Search

Background: Risk prediction models are routinely used to assist in clinical decision making. A small sample size for model development can compromise model performance when the model is applied to new patients. For binary outcomes, the calibration slope (CS) and the mean absolute prediction error (MAPE) are two key measures on which sample size calculations for the development of risk models have been based. CS quantifies the degree of model overfitting while MAPE assesses the accuracy of individual predictions., Methods: Recently, two formulae were proposed to calculate the sample size required, given anticipated features of the development data such as the outcome prevalence and c-statistic, to ensure that the expectation of the CS and MAPE (over repeated samples) in models fitted using MLE will meet prespecified target values. In this article, we use a simulation study to evaluate the performance of these formulae., Results: We found that both formulae work reasonably well when the anticipated model strength is not too high (c-statistic < 0.8), regardless of the outcome prevalence. However, for higher model strengths the CS formula underestimates the sample size substantially. For example, for c-statistic = 0.85 and 0.9, the sample size needed to be increased by at least 50% and 100%, respectively, to meet the target expected CS. On the other hand, the MAPE formula tends to overestimate the sample size for high model strengths. These conclusions were more pronounced for higher prevalence than for lower prevalence. Similar results were drawn when the outcome was time to event with censoring. Given these findings, we propose a simulation-based approach, implemented in the new R package 'samplesizedev', to correctly estimate the sample size even for high model strengths. The software can also calculate the variability in CS and MAPE, thus allowing for assessment of model stability., Conclusions: The calibration and MAPE formulae suggest sample sizes that are generally appropriate for use when the model strength is not too high. However, they tend to be biased for higher model strengths, which are not uncommon in clinical risk prediction studies. On those occasions, our proposed adjustments to the sample size calculations will be relevant., (© 2024. The Author(s).)

Assessing the impact of an intervention by using time-series observational data on multiple units and outcomes is a frequent problem in many fields of scientific research. Here, we propose a novel Bayesian multivariate factor analysis model for estimating intervention effects in such settings and develop an efficient Markov chain Monte Carlo algorithm to sample from the high-dimensional and nontractable posterior of interest. The proposed method is one of the few that can simultaneously deal with outcomes of mixed type (continuous, binomial, count), increase efficiency in the estimates of the causal effects by jointly modeling multiple outcomes affected by the intervention, and easily provide uncertainty quantification for all causal estimands of interest. Using the proposed approach, we evaluate the impact that Local Tracing Partnerships had on the effectiveness of England's Test and Trace programme for COVID-19., (© The Author(s) 2023. Published by Oxford University Press.)

Objectives: To determine how the intrinsic severity of successively dominant SARS-CoV-2 variants changed over the course of the pandemic., Methods: A retrospective cohort analysis in the NHS Greater Glasgow and Clyde (NHS GGC) Health Board. All sequenced non-nosocomial adult COVID-19 cases in NHS GGC with relevant SARS-CoV-2 lineages (B.1.177/Alpha, Alpha/Delta, AY.4.2 Delta/non-AY.4.2 Delta, non-AY.4.2 Delta/Omicron, and BA.1 Omicron/BA.2 Omicron) during analysis periods were included. Outcome measures were hospital admission, ICU admission, or death within 28 days of positive COVID-19 test. We report the cumulative odds ratio; the ratio of the odds that an individual experiences a severity event of a given level vs all lower severity levels for the resident and the replacement variant after adjustment., Results: After adjustment for covariates, the cumulative odds ratio was 1.51 (95% CI: 1.08-2.11) for Alpha versus B.1.177, 2.09 (95% CI: 1.42-3.08) for Delta versus Alpha, 0.99 (95% CI: 0.76-1.27) for AY.4.2 Delta versus non-AY.4.2 Delta, 0.49 (95% CI: 0.22-1.06) for Omicron versus non-AY.4.2 Delta, and 0.86 (95% CI: 0.68-1.09) for BA.2 Omicron versus BA.1 Omicron., Conclusions: The direction of change in intrinsic severity between successively emerging SARS-CoV-2 variants was inconsistent, reminding us that the intrinsic severity of future SARS-CoV-2 variants remains uncertain., Competing Interests: Declaration of Competing Interest The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Thomas C Williams is Principal Investigator for the BronchStart project, which is funded by the Respiratory Syncytial Virus Consortium in Europe (RESCEU), with data collection supported by the National Institute for Health Research., (Copyright © 2023. Published by Elsevier Ltd.)

Longitudinal observational data on patients can be used to investigate causal effects of time-varying treatments on time-to-event outcomes. Several methods have been developed for estimating such effects by controlling for the time-dependent confounding that typically occurs. The most commonly used is marginal structural models (MSM) estimated using inverse probability of treatment weights (IPTW) (MSM-IPTW). An alternative, the sequential trials approach, is increasingly popular, and involves creating a sequence of "trials" from new time origins and comparing treatment initiators and non-initiators. Individuals are censored when they deviate from their treatment assignment at the start of each "trial" (initiator or noninitiator), which is accounted for using inverse probability of censoring weights. The analysis uses data combined across trials. We show that the sequential trials approach can estimate the parameters of a particular MSM. The causal estimand that we focus on is the marginal risk difference between the sustained treatment strategies of "always treat" vs "never treat." We compare how the sequential trials approach and MSM-IPTW estimate this estimand, and discuss their assumptions and how data are used differently. The performance of the two approaches is compared in a simulation study. The sequential trials approach, which tends to involve less extreme weights than MSM-IPTW, results in greater efficiency for estimating the marginal risk difference at most follow-up times, but this can, in certain scenarios, be reversed at later time points and relies on modelling assumptions. We apply the methods to longitudinal observational data from the UK Cystic Fibrosis Registry to estimate the effect of dornase alfa on survival., (© 2023 The Authors. Statistics in Medicine published by John Wiley & Sons Ltd.)

