Preventive interventions to improve older people's health outcomes: systematic review and meta-analysis.

BACKGROUND: Systematic reviews of preventive, non-disease-specific primary care trials for older people often report effects according to what is thought to be the intervention's active ingredient. AIM: To examine the effectiveness of preventive primary care interventions for older people and to identify common components that contribute to intervention success. DESIGN AND SETTING: A systematic review and meta-analysis of 18 randomised controlled trials (RCTs) published in 22 publications from 2009 to 2019. METHOD: A search was conducted in PubMed, MEDLINE, Embase, Web of Science, CENTRAL, CINAHL, and the Cochrane Library. Inclusion criteria were: sample mainly aged ≥65 years; delivered in primary care; and non-disease-specific interventions. Exclusion criteria were: non-RCTs; primarily pharmacological or psychological interventions; and where outcomes of interest were not reported. Risk of bias was assessed using the original Cochrane tool. Outcomes examined were healthcare use including admissions to hospital and aged residential care (ARC), and patient-reported outcomes including activities of daily living (ADLs) and self-rated health (SRH). RESULTS: Many studies had a mix of patient-, provider-, and practice-focused intervention components (13 of 18 studies). Studies included in the review had low-to-moderate risk of bias. Interventions had no overall benefit to healthcare use (including admissions to hospital and ARC) but higher basic ADL scores were observed (standardised mean difference [SMD] 0.21, 95% confidence interval [CI] = 0.01 to 0.40) and higher odds of reporting positive SRH (odds ratio [OR] 1.17, 95% CI = 1.01 to 1.37). When intervention effects were examined by components, better patient-reported outcomes were observed in studies that changed the care setting (SMD for basic ADLs 0.21, 95% CI = 0.01 to 0.40; OR for positive SRH 1.17, 95% CI = 1.01 to 1.37), included educational components for health professionals (SMD for basic ADLs 0.21, 95% CI = 0.01 to 0.40; OR for positive SRH 1.27, 95% CI = 1.05 to 1.55), and provided patient education (SMD for basic ADLs 0.28, 95% CI = 0.09 to 0.48). Additionally, admissions to hospital in intervention participants were fewer by 23% in studies that changed the care setting (incidence rate ratio [IRR] 0.77, 95% CI = 0.63 to 0.95) and by 26% in studies that provided patient education (IRR 0.74, 95% CI = 0.56 to 0.97). CONCLUSION: Preventive primary care interventions are beneficial to older people's functional ability and SRH but not other outcomes. To improve primary care for older people, future programmes should consider delivering care in alternative settings, for example, home visits and phone contacts, and providing education to patients and health professionals as these may contribute to positive outcomes.

Kumanu Tangata: the aftermatch study - protocol to examine the health outcomes of high-level male rugby union players using linked administrative data.

There is increasing interest in the potential long-term outcomes of participation in contact and collision sports, driven by evidence of higher rates of neurodegenerative diseases among former athletes. Recent research has capitalised on large-scale administrative health data to examine health outcomes in contact sport athletes. However, there is limited research on outcomes associated with participation in rugby union, a contact sport with a relatively high incidence of head trauma and musculoskeletal injuries. Additionally, there is scope to investigate a greater range of health outcomes using large, population-based administrative data. The Kumanu Tangata project is a retrospective cohort study that will use linked information from the New Zealand Rugby Register and health records within a comprehensive deidentified whole-population administrative research database known as the Integrated Data Infrastructure. First-class male rugby union players (N=13 227) will be compared with a general population comparison group (N=2 438 484; weighting will be applied due to demographic differences) on a range of mortality and morbidity outcomes (neurodegenerative diseases, musculoskeletal conditions, chronic physical conditions, mental health outcomes). A range of player-specific variables will also be investigated as risk factors. Analyses will consist primarily of Cox proportional hazards models. Ethics approval for the study has been granted by the Auckland Health Research Ethics Committee (Ref. AH23203). Primary research dissemination will be via peer-reviewed journal articles.

Associations Between Gestational Weight Gain, Gestational Diabetes, and Childhood Obesity Incidence.

