Background Health technology assessment bodies increasingly emphasise the importance of preference-weighted health-related quality of life (HRQoL) evidence. However, such measures are often absent in clinical trial publications. It is not yet clear how frequently clinical trials have incorporated these measures over the past five decades, how the use of preference-weighted HRQoL instruments has evolved over time, and how trends differ across disease areas, countries and global regions. This study aims to (1) assess changes over time in the proportions of clinical trials using each preference-weighted HRQoL instrument in adults, and (2) model secular trends in the adoption of these instruments across disease areas, countries and regions. The study will provide a comprehensive, systematic assessment of the use of preference-weighted HRQoL instruments in clinical trials since 1976 and develop a scalable approach for large-scale evidence synthesis.Methods We will identify clinical trials involving humans published in English since 1976 through systematic searches of MEDLINE, Embase, Cochrane Library and Web of Science. We will focus on generic preference-weighted HRQoL instruments for adults, including EQ-5D-3L, EQ-5D-5L, Short Form 6 Dimensions, 12-Item Short Form Health Survey (SF-12), Health Utility Index 2, Health Utility Index 3, Assessment of Quality of Life (AQoL) series (AQoL-4D, AQoL-6D, AQoL-7D, AQoL-8D), Quality of Well-Being Scale (QWB), QWB Self-Administered (QWB-SA), 15D and Patient-Reported Outcomes Measurement Information System (PROMIS) with the Preference Scoring System (PROPr). Screening and data extraction will be automated using natural language processing (NLP) pipeline or large language models (LLMs). To determine the most accurate approach, we will benchmark NLP and LLM performance against a manually curated reference dataset of 5000 randomly sampled articles reviewed independently by three reviewers. Model performance will be evaluated using classification metrics including accuracy, recall and F1-score. Annual counts and proportions of trials using each instrument will be calculated, stratified by disease area, country and region. Trends will be modelled using basis-splines (B-splines) with 2 or 3 degrees of freedom and Bayesian spline regression to estimate secular changes in both absolute numbers and proportions of instrument use over time.Ethics and dissemination This study uses only published literature and does not involve human participants or individual-level data. All results will be reported in aggregate form, with no identifiable information. Formal ethics approval is therefore not required. Findings will be disseminated via peer-reviewed publications and conference presentations, and aggregated data and analysis code will be made publicly available to support transparency and reproducibility.

Health policies must be continually updated as new evidence is generated to ensure the optimal delivery of health interventions and prevention measures. Models are often used to study health problems, but their complexity limits their use by policy-makers. One way to facilitate their use among non-modellers is to develop user-friendly interfaces and make them available online. We conducted a scoping review of journal articles to identify and describe the currently available, interactive, freely available web-based health models that aim to inform health policy relevant to any disease or health issue affecting human populations. This scoping review included 16 web-based models covering 13 diseases or health issues, of which the most common were coronavirus disease 2019 (COVID-19) and malaria. The most common model outputs were epidemiological indicators (14/16), such as case numbers, incidences, or results from diagnostic screening, followed by the cost of implementing the intervention or health measure of interest (10/16). Model validation was performed in 6 of the 16 studies by comparing the model results with the previously published evidence or comparing simulated outcomes with observed data. Sensitivity and scenario analyses were conducted for 62.5

Background. While severe outcomes among hospitalized patients with COVID-19 and influenza are well described, comparative studies are lacking on community transmission and milder illnesses associated with COVID-19 and influenza. Methods. This study is based on a prospective community cohort in Wellington, New Zealand, consisting of participants with acute respiratory illness associated with COVID-19 and influenza, as confirmed by polymerase chain reaction. From 7 February to 2 October 2022, we compared the incidence, risk/protective factors, and clinical features among them. Results. The crude incidence of COVID-19-associated acute respiratory illness was 59 per 100 person-years (PY). The adjusted cumulative incidence for COVID-19 (77/100 PY; 95% CI, 75-80) was 4.5 times higher than for influenza (17/100 PY; 95% CI, 15-19). Among all COVID-19 cases, the proportion of children aged 0 to 17 years with COVID-19 was substantial but smaller than those of influenza (402/1229 [33%] vs 173/255 [68%], P < .0001). The highest incidence of COVID-19 was among adolescents aged 12 to 17 years (109/100 PY; 95% CI, 97-119) and individuals who were European and other ethnicity (83/100 PY; 95% CI, 80-86), whereas the highest influenza incidence was among children aged 1 to 4 years (49/100 PY; 95% CI, 40-58) and M & amacr;ori (35/100 PY; 95% CI, 28-43). Adolescents aged 12 to 17 years had 2.5-times higher peak COVID-19 incidence (5.9/100) than adults aged >= 18 years (2.4/100). Adolescents with 2 doses of the COVID-19 vaccines had 75% greater risk of COVID-19 infection (hazard ratio, 1.75; 95% CI, 1.40-2.20) as compared with adults with 3 doses. Vaccination, age, ethnicity, and household size were independent protective/risk factors for COVID-19 or influenza. Participants with COVID-19, as compared with influenza, were less likely to access health care or experience febrile and severe illnesses but more likely to report sore throat, headache, myalgia, and taste or smell loss. Conclusions. As the world transitions to COVID-19 endemicity, estimating disease burdens in community settings becomes important to understand complete disease pyramids, risk factors, and clinical progression for informing countermeasures.