Importance: There is a need for observational studies to supplement evidence from clinical trials, and the target trial emulation (TTE) framework can help avoid biases that can be introduced when treatments are compared crudely using observational data by applying design principles for randomized clinical trials. Adalimumab (ADA) and tofacitinib (TOF) were shown to be equivalent in patients with rheumatoid arthritis (RA) in a randomized clinical trial, but to our knowledge, these drugs have not been compared head-to-head using routinely collected clinical data and the TTE framework., Objective: To emulate a randomized clinical trial comparing ADA vs TOF in patients with RA who were new users of a biologic or targeted synthetic disease-modifying antirheumatic drug (b/tsDMARD)., Design, Setting, and Participants: This comparative effectiveness study emulating a randomized clinical trial of ADA vs TOF included Australian adults aged 18 years or older with RA in the Optimising Patient Outcomes in Australian Rheumatology (OPAL) data set. Patients were included if they initiated ADA or TOF between October 1, 2015, and April 1, 2021; were new b/tsDMARD users; and had at least 1 component of the disease activity score in 28 joints using C-reactive protein (DAS28-CRP) recorded at baseline or during follow-up., Intervention: Treatment with either ADA (40 mg every 14 days) or TOF (10 mg daily)., Main Outcomes and Measures: The main outcome was the estimated average treatment effect, defined as the difference in mean DAS28-CRP among patients receiving TOF compared with those receiving ADA at 3 and 9 months after initiating treatment. Missing DAS28-CRP data were multiply imputed. Stable balancing weights were used to account for nonrandomized treatment assignment., Results: A total of 842 patients were identified, including 569 treated with ADA (387 [68.0%] female; median age, 56 years [IQR, 47-66 years]) and 273 treated with TOF (201 [73.6%] female; median age, 59 years [IQR, 51-68 years]). After applying stable balancing weights, mean DAS28-CRP in the ADA group was 5.3 (95% CI, 5.2-5.4) at baseline, 2.6 (95% CI, 2.5-2.7) at 3 months, and 2.3 (95% CI, 2.2-2.4) at 9 months; in the TOF group, it was 5.3 (95% CI, 5.2-5.4) at baseline, 2.4 (95% CI, 2.2-2.5) at 3 months, and 2.3 (95% CI, 2.1-2.4) at 9 months. The estimated average treatment effect was -0.2 (95% CI, -0.4 to -0.03; P = .02) at 3 months and -0.03 (95% CI, -0.2 to 0.1; P = .60) at 9 months., Conclusions and Relevance: In this study, there was a modest but statistically significant reduction in DAS28-CRP at 3 months for patients receiving TOF compared with those receiving ADA and no difference between treatment groups at 9 months. Three months of treatment with either drug led to clinically relevant average reductions in mean DAS28-CRP, consistent with remission.

When comparing the risk of a post-infection binary outcome, for example, hospitalisation, for two variants of an infectious pathogen, it is important to adjust for calendar time of infection. Typically, the infection time is unknown and positive test time used as a proxy for it. Positive test time may also be used when assessing how risk of the outcome changes over calendar time. We show that if time from infection to positive test is correlated with the outcome, the risk conditional on positive test time is a function of the trajectory of infection incidence. Hence, a risk ratio adjusted for positive test time can be quite different from the risk ratio adjusted for infection time. We propose a simple sensitivity analysis that indicates how risk ratios adjusted for positive test time and infection time may differ. This involves adjusting for a shifted positive test time, shifted to make the difference between it and infection time uncorrelated with the outcome. We illustrate this method by reanalysing published results on the relative risk of hospitalisation following infection with the Alpha versus pre-existing variants of SARS-CoV-2. Results indicate the relative risk adjusted for infection time may be lower than that adjusted for positive test time.

To investigate if the AY.4.2 sublineage of the SARS-CoV-2 delta variant is associated with hospitalization and mortality risks that differ from non-AY.4.2 delta risks, we performed a retrospective cohort study of sequencing-confirmed COVID-19 cases in England based on linkage of routine health care datasets. Using stratified Cox regression, we estimated adjusted hazard ratios (aHR) of hospital admission (aHR = 0.85; 95% confidence interval [CI], .77-.94), hospital admission or emergency care attendance (aHR = 0.87; 95% CI, .81-.94), and COVID-19 mortality (aHR = 0.85; 95% CI, .71-1.03). The results indicate that the risks of hospitalization and mortality are similar or lower for AY.4.2 compared to cases with other delta sublineages., Competing Interests: Potential conflicts of interest. G. D. declares that his employer, UK Health Security Agency, previously known as Public Health England, received funding from GlaxoSmithKline for a research project related to influenza antiviral treatment. This preceded and had no relation to COVID-19, and G. D. had no role in and received no funding from the project. All other authors report no potential conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed., (© The Author(s) 2022. Published by Oxford University Press for the Infectious Diseases Society of America.)

Time-to-event data are right-truncated if only individuals who have experienced the event by a certain time can be included in the sample. For example, we may be interested in estimating the distribution of time from onset of disease symptoms to death and only have data on individuals who have died. This may be the case, for example, at the beginning of an epidemic. Right truncation causes the distribution of times to event in the sample to be biased towards shorter times compared to the population distribution, and appropriate statistical methods should be used to account for this bias. This article is a review of such methods, particularly in the context of an infectious disease epidemic, like COVID-19. We consider methods for estimating the marginal time-to-event distribution, and compare their efficiencies. (Non-)identifiability of the distribution is an important issue with right-truncated data, particularly at the beginning of an epidemic, and this is discussed in detail. We also review methods for estimating the effects of covariates on the time to event. An illustration of the application of many of these methods is provided, using data on individuals who had died with coronavirus disease by 5 April 2020.

We compare two multi-state modelling frameworks that can be used to represent dates of events following hospital admission for people infected during an epidemic. The methods are applied to data from people admitted to hospital with COVID-19, to estimate the probability of admission to intensive care unit, the probability of death in hospital for patients before and after intensive care unit admission, the lengths of stay in hospital, and how all these vary with age and gender. One modelling framework is based on defining transition-specific hazard functions for competing risks. A less commonly used framework defines partially-latent subpopulations who will experience each subsequent event, and uses a mixture model to estimate the probability that an individual will experience each event, and the distribution of the time to the event given that it occurs. We compare the advantages and disadvantages of these two frameworks, in the context of the COVID-19 example. The issues include the interpretation of the model parameters, the computational efficiency of estimating the quantities of interest, implementation in software and assessing goodness of fit. In the example, we find that some groups appear to be at very low risk of some events, in particular intensive care unit admission, and these are best represented by using 'cure-rate' models to define transition-specific hazards. We provide general-purpose software to implement all the models we describe in the flexsurv R package, which allows arbitrarily flexible distributions to be used to represent the cause-specific hazards or times to events.

Inverse probability of censoring weighting is a popular approach to handling dropout in longitudinal studies. However, inverse probability-of-censoring weighted estimators (IPCWEs) can be inefficient and unstable if the weights are estimated by maximum likelihood. To alleviate these problems, calibrated IPCWEs have been proposed, which use calibrated weights that directly optimize covariate balance in finite samples rather than the weights from maximum likelihood. However, the existing calibrated IPCWEs are all based on the unverifiable assumption of sequential ignorability and sensitivity analysis strategies under non-ignorable dropout are lacking. In this paper, we fill this gap by developing an approach to sensitivity analysis for calibrated IPCWEs under non-ignorable dropout. A simple technique is proposed to speed up the computation of bootstrap and jackknife confidence intervals and thus facilitate sensitivity analyses. We evaluate the finite-sample performance of the proposed methods using simulations and apply our methods to data from an international inception cohort study of systemic lupus erythematosus. An R Markdown tutorial to demonstrate the implementation of the proposed methods is provided.