INTRODUCTION: Excessive maternal gestational weight gain (GWG) is strongly correlated with childhood obesity, yet how excess maternal weight gain and gestational diabetes mellitus (GDM) interact to affect early childhood obesity is poorly understood. The purpose of this study was to investigate whether overall and trimester-specific maternal GWG and GDM were associated with obesity in offspring by age 6 years. METHODS: A cohort of 10,335 maternal-child dyads was established from electronic health records. Maternal weights at conception and delivery were estimated from weight trajectory fits using functional principal components analysis. Kaplan-Meier curves and Cox regression, together with generalized raking, examined time-to-childhood-obesity. RESULTS: Obesity diagnosed prior to age 6 years was estimated at 19.7% (95% CI: 18.3, 21.1). Maternal weight gain during pregnancy was a strong predictor of early childhood obesity (p < 0.0001). The occurrence of early childhood obesity was lower among mothers with GDM compared with those without diabetes (adjusted hazard ratio = 0.58, p = 0.014). There was no interaction between maternal weight gain and GDM (p = 0.55). Higher weight gain during the first trimester was associated with lower risk of early childhood obesity (p = 0.0002) whereas higher weight gain during the second and third trimesters was associated with higher risk (p < 0.0001). DISCUSSION: Results indicated total and trimester-specific maternal weight gain was a strong predictor of early childhood obesity, though obesity risk by age 6 was lower for children of mothers with GDM. Additional research is needed to elucidate underlying mechanisms directly related to trimester-specific weight gain and GDM that impede or protect against obesity prevalence during early childhood.

Excessive maternal gestational weight gain (GWG) and gestational diabetes mellitus (GDM) have been linked to childhood obesity. Yet, research on how excessive total and trimester-specific GWG and GDM interact to affect early childhood obesity remains inconclusive. This study found that inadequate weight gain in the first trimester and excessive weight gain in the second and third trimester were associated with higher risks of childhood obesity by age 6. No significant interaction between maternal GWG and GDM was noted suggesting that these two important maternal conditions do not have a combined effect on the risk of early childhood obesity.

Detection of interferon alpha and beta receptor subunit 1 (IFNAR1) loss-of-function Glu386∗ variant by tri-allelic genotyping.

Mutations of the human interferon alpha and beta receptor subunit 1 (IFNAR1) gene are associated with severe viral infections. Individuals homozygous for the Glu386∗ variant have impaired type I interferon signalling and can suffer severe illness when exposed to certain viruses and live attenuated virus vaccines. Glu386∗ heterozygotes are clinically unaffected, but can pass the variant allele to their descendants. We aimed to develop an assay that can identify IFNAR1 Glu386∗ homozygotes and heterozygotes to support urgent clinical diagnosis, and that can use dried blood spots (DBS) sent at ambient temperature to overcome geographical logistical challenges in the South Pacific region. The tri-allelic genotyping assay interrogates a single nucleotide polymorphism (rs201609461) located in IFNAR1. The reference allele G encodes for wild-type IFNAR1. Minor alleles A (c.1156G>A) and T (c.1156G>T) encode for Glu386Lys and a truncated IFNAR1 protein (p.Glu386∗), respectively. Synthetic oligonucleotides were mixed in equal molar ratio to create six different genotypes which were randomly assigned to 960 genotyping reactions by R software. Three different fluorescence probes were designed to discriminate the three alleles (G, T and A) within a pair of flanking primers in a single genotyping reaction. The assay discriminated all three alleles using DBS from Guthrie cards randomly spiked with synthetic oligonucleotides. We correctly identified all the different genotypes in 960 reactions in these blinded experiments. These findings validate the genotyping assay for rapidly identifying the IFNAR1 Glu386∗ variant from DBS.

Three-phase generalized raking and multiple imputation estimators to address error-prone data.

Validation studies are often used to obtain more reliable information in settings with error-prone data. Validated data on a subsample of subjects can be used together with error-prone data on all subjects to improve estimation. In practice, more than one round of data validation may be required, and direct application of standard approaches for combining validation data into analyses may lead to inefficient estimators since the information available from intermediate validation steps is only partially considered or even completely ignored. In this paper, we present two novel extensions of multiple imputation and generalized raking estimators that make full use of all available data. We show through simulations that incorporating information from intermediate steps can lead to substantial gains in efficiency. This work is motivated by and illustrated in a study of contraceptive effectiveness among 83 671 women living with HIV, whose data were originally extracted from electronic medical records, of whom 4732 had their charts reviewed, and a subsequent 1210 also had a telephone interview to validate key study variables.