Background: Studying the transmissibility of the SARS-CoV-2 and its driving factors is valuable due to the ongoing emergence of new variants. We examined the household transmission characteristics of the Omicron-dominant variant. Methods: The study took place in Wellington, New Zealand, from 7-February to 2-October 2022. When an individual had a confirmed case of SARS-CoV-2, all household members were instructed to take a swab every 3rd day until two consecutive negative swabs were obtained from the last person with SARS-CoV-2. Participants were monitored daily for acute respiratory symptoms until no further symptoms. Results: We enrolled 75 households with confirmed cases of SARS-CoV-2 (total enrolled individuals = 306). The median serial interval was 4 days, and the secondary attack rate (SAR) was 60.6 % (95 % confidence interval (CI) 54.1-66.7). The SAR was higher for older household contacts and lower for contacts of fully vaccinated index cases. The SAR was highest when both the index case and the contacts were unvaccinated or received only one vaccine dose (88 %). Conclusions: In this household cohort, the SARS-CoV-2 Omicron variant showed a high SAR which was modifiable by vaccination; the vaccination status of the index case and household members significantly reduced this.

Artificial intelligence (AI) and machine learning (ML) pose enormous potential for improving quality of life. It can also generate significant social, cultural and other unintended risks. We aimed to explore fairness concepts that can be applied in ML models for disease prediction from key health experts' perspectives in an ethnically diverse high-income country. In-depth interviews with key experts in the health sector in Aotearoa New Zealand (NZ) were implemented between July and December 2022. We invited participants who are key leaders in their ethnic communities, including Maori (Indigenous), Pasifika and Asian. The interview questionnaire comprised six sections: (1) Existing attitudes to healthcare allocation; (2) Existing attitudes to data held at the general practitioner (GP) level; (3) Acceptable data to have at the GP level for disease prediction models; (4) Trade-offs for obtaining benefits vs generating unnecessary concern in deploying these models; (5) Reducing bias in risk prediction models; and (6) Including community consensus into disease prediction models for fair outcomes. The study shows that participants were strongly united in the view that ML models should not create or exacerbate inequities in healthcare due to biased data and unfair algorithms. An exploration of fairness concepts showed that carefully selected data types must be considered for predictive modelling and that trade-offs for obtaining benefits versus generating unnecessary concern produced conflicting opinions. The participants expressed high acceptability for using ML models but expressed deep concerns about inequity issues and how these models might affect the most vulnerable communities (such as Maori in middle-ages and above and those living in deprived communities). Our results could help inform the development of ML models that consider social impacts in an ethnically diverse society. ### Competing Interest Statement The authors have declared no competing interest. ### Funding Statement This work was funded by the Royal Society Te Apārangi. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the report. ### Author Declarations I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained. Yes The details of the IRB/oversight body that provided approval or exemption for the research described are given below: The study was approved by University of Otago ethics approval processes, reference number HD20/012 and D22/101. There were no patients directly involved in this study. I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals. Yes I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance). Yes I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable. Yes All data are confidential and not available for sharing.