Understanding the trajectory of the daily number of COVID-19 deaths is essential to decisions on how to respond to the pandemic, but estimating this trajectory is complicated by the delay between deaths occurring and being reported. In England the delay is typically several days, but it can be weeks. This causes considerable uncertainty about how many deaths occurred in recent days. Here we estimate the deaths per day in five age strata within seven English regions, using a Bayesian model that accounts for reporting-day effects and longer-term changes in the delay distribution. We show how the model can be computationally efficiently fitted when the delay distribution is the same in multiple strata, for example, over a wide range of ages., (© 2022 The Authors. Journal of the Royal Statistical Society: Series C (Applied Statistics) published by John Wiley & Sons Ltd on behalf of Royal Statistical Society.)

Background: The omicron variant (B.1.1.529) of SARS-CoV-2 has demonstrated partial vaccine escape and high transmissibility, with early studies indicating lower severity of infection than that of the delta variant (B.1.617.2). We aimed to better characterise omicron severity relative to delta by assessing the relative risk of hospital attendance, hospital admission, or death in a large national cohort., Methods: Individual-level data on laboratory-confirmed COVID-19 cases resident in England between Nov 29, 2021, and Jan 9, 2022, were linked to routine datasets on vaccination status, hospital attendance and admission, and mortality. The relative risk of hospital attendance or admission within 14 days, or death within 28 days after confirmed infection, was estimated using proportional hazards regression. Analyses were stratified by test date, 10-year age band, ethnicity, residential region, and vaccination status, and were further adjusted for sex, index of multiple deprivation decile, evidence of a previous infection, and year of age within each age band. A secondary analysis estimated variant-specific and vaccine-specific vaccine effectiveness and the intrinsic relative severity of omicron infection compared with delta (ie, the relative risk in unvaccinated cases)., Findings: The adjusted hazard ratio (HR) of hospital attendance (not necessarily resulting in admission) with omicron compared with delta was 0·56 (95% CI 0·54-0·58); for hospital admission and death, HR estimates were 0·41 (0·39-0·43) and 0·31 (0·26-0·37), respectively. Omicron versus delta HR estimates varied with age for all endpoints examined. The adjusted HR for hospital admission was 1·10 (0·85-1·42) in those younger than 10 years, decreasing to 0·25 (0·21-0·30) in 60-69-year-olds, and then increasing to 0·47 (0·40-0·56) in those aged at least 80 years. For both variants, past infection gave some protection against death both in vaccinated (HR 0·47 [0·32-0·68]) and unvaccinated (0·18 [0·06-0·57]) cases. In vaccinated cases, past infection offered no additional protection against hospital admission beyond that provided by vaccination (HR 0·96 [0·88-1·04]); however, for unvaccinated cases, past infection gave moderate protection (HR 0·55 [0·48-0·63]). Omicron versus delta HR estimates were lower for hospital admission (0·30 [0·28-0·32]) in unvaccinated cases than the corresponding HR estimated for all cases in the primary analysis. Booster vaccination with an mRNA vaccine was highly protective against hospitalisation and death in omicron cases (HR for hospital admission 8-11 weeks post-booster vs unvaccinated: 0·22 [0·20-0·24]), with the protection afforded after a booster not being affected by the vaccine used for doses 1 and 2., Interpretation: The risk of severe outcomes following SARS-CoV-2 infection is substantially lower for omicron than for delta, with higher reductions for more severe endpoints and significant variation with age. Underlying the observed risks is a larger reduction in intrinsic severity (in unvaccinated individuals) counterbalanced by a reduction in vaccine effectiveness. Documented previous SARS-CoV-2 infection offered some protection against hospitalisation and high protection against death in unvaccinated individuals, but only offered additional protection in vaccinated individuals for the death endpoint. Booster vaccination with mRNA vaccines maintains over 70% protection against hospitalisation and death in breakthrough confirmed omicron infections., Funding: Medical Research Council, UK Research and Innovation, Department of Health and Social Care, National Institute for Health Research, Community Jameel, and Engineering and Physical Sciences Research Council., Competing Interests: Declaration of interests GD declares that his employer UK Health Security Agency (previously operating as Public Health England) received funding from GlaxoSmithKline for a research project related to influenza antiviral treatment. This preceded and had no relation to COVID-19, and GD had no role in and received no funding from the project. All other authors declare no competing interests., (Copyright © 2022 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license. Published by Elsevier Ltd.. All rights reserved.)

We offer a natural and extensible measure-theoretic treatment of missingness at random. Within the standard missing-data framework, we give a novel characterization of the observed data as a stopping-set sigma algebra. We demonstrate that the usual missingness-at-random conditions are equivalent to requiring particular stochastic processes to be adapted to a set-indexed filtration. These measurability conditions ensure the usual factorization of likelihood ratios. We illustrate how the theory can be extended easily to incorporate explanatory variables, to describe longitudinal data in continuous time, and to admit more general coarsening of observations.

Background: The SARS-CoV-2 delta (B.1.617.2) variant was first detected in England in March, 2021. It has since rapidly become the predominant lineage, owing to high transmissibility. It is suspected that the delta variant is associated with more severe disease than the previously dominant alpha (B.1.1.7) variant. We aimed to characterise the severity of the delta variant compared with the alpha variant by determining the relative risk of hospital attendance outcomes., Methods: This cohort study was done among all patients with COVID-19 in England between March 29 and May 23, 2021, who were identified as being infected with either the alpha or delta SARS-CoV-2 variant through whole-genome sequencing. Individual-level data on these patients were linked to routine health-care datasets on vaccination, emergency care attendance, hospital admission, and mortality (data from Public Health England's Second Generation Surveillance System and COVID-19-associated deaths dataset; the National Immunisation Management System; and NHS Digital Secondary Uses Services and Emergency Care Data Set). The risk for hospital admission and emergency care attendance were compared between patients with sequencing-confirmed delta and alpha variants for the whole cohort and by vaccination status subgroups. Stratified Cox regression was used to adjust for age, sex, ethnicity, deprivation, recent international travel, area of residence, calendar week, and vaccination status., Findings: Individual-level data on 43 338 COVID-19-positive patients (8682 with the delta variant, 34 656 with the alpha variant; median age 31 years [IQR 17-43]) were included in our analysis. 196 (2·3%) patients with the delta variant versus 764 (2·2%) patients with the alpha variant were admitted to hospital within 14 days after the specimen was taken (adjusted hazard ratio [HR] 2·26 [95% CI 1·32-3·89]). 498 (5·7%) patients with the delta variant versus 1448 (4·2%) patients with the alpha variant were admitted to hospital or attended emergency care within 14 days (adjusted HR 1·45 [1·08-1·95]). Most patients were unvaccinated (32 078 [74·0%] across both groups). The HRs for vaccinated patients with the delta variant versus the alpha variant (adjusted HR for hospital admission 1·94 [95% CI 0·47-8·05] and for hospital admission or emergency care attendance 1·58 [0·69-3·61]) were similar to the HRs for unvaccinated patients (2·32 [1·29-4·16] and 1·43 [1·04-1·97]; p=0·82 for both) but the precision for the vaccinated subgroup was low., Interpretation: This large national study found a higher hospital admission or emergency care attendance risk for patients with COVID-19 infected with the delta variant compared with the alpha variant. Results suggest that outbreaks of the delta variant in unvaccinated populations might lead to a greater burden on health-care services than the alpha variant., Funding: Medical Research Council; UK Research and Innovation; Department of Health and Social Care; and National Institute for Health Research., Competing Interests: Declaration of interests GD's employer, Public Health England, has received funding from GlaxoSmithKline for a research project related to seasonal influenza and antiviral treatment; this project preceded and had no relation to COVID-19, and GD had no role in and received no funding from the project. All other authors declare no competing interests., (Crown copyright © 2022 Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license.)