A blueprint for a multi-disease, multi-domain Bayesian adaptive platform trial incorporating adult and paediatric subgroups: the Staphylococcus aureus Network Adaptive Platform trial.

The Staphylococcus aureus Network Adaptive Platform (SNAP) trial is a multifactorial Bayesian adaptive platform trial that aims to improve the way that S. aureus bloodstream infection, a globally common and severe infectious disease, is treated. In a world first, the SNAP trial will simultaneously investigate the effects of multiple intervention modalities within multiple groups of participants with different forms of S. aureus bloodstream infection. Here, we formalise the trial structure, modelling approach, and decision rules that will be used for the SNAP trial. By summarising the statistical principles governing the design, our hope is that the SNAP trial will serve as an adaptable template that can be used to improve comparative effectiveness research efficiency in other disease areas.Trial registration NCT05137119 . Registered on 30 November 2021.

Inferring bacterial transmission dynamics using deep sequencing genomic surveillance data.

Identifying and interrupting transmission chains is important for controlling infectious diseases. One way to identify transmission pairs - two hosts in which infection was transmitted from one to the other - is using the variation of the pathogen within each single host (within-host variation). However, the role of such variation in transmission is understudied due to a lack of experimental and clinical datasets that capture pathogen diversity in both donor and recipient hosts. In this work, we assess the utility of deep-sequenced genomic surveillance (where genomic regions are sequenced hundreds to thousands of times) using a mouse transmission model involving controlled spread of the pathogenic bacterium Citrobacter rodentium from infected to naïve female animals. We observe that within-host single nucleotide variants (iSNVs) are maintained over multiple transmission steps and present a model for inferring the likelihood that a given pair of sequenced samples are linked by transmission. In this work we show that, beyond the presence and absence of within-host variants, differences arising in the relative abundance of iSNVs (allelic frequency) can infer transmission pairs more precisely. Our approach further highlights the critical role bottlenecks play in reserving the within-host diversity during transmission.

Practical considerations for sandwich variance estimation in two-stage regression settings.

We present a practical approach for computing the sandwich variance estimator in two-stage regression model settings. As a motivating example for two-stage regression, we consider regression calibration, a popular approach for addressing covariate measurement error. The sandwich variance approach has been rarely applied in regression calibration, despite it requiring less computation time than popular resampling approaches for variance estimation, specifically the bootstrap. This is likely due to requiring specialized statistical coding. We first outline the steps needed to compute the sandwich variance estimator. We then develop a convenient method of computation in R for sandwich variance estimation, which leverages standard regression model outputs and existing R functions and can be applied in the case of a simple random sample or complex survey design. We use a simulation study to compare the sandwich to a resampling variance approach for both settings. Finally, we further compare these two variance estimation approaches for data examples from the Women's Health Initiative (WHI) and Hispanic Community Health Study/Study of Latinos (HCHS/SOL). The sandwich variance estimator typically had good numerical performance, but simple Wald bootstrap confidence intervals were unstable or over-covered in certain settings, particularly when there was high correlation between covariates or large measurement error.

Adverse Events Following the BNT162b2 mRNA COVID-19 Vaccine (Pfizer-BioNTech) in Aotearoa New Zealand.

INTRODUCTION: In February 2021, New Zealand began its largest ever immunisation programme with the BNT162b2 mRNA coronavirus disease 2019 (COVID-19) vaccine. OBJECTIVE: We aimed to understand the association between 12 adverse events of special interest (AESIs) and a primary dose of BNT162b2 in the New Zealand population aged ≥5 years from 19 February 2021 through 10 February 2022. METHODS: Using national electronic health records, the observed rates of AESIs within a risk period (1-21 days) following vaccination were compared with the expected rates based on background data (2014-2019). Standardised incidence ratios (SIRs) were estimated for each AESI with 95% confidence intervals (CIs) using age group-specific background rates. The risk difference was calculated to estimate the excess or reduced number of events per 100,000 persons vaccinated in the risk period. RESULTS: As of 10 February 2022, 4,277,163 first doses and 4,114,364 second doses of BNT162b2 had been administered to the eligible New Zealand population aged ≥5 years. The SIRs for 11 of the 12 selected AESIs were not statistically significantly increased post vaccination. The SIR (95% CI) for myo/pericarditis following the first dose was 2.3 (1.8-2.7), with a risk difference (95% CI) of 1.3 (0.9-1.8), per 100,000 persons vaccinated, and 4.0 (3.4-4.6), with a risk difference of 3.1 (2.5-3.7), per 100,000 persons vaccinated following the second dose. The highest SIR was 25.6 (15.5-37.5) in the 5-19 years age group, following the second dose of the vaccine, with an estimated five additional myo/pericarditis cases per 100,000 persons vaccinated. A statistically significant increased SIR of single organ cutaneous vasculitis (SOCV) was also observed following the first dose of BNT162b2 in the 20-39 years age group only. CONCLUSIONS: A statistically significant association between BNT162b2 vaccination and myo/pericarditis was observed. This association has been confirmed internationally. BNT162b2 was not found to be associated with the other AESIs investigated, except for SOCV following the first dose of BNT162b2 in the 20-39 years age group only, providing reassurances around the safety of the vaccine.

