Follow-up methods for retrospective cohort studies in New Zealand

OBJECTIVES: To define a general methodology for maximising the success of follow-up processes for retrospective cohort studies in New Zealand, and to illustrate an approach to developing country-specific follow-up methodologies. METHODS: We recently conducted a cohort study of mortality and cancer incidence in New Zealand professional fire fighters. A number of methods were used to trace vital status, including matching with records of the New Zealand Health Information Service (NZHIS), pension records of Work and Income New Zealand (WINZ), and electronic electoral rolls. Non-electronic methods included use of paper electoral rolls and the records of the Registrar of Births Deaths and Marriages. RESULTS: 95% of the theoretical person-years of follow-up of the cohort were traced using these methods. In terms of numbers of cohort members traced to end of follow-up, the most useful tracing methods were fire fighter employment records, the NZHIS, WINZ, and the electronic electoral rolls. CONCLUSIONS: The follow-up process used for the cohort study was highly successful. On the basis of this experience, we propose a generic, but flexible, model for follow-up of retrospective cohort studies in New Zealand. Similar models could be constructed for other countries. IMPLICATIONS: Successful follow-up of cohort studies is possible in New Zealand using established methods. This should encourage the use of cohort studies for the investigation of epidemiological issues. Similar models for follow-up processes could be constructed for other countries.     

职业性回顾队列研究中的接触估计

本文介绍了职业性回顾队列研究中进行接触估计的基本步骤以及针对不同资料特征采用的具体估计方法,为阐明剂量－反应（接触－反应）关系提供证据。     

Growing Up in New Zealand cohort alignment with all New Zealand births

Objective: To compare the birth characteristics of the Growing Up in New Zealand cohort with those of all New Zealand (NZ) births over a similar time period, and to describe cohort alignment to current NZ births. Method: The Growing Up in New Zealand longitudinal study recruited 6,846 children from before birth via their pregnant mothers who were residing in the greater Auckland and Waikato regions during 2009 and 2010. Data were collected from mothers antenatally and six weeks after their expected delivery date, and from routine perinatal health records. These data were compared to Ministry of Health data for all births in NZ between 2007 and 2010. Results: The proportion of males and singleton births were not statistically different to national births. Compared to national births fewer of the cohort children were born low birth weight (4.9% vs. 6.1%, p<0.0001) or preterm (6.4% vs. 7.4%, p=0.001) and the cohort was expected to be more ethnically diverse than national births. Conclusion: Birth parameters for the cohort were generally closely aligned to all NZ births in 2007–2010. Some statistically significant differences reflected small absolute differences, attributable in some part to cohort recruitment requiring survival to six weeks post expected delivery. Implications: The explicit documentation of the alignment of the cohort to national data provides assurance that the study is well placed to deliver findings that can inform policy development relevant to the diversity of the contemporary NZ child population.     

Assessing the feasibility of retrospective cohort studies

While most epidemiologic cohort studies are preceded by some sort of feasibility study, details of such prior investigations are rarely reported. Yet it is during such feasibility studies that critical decisions are made, such as site selection and definition of exposure. Here we present the details on one such feasibility study, conducted to determine the possibility of a cohort mortality study of workers exposed to ethylene oxide. Issues discussed include methods for estimating sample size and power, for estimating levels of exposure, and for assessing the adequacy of personnel records.     

Challenges in comparing the methods and findings of cohort studies of oral health: the Dunedin (New Zealand) and Pelotas (Brazil) studies

Objective: to systematically compare methods and some findings from two prospective cohort studies of oral health. Methods : This paper describes and compares two such population‐based birth cohort studies of younger adults: the Dunedin Multidisciplinary Health and Development Study (conducted in New Zealand); and the 1982 Pelotas Birth Cohort Study (conducted in Brazil). Results: The two cohorts showed socio‐demographic similarities and differences, with their gender mixes being similar, but their ethnic compositions differing markedly. There were some important similarities and differences in methods. Overall dental caries experience was higher among the Dunedin cohort. Each of the studies has examined the association between childhood‐adulthood changes in socio‐economic status and oral health in the mid‐20s. Both studies observed the greatest disease experience among those who were of low SES in both childhood and adulthood, and the least among those who were of high SES in both childhood and adulthood. In each cohort, disease experience in the upwardly mobile and downwardly mobile groups lay between those two extremes. Conclusions and implications: There are important similarities and differences in both methods and findings. While the need for a degree of methodological convergence in future is noted, the two studies are able to use each other as replicate samples for research into chronic oral conditions.     