Risk-prediction models for health outcomes are used in practice as part of clinical decision-making, and it is essential that their performance be externally validated. An important aspect in the design of a validation study is choosing an adequate sample size. In this paper, we investigate the sample size requirements for validation studies with binary outcomes to estimate measures of predictive performance (C-statistic for discrimination and calibration slope and calibration in the large). We aim for sufficient precision in the estimated measures. In addition, we investigate the sample size to achieve sufficient power to detect a difference from a target value. Under normality assumptions on the distribution of the linear predictor, we obtain simple estimators for sample size calculations based on the measures above. Simulation studies show that the estimators perform well for common values of the C-statistic and outcome prevalence when the linear predictor is marginally Normal. Their performance deteriorates only slightly when the normality assumptions are violated. We also propose estimators which do not require normality assumptions but require specification of the marginal distribution of the linear predictor and require the use of numerical integration. These estimators were also seen to perform very well under marginal normality. Our sample size equations require a specified standard error (SE) and the anticipated C-statistic and outcome prevalence. The sample size requirement varies according to the prognostic strength of the model, outcome prevalence, choice of the performance measure and study objective. For example, to achieve an SE < 0.025 for the C-statistic, 60-170 events are required if the true C-statistic and outcome prevalence are between 0.64-0.85 and 0.05-0.3, respectively. For the calibration slope and calibration in the large, achieving SE < 0.15   would require 40-280 and 50-100 events, respectively. Our estimators may also be used for survival outcomes when the proportion of censored observations is high.

Observational longitudinal data on treatments and covariates are increasingly used to investigate treatment effects, but are often subject to time-dependent confounding. Marginal structural models (MSMs), estimated using inverse probability of treatment weighting or the g-formula, are popular for handling this problem. With increasing development of advanced causal inference methods, it is important to be able to assess their performance in different scenarios to guide their application. Simulation studies are a key tool for this, but their use to evaluate causal inference methods has been limited. This paper focuses on the use of simulations for evaluations involving MSMs in studies with a time-to-event outcome. In a simulation, it is important to be able to generate the data in such a way that the correct forms of any models to be fitted to those data are known. However, this is not straightforward in the longitudinal setting because it is natural for data to be generated in a sequential conditional manner, whereas MSMs involve fitting marginal rather than conditional hazard models. We provide general results that enable the form of the correctly specified MSM to be derived based on a conditional data generating procedure, and show how the results can be applied when the conditional hazard model is an Aalen additive hazard or Cox model. Using conditional additive hazard models is advantageous because they imply additive MSMs that can be fitted using standard software. We describe and illustrate a simulation algorithm. Our results will help researchers to effectively evaluate causal inference methods via simulation., (© 2021 The Authors. Biometrical Journal published by Wiley-VCH GmbH.)

Using data from observational studies to estimate the causal effect of a time-varying exposure, repeatedly measured over time, on an outcome of interest requires careful adjustment for confounding. Standard regression adjustment for observed time-varying confounders is unsuitable, as it can eliminate part of the causal effect and induce bias. Inverse probability weighting, g-computation, and g-estimation have been proposed as being more suitable methods. G-estimation has some advantages over the other two methods, but until recently there has been a lack of flexible g-estimation methods for a survival time outcome. The recently proposed Structural Nested Cumulative Survival Time Model (SNCSTM) is such a method. Efficient estimation of the parameters of this model required bespoke software. In this article we show how the SNCSTM can be fitted efficiently via g-estimation using standard software for fitting generalised linear models. The ability to implement g-estimation for a survival outcome using standard statistical software greatly increases the potential uptake of this method. We illustrate the use of this method of fitting the SNCSTM by reanalyzing data from the UK Cystic Fibrosis Registry, and provide example R code to facilitate the use of this approach by other researchers., (© 2021 The Authors. Statistics in Medicine published by John Wiley & Sons Ltd.)

A problem that is frequently encountered in many areas of scientific research is that of estimating the effect of a non-randomized binary intervention on an outcome of interest by using time series data on units that received the intervention ('treated') and units that did not ('controls'). One popular estimation method in this setting is based on the factor analysis (FA) model. The FA model is fitted to the preintervention outcome data on treated units and all the outcome data on control units, and the counterfactual treatment-free post-intervention outcomes of the former are predicted from the fitted model. Intervention effects are estimated as the observed outcomes minus these predicted counterfactual outcomes. We propose a model that extends the FA model for estimating intervention effects by jointly modelling the multiple outcomes to exploit shared variability, and assuming an auto-regressive structure on factors to account for temporal correlations in the outcome. Using simulation studies, we show that the method proposed can improve the precision of the intervention effect estimates and achieve better control of the type I error rate (compared with the FA model), especially when either the number of preintervention measurements or the number of control units is small. We apply our method to estimate the effect of stricter alcohol licensing policies on alcohol-related harms.

We develop and demonstrate methods to perform sensitivity analyses to assess sensitivity to plausible departures from missing at random in incomplete repeated binary outcome data. We use multiple imputation in the not at random fully conditional specification framework, which includes one or more sensitivity parameters (SPs) for each incomplete variable. The use of an online elicitation questionnaire is demonstrated to obtain expert opinion on the SPs, and highest prior density regions are used alongside opinion pooling methods to display credible regions for SPs. We demonstrate that substantive conclusions can be far more sensitive to departures from the missing at random assumption (MAR) when control and intervention nonresponders depart from MAR differently, and show that the correlation of arm specific SPs in expert opinion is particularly important. We illustrate these methods on the iQuit in Practice smoking cessation trial, which compared the impact of a tailored text messaging system versus standard care on smoking cessation. We show that conclusions about the effect of intervention on smoking cessation outcomes at 8 week and 6 months are broadly insensitive to departures from MAR, with conclusions significantly affected only when the differences in behavior between the nonresponders in the two trial arms is larger than expert opinion judges to be realistic., (© 2020 The Authors. Statistics in Medicine published by John Wiley & Sons, Ltd.)