Using Days Alive and Out of Hospital to measure inequities and possible pathways for them after cardiovascular surgery in Aotearoa New Zealand: study protocol for a secondary data analysis.

INTRODUCTION: In Aotearoa New Zealand (NZ), socioeconomic status and being of Maori ethnicity are often associated with poorer health outcomes, including after surgery. Inequities can be partially explained by differences in health status and health system biases are hypothesised as important factors for remaining inequities. Previous work identified inequities between Maori and non-Maori following cardiovascular surgery, some of which have been identified in studies between 1990 and 2012. Days Alive and Out of Hospital (DAOH) is an emerging surgical outcome metric. DAOH is a composite measure of outcomes, which may reflect patient experience and longer periods of DAOH may also reflect extended interactions with the health system. Recently, a 1.1-day difference in DAOH was observed between Maori and non-Maori at a hospital in NZ across a range of operations. METHODS AND ANALYSIS: We will conduct a secondary data analysis using data from the National Minimum Data Set, maintained by the Ministry of Health. We will report unadjusted and risk-adjusted DAOH values between Maori and non-Maori using direct risk standardisation. We will risk adjust first for age and sex, then for each of deprivation (NZDep18), levels of morbidity (M3 score) and rurality. We will report DAOH values across three time periods, 30, 90 and 365 days and across nine deciles of the DAOH distribution (0.1-0.9 inclusive). We will interpret all results from a Kaupapa Maori research positioning, acknowledging that Maori health outcomes are directly tied to the unequal distribution of the social determinants of health. ETHICS AND DISSEMINATION: Ethics approval for this study was given by the Auckland Health Research Ethics Committee. Outputs from this study are likely to interest a range of audiences. We plan to disseminate our findings through academic channels, presentations to interested groups including Maori-specific hui (meetings), social media and lay press.

Using days alive and out of hospital to measure surgical outcomes in New Zealand: a cross-sectional study.

OBJECTIVES: To measure differences at various deciles in days alive and out of hospital to 90 days (DAOH90) and explore its utility for identifying outliers of performance among district health boards (DHBs). METHODS: Days in hospital and mortality within 90 days of surgery were extracted by linking data from the New Zealand National Minimum Data Set and the births and deaths registry between 1 January 2011 and 31 December 2021 for all adults in New Zealand undergoing acute laparotomy (AL-a relatively high-risk group), elective total hip replacement (THR-a medium risk group) or lower segment caesarean section (LSCS-a low-risk group). DAOH90 was calculated without censoring to zero in cases of mortality. For each DHB, direct risk standardisation was used to adjust for potential confounders and presented in deciles according to baseline patient risk. The Mann-Whitney U test assessed overall DAOH90 differences between DHBs, and comparisons are presented between selected deciles of DAOH90 for each operation. RESULTS: We obtained national data for 35 175, 52 032 and 117 695 patients undergoing AL, THR and LSCS procedures, respectively. We have demonstrated that calculating DAOH without censoring zero allows for differences between procedures and DHBs to be identified. Risk-adjusted national mean DAOH90 Scores were 64.0 days, 79.0 days and 82.0 days at the 0.1 decile and 75.0 days, 82.0 days and 84.0 days at the 0.2 decile for AL, THR and LSCS, respectively, matching to their expected risk profiles. Differences between procedures and DHBs were most marked at lower deciles of the DAOH90 distribution, and outlier DHBs were detectable. Corresponding 90-day mortality rates were 5.45%, 0.78% and 0.01%. CONCLUSION: In New Zealand after direct risk adjustment, differences in DAOH90 between three types of surgical procedure reflected their respective risk levels and associated mortality rates. Outlier DHBs were identified for each procedure. Thus, our approach to analysing DAOH90 appears to have considerable face validity and potential utility for contributing to the measurement of perioperative outcomes in an audit or quality improvement setting.

