target population was defined as Planning New Zealand after they were women aged 15 to 44 years who were: awarded the contract in 2019) will be • Community Services Card (CSC) available. holders • living with high levels of socioeco- Eligibility criteria: nomic deprivation, defined as New fair or fragmented? Zealand Index of Deprivation (NZDep) DHBs apply and enforce a range of eligi- quintile 5 areas bility criteria regarding access to funded • at ‘higher risk of unplanned preg- contraception, which varies widely from nancy and poor health and social region to region and even within the same outcomes’.5 city: Auckland’s three DHBs have different The term ‘higher risk’ was not defined, criteria for women accessing LARC in but it was suggested that priority be given the community. Across the country, these to those accessing mental health or alcohol criteria may include: under the care of and drug services, and maternity-service secondary mental health services; termi- users aged under 25 or of Māori or Pacific nation of pregnancy within the last five ethnicity. years; at risk of family violence; aged Services were to be ‘widely available in under 30 with more than four children; primary care settings accessible and conve- substance abuse; Māori or Pacific ethnicity; nient to the target population’,5 and the or currently living in a youth justice resi- initiative was to start in July 2019. dence. Several DHBs have chosen to

fund free contraception for all Māori and states that ‘contraceptive information Pacific women, regardless of income or and services… should not be discriminatory deprivation status, with the intention of and should aim at eliminating stereotypes increasing equity of access and reducing the and discriminatory attitudes that lead to higher rates of unintended pregnancy in forced and coercive practices’.9 Offering these populations. Only one DHB contract free contraception to targeted groups of (Northland) specifically mentioned funding women may cause them to feel that their women with a ‘chronic health condition pregnancies are of less worth to society as that could be worsened with pregnancy’ (eg, a whole, and they may feel under ‘implicit diabetes). pressure’ to choose a particular method.10 The approach to women with mental This is of particular importance when health or substance abuse issues varies by offering LARC methods, which are long region. A number of DHBs will fund contra- lasting and not under the control ception for any women who discloses a of the user to stop, and particularly where mental health or substance abuse issue, there may be cost barriers to having while in other areas only women currently the method removed, such as with an under secondary care mental health and implant or intrauterine contraceptive addiction services can access funding. device (IUCD). Southern DHB reportedly only funds LARC removal if the insertion It is perhaps unsurprising that eligibility took place within that same DHB. Reported criteria remain open to such wide inter- volumes of LARC insertions and removals pretation when the service specification obtained via the Official Information Act identified only women‘ at higher risk of (OIA) show that Auckland and Waitematā unplanned pregnancy and poor health DHBs have had no LARC removals since and social outcomes’,5 without specifying the inception of the contract, while smaller what this would entail within the context DHBs, such as Lakes and Canterbury, of a robust national reproductive health report 324 and 377 LARC removals respec- framework. Many health conditions and tively over the same period of time.11 This their treatments place women at risk of raises the possibility that significant access an adverse outcome in pregnancy. These barriers exist across Auckland and Wait- include: women taking teratogenic medi- ematā, potentially affecting up to 25% of cation; women with a previous preterm New Zealand women. birth; and women with hypertension, renal disease, previous pre-eclampsia, poorly Great care must be taken during contra- controlled diabetes or obesity, all of which ception counselling to ensure that women’s are recognised as contributing to maternal cultural attitudes to pregnancy and and fetal risk.6 childbirth are understood and respected. Unconscious bias can be introduced There appears to be no consideration of when women of specific ethnicities are women at risk due to high body mass index targeted for contraception programmes, (BMI), despite the Royal Australian and New and the possibility of coercion, whether Zealand College of Obstetrics and Gynae- real or perceived, must be avoided. cology’s (RANZCOG’s) 2017 statement on Māori and Pacific families tend to be pre-pregnancy counselling noting that ‘high larger than the families of other ethnicity BMI (>30) is now one of the commonest groups in New Zealand, and targeting and most important risk factors for infer- these groups of women for contraception tility and adverse pregnancy outcomes’.7 could be viewed as an attempt to limit There has also been no consideration of family size.12 women living with disabilities, who can face significant barriers in accessing effective Previous New Zealand government policy contraception.8 has been to fund LARCs for beneficiaries or the children of beneficiaries; this policy Additionally, the application of stringent was withdrawn in 2018 as it was considered eligibility criteria and the use of the term stigmatising.13 Unfortunately, the new ‘target population’ applied to women contract appears to be encouraging a similar fulfilling these criteria is of concern. approach. The World Health Organisation (WHO)

acknowledge that up to 80% of women Service structure already access contraception from their The eighteen DHBs that supplied infor- local primary care practice, via their GP or mation all chose to contract contraception practice nurse.15 It implies a lack of local services to primary care, with some also consultation and stakeholder engagement funding additional community providers about the most appropriate service configu- where they exist, such as Youth One Stop ration. There is also a risk of fragmentation Shops, iwi health providers, nurse-led of patient care unless DHB clinics are services or high-school clinics (see www. part of an integrated system with clear protectedandproud.nz). In some DHBs, all referral pathways and information sharing GP practices are eligible for funding, while between primary and secondary care. in other areas only some practices have DHB clinics may in addition have limited been offered funding. The process by which operational hours or be located centrally, practices are commissioned is reportedly requiring time off work and long journeys unclear in many areas, creating a degree of to access. This poses an additional barrier frustration and confusion among healthcare to accessing care, particularly for young practitioners. Māori and those living in rural areas or on Northland and the Bay of Plenty have low incomes.16 taken a collaborative approach, designing Contraception is a routine part of the services in partnership with local stake- continuum of women’s health services holders and seeking input from end users provided in primary care over a patient’s (women) themselves. The most striking lifetime, along with human papilloma- example of this has been the Bay of Plenty virus (HPV) vaccination, cervical screening, initiative ‘Protected & Proud’, developed preconception counselling, maternity care using ‘wāhine-centred principles that and management of menopause. Primary [inform] service delivery’ and taking care is ideally placed to offer this holistic account of specific focus-group findings.14 approach of providing local, focused, This has resulted in good availability of integrated and intergenerational care for accessible services offered by a range of women and their whānau. Contraception providers. may be addressed opportunistically during Notably, Northland is the only DHB to consultations for other matters. Lack of specifically fund postcoital IUCD insertion. support for primary care to provide contra- Effective up to five days after unprotected ception may lead to a loss of valuable LARC intercourse, postcoital IUCD is the ‘gold insertion and removal skills over time standard’ form of emergency contraception, and further restrict a woman’s choice of with a far lower failure rate than oral emer- providers. gency contraception (EC). Its availability Another largely unaddressed concern is currently highly restricted due to the is access to funded contraception post cost and lack of available practitioners abortion. In New Zealand, the rate of early outside Family Planning New Zealand medical abortion (EMA) as a proportion clinics. of all abortions has risen sharply over the Two DHBs, Auckland and Waitematā, last few years, and it now stands at around have established free community clinics 45% in Auckland, up from under 10% in for LARC insertion and removal admin- 2018.17,18 This proportion is likely to rise istered directly by the DHBs. While this further following the March 2020 change appears to be a positive initiative, local in abortion law.19 With abortion now providers have raised concerns that these decriminalised, women have the ability clinics may in fact exacerbate barriers to to self-refer, and services are being recon- contraception access, restrict choice and figured such that there will be a wider erode the quality of existing primary care choice of community providers, enabling contraceptive services over time. The impo- easier access. In the UK, the rate of EMA sition of DHB-led clinics in areas where as a proportion of all abortions is 71%,20 there is no primary care funding to provide and in Australia’s Northern Territory, contraception introduces an element of following similar legislative change, it is competition for services, and it fails to also over 70%.21

Following an abortion, women are Table 1: Range of funding offered for LARC inser- routinely offered contraception that is free tion and removal across DHBs. if dispensed or inserted at the time within Procedure Lowest Highest the DHB service, and the rate of uptake is funding funding high. Following an EMA, a Jadelle implant amount ($) amount ($) may be inserted at the time of the first visit Jadelle 80 156 to the service, but IUCD insertion must be insertion delayed until abortion is confirmed, and that requires a second visit. In Auckland, Jadelle 60 160 39% of women chose an IUCD post abortion removal in 2019.18 With many women having to IUCD 102 200 travel long distances to access their regional insertion abortion centre, there is a substantial risk IUCD 60 160 that they will not return for funded IUCD removal insertion, increasing the possibility that New Zealand may see a rise in the number of repeat abortions, which has fallen substan- Five DHBs (Canterbury, South Canterbury, tially since the offering of post-abortion MidCentral, Nelson Marlborough and LARC became routine practice. Only Bay of Plenty) offer more than the MoH Northland DHB has included funding for suggested funding, implying that they post-abortion contraception in primary care have taken account of feedback from as part of the new LARC contract. local healthcare practitioners. Some PHOs already fund LARC insertion and removal Finally, it is clear that New Zealand’s for certain enrolled patients (eg, aged current contraceptive-service landscape under 22 years, or ‘high needs’ (defined is highly complex. Women face significant as Māori or Pacific ethnicity or residing in difficulties in navigating to the service of a quintile 5 area)). Local providers have their choice. Family Planning New Zealand’s little incentive to implement a new service website contains clear information on where one already exists with a higher service location and cost, but many of their fee payable, even if it would mean that services are overstretched and wait times more patients would be eligible for of over two months for an IUCD insertion services. are common. There is currently no national website where women can find other qual- Historically poor data collection means ified local contraception providers. If there that it is not possible to compare 2019/20 were, women would find themselves having volumes of LARC insertions in New Zealand to disclose sensitive information in order to with those dating to before the intro- be eligible for funded services in many areas duction of the new contract. However, data of the country. obtained via OIA11 illustrate that more than half of the country’s DHBs report failing to Provider funding achieve the number of insertions for which they were funded. These DHBs include The funding offered by DHBs to providers all three Auckland DHBs, who collectively varies markedly around the country have inserted less than 30% of the LARCs (Table 1). The MoH contract suggested for which they were contracted. This funding LARC procedures at a value ($150) raises significant questions about the that was ‘tested with a selection of PHOs to way in which they have chosen to confirm feasibility’.5 However, this value configure services. does not reflect the true cost of providing services. Many GP practices report that they, By way of contrast, Lakes and Bay of or their PHO, have declined the contract Plenty DHBs have provided far more contra- because it does not cover costs such as ception services than they were originally materials and staff time, and a patient contracted for, which suggests that their co-payment is disallowed. In some provider services are both accessible and acceptable contracts, clauses exist such that providers to women. Further work will need to be can claim only one fee if a LARC is removed done to elucidate the factors contributing to and a new one reinserted at the same visit. these volumes.

that these effects result in savings to the Conclusions NHS of up to £11 for each £1 spent on The implementation of this initiative has contraception.22 been variable and has resulted in marked It is clear that New Zealand urgently needs regional differences in access to services. a national women’s health strategy, which In some areas, women must now disclose should include addressing the variable sensitive personal details in order to access funding and provision of contraception funded contraception—something they services. In addition, DHBs should be held are not expected to do when accessing to account where it is clear that they are abortion services. There are currently no failing to deliver programmes of work for other medical services in New Zealand which they have been contracted. Where that rely on the application of inconsistent DHBs have created successful initiatives for non-clinical criteria in order to access their population, any learnings should be funding. Contraception is an entirely disseminated widely in order to narrow the predictable healthcare cost required by unacceptable gaps in contraception access up to 50% of the population at some point across the country, and to avoid ‘reinventing in their lives, often for over 30 years. It is the wheel’, an all-too-common feature of therefore surprising that it is seen as a low DHB projects. priority for health-policy makers rather than a central part of an integrated, well-re- If success looks like a reduction in unin- sourced and country-wide women’s health tended pregnancy and abortion, then clearly service. Moreover, the benefits of contra- a population-wide approach is required. ception cannot be overstated: they include The issue appears to be one of prioritisation, a reduction in the adverse outcomes with current restrictive funding criteria associated with unintended pregnancy, enabling an unethical form of rationing. improvement in maternal physical and This in turn perpetuates entrenched mental health and retention of women inequities in contraception access and avail- in the workplace. A UK report estimated ability for the women of Aotearoa.