BACKGROUND:Public health surveillance systems need to monitor influenza activity and guide measures to mitigate its high impact on morbidity, mortality and healthcare systems. There is an increasing expectation that surveillance data will support the modeling of future short-term disease scenarios using artificial intelligence (AI) and machine learning (ML). This study examines how influenza surveillance can support AI/ML-based short-term forecasting for influenza at the community and hospital levels in a high-income country setting (Aotearoa/New Zealand). METHODS:This study used a two-phase approach. The first phase involved a comprehensive review of government reports, official websites, and literature to characterize existing influenza surveillance systems. The second phase evaluated systems against eight key attributes-timeliness, sensitivity, specificity, representativeness, coverage, robustness, completeness, and historical data-using a five-level ranking system. Attribute selection was informed by experts' knowledge, ML requirements, and established frameworks. Weighted scores for training and short-term forecasting capabilities were calculated to determine alignment with AI/ML requirements. RESULTS:The Southern Hemisphere Influenza and Vaccine Effectiveness Research and Surveillance (SHIVERS) community cohort and Severe Acute Respiratory Infection (SARI) hospital surveillance emerged as the most useful systems, achieving the highest scores in both training and short-term forecasting in community and hospital settings, respectively. The National Minimum Dataset of hospitalizations and mortality datasets demonstrated strong training potential but are limited in short-term forecasting due to timeliness constraints. Additionally, laboratory-based surveillance performs a useful role in bridging community and hospital datasets. CONCLUSIONS:A set of key attributes is useful for assessing which influenza surveillance systems are best aligned with AI/ML training and short-term forecasting requirements. These attributes distinguished systems that are likely to be the most suitable for modeling future short-term disease scenarios for influenza at the community and hospital levels in New Zealand. Integrating these data sources could enhance influenza forecasts to improve public health responses and intervention planning.

Increases in the use of Bayesian inference in applied analysis, the complexity of estimated models, and the popularity of efficient Markov chain Monte Carlo (MCMC) inference under conjugate priors have led to more scrutiny regarding the specification of the parameters in prior distributions. Impact of prior parameter assumptions on posterior statistics is commonly investigated in terms of local or pointwise assessments, in the form of derivatives or more often multiple evaluations under a set of alternative prior parameter specifications. This paper expands upon these localized strategies and introduces a new approach based on the graph of posterior statistics over prior parameter regions (sensitivity manifolds) that offers additional measures and graphical assessments of prior parameter dependence. Estimation is based on multiple point evaluations with Gaussian processes, with efficient selection of evaluation points via active learning, and is further complemented with derivative information. The application introduces a strategy to assess prior parameter dependence in a multivariate demand model with a high dimensional prior parameter space, where complex prior-posterior dependence arises from model parameter constraints. The new measures uncover a considerable prior dependence beyond parameters suggested by theory, and reveal novel interactions between the prior parameters and the elasticities.

ABSTRACT Background In the age of big data, linked social and administrative health data in combination with machine learning (ML) is being increasingly used to improve prediction in cardiovascular diseases (CVD). We aimed to apply ML methods on extensive national-level health and social administrative datasets to predict future diabetes complications by ethnicity. Methods Five ML models were used to predict CVD events among all people with known diabetes in the population of New Zealand, utilizing national-level administrative data at the individual level. Results The Xgboost ML model had the best predictive power for predicting CVD events three years into the future among the population with diabetes. The optimization procedure also found limited improvement in AUC by ethnicity. The results indicated no trade-off between model predictive performance and equity gap of prediction by ethnicity. The list of variables of importance was different among different models/ethnic groups, for examples: age, deprivation, having had a hospitalization event, and the number of years living with diabetes. Discussion and conclusions We provide further evidence that ML with administrative health data can be used for meaningful future prediction of health outcomes. As such it could be utilized to inform health planning and healthcare resource allocation for diabetes management and the prevention of CVD events. Our results may suggest limited scope for developing prediction models by ethnic group and that the major ways to reduce inequitable health outcomes is probably via improved delivery of prevention and management to those groups with diabetes at highest need.