Temporal validation of a multivariable surgical mortality prediction model (NZRisk): a New Zealand national cohort study.

OBJECTIVES: Clinical risk calculators (CRCs), such as NZRisk, are used daily by clinicians to guide clinical decisions and explain individual risk to patients. The utility and robustness of these tools depends on the methods used to create the underlying mathematical model, as well as the stability of that model in relation to changing clinical practice and patient populations over time. The later should be checked by temporal validation using external data. Few if any of the clinical prediction models in current clinical use have published temporal validation. Here, we use a large external dataset to temporally validate NZRisk; a perioperative risk prediction model used in the New Zealand population. METHODS: A sample of 1 976 362 adult non-cardiac surgical procedures collected over 15 years from the New Zealand Ministry of Health National Minimum Dataset, was used to temporally validate NZRisk. We divided the dataset into 15 single year cohorts and compared 13 of these to our NZRisk model (2 years used for the model building were excluded). We compared the area under the curve (AUC) value, calibration slope and intercept for each single year cohort, to the same values produced by the data used to create NZRisk, by fitting a random effects meta-regression with each year cohort acting as a separate study point. In addition, we used two-sided t-tests to compare each measure across the cohorts. RESULTS: The AUC values for the 30-day NZRisk model applied to our single year cohorts ranged from 0.918 to 0.940 (NZRisk AUC was 0.921). There were eight statistically different AUC values for the following years 2007-2009, 2016 and 2018-2021. The intercept values ranged from -0.004 to 0.007 and 7 years had statistically significant different intercepts during leave-one-out t-tests; 2007-2010, 2012, 2018 and 2021. The slope values ranged from 0.72 to 1.12 and 7 years had statistically significant different slopes during leave-one-out t-tests; 2010, 2011, 2017, 2018 and 2019-2021. The random effects meta-regression upheld our results related to AUC (0.54 (95% CI 0.40 to 0.99), I2 67.57 (95% CI 40.67 to 88.50), Cochran's Q<0.001) and slope (τ 0.14 (95% CI 0.01 to 0.23), I2 98.61 (95% CI 97.31 to 99.50), Cochran's Q<0.001) between year difference. CONCLUSION: The NZRisk model shows differences in AUC and slope but not intercept values over time. The biggest differences were in the calibration slope. The models maintained excellent discrimination over time as shown by the AUC values. These findings suggest we update our model in the next 5 years. To our knowledge, this is the first temporal validation of a CRC in current use.

Atrial fibrillation incidence and outcomes in two cohorts of octogenarians: LiLACS NZ.

BACKGROUND: Atrial fibrillation (AF), the most common cardiac arrhythmia in the general population, has significant healthcare burden. Little is known about AF in octogenarians. OBJECTIVE: To describe the prevalence and incidence rate of AF in New Zealand (NZ) octogenarians and the risk of stroke and mortality at 5-year follow-up. DESIGN: Longitudinal Cohort Study. SETTING: Bay of Plenty and Lakes health regions of New Zealand. SUBJECTS: Eight-hundred-seventy-seven (379 indigenous Maori, 498 non-Maori) were included in the analysis. METHODS: AF, stroke/TIA events and relevant co-variates were established annually using self-report and hospital records (and ECG for AF). Cox proportional-hazards regression models were used to determine the time dependent AF risk of stroke/TIA. RESULTS: AF was present in 21% at baseline (Maori 26%, non-Maori 18%), the prevalence doubled over 5-years (Maori 50%, non-Maori 33%). 5-year AF incidence was 82.6 /1000-person years and at all times AF incidence for Maori was twice that of non-Maori. Five-year stroke/TIA prevalence was 23% (22% in Maori and 24% non- Maori), higher in those with AF than without. AF was not independently associated with 5-year new stroke/TIA; baseline systolic blood pressure was. Mortality was higher for Maori, men, those with AF and CHF and statin use was protective. In summary, AF is more prevalent in indigenous octogenarians and should have an increased focus in health care management. Further research could examine treatment in more detail to facilitate ethnic specific impact and risks and benefits of treating AF in octogenarians.