Competing interests: Nil. Author information: Orna McGinn: General Practitioner & Clinical Director Primary Care Women's Health, Auckland District Health Board. Vicki Mount: General Practitioner, Mission Bay Doctors; MPH candidate, University of Auckland. Helen Fulcher: General Practitioner, Academic Tutor, Auckland Medical School, University of Auckland. Corresponding author: Dr Orna McGinn, General Practitioner & Clinical Director Primary Care Women's Health, Auckland District Health Board [email protected] URL: www.nzma.org.nz/journal-articles/increasing-access-to-contraception-in-new-zealand-assessing-the-impact-of-a-new-funding-initiative

REFERENCES 1. Pollard G. Group Manager, 6. Department of Health. WHO; 2014. Available Child and Community 2018. Clinical Practice from: https://apps.who. Health, Population Health Guidelines: Pregnancy int/iris/bitstream/handle/ and prevention, New Care. Canberra: Australian 10665/133327/978924150 Zealand Ministry of Government Department 7745_eng.pdf;jsessionid=- Health. Personal commu- of Health. Available from: 8577D39EC955EF355799B- nication via email https://www.health.gov.au/ 92CEE2127C9?sequence=1 November 12 2019. sites/default/files/pregnan- 10. Gomez AM, Wapman M. 2. Hohmann-Marriott BE. cy-care-guidelines_0.pdf Under (implicit) pressure: Unplanned pregnancies 7. Women’s Health Commit- young Black and Latina in New Zealand. Aust tee. Pre-pregnancy women’s perceptions N Z J Obstet Gynaecol. counselling Statement. of contraceptive care. 2018 Apr;58(2):247-250. Royal Australian and Contraception. 2017 3. McGinn O, Fulcher HJ, New Zealand College of Oct;96(4):221-226. Arroll B, McCowan L. Obstetricians and Gynae- 11. Number of long acting Barriers to the prescrip- cologists; 2017. Available reversible contraceptives tion of LARCs in general from: https://ranzcog. (LARCs) delivered under practice in New Zealand-a edu.au/RANZCOG_SITE/ DHB contraceptive access qualitative research media/RANZCOG-ME- contracts for the period study. N Z Med J. 2019 DIA/Women%27s%20 1/7/19 to 31/6/20. Response Dec 13;132(1507):63-69. Health/Statement%20 to request under the and%20guidelines/ 4. Genter JA. Thousands of Official Information Act Clinical-Obstetrics/ women to benefit from 1982 to the Ministry of Pre-pregnancy-Counsel- contraceptive changes. Health, September 2020 ling-(C-Obs-3a)-review-Ju- (Internet) Releases: New 12. Urale PWB, O’Brien MA, ly-2017_1.pdf?ext=.pdf Zealand Government; Fouché CB. The relation- 2019 April 4. Available 8. Dickson J, Thwaites A, ship between ethnicity and from: https://www. Bacon L. Contraception fertility in New Zealand. beehive.govt.nz/release/ for adolescents with Kōtuitui: New Zealand thousands-women-bene- disabilities: taking control Journal of Social Sciences fit-contraceptive-changes. of periods, cycles and Online. 2019;14(1):80-94. conditions. BMJ Sex Reprod 5. Service Specification: Available from: https:// Health. 2018 Jan;44(1):7-13. Increasing Equity of doi.org/10.1080/11770 Access to Contraception. 9. World Health Organization. 83X.2018.1534746. Ministry of Health; 2019. Framework for ensuring 13. Hutton C. Beneficiary Response to request under human rights in the contraceptive scheme the Official Information provision of contraceptive comes to an end. (Internet) Act 1982 to the Ministry of information and services Radio New Zealand; 2018 Health, September 2020. 2014. Geneva, Switzerland: August 8. Available from:

https://www.rnz.co.nz/ Board; 2019. Available of Health & Social Care; news/national/363603/ from: http://national- published 2019 13 June. beneficiary-contra womenshealth.adhb. Available from: https:// ceptive-scheme- govt.nz/assets/Womens- assets.publishing.service. comes-to-an-end. health/Documents/ACR/ gov.uk/government/ 14. Protected&Proud: Me aro NWH-Annual-Clini- uploads/system/uploads/ koe ki te hā o Hiueahuoue cal-Report-2019.pdf. attachment_data/ (Internet). Bay of Plenty 18. National Women’s Annual file/808556/Abortion_Statis- District Health Board; Clinical Report. Auckland: tics__England_and_ c2019-2020. Available from: National Women’s Health, Wales_2018__1_.pdf https://www.protectedan- Auckland DIstrict Health 21. Belton S, McQueen G, Ali E. dproud.nz/services. Board; 2018. Available Impact of legislative change 15. Ministry of Health. from: https://national- on waiting time for women Contraception: Findings womenshealth.adhb.govt. accessing surgical abortion from the 2014/15 New nz/assets/Womens- services in a rural hospital Zealand Health Survey. health/Documents/ACR/ in the Northern Territory. Wellington: Ministry NWH-Annual-Clinical-Re- Aust N Z J Obstet Gynaecol. of Health; published port-2018-final.pdf 2020 Jun;60(3):459-464. online November 2019. 19. Ministry of Justice. Abor- 22. Department of Health. A 16. Lawton B, Makowha- tion Legislation Act 2020. Framework for Sexual remahihi C, Cram F, et Ministry of Justice: New Health Improvement in al. Pounamu: E Hine: Zealand Government; 2020 England. England: Youth access to contraception for March 23. Available from: Justice Board for England indigenous Māori teenage http://www.legislation.govt. and Wales; 2013 March. mothers. J Prim Health nz/act/public/2020/0006/ Available from: https:// Care. 2016 Mar; 8(1):52–9. latest/whole.html www.gov.uk/government/ publications/a-frame- 17. National Women’s Annual 20. Department of Health work-for-sexual-health-im- Clinical Report. Auckland: & Social Care. Abortion provement-in-england National Women’s Health, Statistics, England and Auckland District Health Wales: 2018. Department

Think about cats in acute vision loss Aaron Yap, Moaz Alshaikhi, Kay Evans

at-scratch disease (CSD) is a systemic one month earlier, following a cat scratch in illness caused by the gram-negative the Philippines. His right visual acuity (VA) C bacteria Bartonella henselae (B. hense- was 6/60 and his left VA was 6/6. He had a lae). Cats serve as the primary host reservoir right relative afferent pupillary defect. On slit for B. henselae, with cat fleas as the horizon- lamp examination, there was mild cellular tal vector of transmission. It is transmitted activity in the anterior chamber and vitreous to humans through close contact with cats on both sides. Both optic discs were swollen and cat scratches and bites. We describe two with subretinal fluid tracking towards cases of bilateral Bartonella neuroretinitis the macula. There was early macular star that were presented to Palmerston North formation on both sides (Figure 1). Hospital Eye Clinic. Positive results returned for Bartonella IgG (1:1024), Toxoplasma IgM and Quanti- Case reports FERON-TB Gold. He presented a diagnostic Case 1 dilemma because of the high IgM for toxoplasmosis. However, a diagnosis of A 22-year-old male presented with a cat-scratch disease was the top of the differ- five-day history of blurred, right-central ential because of the history, the appearance vision, headaches and light sensitivity. He of bilateral macular stars (Figure 1) and the had a febrile illness with lymphadenopathy high level of IgG antibody to Bartonella.

Figure 1: Fundus photographs show bilateral optic disc swelling with early macular star formation (see arrows).

Case 2 Discussion A 21-year-old female presented with a Optic disc swelling with macular star is the one-week history of headaches, blurred characteristic presentation of CSD neurore- left-central vision, light sensitivity, night tinitis. The deep white chorioretinal lesions sweats and myalgia. She had two pet cats and optic nerve head lesions exhibited by but denied any recent cat scratches. Her both cases are well documented in the liter- right VA was 6/5 and left VA was hand ature.1,2 Both patients tested positive for B. motions. On slit lamp examination, the henselae IgG, which carries a high sensitivity anterior chamber and vitreous were quiet (88%) and specificity (94%) for the disease.3 on both sides. There was bilateral optic It is thought that IgM antibodies need not nerve head swelling, a prominent left necessarily be raised to make a definitive macular star and multifocal retinal spot diagnosis.4 Although both cases had bilateral lesions (Figure 2). Serology demonstrated disease, ocular CSD is typically a unilateral high levels of IgG antibody (1:1024) to disease (76–83%).2,5 Blood pressure was Bartonella but was negative for IgM. measured to be normal in both patients at initial presentation, as malignant hyper- Treatment tension was an important differential to rule Both cases were treated with doxycycline out in bilateral optic nerve head swelling. 100mg twice daily, rifampicin 300mg Treatment for Bartonella neuroretinitis twice daily and prednisone (1mg per remains controversial. A retrospective kilogram daily). Co-trimoxazole was analysis of 107 eyes found that, for patients initially given to the first case because of with VA worse than 6/9, the combination of the concern regarding toxoplasmosis, but steroids and antibiotics resulted in a better it was subsequently discontinued. The total visual outcome compared with antibiotics duration of treatment was six weeks for alone. The most common antibiotic combi- both patients. Both cases improved nation included doxycycline with rifampicin. on treatment. The first case had a final It also found that clinicians tend to reserve VA of right 6/18 and left 6/4.5. The second the use of steroids for those with moderate to case had a final VA of right 6/4.5 and severe vision loss.2 Both patients in our series left 6/9. had severe vision loss and responded well to oral steroids and antibiotics.

Figure 2: Fundus fluorescein angiogram of the left eye show a tuft of telangiectatic vessels overlying the optic disc lesion. Yellow-white multifocal retinal lesions become hyperfluorescent towards the late phase of the study (see arrows).

Competing interests: Nil. Acknowledgements: Credits to Erina Wood and Palmerston North Eye Clinic nurses for the clinical photographs. Author information: Aaron Yap, MBChB: Palmerston North Eye Department, MidCentral District Health Board. Moaz Alshaikhi, MBChB: Palmerston North Eye Department, MidCentral District Health Board. Kay Evans, MB., B.Chir., FRANZCO: Palmerston North Eye Department, MidCentral District Health Board. Corresponding author: Dr Aaron Yap, Palmerston North Eye Department, 50 Ruahine Street, Roslyn, Palmerston North 4414, +64211215528 [email protected] URL: www.nzma.org.nz/journal-articles/think-about-cats-in-acute-vision-loss

REFERENCES 1. Oray M, Onal S, Koc 3. Suhler EB, Lauer AK, 5. Chi SL, Stinnett S, Eggen- Akbay A, Tugal Tutkun I. Rosenbaum JT. Prevalence berger E, Foroozan R, Diverse Clinical Signs of of serologic evidence of Golnik K, Lee MS, et al. Ocular Involvement in Cat cat scratch disease in Clinical characteristics in Scratch Disease. Turkish patients with neurore- 53 patients with cat scratch journal of ophthalmol- tinitis. Ophthalmology. optic neuropathy. Ophthal- ogy. 2017;47(1):9-17. 2000;107(5):871-6. mology. 2012;119(1):183-7. 2. Habot-Wilner Z, Trivizki 4. Curi AL, Machado D, O, Goldstein M, Kesler A, Heringer G, Campos WR, Shulman S, Horowitz J, Lamas C, Rozental T, et et al. Cat-scratch disease: al. Cat-scratch disease: ocular manifestations ocular manifestations and treatment outcome. and visual outcome. Acta ophthalmologica. International ophthalmol- 2018;96(4):e524-e32. ogy. 2010;30(5):553-8.