Impact of rotavirus vaccine on paediatric rotavirus hospitalisation and intussusception in New Zealand: A retrospective cohort study

BackgroundRotavirus results in a significant burden of hospitalisations and deaths globally. Rotavirus vaccine has been used in New Zealand since July 2014. The aim of this study was to assess the safety and effectiveness of RotaTeq® vaccine in New Zealand between 2006 and 2016.MethodsA national cohort study of 723,695 children aged less than 6 years was carried out using linked administrative datasets. Study outcomes were hospitalisation for intussusception, rotavirus, and all-cause gastroenteritis. Intussusception hospitalisation rates were calculated from 2006 to 2016, and rotavirus and all-cause gastroenteritis hospitalisation rates from 2011 to 2016. We examined the effect of RotaTeq® vaccination on rotavirus and all-cause gastroenteritis hospitalisation rates using Poisson regression. Adjusted incidence rate ratios controlled for sex, year of birth, ethnicity, socioeconomic deprivation, and district health board area.ResultsSignificant reductions in the incidence of rotavirus hospitalisation were seen in all age groups, ethnicities, and deprivation following the introduction of RotaTeq®. There was a 92.6% reduction in hospitalisation incidence in the vaccinated cohort (p < 0.0001). There was also a 48% reduction in all-cause gastroenteritis hospitalisation incidence in the vaccinated cohort (p < 0.0001). The average annual intussusception rate in children aged less than 3 years was 26.2 per 100,000, with no significant change over time (p = 0.847).ConclusionsIn New Zealand the introduction of RotaTeq® resulted in a significant reduction in rotavirus hospitalisation, and a halving in all-cause gastroenteritis hospitalisation. There has been no change in the overall incidence of intussusception or clear change in patterns of cases, although intussusception cases did occur within risk period immediately post vaccine.     

队列研究分层分析方法

分层分析方法是调整混淆因素的重要方法之一.本文介绍了队列研究双暴露水平分层分析方法,重点讨论Mantel—Haenszel检验方法、不均衡性检验和率比的趋势性检验的统计学方法.为了便于读者应用掌握,在计算方法上作了适当的调整和改进.这些方法在慢性病病因研究中有重要的应用价值.     

Secondary enuresis in a birth cohort of New Zealand children

The development of secondary enuresis was examined in a birth cohort of New Zealand children studied to the age of 10 years. By this age, 7.9% of children had developed secondary enuresis. Analysis suggested two risk factors were involved in secondary enuresis. First, children who were late to attain nocturnal bladder control had significantly higher risks of secondary enuresis than children who attained early nocturnal bladder control. Secondly, the child's level of exposure to adverse life events was associated with the onset of secondary enuresis. Children who attained bladder control after 5 years were 3.39 (1.76-6.56) times more likely to develop secondary enuresis than those who attained bladder control before the age of 3 years (P less than 0.001); children who were exposed to four or more life events in a given year were 2.56 (1.18-5.50) times more likely to develop secondary enuresis in that year (P less than 0.05). These results suggest that the rate at which the child acquires primary bladder control acts as a vulnerability factor which determines the child's susceptibility to developing secondary enuresis when exposed to stress.     