Prognostic modelling of clinical outcomes after first-time acute coronary syndrome in New Zealand.

OBJECTIVE: The Multi-Ethnic New Zealand Study of Acute Coronary Syndromes (MENZACS) was established to investigate the drivers of secondary events after first-time acute coronary syndrome (ACS), including addressing inequitable outcomes by ethnicity. Herein, the first clinical outcomes and prognostic modelling approach are reported. METHODS: First, in 28 176 New Zealanders with first-time ACS from a national registry, a clinical summary score for predicting 1-year death/cardiovascular readmission was created using Cox regression of 20 clinical variables. This score was then calculated in the 2015 participant MENZACS study to represent clinical risk. In MENZACS, Cox regression was used to assess N-terminal pro-B-type natriuretic peptide (NT-proBNP) as a prognostic marker for death/cardiovascular readmission in four models, adjusting for (1) age and sex; (2) age, sex, ethnicity; (3) clinical summary score; (4) clinical summary score and ethnicity. RESULTS: Of the 2015 MENZACS participants (mean age 61 years, 79% male, 73% European, 14% Maori, 5% Pacific people), 2003 were alive at discharge. Of the 2003, 416 (20.8%) experienced all-cause death/cardiovascular readmission over a median of 3.5 years. In a simple model, age, male sex, Maori ethnicity and NT-proBNP levels were significant predictors of outcome. After adjustment for the clinical summary score, which includes age and sex, NT-proBNP and ethnicity were no longer statistically significant: log2(NT-proBNP) hazard ratio (HR) 1.03, 95% confidence interval (95% CI) 0.98 to 1.08, p=0.305; Maori ethnicity HR 1.26, 95% CI 0.97 to 1.62, p=0.084. CONCLUSIONS: In 2015 patients with first-time ACS, recurrent events were common (20.8%). Increasing NT-proBNP levels and Maori ethnicity were predictors of death/cardiovascular readmission, but not after adjustment for the 20 clinical risk factors represented by the clinical summary score. TRIAL REGISTRATION NUMBER: ACTRN12615000676516.

Thrombotic events following the BNT162b2 mRNA COVID-19 vaccine (Pfizer-BioNTech) in Aotearoa New Zealand: A self-controlled case series study.

BACKGROUND: An association between thrombotic events and SARS-CoV-2 infection and the adenovirus-based COVID-19 vaccines has been established, leading to concern over the risk of thrombosis after BNT162b2 COVID-19 vaccination. OBJECTIVES: To evaluate the risk of arterial thrombosis, cerebral venous thrombosis (CVT), splanchnic thrombosis, and venous thromboembolism (VTE) following BNT162b2 vaccination in New Zealand. METHODS: This was a self-controlled case series using national hospitalisation and immunisation records to calculate incidence rate ratios (IRR). The study population included individuals aged ≥12 years, unvaccinated, or vaccinated with BNT162b2, who were hospitalised with one of the thrombotic events of interest from 19 February 2021 through 19 February 2022. The risk period was 0-21 days after receiving a primary or booster dose of BNT162b2. RESULTS: 6039 individuals were hospitalised with one of the thrombotic events examined, including 5127 with VTE, 605 with arterial thrombosis, 272 with splanchnic thrombosis, and 35 with CVT. The proportion of individuals vaccinated with at least one dose of BNT162b2 ranged from 82.7 % to 91.4 %. Compared with the control unexposed period, the IRR (95 % CI) of VTE, arterial thrombosis, splanchnic thrombosis, and CVT were 0.87 (0.76-1.00), 0.73 (0.56-0.95), 0.71 (0.43-1.16), and 0.87 (0.31-2.50) in the 21 days after BNT162b2 vaccination, respectively. There was no statistically significant increased risk of thrombosis following BNT162b2 in different ethnic groups in New Zealand. CONCLUSION: The BNT162b2 vaccine was not found to be associated with thrombosis in the general population or different ethnic groups in New Zealand, providing reassurance for the safety of the BNT162b2 vaccine.

Multiwave validation sampling for error-prone electronic health records.