Updated algorithms are required to differentiate type 1 from type 2 diabetes using the Virtual Diabetes Register Lynne Chepulis, Christopher Mayo, Ryan Paul

he Virtual Diabetes Register (VDR) is data to differentiate between T1D and T2D an annually updated national register are being developed to define these subset T of all patients with diabetes mellitus. populations from the VDR. However, The VDR was designed to monitor the prev- differentiating between T1D and T2D can be alence of diabetes in New Zealand and sup- difficult, as T2D is becoming more common port quality improvement initiatives.1,2 The in youth and young adults and the onset of register is compiled from publicly available T1D in adulthood is becoming increasingly health data based on patients’ use of health recognised.5 Further, approximately 10% of services, including hospitalisation records, patients with T1D will have negative islet laboratory test results and pharmaceutical cell antibody titres, and a similar proportion dispensing information. As of 2018, the VDR of patients with T2D will have mildly estimates that there are 253,480 patients elevated autoantibody titres.5 Consequently, with diabetes in New Zealand.1 However, misclassification of the type of diabetes is this is lower than the number forecast prior common in the community. Approximately to 2017, due to a change in the algorithm one third of patients with T2D are misclas- used to define the VDR population.3 sified as T1D,6 and approximately 40% of 7 Although the VDR is a valuable tool in adults with T1D are misdiagnosed with T2D. that it allows for the description of diabetes One algorithm designed to identify prevalence by geography, age, ethnicity patients with T1D was published by and gender,3,4 it does also have limitations. McKergow et al in 2017,8 and it has been Currently, the VDR does not contain diag- used to define the T1D patient population nosis information, and thus it is unable to in further work with VDR data.9 However, discriminate between the different types of the diagnostic accuracy of the McKergow diabetes, namely type 1 diabetes (T1D) and algorithm has not been validated against the more common type 2 diabetes (T2D). a confirmed T1D clinical register. Hence, This poses a problem if VDR data is to be the aim of this study was to determine the used for research purposes, as often times accuracy of the McKergow algorithm by a clinically defined population (eg, patients using it to predict which patients in the with T1D) is required for analysis. In some 2017 VDR dataset had T1D as compared cases, this can be circumvented via the use to a known population of patients with of clinically confirmed patient registers, confirmed T1D at the Waikato Regional though these are generally restricted to a Diabetes Services at Waikato District Health few individual district health board (DHB) Board (WDHB) for the same time period. databases that contain limited current up Ethics approval was granted for this study to date data rather than a national dataset. by the Health and the Disability Ethics Hence algorithms using additional clinical Committee (ref: 17/NTB/222).

The Waikato VDR dataset used in this were not (true negatives) (Figure 1). This study included all patients registered in gives the McKergow algorithm a sensitivity the national dataset and domiciled in the of 80%, a specificity of 98% and an overall Waikato DHB region during 2017. As per accuracy of 97%—but a positive predictive the McKergow algorithm,8 this included value (PPV) of only 66%. patients who died during 2017 and those not Although these values of specificity and enrolled in a primary health organisation accuracy look good, they are falsely repre- (PHO) (n=23,211). As such, this population is senting the validity of this algorithm, as the larger than the 21,767 reported on the VDR overall accuracy would have still been >95% 1 website, as the latter excludes patients who even if the algorithm had detected no cases are not alive or not enrolled in a PHO as of of T1D, due to the fact that they comprise 31 December 2017. such a small proportion (~3.5%) of the The WDHB diabetes clinical register VDR dataset. The currently low PPV of the included 1,303 patients who have all had McKergow algorithm is clinically significant, T1D diagnosed by an endocrinologist.9 as shown by Wheeler et al, whose use of the Patients in the initial WDHB dataset who algorithm underestimated insulin pump were not part of the VDR extract were therapy by one third in Waikato patients excluded (n=85), such that 1,218 patients with T1D.9,10 Importantly, the algorithm were included in this study. excluded 229 of the 247 false negatives The McKergow algorithm was applied (patients with confirmed T1D) from the to the VDR data as described previously8 predicted T1D population because of their but with the inclusion of additional data use of oral hypoglycaemic agents (Figure 1). for 2014–2017 to account for the time This suggests that this may be the step in the period of study. This included diabetes algorithm that requires further fine tuning medication dispensing data (1 January to improve the PPV. 2006–31 December 2017), hospital diag- However, it must be noted that, although nosis and discharge data (1 January 1988–31 we used the WDHB as a ground truth December 2017) and death records (2017) dataset, our study does have limitations. to determine the number of patients with It is possible that a few of the false posi- T1D (Figure 1). Although not defined in the tives may have been true positives (eg, if McKergow study, dispensed insulin in our they had never been to the WDHB regional study included insulin glargine, insulin diabetes clinic or hospital before and/or isophane with insulin neutral, insulin aspart had been diagnosed in a different region with insulin aspart protamine, insulin before moving to the Waikato region). neutral, insulin zinc suspension, insulin Further, there may be differences in the neutral, insulin lispro, insulin glulisine, diabetes medications used in the algorithm insulin aspart, insulin isophane and insulin in our study, as the medications used in the lispro with insulin lispro protamine. Simi- McKergow study were not defined,8 and this larly, oral hypoglycaemics and alpha may have influenced the results. Regardless, glucosidase inhibitors included metformin our results do suggest that the current algo- hydrochloride, vildagliptin with metformin rithm does not appear to be accurate for hydrochloride, glibenclamide, gliclazide, differentiating T1D from T2D in the VDR. glipizide, vildagliptin, pioglitazone and In conclusion, due to the small proportion acarbose. Linkage between the datasets was of patients with T1D in the VDR dataset and undertaken using master National Health the large number of false positives and false Index (NHI) numbers. negatives that were detected, the McKergow Upon running the McKergow algorithm, algorithm does not appear to be an overly a total of 1,467 of 23,211 patients were effective or precise method for deter- predicted to have T1D. Of these, 971 had mining patients with T1D from the VDR. To confirmed T1D as per the WDHB dataset ensure that ongoing T1D research in New (true positives), and 496 did not (false Zealand is effective, further work is needed positives). Of the 21,744 patients who to define a suitable algorithm for defining were predicted to not have T1D, 247 were subsets of patients from the VDR, including recorded with confirmed T1D in the WDHB benchmarking against other known clinical dataset (false negatives) whereas 21,497 datasets.

Figure 1: Predicted population of type 1 diabetes mellitus (T1D) from the 2017 VDR register (n=23,211) compared to a known WDHB clinical register of patients with confirmed T1D (d=1,218).

Listed on 2017 Virtual Diabetes Register n = 23211 (total number in 2017 VDR), d = 1218 (WDHB T1D patients in VDR)

No (n = 16638, d = 3) Insulin dispensed during study period? (September 1, 2015 to December 31, 2017)

Yes (n = 6573, d = 1215)

Oral hypoglycaemics or alpha glucosidase inhibitors Yes (n = 4843, d = 229) dispensed between January 1, 2006 & December 31,

2017?

No (n = 1730, d = 986)

Any hospital discharges with a diagnosis of T2D and Yes (n = 240, d = 15) no hospital discharges with a diagnosis of T1D

between 1988 & December 31, 2017? Not predicted T1D (n = 21744, d = 247)

No (n = 1490, d = 971)

Yes (n = 12, d = 0) Died during study period and a diagnosis of type 2 diabetes mellitus recorded in death records?

No (n = 1478, d = 971)

Yes (n = 11, d = 0) Any hospital discharges with a diagnosis of cystic fibrosis, pancreatectomy, or neonatal diabetes mellitus prior to insulin dispensing?

No (n = 1467, d = 971)

Predicted T1D population (n = 1467, d = 971)

Figure 1: Predicted population of Type 1 diabetes mellitus (T1D) from the 2017 VDR register (n=23211) compared to a known WDHB clinical register of patients with confirmed T1D (d=1218).

Competing interests: Nil. Author information: Lynne Chepulis: Senior Research Fellow, Waikato Medical Research Centre, University of Waikato, Hamilton. Christopher Mayo: Medical Student, Faculty of Medical and Health Sciences, University of Auckland, Auckland. Ryan Paul: Endocrinologist, Waikato Regional Diabetes Service, Waikato District Health Board, Hamilton. Corresponding author: Lynne Chepulis, Medical Research Centre, University of Waikato, Private Bag 3015, Hamilton [email protected] URL: www.nzma.org.nz/journal-articles/updated-algorithms-are-required-to-differentiate-type-1-from-type-2-diabetes-using-the-virtual-diabetes-register

REFERENCES 1. Ministry of Health. Virtual 5. Oram RA, Patel K, Hill A 9. Tamatea J, Chepulis L, Gold- Diabetes Register. Available et al., A type 1 diabetes smith J, et al., Glycaemic from https://www.health. genetic risk score can control across the lifespan govt.nz/our-work/diseas- aid discrimination in a cohort of New Zealand es-and-conditions/diabetes/ between type 1 and patients with type 1 about-diabetes/virtual-dia- type 2 diabetes in young diabetes mellitus. Internal betes-register-vdr [accessed adults. Diabetes Care, Medicine Journal (online) April 2020]. 2019. 2016. 39(3): p. 337-344. https://doi.org/10.1111/ 2. Jo E., Drury P. Devel- 6. De Lusignan S, Sadek imj.14816, 2020. opment of a virtual N, Mulnier H et al., 10. Wheeler BJ, Braund R, diabetes register using Miscoding, misclassifi- Galland B et al., District information technology in cation and misdiagnosis health board of residence, New Zealand. Healthcare of diabetes in primary ethnicity and socioeco- Informatics Research, care. Diabetic Medicine, nomic status all impact 2015. 21(1): p. 49-55. 2012. 29(2): p. 181-189. publicly funded insulin 3. Chan WC, Papaconstan- 7. Thomas NJ, Lynam A, Hill pump uptake in New tinou D, Lee M et al., Can AV et al., Type 1 diabetes Zealand patients with type administrative health defined by severe insulin 1 diabetes. New Zealand utilisation data provide deficiency occurs after Medical Journal 2019. an accurate diabetes 30 years of age and is 132(1491): p. 78-89. prevalence estimate for a commonly treated as type geographical region? Diabe- 2 diabetes. Diabetologia, tes Research and Clinical 2019. 62(7): p. 1167-1172. Practice, 2018. 139: p. 59-71. 8. McKergow E, Parkin L, 4. Coppell KJ, Mann J, Willians Barson DJ et al., Demo- SM et al., Prevalence of graphic and regional diagnosed and undiagnosed disparities in insulin pump diabetes and prediabetes utilization in a setting in New Zealand: findings of universal funding: a from the 2008/09 Adult New Zealand nationwide Nutrition Survey. New study. Acta Diabetologica, Zealand Medical Journal, 2017. 54(1): p. 63-71. 2013. 126(1370): p. 23-42.