Aim We aimed to combine Global Burden of Disease (GBD) Study data and local data to identify the highest priority intervention domains for preventing cardiovascular disease (CVD) in the case study country of Aotearoa New Zealand (NZ). Methods Risk factor data for CVD in NZ were extracted from the GBD using the “GBD Results Tool.” We prioritized risk factor domains based on consideration of the size of the health burden (disability-adjusted life years [DALYs]) and then by the domain-specific interventions that delivered the highest health gains and cost-savings. Results Based on the size of the CVD health burden in DALYs, the five top prioritized risk factor domains were: high systolic blood pressure (84,800 DALYs; 5400 deaths in 2019), then dietary risk factors, then high LDL cholesterol, then high BMI and then tobacco (30,400 DALYs; 1400 deaths). But if policy-makers aimed to maximize health gain and cost-savings from specific interventions that have been studied, then they would favor the dietary risk domain (e.g., a combined fruit and vegetable subsidy plus a sugar tax produced estimated lifetime savings of 894,000 health-adjusted life years and health system cost-savings of US$11.0 billion; both 3% discount rate). Other potential considerations for prioritization included the potential for total health gain that includes non-CVD health loss and potential for achieving relatively greater per capita health gain for Māori (Indigenous) to reduce health inequities. Conclusions We were able to show how CVD risk factor domains could be systematically prioritized using a mix of GBD and country-level data. Addressing high systolic blood pressure would be the top ranked domain if policy-makers focused just on the size of the health loss. But if policy-makers wished to maximize health gain and cost-savings using evaluated interventions, dietary interventions would be prioritized, e.g., food taxes and subsidies.

ObjectivesTo optimise planning of public health services, the impact of high-cost users needs to be considered. However, most of the existing statistical models for costs do not include many clinical and social variables from administrative data that are associated with elevated health care resource use, and are increasingly available. This study aimed to use machine learning approaches and big data to predict high-cost users among people with cardiovascular disease (CVD).MethodsWe used nationally representative linked datasets in New Zealand to predict CVD prevalent cases with the most expensive cost belonging to the top quintiles by cost. We compared the performance of four popular machine learning models (L1-regularised logistic regression, classification trees, k-nearest neighbourhood (KNN) and random forest) with the traditional regression models.ResultsThe machine learning models had far better accuracy in predicting high health-cost users compared with the logistic models. The harmony score F1 (combining sensitivity and positive predictive value) of the machine learning models ranged from 30.6% to 41.2% (compared with 8.6-9.1% for the logistic models). Previous health costs, income, age, chronic health conditions, deprivation, and receiving a social security benefit were among the most important predictors of the CVD high-cost users.ConclusionsThis study provides additional evidence that machine learning can be used as a tool together with big data in health economics for identification of new risk factors and prediction of high-cost users with CVD. As such, machine learning may potentially assist with health services planning and preventive measures to improve population health while potentially saving healthcare costs.

New Zealanders can cross borders freely, work and live in Australia indefinitely thanks to the Trans-Tasman Travel Agreement. This paper uses a recently developed decomposition method to decompose the weekly wage gap at various quantiles on the wage distribution between New Zealand-born (NZ-born) and Australian-born workers, and between NZ-born workers, migrants from other English speaking countries (OESC), and from non-English speaking countries (NESC) to determine how free and regulated migration influences migrants' performance in the Australian labour market. We found that NZ-born workers earned higher weekly wages than both Australian-born and NESC workers but earned lower wages than OESC migrants. Differences in endowment were primarily responsible for the wage gaps between NZ-born and Australian-born workers and between NZ-born and OESC migrants. However, differences in returns to worker and job characteristics are mainly responsible for the wage gap between NZ-born and NESC migrants.

Poor diet is a major risk factor for excess weight gain and obesity-related diseases, including cardiovascular diseases, type 2 diabetes mellitus, osteoarthritis and several cancers. This paper aims to assess the potential impacts of real-world food and beverage taxes on change in dietary risk factors, health gains (in quality-adjusted life years (QALYs)), health system costs and greenhouse gas (GHG) emissions as if they had all been implemented in New Zealand (NZ). Ten taxes or tax packages were modelled. A proportional multistate life table model was used to predict resultant QALYs and costs over the remaining lifespan of the NZ population alive in 2011, as well as GHG emissions. QALYs ranged from 12.5 (95% uncertainty interval (UI) 10.2 to 15.0; 3% discount rate) per 1000 population for the import tax on sugar-sweetened beverages (SSB) in Palau to 143 (95% UI 118 to 171) per 1000 population for the excise duties on saturated fat, chocolate and sweets in Denmark, while health expenditure savings ranged from 2011 NZ$245 (95% UI 188 to 310; 2020 US$185) per capita to NZ$2770 (95% UI 2140 to 3480; US$2100) per capita, respectively. The modelled taxes resulted in decreases in GHG emissions from baseline diets, ranging from -0.2% for the tax on SSB in Barbados to -2.8% for Denmark's tax package. There is strong evidence for the implementation of food and beverage tax packages in NZ or similar high-income settings.