Acid rain studies in Australia and New Zealand

This paper reviews the results of the few scattered acid rain studies in Australia and New Zealand. Large scale spatial discussions are not possible, so discussion, focussing on wet deposition, is limited to rainwater acidity and chemistry on a regional and local basis. Rainwater samples were mainly collected on a daily or event basis, and some attempt was made to ensure contamination from dry fallout in all cases. Rainwater acidity can be divided into three regimes: non-tropical urban, non-tropical rural and tropical rural areas with the Tropic of Capricorn the dividing line between tropical and non-tropical. In Sydney, representative of non-tropical urban, local emissions of acid gases, particularly sulphate and nitrate, resulted in an average pH of 4.4. At several non-tropical rural locations, pHs average between 5.0 and 5.7, indicative of global background levels. Tropical rural pHs average about 4.5, due mainly to natural acidity caused by vegetation release of volatile organic acids. These values indicate that rainwater in this region contains from 2 to 12 times less acidity than in the northern hemisphere where acid rain is a problem. Rainwater chemistry is dominated by ocean influences along the coast and soil and vegetation influences inland. Elevated levels of sulphate and nitrate in rainfall of the Latrobe Valley and the Hunter Valley may be due to power station and industrial sources located there, but do not prove to be a problem.     

Comparison of prospective and retrospective methods for haplotype inference in case-control studies

We compare bias and power of three methods for haplotype inference on disease risk using unphased genotype data from a case-control study. We examine the prospective score test of Schaid et al., a novel modification of the prospective estimating equations of Zhao et al. and the retrospective likelihood of Epstein and Satten. We find that all three approaches are roughly comparable when the haplotype effect on disease odds follows a multiplicative model. However, for dominant and recessive models of haplotype effect, the retrospective-likelihood method has increased efficiency with respect to the prospective methods. As all three methods assume haplotype frequencies are in Hardy-Weinberg Equilibrium (HWE), we compare the robustness of each procedure to departures from HWE. We find the prospective methods are robust to departure from HWE, while the retrospective-likelihood method is biased for dominant and recessive models of haplotype effect. To remedy this limitation of the retrospective-likelihood method, we propose a modification that allows for a non-negative fixation index (common to all haplotype pairs) and show it dramatically reduces the bias of the retrospective likelihood when HWE is violated.     

A retrospective cohort study investigating the comparative effectiveness of pneumococcal vaccines against hospitalisation with otitis media and pneumonia in New Zealand

BackgroundSince 2008 New Zealand has used three different formulations of pneumococcal vaccines on the national infant schedule, PCV7, PCV10 and PCV13, switching between PCV10 and PCV13 twice in 10 years. We have used New Zealand’s linkable, administrative health data to examine the comparative risk of otitis media (OM) and pneumonia hospitalisations among children receiving three different pneumococcal conjugate vaccines (PCV).MethodsThis was a retrospective cohort study using linked administrative data. Outcomes were otitis media, all cause pneumonia and bacterial pneumonia related hospitalisation for children in three cohorts representing periods where PCVs transitioned between PCV7, PCV10, PCV13 and back to PCV10 between 2011 and 2017. Cox’s proportional hazard regression was used to provide hazard ratio estimates to compare outcomes for children vaccinated with different vaccine formulations and to adjust for different sub population characteristics.ResultsEach observation period, where different vaccine formulations coincided, and therefore comparable with respect to age and the environment, included over fifty-thousand infants and children. PCV10 was associated with a reduced risk for OM compared with PCV7 (Adjusted HR 0.89, 95 %CI 0.82–0.97). There were no significant differences between PCV10 and PCV13 in risk of hospitalisation with either otitis media or all-cause pneumonia amongst the transition 2 cohort. In the 18 -month follow-up, after transition 3, PCV13 was associated with a marginally higher risk of all-cause pneumonia and otitis media compared to PCV10.ConclusionThese results should offer reassurance about the equivalence of these pneumococcal vaccines against the broader pneumococcal disease outcomes OM and pneumonia.     