Are patients diagnosed with functional symptoms during ‘code stroke’ receiving best medical care? Jaron Huang, Karim M Mahawish

troke is the leading cause of disability of cases presented with left-sided senso- and the third leading cause of death rimotor symptoms (11 cases out of 17), and Sin New Zealand.1 Early, rapid and the remainder had right-sided or bilateral accurate diagnosis is required during ‘code sensorimotor symptoms. Seven cases had stroke’ in order to deliver thrombolysis and dysphasia as a component of their presen- thrombectomy, therapies proven to reduce tation. Clinical examination findings disability and mortality. ‘Stroke mimics’ are suggestive of a functional presentation presentations that are attributed to causes included give-way weakness and incon- other than stroke and account for one-quar- sistent examination findings, which may ter of stroke-like presentations.2 Fifteen improve with encouragement. All patients percent of patients with stroke mimics have underwent brain imaging with CT and/or no identifiable organic disease to explain MRI. their presentation and are diagnosed as Out of the 17 presentations, 12 had a final 2 functional neurological disorders (FND). It diagnosis of FND. A summary of the care is is important to identify patients with FND displayed in Figure 1. to avoid administering potentially harmful The majority of patients had clear docu- medication. mentation of their clinical history and In this audit, we looked at the standard of examination findings in the medical records medical care that we provided to patients (16 out of 17). A clear diagnosis of FND was diagnosed with FND during code stroke at documented in 8 out of 12 confirmed cases. Palmerston North Hospital, a medium sized We performed poorly against standard institution, where we serve a population three, as evidence of clear communication of 174,340 with approximately 330 stroke occurred in only 2 out of the 12 FND cases. admissions per annum. Standard four was met in 15 out of 17 cases, Though there are no guidelines on how to and multidisciplinary input occurred in 6 manage FND, a literature search identified out of 12. Among the FND presentations, several cohort studies and one randomised only 2 out of 12 met all five standards, and controlled trial that describe specific the mean number of standards met was 3 factors that were associated with improved out of 5. symptoms and functional outcomes for MR brain imaging was performed in 3 of this population. Based on these factors, we the 12 patients with a final diagnosis of FND. formulated the standards in Table 1. The final diagnosis of the remaining five We applied the standards to patients who presentations included orthostatic hypo- presented to Palmerston North Hospital tension, cognitive impairment, posterior between 1 January 2018 and 31 December reversible encephalopathy syndrome and 2019 and who triggered a code stroke (or two ischaemic strokes. The two patients with ‘thrombolysis call’) and were diagnosed a final diagnosis of ischaemic stroke were with FND at initial presentation. Seventeen young (aged 40s–50s) with no vascular risk presentations involving 16 patients met our factors; further, their examination findings inclusion criteria; their baseline character- fluctuated, which was later discovered to be istics are shown in Table 2. The majority due to multi-territory cerebral infarcts.

Table 1: Standards of care.

Standards 1 Patients should have clear documentation of their clinical history and examination findings.

Rationale: We considered this to be an important aspect of medical care. 2 Patients diagnosed with functional symptoms should have a documented diagnosis that includes the word ‘functional’ or equivalent (eg, ‘psychogenic’, ‘conversion disorder’, ‘pseudo- seizure’ or ‘non-epileptic seizure’) in their final diagnosis.

Rationale: In a study looking at patients’ responses to various terms for medically unexplained diseases, ‘functional weakness’ was found to accurately reflect the diagnosis and was the most acceptable term for patients.3

3 Patients diagnosed with functional symptoms should have documented evidence that their diagnosis was explained to them clearly with specific use of the word ‘functional’ or equivalent.

Rationale: Clear, early communication of FND is associated with improved patient outcomes.4 As an analogy, in one study of patients presenting with psychogenic non-epileptic seizures (PNES), 44% of patients had immediate cessation of symptoms upon communication of their diagnosis and remained seizure-free at three months.5

4 Ischaemic stroke should not be misdiagnosed.

Rationale: Ischaemic stroke requires targeted therapy in order to improve patient outcomes. 5 There should be documented evidence of multidisciplinary team input (eg, physiotherapist, occupational therapist, speech language therapist, etc) or outpatient clinic referrals made.

Rationale: Physiotherapy has an important role in helping patients with functional movement disorders to regain mobility and function, with trials demonstrating better functional outcomes compared to control.6

Table 2: Baseline characteristics.

Patient Characteristics n=16 Mean age (range) 57.6 (31–80)

Female sex (%) 8 (50.0)

Ethnicity (%)

New Zealand European 10 (62.5)

Other European 3 (18.8)

Māori 3 (18.8)

Vascular risk factors (%) 10 (62.5)

Previous vascular events (%) 4 (25.0)

Previous psychiatric history (%)

Depression 2 (12.5)

Previous panic attacks 1 (6.3)

Previous presentation with functional symptoms (%) 9 (56.3)

may have played a role in their initial Discussion misdiagnosis, given their young age, lack In this audit, we defined objective, of vascular risk factors and fluctuating evidence-based standards with which to symptoms. These atypical stroke presenta- assess the medical care received by patients tions have led us to adopt a low threshold presenting acutely with suspected functional for performing immediate MRI brain symptoms during code stroke. Applying imaging in younger patients to help with these standards to presentations with FND at diagnostic clarification. our institution has shown that the majority The findings presented by this audit high- did not meet all the standards (ie, what light areas of weakness at our institution, should constitute ‘best medical care’); in including missed therapy and miscommuni- only 2 out of 12 instances were all five stan- cation of diagnosis, which may have led to dards met. It is disappointing that evidence worse patient outcomes and longer hospital of clear communication was only recorded stays. These factors could be improved by in two patients’ clinical notes. This is likely incorporating these standards into a FND to be under-representative of the true figure, care pathway. A subsequent re-audit could as not all communication is documented. then be conducted to assess the impact of This audit finding has led us to proactively this intervention on care processes and communicate the diagnosis of FND and patient outcomes. To our knowledge, this document the discussion. is the first set of evidence-based standards Two patients with ischaemic stroke for FND and could be applied to assess the were misdiagnosed, meaning acute ther- standard of care at other institutions. apies were not delivered. Cognitive biases

Figure 1: Percentages of applicable presentations meeting each care standard 1–5, and percentage of FND presentations meeting all care standards.

Competing interests: Nil. Acknowledgements: We thank Dr Almond Leung, who assisted with data collection in this audit. Author information: Jaron Huang: House Officer, MidCentral District Health Board. Karim M Mahawish: Senior Medical Officer in General, Geriatric & Stroke Medicine, Department of Internal Medicine, MidCentral District Health Board. Corresponding author: Dr Jaron Huang, House Officer, MidCentral District Health Board, 50 Ruahine Street, Palmerston North 4442, New Zealand, +64 27 692 4985 [email protected] URL: www.nzma.org.nz/journal-articles/are-patients-diagnosed-with-functional-symptoms- during-code-stroke-receiving-best-medical-care

REFERENCES 1. Ministry of Health. (2019). 3. Stone J, Wojcik W, Durrance seizures: incidence, Mortality 2017 data tables. D, et al. What should population characteristics, Available online at: https:// we say to patients with and early outcome from www.health.govt.nz/ symptoms unexplained a prospective audit of a system/files/documents/ by disease? The “number first seizure clinic. Epilepsy publications/mort-2017- needed to offend”. BMJ. Behav. 2011;20(2):308-311. pub-20191218-final.xlsx. 2002;325(7378):1449-1450. 6. Nielsen G, Buszewicz Accessed 6 September 2020. 4. Stone J. Functional M, Stevenson F, et al. 2. Jones AT, O’Connell NK, neurological disorders: Randomised feasibility David AS. Epidemiology the neurological assess- study of physiotherapy for of functional stroke ment as treatment. Pract patients with functional mimic patients: a Neurol. 2016;16(1):7-17. motor symptoms. J Neurol systematic review and 5. Duncan R, Razvi S, Mulhern Neurosurg Psychiatry. meta-analysis. Eur J S. Newly presenting 2017;88(6):484-490. Neurol. 2020;27(1):18-26. psychogenic nonepileptic

DNA information: access, use and implications for healthcare in Aotearoa New Zealand Sara K Filoche, Jon Cornwall

recently published study on public at University of Canberra, raised concerns attitudes towards genomic data about the uploading of genomic data onto A sharing has highlighted issues My Health Records, saying: ‘It’s outrageous. around whether and how people are willing Why didn’t we know about this? The lack of to share their DNA information.1 The study information only fosters fear and distrust.’”2 is the largest global study of its kind, and it Concern around the misuse of DNA infor- showed that many populations of people are mation is not unfounded, as improper use unwilling to share DNA or genomic informa- of, or access to, such data can promote tion and have issues with how this informa- mistrust of health providers and contribute tion is shared, who can use it and whether to inequity in health outcomes.3 In reference it can be used for commercial purposes.1 to an indigenous context with signif- This research highlights many issues that icant relevance to Aotearoa New Zealand, will have to be addressed in Aotearoa New Hudson and colleagues recently stated that Zealand, including how the use of genomic “Numerous research projects, genomic or information in a healthcare context will otherwise, exhibit enduring negative effects garner the trust of the population to ensure on Indigenous Peoples, minority populations that genomic technologies are used effec- or socially disadvantaged groups owing tively and equitably. In concert with recent to under-representation, lack of informed worldwide developments in healthcare-data consent, lack of consultation, misinterpre- access, where it is becoming more common tation and/or misuse of samples and data.” for genomic information about individual These authors call for “Transparency about persons to be linked to their healthcare the origins and provenance of genomic records, the study raises an important and data, the ways in which the genomic data contemporary question: In Aotearoa New are accessed and used, and how benefits Zealand, do we know how we are going to are generated and distributed equitably responsibly and sensitively utilise the DNA are central to maintaining a high level of information of our citizens in a healthcare integrity within the research enterprise.”4 setting? What sociocultural licence currently Untoward repercussions surrounding the exists to support the implementation of a misuse of genomic information have already framework that allows access to the DNA been observed in Aotearoa New Zealand. information of our citizens for healthcare Take the negative publicity associated with purposes? press releases focused on the identification In 2018, it became possible for Australia to of the ‘warrior gene’. This involved a case link the genomic information of its citizens of unethical research and racial profiling to their health records using the My Health whereby race-based behaviour (aggression Record system. However, as reported in the in Māori men) was linked to a certain gene 5 Sydney Morning Herald, appropriate consul- (MAO-A) . Now, a US-based direct-to-con- tation and governance are not currently sumer company has made commercial gains in place: “Bruce Arnold, a privacy expert by developing a product that tests for the