Accurate prognosis information after a diagnosis of chronic obstructive pulmonary disease (COPD) would facilitate earlier and better informed decisions about the use of prevention strategies and advanced care plans. We therefore aimed to develop and validate an accurate prognosis model for incident COPD cases using only information present in general practitioner (GP) records at the point of diagnosis. Incident COPD patients between 2004-2012 over the age of 35 were studied using records from 396 general practices in England. We developed a model to predict all-cause five-year mortality at the point of COPD diagnosis, using 47,964 English patients. Our model uses age, gender, smoking status, body mass index, forced expiratory volume in 1-second (FEV1) % predicted and 16 co-morbidities (the same number as the Charlson Co-morbidity Index). The performance of our chosen model was validated in all countries of the UK (N = 48,304). Our model performed well, and performed consistently in validation data. The validation area under the curves in each country varied between 0.783-0.809 and the calibration slopes between 0.911-1.04. Our model performed better in this context than models based on the Charlson Co-morbidity Index or Cambridge Multimorbidity Score. We have developed and validated a model that outperforms general multimorbidity scores at predicting five-year mortality after COPD diagnosis. Our model includes only data routinely collected before COPD diagnosis, allowing it to be readily translated into clinical practice, and has been made available through an online risk calculator (https://skiddle.shinyapps.io/incidentcopdsurvival/)., Competing Interests: Dr. Kiddle reports grants from Medical Research Council, during the conduct of the study; personal fees from Roche Diagnostics and DIADEM, outside the submitted work. After completing this work, but before manuscript submission Dr. Kiddle became an employee of AstraZeneca. Ms. Whittaker reports grants from GlaxoSmithKline, during the conduct of the study. Dr. Seaman has nothing to disclose. Dr. Quint reports grants from MRC, grants from The Health Foundation, grants from BLF, grants and personal fees from GSK, grants and personal fees from BI, grants and personal fees from Insmed, grants and personal fees from AZ, personal fees from Chiesi, personal fees from Teva, outside the submitted work. This does not alter our adherence to PLOS ONE policies on sharing data and materials.

It is often of interest to explore how dose affects the toxicity and efficacy properties of a novel treatment. In oncology, efficacy is often assessed through response, which is defined by a patient having no new tumour lesions and their tumour size shrinking by 30%. Usually response and toxicity are analysed as binary outcomes in early phase trials. Methods have been proposed to improve the efficiency of analysing response by utilising the continuous tumour size information instead of dichotomising it. However, these methods do not allow for toxicity or for different doses. Motivated by a phase II trial testing multiple doses of a treatment against placebo, we propose a latent variable model that can estimate the probability of response and no toxicity (or other related outcomes) for different doses. We assess the confidence interval coverage and efficiency properties of the method, compared to methods that do not use the continuous tumour size, in a simulation study and the real study. The coverage is close to nominal when model assumptions are met, although can be below nominal when the model is misspecified. Compared to methods that treat response as binary, the method has confidence intervals with 30-50% narrower widths. The method adds considerable efficiency but care must be taken that the model assumptions are reasonable.

Background: Cystic fibrosis (CF) is an inherited, chronic, progressive condition affecting around 10,000 individuals in the United Kingdom and over 70,000 worldwide. Survival in CF has improved considerably over recent decades, and it is important to provide up-to-date information on patient prognosis., Methods: The UK Cystic Fibrosis Registry is a secure centralized database, which collects annual data on almost all CF patients in the United Kingdom. Data from 43,592 annual records from 2005 to 2015 on 6181 individuals were used to develop a dynamic survival prediction model that provides personalized estimates of survival probabilities given a patient's current health status using 16 predictors. We developed the model using the landmarking approach, giving predicted survival curves up to 10 years from 18 to 50 years of age. We compared several models using cross-validation., Results: The final model has good discrimination (C-indexes: 0.873, 0.843, and 0.804 for 2-, 5-, and 10-year survival prediction) and low prediction error (Brier scores: 0.036, 0.076, and 0.133). It identifies individuals at low and high risk of short- and long-term mortality based on their current status. For patients 20 years of age during 2013-2015, for example, over 80% had a greater than 95% probability of 2-year survival and 40% were predicted to survive 10 years or more., Conclusions: Dynamic personalized prediction models can guide treatment decisions and provide personalized information for patients. Our application illustrates the utility of the landmarking approach for making the best use of longitudinal and survival data and shows how models can be defined and compared in terms of predictive performance.

Inverse probability of treatment weighting is a popular propensity score-based approach to estimate marginal treatment effects in observational studies at risk of confounding bias. A major issue when estimating the propensity score is the presence of partially observed covariates. Multiple imputation is a natural approach to handle missing data on covariates: covariates are imputed and a propensity score analysis is performed in each imputed dataset to estimate the treatment effect. The treatment effect estimates from each imputed dataset are then combined to obtain an overall estimate. We call this method MIte. However, an alternative approach has been proposed, in which the propensity scores are combined across the imputed datasets (MIps). Therefore, there are remaining uncertainties about how to implement multiple imputation for propensity score analysis: (a) should we apply Rubin's rules to the inverse probability of treatment weighting treatment effect estimates or to the propensity score estimates themselves? (b) does the outcome have to be included in the imputation model? (c) how should we estimate the variance of the inverse probability of treatment weighting estimator after multiple imputation? We studied the consistency and balancing properties of the MIte and MIps estimators and performed a simulation study to empirically assess their performance for the analysis of a binary outcome. We also compared the performance of these methods to complete case analysis and the missingness pattern approach, which uses a different propensity score model for each pattern of missingness, and a third multiple imputation approach in which the propensity score parameters are combined rather than the propensity scores themselves (MIpar). Under a missing at random mechanism, complete case and missingness pattern analyses were biased in most cases for estimating the marginal treatment effect, whereas multiple imputation approaches were approximately unbiased as long as the outcome was included in the imputation model. Only MIte was unbiased in all the studied scenarios and Rubin's rules provided good variance estimates for MIte. The propensity score estimated in the MIte approach showed good balancing properties. In conclusion, when using multiple imputation in the inverse probability of treatment weighting context, MIte with the outcome included in the imputation model is the preferred approach.

We propose semi-parametric methods to model cohort data where repeated outcomes may be missing due to death and non-ignorable dropout. Our focus is to obtain inference about the cohort composed of those who are still alive at any time point (partly conditional inference). We propose: i) an inverse probability weighted method that upweights observed subjects to represent subjects who are still alive but are not observed; ii) an outcome regression method that replaces missing outcomes of subjects who are alive with their conditional mean outcomes given past observed data; and iii) an augmented inverse probability method that combines the previous two methods and is double robust against model misspecification. These methods are described for both monotone and non-monotone missing data patterns, and are applied to a cohort of elderly adults from the Health and Retirement Study. Sensitivity analysis to departures from the assumption that missingness at some visit t is independent of the outcome at visit t given past observed data and time of death is used in the data application., (© 2018, The Authors. Biometrics published by Wiley Periodicals, Inc. on behalf of International Biometric Society.)