Evidence of suboptimal maternal vaccination coverage in pregnant New Zealand women and increasing inequity over time: A nationwide retrospective cohort study

BackgroundAdequate maternal vaccination coverage is critical for the prevention and control of infectious disease outbreaks such as pertussis, influenza, and more recently COVID-19. To guide efforts to increase vaccination coverage this study examined the extent of vaccination coverage in pregnant New Zealand women over time by area-level deprivation and ethnicity.MethodsA retrospective cohort study was used consisting of all pregnant women who delivered between 01 January 2013 and 31 December 2018, using administrative health datasets. Outcomes were defined as receipt of influenza or pertussis vaccination in any one of the relevant data sources (National Immunisation Register, Proclaims, or Pharmaceutical collection) during their eligible pregnancy. Ethnicity was prioritised as Māori (NZ indigenous), Pacific, Asian, and Other or NZ European and deprivation was defined using New Zealand Index of Multiple Deprivation (IMD).ResultsBetween 2013 and 2018, Asian women had the highest maternal vaccination coverage (36%) for pertussis, while Māori and Pacific women had the lowest, 13% and 15% respectively. Coverage of pertussis vaccination during pregnancy in low deprivation Māori women was 24% and 28% in Pacific women. This is in comparison to 30% and 25% in high deprivation Asian and European/Other women, respectively. Similar trends were seen for influenza.ConclusionBetween 2013 and 2018 maternal vaccination coverage increased for pertussis and influenza. Despite this coverage remains suboptimal, and existing ethnic and deprivation inequities increased. There is an urgent need to focus on equity, to engage and support ethic communities by creating genuinely accessible, culturally appropriate health services.     

Case-control studies of cancer in New Zealand electrical workers

A series of reports, including a New Zealand case-control study, have suggested that electrical workers are at increased risk of leukaemia. We report here a further series of case-control studies based on the New Zealand Cancer Registry. These involved 19,904 male patients registered with cancer for the period 1980-1984 who were aged 20 years or more at time of registration. For each cancer site, the registrations for other sites formed the control group. Three main findings emerged. First, there is an elevated leukaemia risk in New Zealand electrical workers (odds ratio (OR) = 1.62, 95% confidence interval (Cl) 1.04-2.52), but little evidence of increased risks for other cancer sites. Second, contrary to other published studies, the increased risk was primarily for chronic leukaemia (OR = 2.12) rather than acute leukaemia (OR = 1.25), and for lymphatic leukaemia (OR = 1.73) rather than myeloid leukaemia (OR = 1.22). Third, the increased risk was strongest for certain categories of electrical work including radio and television repairers (OR = 7.86, 95% CI 2.20-28.09), electricians (OR = 1.68, 95% Cl = 0.75-3.79), linemen (OR = 2.35, 95% Cl 0.97-5.70) and power station operators (OR = 3.89, 95% Cl 1.00-15.22).     

队列研究中纵向缺失数据填补方法的模拟研究

目的 数据缺失是队列研究中几乎无法避免的问题。本文旨在通过模拟研究,比较当前常见的8种缺失数据处理方法在纵向缺失数据中的填补效果,为纵向缺失数据的处理提供有价值的参考。方法 模拟研究基于R语言编程实现,通过Monte Carlo方法产生纵向缺失数据,通过比较不同填补方法的平均绝对偏差、平均相对偏差和回归分析的Ⅰ类错误,评价不同填补方法对于纵向缺失数据的填补效果及对后续多因素分析的影响。结果 均值填补、k近邻填补（KNN）、回归填补和随机森林的填补效果接近,且表现稳定;多重插补和热卡填充次于以上填补方法;K均值聚类和EM算法填补效果最差,表现也最不稳定。均值填补、EM算法、随机森林、KNN和回归填补可较好地控制Ⅰ类错误,多重插补、热卡填充和K均值聚类不能有效控制Ⅰ类错误。结论 对于纵向缺失数据,在随机缺失机制下,均值填补、KNN、回归填补和随机森林均可作为较好的填补方法,当缺失比例不太大时,多重插补和热卡填充也表现较好,不推荐K均值聚类和EM算法。