warrior gene. Unless we consider, anticipate Genomics Aotearoa Rakeiora programme and address the potential for harm asso- recently received significant funding ciated with the misuse of DNA information, to assist the achievement of outcomes there is the potential for similar adverse including, among other things, “Data [that events to occur. are] linkable and protected with careful What are the considerations for people governance and approval to primary and having their DNA information available to secondary care health data and National 10 clinicians or researchers and linked to their Health Datasets.” This is in conjunction health records? In essence, people who with other complementary aims of the undergo some form of DNA testing need to Genomics Aotearoa Rakeiora programme, know what their DNA information may be including developing frameworks that used for and that their data is protected, specifically identify and cater for Māori safe and will not lead to personal harm.4,6 interests in regard to genomic data acqui- In addition, healthcare practitioners need sition and use for health benefits. However, to be aware that genomic data sharing although local funding and research initia- will need to be considered as part of the tives point to an increase in research informed consent process. As an example, momentum and a recognition of the impor- at an operational level for a general practi- tance of this topic, there remains a lack of tioner, there will need to be appropriate and empirically derived data that could assist considered policy and practice to adequately with a way forward in establishing precisely inform patients about how DNA information how local people and their whānau wish to may be used.3,7 For example, who is to be see their genetic information being used in a provided information about their family healthcare context. members?8 How will their family history While DNA-linked health records are not be collated?9 For individual healthcare yet a reality in Aotearoa New Zealand, it is providers, such systems will likely change expected that they will be at some point in patient management, and it is possible that the near future. There are many benefits to healthcare practitioners could be expected having such a system, including non-clinical to be aware of new diseases or health use of DNA information by researchers variants that are discovered and relate these and pharmaceutical companies to assist in to patients on a case-by-case basis. Certainly the development of more efficacious thera- the expectations of patient populations in peutics and to improve our understanding their exchanges with healthcare profes- of both health and disease.6 With technol- sionals will need to be taken into account ogies such as pharmacogenomic testing if DNA-linked healthcare records become a being currently available, this will certainly reality. Currently, it is not clear what these push the agenda forward, whether we— expectations are. the public and healthcare professionals of There have been, and continue to be, Aotearoa New Zealand—are ready or not. several initiatives that explore the impact of In the meantime, it is up to health agencies genomic technologies on both society and to continue to establish what sociocultural healthcare in Aotearoa New Zealand. The licence exists for using genomic healthcare Human Genome Research Project, funded by information within the Aotearoa New the New Zealand Law Foundation, was one Zealand population, in order to understand of the first large-scale investigations into the how genomic medicine can be equitably impact of genetic technology.3 Subsequent integrated into healthcare practice (ie, explorations in this space have included within the framework of Whakamaua: 11 the Ministry of Health and National Health Māori Health Action Plan 2020–2025). Committee initiative The Introduction of In supporting such data use, we require Fit-for-Purpose Omics-based Technologies, discussion around how we think about which operated as a ‘think-piece’ on the clinical data ownership, including perhaps introduction of omics-based technology whether individuals also need or expect in Aotearoa New Zealand. More recently, financial compensation as a consequence 12,13 the Ministry of Business Innovation and of sharing their genetic information. Employment-funded Genomics Aotearoa At present, it is not clear whether there is infrastructure has been established, and the sufficient information surrounding local

sociocultural licence to safely implement Zealand data have indicated, when DNA wider access to DNA or genomic healthcare data sharing becomes more prevalent, that records. Such information is desperately information places the trust-based rela- required to appropriately inform the imple- tionship between the healthcare system, mentation of DNA data sharing, so that healthcare providers and patients in some healthcare is respectful, culturally safe and level of jeopardy. relevant.14–17 As recent global and other New

Competing interests: Nil. Author information: Sara K Filoche: Department of Obstetrics, Gynaecology and Women’s Health and Department of Pathology and Molecular Medicine, University of Otago Wellington, New Zealand. Jon Cornwall: Centre for Early Learning in Medicine, University of Otago, Dunedin, New Zealand. Corresponding author: Dr Sara K Filoche: Department of Obstetrics, Gynaecology and Women’s Health and Department of Pathology and Molecular Medicine, University of Otago Wellington, New Zealand, 04 918 6888 [email protected] URL: www.nzma.org.nz/journal-articles/dna-information-access-use-and-implications-for-healthcare-in-aotearoa-new-zealand

REFERENCES 1. Middleton A, Milne R, on unrestricted access 9. The Royal Australian Almarri MA, Anwer S, to genomic data. Nature College of General Prac- Atutornu J, Baranova Reviews Genetics. titioners. Genomics in EE, et al. Global public 2020;21(6):377-84. general practice. East perceptions of genomic 5. Crampton P, Parkin C. Melbourne, Victo- data sharing: What shapes Warrior genes and risk-tak- ria RACGP, 2020. the willingness to donate ing science. The New 10. Genomics Aotearoa. DNA and health data? Zealand Medical Journal. https://www.genom- The American Journal 2007;120(1250):U2439. ics-aotearoa.org.nz/ of Human Genetics. 6. Middleton A. Society and projects/rakeiora-path- 2020;107(4):743-52. personal genome data. finder-genomic-medicine/ 2. Genome Web. Genomics Human Molecular Genetics. rakeiora-announce- in the records. https:// 2018;27(R1):R8-R13. ment. Accessed 11 www.genomeweb.com/ December, 2020. 7. 7Badzek L, Henaghan scan/genomics-records#. M, Turner M, Monsen R. 11. Ministry of Health. X40AfNAzaUk. Accessed Ethical, legal, and social Whakamaua: Māori 20 September, 2020. issues in the translation of Health Action Plan 3. Hill G. Pharmacogenetics: A genomics into health care. 2020–2025. Wellington: Review of the ethical, Social Journal of Nursing Schol- Ministry of Health; 2020. and Policy Implications of arship. 2013;45(1):15-24. 12. Ballantyne A. How should ‘Personalised Medicine’. 8. Dheensa S, Fenwick A, we think about clinical In: Genes, Society, and Lucassen A. Approach- data ownership? Journal of the Future. Volume 3. ing confidentiality at a Medical Ethics. 2020:mede- Henaghan M (ed). 2009. familial level in genomic thics-2018-105340. 4. Hudson M, Garrison NA, medicine: A focus group 13. Briscoe F, Ajunwa I, Gaddis Sterling R, Caron NR, Fox study with healthcare A, McCormick J. Evolving K, Yracheta J, et al. Rights, professionals. BMJ Open. public views on the value interests and expectations: 2017;7(2):e012443. of one’s DNA and expecta- Indigenous perspectives

tions for genomic database 15. Hudson M, Beaton A, society ready for advanced governance: Results from Milne M, Port W, Russell genomic medicine? a national survey. PloS K, Smith B, et al. Te Mata Australasian Medical one. 2020;15(3):e0229044. Ira: Guidelines for genomic Journal. 2014;7(4):200-202. 14. Caron NR, Chongo M, research with Māori. Māori 17. Jenkins K. ‘Can I see Hudson M, Arbour L, and Indigenous Gover- your social licence Wasserman WW, Robertson nance Centre, University please?’ Policy Quarterly. S, et al. Indigenous genomic of Waikato; 2016. 2018;14(4):27-35. databases: Pragmatic 16. Cornwall J, Slatter T, Print considerations and cultural C, Guilford P, Henaghan M, contexts. Frontiers in Wee R. Culture, law, ethics, Public Health. 2020;8:111. and social implications: Is

Remembering past pandemics to ensure we plan for the future: updated New Zealand survey of memorials to the 1918 influenza pandemic Nick Wilson, Geoffrey Rice, George Thomson, Michael G Baker

ew Zealand has had a problem with insufficient pandemic plan- Methods ning, as indicated by its low score We followed the methods of a previous N 1 5 on a validated Global Health Security New Zealand survey with the same defi- Index in 2019.2 This inadequate planning nition of a pandemic-related memorial. We probably meant that the country needed conducted internet searches for new memo- to have a national lockdown to successful- rials (the last search was on 8 August 2020), ly control COVID-19, in contrast to a very along with field research for all the new well-prepared jurisdiction such as Taiwan memorials we identified. Field visits were (which achieved elimination of COVID-19 conducted between December 2018 and without needing a lockdown).3 This plan- August 2020. ning deficit is even more notable when considering that New Zealand may be the Results second-most-optimal island refuge for A national memorial was identified, humanity in the case of a pandemic that the first such national-level memorial for 4 threatens human extinction. this pandemic created in New Zealand Memorialisation of past pandemic (Table 1). In addition, four new local memo- disasters may serve as a visual public rials were identified: two in Christchurch, reminder of persisting pandemic threats one in Waimate and one in New Plymouth. and the need to invest in prevention and pandemic planning (as argued previ- Discussion ously).5 With this in mind, we aimed to This survey now brings the total number provide a best current estimate of the of publicly accessible memorials that refer number of physical memorials to the 1918 to the 1918 influenza pandemic in New influenza pandemic in New Zealand. This Zealand to 12, which includes one new pandemic killed an estimated 9,000 New national memorial. Worldwide, it appears Zealanders,6,7 and, at its peak, an estimated very rare to have a national memorial to 440 people died in a single day, equivalent the 1918 influenza pandemic. In addition to to 1,820 deaths in today’s numbers this new one in New Zealand, we only iden- (Figure 1). We also aimed to extend a tified one other national memorial, which previous analysis5 to include the 2018 was in Samoa.11 However, there are a few centennial year to see whether that anni- memorials to specific groups (eg, the Lueg versary had generated additional interest Monument in Switzerland to Bern caval- in memorialising this event. rymen who died in the pandemic).12

The relative lack of memorials to the 1918 announced approval for a large, multi-mil- pandemic contrasts markedly with the high lion-dollar national Erebus memorial to prevalence of war memorials in the New be built in a Parnell park to commemorate Zealand setting. That is, there are now 1.3 the 257 people who were killed when an publicly accessible pandemic memorials per Air New Zealand plane crashed into Mount 1,000 deaths, in contrast to 31 memorials per Erebus in 1979.15 1,000 deaths for World War I and World War Nevertheless, it may be desirable to II combined and 213 memorials per 1,000 move away from the conventional idea 5 deaths for the South African War. Interna- of the physical memorial, as argued for tionally, the lack of memorialisation of past previously in New Zealand, and towards 13 pandemics has also been noted. a research fund as a ‘living memorial’.8 The national war memorial at Pukeahu Such alternatives are also being considered National War Memorial Park in Wellington internationally (eg, various digital projects also has qualitative differences compared to relating to the COVID-19 pandemic).13 the national memorial to the 1918 pandemic: In summary, we identified an addi- it completely dwarfs the scale of the national tional five memorials to the 1918 influenza memorial to the 1918 pandemic, which is in pandemic, bringing the total number a corner of this same national war memorial of publicly accessible memorials to 12. site. Similarly, the memorial to the Christ- Nevertheless, New Zealand needs to church earthquake of 2011 that caused 185 further consider how it memorialises past deaths (a photo of which is in another New pandemics—and to make better use of such 14 Zealand disaster memorial study) is vastly memorials to constantly remind the public greater in scale than the national memorial and the government to invest adequately to the 1918 pandemic. Continuing the same in pandemic prevention and pandemic pattern, Auckland Council has recently planning.

Figure 1: Photograph of 440 students at Wellington College symbolically representing the worst day for deaths from the 1918 pandemic in New Zealand (photo by Luke Pilkinton-Ching, University of Otago, Wellington, 2018).8

Table 1: Identified physical memorials to the 1918 influenza pandemic in New Zealand.

Memorial Further details New memorials identified in this updated survey (n=5)

National influenza The memorial was unveiled by the Prime Minister, the Rt Hon Jacinda Ardern, on 6 pandemic memorial at November 2019.9 A specific feature is that, in addition to the memorial being dedicated Pukeahu National War to the victims, the role of others is acknowledged: “The service of the health profes- Memorial Park, Wellington sionals and many volunteers who risked their own lives to care for their communities is gratefully acknowledged.” It also states that “This disaster shaped modern approaches to managing infectious diseases, helping to protect future generations.” The memorial is adjacent to a building that functions as an educational centre, which had additional information on the pandemic (at least at the time of the memorial’s unveiling event).

Memorial to Christchurch The memorial titled Remembering Black November is in the form of an information victims (sited in central board with the image of a grave headstone that refers to the 462 Christchurch resi- Christchurch City) dents who died from the pandemic. The information includes mention of the strong volunteer response in this city, which potentially contributed to the relatively low total mortality rate in Christchurch compared with other New Zealand cities. It was erected by the Christchurch City Council in the centennial year of the pandemic (2018). The information board is located just south-east of the intersection of Hereford Street and Cambridge Terrace.