The nested case-control and case-cohort designs are two main approaches for carrying out a substudy within a prospective cohort. This article adapts multiple imputation (MI) methods for handling missing covariates in full-cohort studies for nested case-control and case-cohort studies. We consider data missing by design and data missing by chance. MI analyses that make use of full-cohort data and MI analyses based on substudy data only are described, alongside an intermediate approach in which the imputation uses full-cohort data but the analysis uses only the substudy. We describe adaptations to two imputation methods: the approximate method (MI-approx) of White and Royston (2009) and the "substantive model compatible" (MI-SMC) method of Bartlett et al. (2015). We also apply the "MI matched set" approach of Seaman and Keogh (2015) to nested case-control studies, which does not require any full-cohort information. The methods are investigated using simulation studies and all perform well when their assumptions hold. Substantial gains in efficiency can be made by imputing data missing by design using the full-cohort approach or by imputing data missing by chance in analyses using the substudy only. The intermediate approach brings greater gains in efficiency relative to the substudy approach and is more robust to imputation model misspecification than the full-cohort approach. The methods are illustrated using the ARIC Study cohort. Supplementary Materials provide R and Stata code., (© 2018 The Authors. Biometrics published by Wiley Periodicals, Inc. on behalf of International Biometric Society.)

Cohort data are often incomplete because some subjects drop out of the study, and inverse probability weighting (IPW), multiple imputation (MI), and linear increments (LI) are methods that deal with such missing data. In cohort studies of ageing, missing data can arise from dropout or death. Methods that do not distinguish between these reasons for missingness typically provide inference about a hypothetical cohort where no one can die (immortal cohort). It has been suggested that inference about the cohort composed of those who are still alive at any time point (partly conditional inference) may be more meaningful. MI, LI, and IPW can all be adapted to provide partly conditional inference. In this article, we clarify and compare the assumptions required by these MI, LI, and IPW methods for partly conditional inference on continuous outcomes. We also propose augmented IPW estimators for making partly conditional inference. These are more efficient than IPW estimators and more robust to model misspecification. Our simulation studies show that the methods give approximately unbiased estimates of partly conditional estimands when their assumptions are met, but may be biased otherwise. We illustrate the application of the missing data methods using data from the 'Origins of Variance in the Old-old' Twin study.

Estimating the parameters of a regression model of interest is complicated by missing data on the variables in that model. Multiple imputation is commonly used to handle these missing data. Joint model multiple imputation and full-conditional specification multiple imputation are known to yield imputed data with the same asymptotic distribution when the conditional models of full-conditional specification are compatible with that joint model. We show that this asymptotic equivalence of imputation distributions does not imply that joint model multiple imputation and full-conditional specification multiple imputation will also yield asymptotically equally efficient inference about the parameters of the model of interest, nor that they will be equally robust to misspecification of the joint model. When the conditional models used by full-conditional specification multiple imputation are linear, logistic and multinomial regressions, these are compatible with a restricted general location joint model. We show that multiple imputation using the restricted general location joint model can be substantially more asymptotically efficient than full-conditional specification multiple imputation, but this typically requires very strong associations between variables. When associations are weaker, the efficiency gain is small. Moreover, full-conditional specification multiple imputation is shown to be potentially much more robust than joint model multiple imputation using the restricted general location model to mispecification of that model when there is substantial missingness in the outcome variable.

Most methods for handling incomplete data can be broadly classified as inverse probability weighting (IPW) strategies or imputation strategies. The former model the occurrence of incomplete data; the latter, the distribution of the missing variables given observed variables in each missingness pattern. Imputation strategies are typically more efficient, but they can involve extrapolation, which is difficult to diagnose and can lead to large bias. Double robust (DR) methods combine the two approaches. They are typically more efficient than IPW and more robust to model misspecification than imputation. We give a formal introduction to DR estimation of the mean of a partially observed variable, before moving to more general incomplete-data scenarios. We review strategies to improve the performance of DR estimators under model misspecification, reveal connections between DR estimators for incomplete data and 'design-consistent' estimators used in sample surveys, and explain the value of double robustness when using flexible data-adaptive methods for IPW or imputation.

Cannabinoids are the most commonly used illegal substances in the world [1]. Synthetic Cannabinoids (SCB) are also known as "Spice", "K2", "Spike", "herbal incense", "Cloud 9", "Mojo" and many others are becoming a large public health concern due to their increasing use, unpredictable toxicity, and abuse potential [2]. The most common reported toxicities with SCB use based on studies using Texas Poison control record are tachycardia, agitation and irritability, drowsiness, hallucinations, delusions, hypertension, nausea, confusion, dizziness, vertigo, chest pain, acute kidney injury, seizures, heart attacks and both ischemic and hemorrhagic strokes [3]. The Emergency Department (ED) here at Lincoln Medical Center has certainly seen a sizeable volume of K2 abusers who present displaying a spectrum of symptoms as noted above. However, during a concentrated outbreak of K2 use in the summer of 2015, a large cohort of patients presented with a toxidrome not previously described in any published literature. This included marked bradycardia and hypotension while maintaining global neurologic function. Although these patients were drowsy and sleepy at presentation, tactile stimuli would arouse these patients to awaken and participate in an interview. The patients described in this case series, appeared to be on the brink of cardiovascular collapse. The vital signs however normalized with intravenous fluid (IVF) hydration only, over the course of 6 to 7h, allowing a safe discharge from the ED., (Copyright © 2017 Elsevier Inc. All rights reserved.)

Linear increments (LI) are used to analyse repeated outcome data with missing values. Previously, two LI methods have been proposed, one allowing non-monotone missingness but not independent measurement error and one allowing independent measurement error but only monotone missingness. In both, it was suggested that the expected increment could depend on current outcome. We show that LI can allow non-monotone missingness and either independent measurement error of unknown variance or dependence of expected increment on current outcome but not both. A popular alternative to LI is a multivariate normal model ignoring the missingness pattern. This gives consistent estimation when data are normally distributed and missing at random (MAR). We clarify the relation between MAR and the assumptions of LI and show that for continuous outcomes multivariate normal estimators are also consistent under (non-MAR and non-normal) assumptions not much stronger than those of LI. Moreover, when missingness is non-monotone, they are typically more efficient.