This study aimed to identify dietary trends in Aotearoa New Zealand (NZ) and whether inequities in dietary patterns are changing. We extracted data from the Household Economic Survey (HES), which was designed to provide information on impacts of policy-making in NZ, and performed descriptive analyses on food expenditures. Overall, total household food expenditure per capita increased by 0.38% annually over this period. Low-income households spent around three quarters of what high-income households spent on food per capita. High-income households experienced a greater increase in expenditure on nuts and seeds and a greater reduction in expenditure on processed meat. There was increased expenditure over time on fruit and vegetables nuts and seeds, and healthy foods in Māori (Indigenous) households with little variations in non-Māori households. But there was little change in processed meat expenditure for Māori households and expenditure on less healthy foods also increased over time. Routinely collected HES data were useful and cost-effective for understanding trends in food expenditure patterns to inform public health interventions, in the absence of nutrition survey data. Potentially positive expenditure trends for Māori were identified, however, food expenditure inequities in processed meat and less healthy foods by ethnicity and income continue to be substantial.

Background: Health demoting consumption of alcohol and tobacco are some of the most important risk factors for health loss worldwide, however there is limited information on these consumption risk factors in New Zealand (NZ) and whether inequities in the risk factors are ethnically patterned. Methods: We used three nationally representative Household Economic Survey waves (2006/07, 2009/10, 2012/ 13) (n = 9030) in NZ to examine household expenditure for key health risk-related components of consumption by ethnicity, and its contributors to the differences using non-parametric, parametric and decomposition methods. Results: Ma over bar ori households (NZ indigenous population) were significantly poorer (25% less) than non-Ma over bar ori households in terms of household per capita expenditure. However, our various econometric estimations suggested that, in relative terms, Ma over bar ori spent more on tobacco and alcohol, and less on healthcare. The gaps become larger at upper quantiles of the budget share distributions; the composition effect (the gap due to differences in individual and household characteristics between Ma over bar ori and non-Ma over bar ori) explains most of the tobacco and alcohol budget share gap between the two groups, and less for healthcare. The structure effect (the gap due to returns to/ or effect of individual and household characteristics) contributes very little to the budget share gap for tobacco and drink, but increasingly and predominantly when moving along the distribution of healthcare budget share. The differences between Ma over bar ori and non-Ma over bar ori in household ownership, education, and income negatively affect budget share on these health demoting consumption (tobacco and alcohol). The household head's age, education, and employment contributed most to the structure effect. Conclusions: Our study suggested ethnic inequities in the health risk consumption behaviour are evidenced in NZ. Interventions targeting education and employment that significantly affect household budget shares on risk factors (i.e., harmful consumption) for health loss may help narrow the gaps.

Policies to mitigate climate change are essential. The objective of this paper was to estimate the impact of greenhouse gas (GHG) food taxes and assess whether such a tax could also have health benefits in Aotearoa NZ. We undertook a systemised review on GHG food taxes to inform four tax scenarios, including one combined with a subsidy. These scenarios were modelled to estimate lifetime impacts on quality-adjusted health years (QALY), health inequities by ethnicity, GHG emissions, health system costs and food costs to the individual. Twenty-eight modelling studies on food tax policies were identified. Taxes resulted in decreased consumption of the targeted foods (e.g., −15.4% in beef/ruminant consumption, N = 12 studies) and an average decrease of 8.3% in GHG emissions (N = 19 studies). The “GHG weighted tax on all foods” scenario had the largest health gains and costs savings (455,800 QALYs and NZD 8.8 billion), followed by the tax—fruit and vegetable subsidy scenario (410,400 QALYs and NZD 6.4 billion). All scenarios were associated with reduced GHG emissions and higher age standardised per capita QALYs for Māori. Applying taxes that target foods with high GHG emissions has the potential to be effective for reducing GHG emissions and to result in co-benefits for population health.