Memorial to those who This memorial is the only one in the country that refers to deaths in a whole region. It died in the Taranaki region specifically mentions the relatively high mortality burden borne by Māori and includes (sited in New Plymouth text in te reo Māori. It was erected on the 100th anniversary of the pandemic in City) November 2018.10

Memorial to Dr Margaret A giant portrait of this doctor was painted on a grain silo in the town of Waimate in 2018. Cruickshank (Waimate) Near the base of the silo is an associated and publicly accessible information board that discusses her role as a doctor in the 1918 pandemic (which ultimately killed her). This memorial is in addition to the two other memorials to Dr Cruickshank (see Previously identified memorials in Table 1). However, all these memorials are also likely to reflect her contribution as a hard-working local doctor and New Zealand’s first woman general practitioner.

Memorial plaque for those This memorial plaque is attached to a seat by the Avon River, Christchurch (near 24 Ox- who died in Christchurch ford Terrace). It was erected in July 2020 by the Canterbury History Foundation. It is in memory “of the 460 Christchurch victims.” It also notes that “over 9,000 New Zealanders died in this pandemic, including 2,500 Māori.”

Previously identified memorials5

Publicly accessible (i) Granite memorial to the 1,128 citizens of Auckland who died in the pandemic memorials (n=7)5 (Waikumete Cemetery, Glen Eden, Auckland); (ii) headstone on grave raised by public subscriptions for the nurse Jessie Linton (Shortland Historic Cemetery, Thames); (iii) the Featherston Military Camp memorial obelisk and adjacent memorial wall (Featherston Cemetery, Featherston); (iv) marble statue of Dr Margaret Cruickshank with information board (Seddon Square, Waimate, South Canterbury); (v) grave of Dr Margaret Cruick- shank and associated information board (Waimate Old Cemetery, Waimate); (vi) plaque for nurses and a doctor who died and a detailed information board in the Christchurch Nurses’ Memorial Chapel (Christchurch Hospital site, Christchurch); (vii) plaque for nurses who died (Chapel in Dunedin Hospital, Dunedin).

Other memorials This grouping includes pandemic-related memorials with restricted access because they are in private settings, or which do not specifically refer to the pandemic in the text on the memorial. There are another 11 memorials in this category, as detailed in a previous survey.5

Competing interests: Nil. Author information: Nick Wilson: Professor, University of Otago, Wellington. Geoffrey Rice: Emeritus Professor, University of Canterbury, Christchurch. George Thomson: Associate Professor, University of Otago, Wellington. Michael G Baker: Professor, University of Otago, Wellington. Corresponding author: Professor Nick Wilson, University of Otago, Mein St, Newtown, Wellington, New Zealand. 64+ (4) 021 2045523 [email protected] URL: www.nzma.org.nz/journal-articles/remembering-past-pandemics-to-ensure-we-plan-for-the- future-updated-new-zealand-survey-of-memorials-to-the-1918-influenza-pandemic

REFERENCES 1. Boyd MJ, Wilson N, Nelson Health Disaster. Christ- New Zealand (24 August C. Validation analysis of church: Canterbury 2019). https://www.rnz. Global Health Security University Press, 2017. co.nz/international/ Index (GHSI) scores 2019. 7. Summers JA, Baker M, pacific-news/397372/ BMJ Glob Health 2020;5. Wilson N. New Zealand’s remember- 2. Boyd M, Baker MG, experience of the 1918- ing-the-lives-lost-in-the-in- Nelson C, Wilson N. 19 influenza pandemic: fluenza-epidemic-in-samoa. The 2019 Global Health a systematic review 12. Whitmarsh A. Swiss Security Index (GHSI) after 100 years. N Z memorials of the First and its implications for Med J 2018;131:54-69. World War. 2015;(5 New Zealand and Pacific 8. University of Otago. Call for November). http://www. regional health security. pandemic research fund switzerland1914-1918. N Z Med J 2020;133:83-92. as “living memorial” to net/blog/swiss-memorials- 3. Summers J, Lin H-H, 1918 flu victims. University of-the-first-world-war. Cheng H-Y, Telfar Barnard of Otago Media Release 13. Shoot B. The rare challenge L, Kvalsvig A, Wilson N, 2018;(14 December). https:// of building a pandemic Baker M. Potential lessons www.otago.ac.nz/news/ memorial. City Monitor from the Taiwan and New news/otago701754.html. 2020;(9 June, updated 19 Zealand health responses 9. Manch T. Influenza October). https://city- to the COVID-19 pandemic. pandemic that killed monitor.ai/community/ Lancet Regional Health 9000 memorialised in the-rare-challenge-of-build- Western Pacific 2020;https:// Wellington. Stuff 2019;(6 ing-a-pandemic-memorial. doi.org/10.1016/j. November). https://i. 14. Wilson N, Jones AC, Rice G, lanwpc.2020.100044. stuff.co.nz/national/ Thomson G. Epidemiology 4. Boyd M, Wilson N. The politics/117209806/ of major disasters in New Prioritization of Island influenza-pandem- Zealand as revealed by Nations as Refuges from ic-that-killed-9000-memo- disaster memorials. N Z Extreme Pandemics. Risk rialised-in-wellington. Med J 2019;132:104-07. Anal 2020;40:227-39. 10. Groenestein C. Forgotten 15. Auckland Council. Local 5. Wilson N, Ferguson C, Rice 1918 flu victims now board gives approval for G, Baker MG, Schrader B, have a memorial in New national Erebus Memorial Clement C, Thomson G. Plymouth cemetery. Stuff in Dove-Myer Robinson Remembering the 1918 2018(10 November). Park. Auckland Council influenza pandemic: https://i.stuff.co.nz/ 2020;(17 November). national survey of memori- national/health/108506823/ https://ourauckland. als and scope for enhancing forgotten-1918-flu-victims- aucklandcouncil.govt. educational value around now-have-a-memorial-in- nz/articles/news/2020/11/ pandemic preparedness. new-plymouth-cemetery. local-board-gives-ap- N Z Med J 2017;130:53-70. 11. Radio New Zealand. proval-for-national-ere- 6. Rice GW. Black Flu Remembering the lives bus-memorial-in-dove- 1918: The Story of New lost in the influenza myer-robinson-park/. Zealand’s Worst Public epidemic in Samoa. Radio

Cheese Mite Itch and Conjunctivitis—A “Minor Horror” of the Great War 1920

o doubt many of the readers of this their depredations to cheese, but live free journal will have read with pleasure and from choice upon slowly-decomposing Nthat most instructive and excellent- vegetable and animal matter (cheese, ly-written book, “The Minor Horrors of cereals, flour, sugar, preserves, dried War,” and also its sequel, “More Minor Hor- anatomical specimens, bacon, dried fruits rors.” The author, Professor A. E. Shipley, in and fungi), and in the corners of dwellings. his preface, states “there are whole ranges Incidentally they get on to man. of Minor Horrors of War left untouched” The family Tyroglyphidæ are very small in his pages. It does seem weird that this mites devoid of eyes and without tracheæ. Dominion, so far distant from the centre of Tyroglyphus Siro is a somewhat sluggish warfare, should have added yet another to species with heavy body, short legs, long the list. Yet this would appear to be the case. hairs, four long bristles, claws and suckers. In an exhaustive search of the literature I Tyroglyphus Longior, Gervais, is white have been unable to find any reference to or yellowish with two black spots on the the condition occurring amongst those who abdomen. Length 0.55–0.61 mm, breadth handle cheese, and no mention is made of 0.28 mm. Siro and Longior Gervais have been disease being produced by the species of found accidentally in the fæces, urine, in mite which infested cheese in this country. pus, and on the skin. Tyroglyphus Longior, Undoubtedly it is a “Minor Horror” of war, Castellani, was found by Castellani in copra because, but for the war and the consequent and on people affected by copra itch in prolonged storage of cheese in New Zealand, Ceylon. In this variety, in contrast to Tyro- it is at least certain that the affection would glyphus Longior, Gervais, there is no pair never have reached the stage of being dig- of short hairs on the ventral surface behind nified by a special investigation. The condi- the anal suckers. Carpoglyphus is readily tion did exist before 1914, but the advent of distinguished by its greater opacity. It is of a cold storage and quick transport from the deep cream colour when found on cheese. factories to the market in Great Britain had Aleurobius farinæ is sometimes known as caused it to practically disappear. However, the flour mite and has deep reddish-brown during hostilities it appeared in several of legs and short hairs. Glypciphagus is usually the stores throughout the country, and in found in sugar and is distinguished by the 1920 I made an investigation on behalf of fact that its dorsum is hairy or plumose. the Department of Health. Probably not two Rhizoglyphus has short legs armed with dozen well-marked cases have occurred spines. The tarsi end in a claw. altogether, but undoubtedly the disease is a definite entity with a specific cause, and, as such, is worth of chronicle. Life History This, which is similar in all the species, The Cheese Mite—Family: Tyroglyphidæ. consists of four stages: the egg, the larvæ, Genera: Stilton and Cheddar Mite: Tyro- the nymph, and the adult. From egg to adult glyphus, Siro and Longior, Glyciphagus, stage occupies from four to five weeks. The Rhizoglyphus, Aleurobius Farinæ. Cheddar eggs are white ovoid bodies just visible to Mites: Carpoglyphus. Old cheese of any kind the naked eye. They hatch in ten to twelve is liable to attack by any of the above-men- days. The larvæ are colourless and have tioned. The various genera do not confine three pairs of legs. They feed actively for a

week, become quiescent, cast their skin and tival mucosa irritated, but also the bronchial emerge as the first nymph. The first nymph passages of those who work in the stores. has four pairs of legs. It moults again and The affection presents no distinctive becomes the second nymph, which is larger features, but appears to be slow in reacting than the first. After the third moult the mite to treatment. emerges as the adult. Diagnosis Cheese mites may be easily demonstrated On superficial examination the dermatitis in affected cheese by the examination with may be mistaken for scabies, but burrows an ordinary magnifying-glass of a little dust are not present, and the two parasites are or mould which collects upon the surface. very different. The cause of the conjunc- It will be at once apparent that this fine tivitis may be settled by questioning the dust is a mass of moving life. Much work patient as to his employment. Some of those upon these parasites has been done in the affected especially complained of the irri- Zoology Department, University College, tation on the back, due to the mites getting Reading. Investigations there have proved down at the back between collar and neck. that non-infected buildings can become infected by flies and by moths carrying the Treatment The application of ordinary soothing mites upon their legs. They are lovers of ointments to the dermatitis is effective. Beta- darkness, but this preference is excelled by naphthol ointment (5–10 per cent.) is useful. their love of cracks, crevices, and corners. Routine conjunctival treatment eventually Boiling water will not dislodge nor kill them. has prompt effect upon the conjunctivitis. In England it is estimated that damage to the Obviously, as part of the treatment, the extent of £2183 is done annually to Stilton patient must be taken off work associated cheese alone. In New Zealand at the end of with the cheese mites. every loading season the pest may appear in the sheds holding cheese awaiting export. Prevention Since the introduction of cold storage there Disease in man only occurs when the has been great improvement. stores are badly infested; this happens when there is unusually prolonged storage. Under Disease normal conditions of export trouble rarely occurs, and probably we shall hear no more Produced in of the affection in the Dominion, provided no Man in New Zealand interference occurs with regular shipping. It is found that if the temperature of the sheds The species of mite which infested the can be kept constantly at 35degs. F. or lower stores in the Dominion has been identified the mites do not become troublesome or as Tyroglyphus Longior, Gervais. Dermatitis active, although they are not killed. There is and conjunctivitis are the two manifesta- difficulty in maintaining that temperature tions. The skin condition is caused by the at the suitable level. It usually varies from bites of the mites. Thus irritation is caused 35deg. to 40 deg. F. All windows should be and consequent scratching. A dermatitis netted to prevent the entrance of flies and results which resembles the various lesions moths. If doors are kept open a netted door produced in scabies as the result of irri- should be provided. Movable shelves are an tation. Of course, no burrows are present. advantage, and floors are best of concrete. Usually only the exposed parts of the body, When cheese is attacked the damage is the arms and faces, are affected. Prurig- lessened very considerably by brushing the inous papules, papulo-papules and pustules surface daily and removing the dust. are generally present. The conjunctivitis is caused by the fine dust from the cheese Formalin has no value in connection with being raised during the handling of the the pest. Carbolic acid, of course, cannot produce and so gaining access to the eyes. be used upon cheese, but it may be utilised Whether or not the conjunctivitis, which upon the shelves and, especially, the corners in most cases was severe, is produced by of infested stores. A five-per-cent. solution the mites present in the dust or merely by kills the mite in two minutes; increase some irritative property of the dust, it is in strength does not augment efficiency. impossible to say. Not only is the conjunc- Carbon-disulphide: This is employed upon