Analysis of matched case-control studies is often complicated by missing data on covariates. Analysis can be restricted to individuals with complete data, but this is inefficient and may be biased. Multiple imputation (MI) is an efficient and flexible alternative. We describe two MI approaches. The first uses a model for the data on an individual and includes matching variables; the second uses a model for the data on a whole matched set and avoids the need to model the matching variables. Within each approach, we consider three methods: full-conditional specification (FCS), joint model MI using a normal model, and joint model MI using a latent normal model. We show that FCS MI is asymptotically equivalent to joint model MI using a restricted general location model that is compatible with the conditional logistic regression analysis model. The normal and latent normal imputation models are not compatible with this analysis model. All methods allow for multiple partially-observed covariates, non-monotone missingness, and multiple controls per case. They can be easily applied in standard statistical software and valid variance estimates obtained using Rubin's Rules. We compare the methods in a simulation study. The approach of including the matching variables is most efficient. Within each approach, the FCS MI method generally yields the least-biased odds ratio estimates, but normal or latent normal joint model MI is sometimes more efficient. All methods have good confidence interval coverage. Data on colorectal cancer and fibre intake from the EPIC-Norfolk study are used to illustrate the methods, in particular showing how efficiency is gained relative to just using individuals with complete data., (© 2015 The Authors Biometrics published by Wiley Periodicals, Inc. on behalf of International Biometric Society.)

Background: Informative birth size occurs when the average outcome depends on the number of infants per birth. Although analysis methods have been proposed for handling informative birth size, their performance is not well understood. Our aim was to evaluate the performance of these methods and to provide recommendations for their application in randomised trials including infants from single and multiple births., Methods: Three generalised estimating equation (GEE) approaches were considered for estimating the effect of treatment on a continuous or binary outcome: cluster weighted GEEs, which produce treatment effects with a mother-level interpretation when birth size is informative; standard GEEs with an independence working correlation structure, which produce treatment effects with an infant-level interpretation when birth size is informative; and standard GEEs with an exchangeable working correlation structure, which do not account for informative birth size. The methods were compared through simulation and analysis of an example dataset., Results: Treatment effect estimates were affected by informative birth size in the simulation study when the effect of treatment in singletons differed from that in multiples (i.e. in the presence of a treatment group by multiple birth interaction). The strength of evidence supporting the effectiveness of treatment varied between methods in the example dataset., Conclusions: Informative birth size is always a possibility in randomised trials including infants from both single and multiple births, and analysis methods should be pre-specified with this in mind. We recommend estimating treatment effects using standard GEEs with an independence working correlation structure to give an infant-level interpretation., (© 2015 John Wiley & Sons Ltd.)

Missing covariate data commonly occur in epidemiological and clinical research, and are often dealt with using multiple imputation. Imputation of partially observed covariates is complicated if the substantive model is non-linear (e.g. Cox proportional hazards model), or contains non-linear (e.g. squared) or interaction terms, and standard software implementations of multiple imputation may impute covariates from models that are incompatible with such substantive models. We show how imputation by fully conditional specification, a popular approach for performing multiple imputation, can be modified so that covariates are imputed from models which are compatible with the substantive model. We investigate through simulation the performance of this proposal, and compare it with existing approaches. Simulation results suggest our proposal gives consistent estimates for a range of common substantive models, including models which contain non-linear covariate effects or interactions, provided data are missing at random and the assumed imputation models are correctly specified and mutually compatible. Stata software implementing the approach is freely available., (© The Author(s) 2014.)

The C statistic is a commonly reported measure of screening test performance. Optimistic estimation of the C statistic is a frequent problem because of overfitting of statistical models in small data sets, and methods exist to correct for this issue. However, many studies do not use such methods, and those that do correct for optimism use diverse methods, some of which are known to be biased. We used clinical data sets (United Kingdom Down syndrome screening data from Glasgow (1991-2003), Edinburgh (1999-2003), and Cambridge (1990-2006), as well as Scottish national pregnancy discharge data (2004-2007)) to evaluate different approaches to adjustment for optimism. We found that sample splitting, cross-validation without replication, and leave-1-out cross-validation produced optimism-adjusted estimates of the C statistic that were biased and/or associated with greater absolute error than other available methods. Cross-validation with replication, bootstrapping, and a new method (leave-pair-out cross-validation) all generated unbiased optimism-adjusted estimates of the C statistic and had similar absolute errors in the clinical data set. Larger simulation studies confirmed that all 3 methods performed similarly with 10 or more events per variable, or when the C statistic was 0.9 or greater. However, with lower events per variable or lower C statistics, bootstrapping tended to be optimistic but with lower absolute and mean squared errors than both methods of cross-validation., (© The Author 2014. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health.)

Clustered data commonly arise in epidemiology. We assume each cluster member has an outcome Y and covariates X. When there are missing data in Y, the distribution of Y given X in all cluster members ("complete clusters") may be different from the distribution just in members with observed Y ("observed clusters"). Often the former is of interest, but when data are missing because in a fundamental sense Y does not exist (e.g., quality of life for a person who has died), the latter may be more meaningful (quality of life conditional on being alive). Weighted and doubly weighted generalized estimating equations and shared random-effects models have been proposed for observed-cluster inference when cluster size is informative, that is, the distribution of Y given X in observed clusters depends on observed cluster size. We show these methods can be seen as actually giving inference for complete clusters and may not also give observed-cluster inference. This is true even if observed clusters are complete in themselves rather than being the observed part of larger complete clusters: here methods may describe imaginary complete clusters rather than the observed clusters. We show under which conditions shared random-effects models proposed for observed-cluster inference do actually describe members with observed Y. A psoriatic arthritis dataset is used to illustrate the danger of misinterpreting estimates from shared random-effects models., (© 2014 The Authors. Biometrics published by Wiley Periodicals, Inc. on behalf of International Biometric Society.)

Background: Chained equations imputation is widely used in medical research. It uses a set of conditional models, so is more flexible than joint modelling imputation for the imputation of different types of variables (e.g. binary, ordinal or unordered categorical). However, chained equations imputation does not correspond to drawing from a joint distribution when the conditional models are incompatible. Concurrently with our work, other authors have shown the equivalence of the two imputation methods in finite samples., Methods: Taking a different approach, we prove, in finite samples, sufficient conditions for chained equations and joint modelling to yield imputations from the same predictive distribution. Further, we apply this proof in four specific cases and conduct a simulation study which explores the consequences when the conditional models are compatible but the conditions otherwise are not satisfied., Results: We provide an additional "non-informative margins" condition which, together with compatibility, is sufficient. We show that the non-informative margins condition is not satisfied, despite compatible conditional models, in a situation as simple as two continuous variables and one binary variable. Our simulation study demonstrates that as a consequence of this violation order effects can occur; that is, systematic differences depending upon the ordering of the variables in the chained equations algorithm. However, the order effects appear to be small, especially when associations between variables are weak., Conclusions: Since chained equations is typically used in medical research for datasets with different types of variables, researchers must be aware that order effects are likely to be ubiquitous, but our results suggest they may be small enough to be negligible.