AIM:To provide preliminary high-level modelling estimates of the impact of denicotinisation of tobacco on changes in smoking prevalence in Aotearoa New Zealand relative to the New Zealand Government's Smokefree 2025 goal. METHODS:An Excel spreadsheet was populated with smoking and vaping prevalence data from the New Zealand Health Survey and we projected business-as-usual trends. Using various parameters from the literature (New Zealand trial data, New Zealand EASE-ITC Study results), we modelled the potential impact of denicotinisation of tobacco (with no other tobacco permitted for sale) out to 2025. In addition to the base case (considered most likely), Scenario 1 used estimates from a published expert knowledge elicitation process, and Scenario 2 considered the addition of extra mass-media campaign and Quitline support to the base case. RESULTS:With the denicotinisation intervention, adult daily smoking prevalences were estimated to decline to under 5% by 2025 for the European/Other ethnic grouping (in the base case and both scenarios) and in one scenario (Scenario 1) for Māori (2.5%). However, prevalence did not fall below 5% in the base case for Māori (7.7%) or in Scenario 2 (5.2%). In the base case, vaping was estimated to increase to 7.9% in the adult population by 2025, and up to 10.7% in one scenario (Scenario 1). CONCLUSIONS:This preliminary high-level modelling suggests that mandated denicotinisation has a plausible chance of achieving the New Zealand Government's Smokefree 2025 goal. The probability of success would increase if supplemented with interventions such as mass-media campaigns offering Quitline support (especially if predominantly designed for a Māori audience). Nevertheless, there is much uncertainty with these results and more sophisticated modelling is forthcoming.

Background Myocardial infarction mortality has declined since the 1970s, but contemporary drivers of this trend remain unexplained. The aim of this study was to compare the contribution of trends in event rates and case fatality to declines in myocardial infarction mortality in four high-income jurisdictions from 2002-15. Methods Linked hospitalisation and mortality data were obtained from New South Wales (NSW), Australia; Ontario, Canada; New Zealand; and England, UK. People aged between 30 years and 105 years were included in the study. Age-adjusted trends in myocardial infarction event rates and case fatality were estimated from Poisson and binomial regression models, and their relative contribution to trends in myocardial infarction mortality calculated. Findings 1 947 895 myocardial infarction events from a population of 80.4 million people were identified in people aged 30 years or older. There were significant declines in myocardial infarction mortality, event rates, and case fatality in all jurisdictions. Age-standardised myocardial infarction event rates were highest in New Zealand (men 893/100 000 person-years in 2002, 536/100 000 person-years in 2015; women 482/100 000 person-years in 2002, 271/100 000 person-years in 2015) and lowest in England (men 513/100 000 person-years in 2002, 382/100 000 person-years in 2015; women 238/100 000 person-years in 2002, 173/100 000 person-years in 2015). Annual age-adjusted reductions in event rates ranged from -2.6% (95% CI -3.0 to -2.3) in men in England to -4.3% (-4.4 to -4.1) in women in Ontario. Age-standardised case fatality was highest in England in 2002 (48%), but declined at a greater rate than in the other jurisdictions (men -4.1%/year, 95% CI -4.2 to -4.0%; women -4.4%/year, -4.5 to -4.3%). Declines in myocardial infarction mortality rates ranged from -6.1%/year to -7.6%/year. Event rate declines were the greater contributor to myocardial infarction mortality reductions in Ontario (69.4% for men and women), New Zealand (men 68.4%; women 67.5%), and NSW women (60.1%), whereas reductions in case fatality were the greater contributor in England (60% in men and women) and for NSW men (54%). There were greater contributions from case fatality than event rate reductions in people younger than 55 years in all jurisdictions, with contributions to mortality declines varying by country in those aged 55-74 years. Event rate declines had a greater impact than changes in case fatality in those aged 75 years and older. Interpretation While the mortality burden of myocardial infarction has continued to fall across these four populations, the relative contribution of trends in myocardial infarction event rates and case fatality to declining mortality varied between jurisdictions, including by age and sex. Understanding the causes of this variation will enable optimisation of prevention and treatment efforts. Copyright (C) 2022 The Author(s). Published by Elsevier Ltd.