flour, grain, and many other foodstuffs. It destroys all forms of animal life, but is Tyroglyphus highly inflammable, and is dangerous to as a Cause of human life in large amounts. The vapour from this liquid is effective almost instan- Disease in Other taneously, but, although it is a heavy gas, it Parts of the World does not work underneath the cheese, nor A short description of the diseases in man does it kill the eggs of Tyroglyphus. It may attributable to other species of the mite may be used in various ways: (a) By suspension be of interest. in shallow vessels from which it evapo- Vanillismus—This is stated to be caused rates and diffuses as a gas. This method by Tyroglyphus sero and Aleurobius farinæ. is successful in killing the majority of the Those who handle vanilla beans may be mites, but some still survive at the base of attacked by a pruriginous eruption on the the cheese. In New Zealand good results exposed parts. Desquamation follows as have been obtained with a strength of a rule. The affection is usually known as 13lbs. of carbon disulphide to 5000 cubic “Vanilla Itch.” feet of space; freedom from trouble was ensured for six months by this means. (b) Grocer’s Itch—Here Glyciphagus is the The affected cheese may be sprayed with irritant; sugar merchants and grocers are disulphide by the use of an ordinary garden frequently troubled by swarms of this spray. This is a wasteful method. (c) The species. I have not heard of the disease in cheese may be painted with the liquid. This New Zealand, and I cannot discover any is the most effective means, though it would description of the clinical appearances. be difficult of application on a large scale, Copra Itch—Tyroglyphus Longior; var. and the cost might be prohibitive. Castellani, is the cause of copra itch. Intense To absolutely eradicate the pest it is itching attends this dermatitis. Castellani, necessary to treat the cheese three times: working in Ceylon, first described the once to kill the adults, again after an interval condition in 1911; investigation proved it of twelve days to kill the larvæ, and, lastly, a to be produced by the above species, which third treatment is necessary after a similar was found swarming in many samples of period to ensure that all mites and larvæ are copra. Copra dust containing the mite, when destroyed. The best method is to fumigate rubbed on the skin, causes a pruriginous thoroughly with carbon-disulphide a shed or eruption within 48 hours; similarly, the dairy between seasons, when no cheese is in mites alone, without the dust, will produce store. In addition, corners, window ledges, the characteristic lesions. As in cheese mite crevices, shelves, and all woodwork should itch, only the uncovered parts of the body be thoroughly treated with five-per-cent. are bitten by the parasite. carbolic. The employment of suitable face- In conclusion, I desire to express my masks and close cover for the arms would thanks to the Director-General of Health prevent the Tyroglyphus from biting, but for permission to publish this account; to in actual practice I fear that any rule to Professor Kirk, Dr. Frengley, Mr. Cuddie compel the use of such would be more (Department of Agriculture), and Mr. Hurley often honoured in the breach than in the (Government Bacteriology) for help and observance. facilities afforded to me in the investigation.

URL: www.nzma.org.nz/journal-articles/cheese-mite-itch-and-conjunctivitis-a-minor-horror-of- the-great-war

Proceedings of the Waikato Clinical Campus Research Seminar, Wednesday 14 October 2020

Retrospective review of at the Neurosurgery Unit of statistical analyses and visual- meningioma treated at Waikato Hospital during the ization. Descriptive analysis of period 2007–2019. the patient cohort was carried Waikato Hospital during out using the demographic and the period 2007–2019 Aims 1. To evaluate patient tumour characteristics data. David Wang,1 Fouzia Ziad,2 characteristics in those Chi-square test for categorical Ziad Thotathil,3 Michael Jameson4,5 who had meningiomas variables (with continuity correction) and one-way test 1Dept of Pathology, University of resected at Waikato for continuous variables (with Otago, Dunedin, 2Dept of Pathology, Hospital from 2007 to Waikato Hospital, Hamilton, 3Dept 2019. equal variance assumption) of Radiation Oncology, Waikato were used to evaluate for statis- Hospital, Hamilton, 4Dept of 2. To characterise tical significance. Population Medical Oncology, Waikato Hospital, frequency of menin- data was obtained for these Hamilton, 5Waikato Clinical Campus, gioma histologic subtypes district health boards from the University of Auckland, Hamilton and grades. census for the Midland region Background between 2007 and 2019. Meningiomas are the most 3. To determine the overall Results common primary intracranial and age-standardised A total of 393 patients with and spinal tumours. Although incidence of meningioma histologically confirmed menin- 80% of meningiomas show in the Midland region giomas were identified. The benign clinical behaviour and from 2010 to 2019. incidence of meningioma in can be cured by resection alone, 4. To evaluate disparities in the Midland region has shown about 20% recur after resection incidence between Māori a rising trend over the study and need additional treatment. and other ethnicities. period. The crude incidence Histopathological assessment each year ranged from 2.69 per aims to identify patients at risk 5. To evaluate changes in 100,000 in 2008 to the highest of recurrence. There is a lack of annual age-standardised of 5.14 per 100,000 in 2018. information on the incidence of incidence of meningioma The age adjusted annualized meningiomas in New Zealand. from 2007 to 2019. incidence based on the WHO Being a benign tumour, the Methods standard population for each National Cancer Registry does A retrospective search was ethnicity group showed Māori not collect data except for the carried out with the Waikato had the highest age adjusted uncommon malignant variants. Hospital enterprise software annualised incidence at 6.90 Through this retrospective study using the keyword ‘menin- per 100,000 followed by Pacific of consecutive cases treated at gioma’ in the pathology reports Island ethnic group at 6.16 per Waikato Hospital, we hoped to during the period between 1 100,000. This rate is 2.8 times gain a better understanding of June 2007 and 1 June 2019. higher than the New Zealand the epidemiological and clinical Data on age, gender, ethnicity, European ethnicity category features that characterise number and location of lesions, and 3.75 times higher than meningiomas in New Zealand. grade of lesion (reclassified as Asian population. Female We also aimed to study ethnic per WHO grading criteria 2016), patients were the majority variations to see if there is a extent of surgery, adjuvant accounting for 73.4% of all greater incidence of menin- therapy and current status or patients. Furthermore, there gioma in Māori/Pacific Island status at last follow-up was was a statistically significant population as suggested by retrieved from the electronic difference in the proportion previous studies from Auckland. records. The R language and of females between Māori and Study design Environment for Statistical non-Māori (80.1% and 69.8% A retrospective review of Computing (release 3.5.1) was respectively, p=0.036). The mean meningioma cases operated used for data preparation, age at surgery for all patients

was 57.9 with a standard devi- University of Auckland, Auckland, The adjusted hazard ratio of ation of 13.4. There was also a 3 Waikato District Health Board, having metastatic recurrence 4 significant difference between Hamilton, School of Medicine, The for post-menopausal women the mean ages of surgery in University of Auckland, Auckland compared to pre-menopausal Māori (54.2) vs non-Māori Aims women was 1.39 (95% CI: 1.06- (59.8). Transitional menin- This study aims to examine 1.82, p-value=0.019) for ER+ and/ giomas were the most common the association of menopausal or PR+ cases. Age did not affect subtype. Out of all the spec- status and risk of metastatic the risk of metastatic relapse for imens, 79.5% were WHO grade relapse for women diagnosed ER+ and/or PR+ cases but affect I, 19.1% were WHO grade II and with stage I–III breast cancer at the risk for ER- and PR- cases 1.4% was WHO grade III. There the age of 45–55 years in New with an adjusted hazard ratio of was a significant difference Zealand. 0.94 per year (95% CI: 0.89–1.00, between the WHO grades by Methods p-value=0.045). sex (p=0.004 for trend), 29.5% We included women diag- Conclusions of meningioma cases in males nosed with stage I–III breast Menopausal status has an were WHO grade II compared cancer at the age of 45–55 years impact on the risk of metastatic to 15.2% in females. However, with recorded menopausal recurrence for women diag- there was no significant status at time of diagnosis in nosed with ER+ and/or PR+ stage difference between WHO grades the Auckland and Waikato I–III breast cancer at the age of and overall survival with regard Breast Cancer Register. Cumu- 45–55 years, but no impact on to ethnicity. lative incidence of distant those with ER- and PR- disease. Conclusion metastatic recurrence was On the other hand, age is an The data derived from this examined with the Kaplan– important influence factor study provides an overview of Meier method by age group on risk of metastatic relapse incidence and demographics (45–49 years and 50–55 years) for women with ER- and PR- of meningioma in the Midland and by menopausal status disease, but not for ER+ and/PR+ region. It should be noted that (pre-menopausal, peri-meno- cancers. the study cohort comprised pausal and post-menopausal) of only the histologically after stratifying the data by hormone receptor status: (1) Where has all the flour confirmed meningiomas and gone? An analysis of does not include the patients estrogen receptor positive (ER+) who had imaging findings and/or progesterone receptor consumer purchasing consistent with meningioma positive (PR+) and (2) ER- and of flour products during who did not have surgery. PR-. Cox proportional hazards the COVID-19 pandemic The study shows comparable model was used to estimate 1 2 trends with data derived the adjusted hazard ratio of Leandrie Young, Judith McCool, 2 from studies from Auckland developing recurrent metastatic Alistair Woodward other parts of the world. This breast cancer by menopausal 1Medical student, University of study also confirmed a large status after adjustment for age, Auckland, Auckland, 2School of difference between the annu- ethnicity, year of diagnosis, Population Health, University alised age-adjusted incidence socioeconomic status, public/ of Auckland, Auckland between Māori and non-Māori private hospital treatment, Aim population. The obvious health mode of detection, cancer stage In December 2019, a new inequality between Māori and cancer grade. coronavirus (SARS-CoV-2) was reported in Wuhan, China. and non-Māori in this regard Results On 25 March 2020, a state of needs to be explored further We have identified 5,309 emergency was declared in to determine the likely causal women diagnosed with stage New Zealand and the country factor—whether sociocultural I–III breast cancer at the age of would be in lockdown for a and/or biological. The results 45–55 years: 2,799 pre-meno- minimum of four weeks. Media also create the foundation pausal, 929 peri-menopausal reported shelves being emptied for further studies exploring and 1,581 post-menopausal. of various products, including genetic and epigenetic charac- The Kaplan–Meier curves flour. Our aim was to study the teristics of meningioma in New showed significant difference demand in New Zealand and to Zealand. in risk of metastatic recurrence consider possible reasons for between different meno- observed changes. Association of pausal status for ER+ and/ menopausal status or PR+ cases (p-value=0.047), Method with a 10-year cumulative A comprehensive literature and risk of metastatic incidence of metastatic recur- search was undertaken as relapse of breast cancer rence of 11.2% (95% confidence a student research elective interval (CI): 9.6%–12.9%) during lockdown. Sales data Chunhuan Lao,1 Mark Elwood,2 for pre-menopausal women, were obtained from New Marion Kuper-Hommel,3 12.4% (95% CI: 9.4%–15.3%) for Zealand’s largest grocery Ian Campbell,3,4 Ross Lawrenson1,3 peri-menopausal women and distributor, Foodstuffs, 1 Medical Research Centre, The 15.6% (95% CI: 13.0%–18.1%) including flour, in store baking University of Waikato, Hamilton, for post-menopausal women. and proprietary bread sales 2 School of Population Health, The