Electronic health record (EHR) data are increasingly used for biomedical research, but these data have recognized data quality challenges. Data validation is necessary to use EHR data with confidence, but limited resources typically make complete data validation impossible. Using EHR data, we illustrate prospective, multiwave, two-phase validation sampling to estimate the association between maternal weight gain during pregnancy and the risks of her child developing obesity or asthma. The optimal validation sampling design depends on the unknown efficient influence functions of regression coefficients of interest. In the first wave of our multiwave validation design, we estimate the influence function using the unvalidated (phase 1) data to determine our validation sample; then in subsequent waves, we re-estimate the influence function using validated (phase 2) data and update our sampling. For efficiency, estimation combines obesity and asthma sampling frames while calibrating sampling weights using generalized raking. We validated 996 of 10,335 mother-child EHR dyads in six sampling waves. Estimated associations between childhood obesity/asthma and maternal weight gain, as well as other covariates, are compared to naïve estimates that only use unvalidated data. In some cases, estimates markedly differ, underscoring the importance of efficient validation sampling to obtain accurate estimates incorporating validated data.

Discussion on "A formal causal interpretation of the case-crossover design" by Zach Shahn, Miguel A. Hernán, and James M. Robins.

It has always been clear that the case-crossover design works, for some definition of "works," but some of the details have been surprisingly elusive, and it is good to see more of them nailed down by Shahn et al. My interest in case-crossover analyses has mostly been in the context of air pollution epidemiology mentioned at the end of the paper. The air pollution setting is distinctive for several reasons: as the exposure variable is plausibly exogenous, it is possible to use control times after the case time, the effects of interest are quite small, and the same measured exposure series is shared over many-perhaps all-of the cohort.

Optimal sampling for design-based estimators of regression models.

Two-phase designs measure variables of interest on a subcohort where the outcome and covariates are readily available or cheap to collect on all individuals in the cohort. Given limited resource availability, it is of interest to find an optimal design that includes more informative individuals in the final sample. We explore the optimal designs and efficiencies for analyses by design-based estimators. Generalized raking is an efficient class of design-based estimators, and they improve on the inverse-probability weighted (IPW) estimator by adjusting weights based on the auxiliary information. We derive a closed-form solution of the optimal design for estimating regression coefficients from generalized raking estimators. We compare it with the optimal design for analysis via the IPW estimator and other two-phase designs in measurement-error settings. We consider general two-phase designs where the outcome variable and variables of interest can be continuous or discrete. Our results show that the optimal designs for analyses by the two classes of design-based estimators can be very different. The optimal design for analysis via the IPW estimator is optimal for IPW estimation and typically gives near-optimal efficiency for generalized raking estimation, though we show there is potential improvement in some settings.

Model averaging with the hybrid model: An asymptotic study and demonstration.

In this paper, we present a new model averaging technique that can be applied in medical research. The dataset is first partitioned by the values of its categorical explanatory variables. Then for each partition, a model average is determined by minimising some form of squared errors, which could be the leave-one-out cross-validation errors. From our asymptotic optimality study and the results of simulations, we demonstrate under several high-level assumptions and modelling conditions that this model averaging procedure may outperform jackknife model averaging, which is a well-established technique. We also present an example where a cross-validation procedure does not work (that is, a zero-valued cross-validation error is obtained) when determining the weights for model averaging.

Combining multiple imputation with raking of weights: An efficient and robust approach in the setting of nearly true models.

Multiple imputation (MI) provides us with efficient estimators in model-based methods for handling missing data under the true model. It is also well-understood that design-based estimators are robust methods that do not require accurately modeling the missing data; however, they can be inefficient. In any applied setting, it is difficult to know whether a missing data model may be good enough to win the bias-efficiency trade-off. Raking of weights is one approach that relies on constructing an auxiliary variable from data observed on the full cohort, which is then used to adjust the weights for the usual Horvitz-Thompson estimator. Computing the optimally efficient raking estimator requires evaluating the expectation of the efficient score given the full cohort data, which is generally infeasible. We demonstrate MI as a practical method to compute a raking estimator that will be optimal. We compare this estimator to common parametric and semi-parametric estimators, including standard MI. We show that while estimators, such as the semi-parametric maximum likelihood and MI estimator, obtain optimal performance under the true model, the proposed raking estimator utilizing MI maintains a better robustness-efficiency trade-off even under mild model misspecification. We also show that the standard raking estimator, without MI, is often competitive with the optimal raking estimator. We demonstrate these properties through several numerical examples and provide a theoretical discussion of conditions for asymptotically superior relative efficiency of the proposed raking estimator.