Bone age is an important index in the measurement of the biological maturity in children. Although many machine learning methods have been developed to classify bone radiographs, a number of limitations like the need for large datasets still exists. In this paper, we propose an efficient method based on convolutional neural networks to automatically predict bone age given radiographs. The proposed approach considers the differences between male and female skeletal development in order to classify the sex from radiographs, in conjunction with the use of a bone age assessment (BAA) network to significantly improve the prediction performance of bone age. The dataset used is from the 2017 Pediatric Bone Age Challenge, which contains 12,611 left-hand radiographs. Pre-processing of the radiographs is first performed, followed by detection of key-points of the hand for the application of affine transformation to register the hand pose. Transfer learning is then applied to the sex determination and BAA models with the pre-trained weights on ImageNet, to leverage on the generic features from millions of real-world images. Three BAA models were trained on three separated portions of the dataset, and utilized with the sex information to obtain the bone age. Results indicate that the method outperforms existing BAA methods as the mean absolute error (MAE) obtained was 5.31 months (0.44 years) while an accuracy of 94.7% was achieved for the sex determination. When sex labels were provided for the testing radiographs, the MAE of our method was further reduced to 4.68 months (0.39 years). The advantage of our method is that it eliminates the need for hand crafting features and utilizes sex as an important information in the prediction of bone age.

Evidence suggests that smartphone apps can be effective in the self-management of weight. Given the low cost, broad reach, and apparent effectiveness of weight loss apps, governments may seek to encourage their uptake as a tool to reduce excess weight in the population. Mass media campaigns are 1 mechanism for promoting app use. However, the cost and potential cost-effectiveness are important considerations. The aim of our study was to use modeling to assess the health impacts, health system costs, cost-effectiveness, and health equity of a mass media campaign to promote high-quality smartphone apps for weight loss in New Zealand. We used an established proportional multistate life table model that simulates the 2011 New Zealand adult population over the lifetime, subgrouped by age, sex, and ethnicity (Māori [Indigenous] or non-Māori). The risk factor was BMI. The model compared business as usual to a one-off mass media campaign intervention, which included the pooled effect size from a recent meta-analysis of smartphone weight loss apps. The resulting impact on BMI and BMI-related diseases was captured through changes in health gain (quality-adjusted life years) and in health system costs. The difference in total health system costs was the net sum of intervention costs and downstream cost offsets because of altered disease rates. An annual discount rate of 3% was applied to health gains and health system costs. Multiple scenarios and sensitivity analyses were conducted, including an equity adjustment. Across the remaining lifetime of the modeled 2011 New Zealand population, the mass media campaign to promote weight loss app use had an estimated overall health gain of 181 (95% uncertainty interval 113-270) quality-adjusted life years and health care costs of –NZ $606,000 (–US $408,000; 95% uncertainty interval –NZ $2,540,000 [–US $1,709,000] to NZ $907,000 [US $610,000]). The mean health care costs were negative, representing overall savings to the health system. Across the outcomes examined in this study, the modeled mass media campaign to promote weight loss apps among the general population would be expected to provide higher per capita health gain for Māori and hence reduce health inequities arising from high BMI, assuming that the intervention would be as effective for Māori as it is for non-Māori. A modeled mass media campaign to encourage the adoption of smartphone apps to promote weight loss among the New Zealand adult population is expected to yield an overall gain in health and to be cost-saving to the health system. Although other interventions in the nutrition and physical activity space are even more beneficial to health and produce larger cost savings (eg, fiscal policies and food reformulation), governments may choose to include strategies to promote health app use as complementary measures.

Population diets have impacts on both human and planetary health. This research aims to optimise a New Zealand (NZ) version of the EAT-Lancet diet and to model the impact of this diet on population health if it was adopted in NZ. The optimisation methods used mathematical equations in Excel to ensure: population diets met the nutritional recommendations; diet-related greenhouse gas (GHG) emissions did not exceed the NZ GHG boundary; and diet costs did not exceed baseline costs of the average diet. The EAT-Lancet diet was also directly mapped onto the NZ adult nutrition survey food groups, as another estimate of a NZ EAT-Lancet diet. Both diets were modelled using a DIET multi-state life-table model to estimate lifetime impacts on quality adjusted life years (QALYs), ethnic health inequities and health system costs. The optimised diet differed greatly from baseline intake with large amounts of fruits and vegetables, some fish but no beef, lamb, pork or poultry. Modelling nationwide adoption of the NZ EAT-Lancet diets generated large health savings (approximately 1.4 million QALYs), and health system cost savings (around NZD 20 billion). A healthy, climate-friendly, cost-neutral diet is possible for NZ and, if adopted, could provide large health gain, cost savings and reductions in ethnic health inequities.