leading up to and during the abdomen from a bitumen truck Methods COVID-19 lockdown (and for backfire that poured bitumen Patients with a large AAA same period in 2019). over him. (5cm for females, 5.5cm for males) or those that had an Results In the absence of peanut oil, open aneurysm repair (OAR) or An increase of approximately butter was initially used but endovascular aneurysm repair 136% in sales of flour occurred application was difficult due to (EVAR) were included in the during the week preceding the solid nature of the agent, analysis. Three international lockdown. Bread sales peaked and, subsequently, when the datasets were used for external at the same time. In-store butter liquefies, the bedside and validation: New Zealand (two baking sales peaked on the patient’s clothing were covered centres: OAR, EVAR and high-risk 22 March followed by a rapid with the agent. non-operated patients), Nether- decline. An alternative, using peanut lands (one centre: EVAR patients Conclusion butter, was recommended by only) and Singapore (one Consistent with interna- our senior author, and the semi- centre: OAR and EVAR patients). tional evidence, people in New liquid nature of this agent made Discrimination using receiver Zealand engaged in a form application much easier, with operating curves (ROC) and of panic buying during the removal using a wooden spatula c-statistics>0.7 was considered lockdown pandemic. Panic simple, painless and neater sufficiently accurate. Patients buying is likely fuelled by stress than butter without peripheral survival was confirmed from and previous research suggests waste. national mortality databases and it is driven by self-preser- all patients have completed at vation. Public and social media We have received much least five years’ follow-up. function as an amplifier of this positive feedbacks from emer- Results stress and, as a result, a similar gency nurses, burn clinical There were 1,004 patients in trend in consumer purchasing nurse specialist and the patient. the New Zealand cohort, and the has occurred globally. A We would like to share this ROC for the model in predicting different degree of response positive experience for dissem- 30 days and two, four and five has been observed in different ination of a safe and effective year survival was 0.76, 0.67, supermarket chains, suggesting way to manage bitumen burn in 0.74 and 0.76, respectively. that changes in consumer the medical community. The corresponding ROC for the purchasing are greater in Singaporean database (n=301 certain populations. The data An international patients) was 0.88, 0.80, 0.80 and also indicate a global interest validation of 0.82. There were 381 patients in baking. However, whether from Netherlands, and ROC for the flour was used remains an abdominal 30 days and one, two and five unclear. aortic aneurysm years was 0.73, 0.67,0.70 and survival model 0.66, respectively. An unusual use of 1,2 3 Khashram M, Williman JA, Conclusions 4 5 6 peanut butter to treat Khashram Z, Vincent ZL, Kian CJ, The AAA DES model appears 6 7 5 bitumen burns Xuan ZTZ, Oliveira-Pinto J, Hill AA, to predict 30-day survival well Verhagen HJ7 Dr Sarah Cox,1 Mr Eric Tan2 in all datasets and accurate 1Department of Surgery, Univer- medium-term survival in the 1 Registrar, Plastics Department, 2 sity of Auckland, Department of New Zealand and Singaporean Waikato Hospital, Hamilton, 2Plastic Vascular Surgery, Waikato Hospital, cohorts. Despite differences in and Reconstructive Surgery Depart- 3 Hamilton, Biostatistics and Compu- study population, the model ment, Waikato Hospital, Hamilton tational Biology Unit, University predicted early survival well Bitumen burn is a unique of Otago, Christchurch, 4Phoenix in the Dutch population, but chemical burn, usually Creative Solutions, Inc., Toronto, 5 medium-term survival was not sustained as a result of work- Canada , Department of Vascular accurate. This might be due to place related occupational Surgery, Auckland City Hospital, Auckland, 6Department of Vascular patient selection and type of hazard. Surgery, Singapore General Hospital, AAA repair. This model requires 7 The traditional mainstay of Singapore, Department of Vascular further external validation to removal in Australasia is with Surgery, Erasmus University Medical test the broader clinical utility. Center, Rotterdam, The Netherlands peanut oil, but with heightened Background Assessing validity of questions awareness of peanut allergy in the Harti Hauora Tamariki at workplace, it is no longer Management of patients Tool, a holistic, whanau-centred available in our district health with large abdominal aortic (family-centred) health screening tool, used in the Harti-Hauora board. aneurysms can be challenging despite international guidelines randomised controlled trial We present our experience that recommend treatment at Abigail Weaver,1 Dr Amy Jones,2 of managing an acute bitumen certain diameter thresholds. Dr Nina Scott,2 Dr Polly-Atatoa Carr,2 burn in a 70-year-old man The aim of this study was to Dr Bridgette Masters-Awatere2 with short bowel syndrome externally validate a previously 1Medical student, University of who sustained this injury over validated a discrete simulation Auckland, Auckland, 2Te Puna his mouth, face, chest and AAA predictive model. Oranga (Maori Health Unit), Waikato District Health Board, Hamilton

Aims Giant cell BCC: an or daughters and was divorced This study assessed the unusual case report from his wife two years prior. validity of questions in a He only has his more elderly screening tool called Harti highlighting the sister to check in with, but, Hauora Tamariki tool. These perils of neglecting unfortunately, she had her own questions aim to identify modi- low-risk skin cancer health issues to contend with fiable determinants of health and could not give him more Luxi Sun,1 Eric xTan1 and provide enhanced access careful attention. Remarkably, 1 to care. Plastic and Reconstructive he has a fully clinically treatable Surgery Department, Waikato cancer, for which earlier identi- Methods Hospital, Hamilton Four hundred and eighty-six fication and encouragement to Background engage with healthcare services participants were collated by Basal cell carcinomas would have significantly ethnicity (Māori and non-Māori). (BCCs) are the most common reduced his risk of morbidity. The analysis uses comparative type of non-melanomatous Although surgical planning is of population data to identify esti- skin cancers, typically with a great relief to the patient, upon mated positive predictive values slow and indolent course of follow-up, he was far more and compares these values to the progression. The incidence of grateful for his placement in a Harti data to estimate validity. BCC was estimated as 299 per rest home ultimately. As clini- Yield,which is the proportion of 100,000. Giant BCCs account for cians we are often more focused Harti participants who identified less than 1% of BCC with 50% on the medical or surgical need, was also measured. risk of regional spread and 30% process of relieving a patient distant metastasis risk when Results of their disease burden, but we they are more than 15cm. Out of the 12 questions cannot lose sight of the huge analysed, 8 were identified Our patient is a 74-year-old impact that emotional wellbeing as valid, 3 as non-valid and Caucasian gentleman who can play to a patient’s quality 1 question was classified as presented with two fungating of life. having uncertain validity. The lesions over his left back three low validity questions (15cm in diameter) and left were regarding alcohol, drug The unmet need of posterior elbow (6cm) that have use and GP enrolment. High-va- diabetic foot disease steadily grown over the last lidity and high-yield questions two years. On initial referral, in New Zealand were smoking, immunisations he also complained of getting 1,2 and oral-health categories. The Dr Odette Hart, weight loss, night sweats and 3 3 three question topics considered Ms Tass Borman, Ms Claire O’Shea, poor appetite with anaemia on 4 valid due to high yield were A/Prof Michael Jameson, investigation. He lives alone 1,2 enrolment to Well Child Mr Manar Khashram without support. A giant BCC 1 providers, participation in B4 Department of Vascular Surgery, was confirmed, and the patient School Checks and disclosure of Waikato District Health Board, underwent wide local resection 2 domestic abuse. Hamilton, Department of Surgery, of a tumour resulting in a skin University of Auckland, Auckland, Discussion deficit of almost 30cm—some- 3Department of Diabetology, Validating this tool identified thing that could have been Waikato District Health Board, 4 questions applicable across a avoided if the BCC could have Hamilton, Waikato Clinical Campus, range of settings. Importantly, been primarily excised when it University of Auckland, Hamilton this study identified a simple was much smaller in dimension. Introduction method for assessing the It is estimated that 25% of Conclusion validity of screening questions. patients with diabetes will Giant BCCs are rare subtypes This method can be included develop diabetic foot ulcers of BCC that are commonly in quality improvement and during their lifetime, with a reported with significant ensure that screening ques- high rate of recurrence (>50% local, regional and distant tions are able to identify unmet after three years). The sequel of implications. Patients and the needs, while at the same time major limb amputation due to public should be made aware avoiding asking questions that diabetic foot disease is common, that, although the majority are of little value and add time and has a major psychological, of BCCs are relatively slow to to the screening process. functional and financial impact. grow and do not metastasise, Conclusion if left untreated, these BCCs Methods The majority of questions can become giant BCCs with This prospective observa- were valid. They were effective different prognostic profiles and tional study collected patient, in identifying needs allowing treatment management. wound and outcome data from for referrals. Some questions March to September 2020 will be reconsidered and altered Upon discussion with the (excluding stage 4 COVID-19-re- for future versions of the tool. patient, he stated that the main lated lockdown) on all diabetic The two questions requiring reason he did not attend for patients reviewed at the diabetic alteration are alcohol and drug review to his GP was a lack foot ward round at Waikato concern. These questions will be of social support and embar- Hospital, New Zealand, with at altered to ask quantity of use. rassment. He does not have sons least one month follow-up. This

weekly multidisciplinary ward days, and 30% resided outside Conclusion round comprises a vascular the Waikato District Health This prospective pilot study surgeon and team, an endocri- Board region. One third of highlights the ethnic inequity nologist and team, a podiatrists patients were re-admitted with and management challenges specialised in diabetic foot an index limb issue. There were of diabetic foot disease within disease, a diabetes nurse, a 19 patients on dialysis; of those, the Waikato region. Further orthoptist and a vascular charge 18 patients identified as Māori. research is required define nurse. these issues and facilitate Twenty-one patients improvements in reducing Results underwent a major limb ampu- major limb amputation and There were 71 patients and tation, with a higher proportion mortality associated with 121 limbs treated over 104 of Māori (13/21) than non-Māori diabetic foot disease. presentations. The average age (8/21) patients. Ten (14%) was 64.5 years, 53 (75%) were patients died during the study male and 36 (51%) patients period, with 60% of deaths identified as Māori. The average occurring within 30 days of length of stay under the admission. vascular surgery team was 12.9

URL: www.nzma.org.nz/journal-articles/proceedings-of-the-waikato-clinical-campus-research-seminar-wednesday-14-october-2020

Consensus statement on the treatment of transplant- eligible patients with newly diagnosed multiple myeloma in New Zealand Nicole Chien, Ken Romeril, Bart Baker, Hugh Goodman, Henry Chan, on behalf of the Myeloma Interest Group Published: 18 December 2020 (Vol 133 No 1527)

In Table 1, the recommended dosage of Bortezomib for the VTD induction regimen was incorrectly reported as 1.3mg/m2. The actual recommended dosage is 1.5mg/m2. This error was resolved online and in the PDF on 13 January 2021.

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]