We are concerned with multiple imputation of the ratio of two variables, which is to be used as a covariate in a regression analysis. If the numerator and denominator are not missing simultaneously, it seems sensible to make use of the observed variable in the imputation model. One such strategy is to impute missing values for the numerator and denominator, or the log-transformed numerator and denominator, and then calculate the ratio of interest; we call this 'passive' imputation. Alternatively, missing ratio values might be imputed directly, with or without the numerator and/or the denominator in the imputation model; we call this 'active' imputation. In two motivating datasets, one involving body mass index as a covariate and the other involving the ratio of total to high-density lipoprotein cholesterol, we assess the sensitivity of results to the choice of imputation model and, as an alternative, explore fully Bayesian joint models for the outcome and incomplete ratio. Fully Bayesian approaches using Winbugs were unusable in both datasets because of computational problems. In our first dataset, multiple imputation results are similar regardless of the imputation model; in the second, results are sensitive to the choice of imputation model. Sensitivity depends strongly on the coefficient of variation of the ratio's denominator. A simulation study demonstrates that passive imputation without transformation is risky because it can lead to downward bias when the coefficient of variation of the ratio's denominator is larger than about 0.1. Active imputation or passive imputation after log-transformation is preferable., (© 2013 The Authors. Statistics in Medicine published by John Wiley & Sons, Ltd.)

In phase II cancer trials, tumour response is either the primary or an important secondary endpoint. Tumour response is a binary composite endpoint determined, according to the Response Evaluation Criteria in Solid Tumors, by (1) whether the percentage change in tumour size is greater than a prescribed threshold and (2) (binary) criteria such as whether a patient develops new lesions. Further binary criteria, such as death or serious toxicity, may be added to these criteria. The probability of tumour response (i.e. 'success' on the composite endpoint) would usually be estimated simply as the proportion of successes among patients. This approach uses the tumour size variable only through a discretised form, namely whether or not it is above the threshold. In this article, we propose a method that also estimates the probability of success but that gains precision by using the information on the undiscretised (i.e. continuous) tumour size variable. This approach can also be used to increase the power to detect a difference between the probabilities of success under two different treatments in a comparative trial. We demonstrate these increases in precision and power using simulated data. We also apply the method to real data from a phase II cancer trial and show that it results in a considerably narrower confidence interval for the probability of tumour response., (© 2013 The authors. Statistics in Medicine published by John Wiley & Sons, Ltd.)

The simplest approach to dealing with missing data is to restrict the analysis to complete cases, i.e. individuals with no missing values. This can induce bias, however. Inverse probability weighting (IPW) is a commonly used method to correct this bias. It is also used to adjust for unequal sampling fractions in sample surveys. This article is a review of the use of IPW in epidemiological research. We describe how the bias in the complete-case analysis arises and how IPW can remove it. IPW is compared with multiple imputation (MI) and we explain why, despite MI generally being more efficient, IPW may sometimes be preferred. We discuss the choice of missingness model and methods such as weight truncation, weight stabilisation and augmented IPW. The use of IPW is illustrated on data from the 1958 British Birth Cohort.

Background: Multiple imputation is often used for missing data. When a model contains as covariates more than one function of a variable, it is not obvious how best to impute missing values in these covariates. Consider a regression with outcome Y and covariates X and X2. In 'passive imputation' a value X* is imputed for X and then X2 is imputed as (X*)2. A recent proposal is to treat X2 as 'just another variable' (JAV) and impute X and X2 under multivariate normality., Methods: We use simulation to investigate the performance of three methods that can easily be implemented in standard software: 1) linear regression of X on Y to impute X then passive imputation of X2; 2) the same regression but with predictive mean matching (PMM); and 3) JAV. We also investigate the performance of analogous methods when the analysis involves an interaction, and study the theoretical properties of JAV. The application of the methods when complete or incomplete confounders are also present is illustrated using data from the EPIC Study., Results: JAV gives consistent estimation when the analysis is linear regression with a quadratic or interaction term and X is missing completely at random. When X is missing at random, JAV may be biased, but this bias is generally less than for passive imputation and PMM. Coverage for JAV was usually good when bias was small. However, in some scenarios with a more pronounced quadratic effect, bias was large and coverage poor. When the analysis was logistic regression, JAV's performance was sometimes very poor. PMM generally improved on passive imputation, in terms of bias and coverage, but did not eliminate the bias., Conclusions: Given the current state of available software, JAV is the best of a set of imperfect imputation methods for linear regression with a quadratic or interaction effect, but should not be used for logistic regression.

Two approaches commonly used to deal with missing data are multiple imputation (MI) and inverse-probability weighting (IPW). IPW is also used to adjust for unequal sampling fractions. MI is generally more efficient than IPW but more complex. Whereas IPW requires only a model for the probability that an individual has complete data (a univariate outcome), MI needs a model for the joint distribution of the missing data (a multivariate outcome) given the observed data. Inadequacies in either model may lead to important bias if large amounts of data are missing. A third approach combines MI and IPW to give a doubly robust estimator. A fourth approach (IPW/MI) combines MI and IPW but, unlike doubly robust methods, imputes only isolated missing values and uses weights to account for remaining larger blocks of unimputed missing data, such as would arise, e.g., in a cohort study subject to sample attrition, and/or unequal sampling fractions. In this article, we examine the performance, in terms of bias and efficiency, of IPW/MI relative to MI and IPW alone and investigate whether the Rubin's rules variance estimator is valid for IPW/MI. We prove that the Rubin's rules variance estimator is valid for IPW/MI for linear regression with an imputed outcome, we present simulations supporting the use of this variance estimator in more general settings, and we demonstrate that IPW/MI can have advantages over alternatives. IPW/MI is applied to data from the National Child Development Study., (© 2011, The International Biometric Society.)

Background: Hard-to-reach population subgroups are typically investigated using convenience sampling, which may give biased estimates. Combining information from such surveys, a probability survey and clinic surveillance, can potentially minimize the bias. We developed a methodology to estimate the prevalence of undiagnosed HIV infection among men who have sex with men (MSM) in England and Wales aged 16-44 years in 2003, making fuller use of the available data than earlier work., Methods: We performed a synthesis of three data sources: genitourinary medicine clinic surveillance (11 380 tests), a venue-based convenience survey including anonymous HIV testing (3702 MSM) and a general population sexual behaviour survey (134 MSM). A logistic regression model to predict undiagnosed infection was fitted to the convenience survey data and then applied to the MSMs in the population survey to estimate the prevalence of undiagnosed infection in the general MSM population. This estimate was corrected for selection biases in the convenience survey using clinic surveillance data. A sensitivity analysis addressed uncertainty in our assumptions., Results: The estimated prevalence of undiagnosed HIV in MSM was 2.4% [95% confidence interval (95% CI 1.7-3.0%)], and between 1.6% (95% CI 1.1-2.0%) and 3.3% (95% CI 2.4-4.1%) depending on assumptions; corresponding to 5500 (3390-7180), 3610 (2180-4740) and 7570 (4790-9840) men, and undiagnosed fractions of 33, 24 and 40%, respectively., Conclusions: Our estimates are consistent with earlier work that did not make full use of data sources. Reconciling data from multiple sources, including probability-, clinic- and venue-based convenience samples can reduce bias in estimates. This methodology could be applied in other settings to take full advantage of multiple imperfect data sources.