Search

Reference-based multiple imputation methods have become popular for handling missing data in randomised clinical trials. Rubin’s variance estimator is well known to be biased compared to the refere...

The National Research Council (NRC) Expert Panel Report on Prevention and Treatment of Missing Data in Clinical Trials highlighted the need for clearly defining objectives and estimands. That report sparked considerable discussion and literature on estimands and how to choose them. Importantly, consideration moved beyond missing data to include all postrandomization events that have implications for estimating quantities of interest (intercurrent events, aka ICEs). The ICH E9(R1) draft addendum builds on that research to outline key principles in choosing estimands for clinical trials, primarily with focus on confirmatory trials. This paper provides additional insights, perspectives, details, and examples to help put ICH E9(R1) into practice. Specific areas of focus include how the perspectives of different stakeholders influence the choice of estimands; the role of randomization and the intention-to-treat principle; defining the causal effects of a clearly defined treatment regimen, along with the implications this has for trial design and the generalizability of conclusions; detailed discussion of strategies for handling ICEs along with their implications and assumptions; estimands for safety objectives, time-to-event endpoints, early-phase and one-arm trials, and quality of life endpoints; and realistic examples of the thought process involved in defining estimands in specific clinical contexts.

Background A complete case logistic regression will give a biased estimate of the exposure odds ratio only if there is a multiplicative interaction between the exposure and outcome with respect to the probability of missingness – whereas linear regression with a continuous outcome is biased in more scenarios, including when only the outcome causes missingness. It is not clear whether a complete case logistic regression will give a biased estimate of the odds ratio if missingness depends on a continuous outcome but this outcome is dichotomised for the analysis – a common situation in epidemiology. Methods We investigated this using a simulation study and data from the Avon Longitudinal Study of Parents and Children (ALSPAC), a UK birth cohort. We also examined whether any bias could be reduced by including a proxy for the binary outcome as an auxiliary variable in multiple imputation. Results There was negligible bias in the exposure odds ratio when the probability of being a complete case was independently associated with the exposure and (continuous) outcome but important bias in the presence of an interaction, particularly at high levels of missing data. Inclusion of the proxy led to significant bias reductions when this had high sensitivity and specificity in relation to the study outcome. Conclusions The robustness of logistic regression to missing data is maintained even when the outcome is a binary version of a continuous outcome. Bias due to an interaction between the exposure and outcome in their effect on selection could be reduced by including proxies for the missing outcome as auxiliary variables in MI. If such proxies are available, we would recommend using MI over a complete case analysis because, in practice, it would be difficult to rule out an interaction.

Multiple imputation (MI) is a method for repairing and analyzing data with missing values. MI replaces missing values with a sample of random values drawn from an imputation model. The most popular form of MI, which we call posterior draw multiple imputation (PDMI), draws the parameters of the imputation model from a Bayesian posterior distribution. An alternative, which we call maximum likelihood multiple imputation (MLMI), estimates the parameters of the imputation model using maximum likelihood (or equivalent). Compared to PDMI, MLMI is less computationally intensive, faster, and yields slightly more efficient point estimates. A past barrier to using MLMI was the difficulty of estimating the standard errors of MLMI point estimates. We derive, implement, and evaluate three consistent standard error formulas: (1) one combines variances within and between the imputed datasets, (2) one uses the score function, and (3) one uses the bootstrap to estimate variance components due to sampling and imputation. Formula (1) modifies for MLMI a formula that has long been used under PDMI, while formulas (2) and (3) can be used without modification under either PDMI or MLMI. We have implemented MLMI and the standard error estimators in the mlmi and bootImpute packages for R., 38 pages, 7 tables, 1 figure

Testing for SARS-CoV-2 internationally has focused on COVID-19 diagnosis among symptomatic individuals using reverse transcriptase polymerase chain reaction (PCR) tests. Recently, however, SARS-CoV-2 antigen rapid lateral flow tests (LFT) have been rolled out in several countries for testing asymptomatic individuals in public health programmes. Validation studies for LFT have been largely cross-sectional, reporting sensitivity, specificity and predictive values of LFT relative to PCR. However, because PCR detects genetic material left behind for a long period when the individual is no longer infectious, these statistics can under-represent sensitivity of LFT for detecting infectious individuals, especially when sampling asymptomatic populations. LFTs (intended to detect individuals with live virus) validated against PCR (intended to diagnose infection) are not reporting against a gold standard of equivalent measurements. Instead, these validation studies have reported relative performance statistics that need recalibrating to the purpose for which LFT is being used. We present an approach to this recalibration.We derive a formula for recalibrating relative performance statistics from LFT vs PCR validation studies to give likely absolute sensitivity of LFT for detecting individuals with live virus. We show the differences between widely reported apparent sensitivities of LFT and its absolute sensitivity as a test of presence of live virus. After accounting for within-individual viral kinetics and epidemic dynamics within asymptomatic populations we show that a highly performant test of live virus should show a LFT-to-PCR relative sensitivity of less than 50% in conventional validation studies, which after re-calibration would be an absolute sensitivity of more than 80%.Further studies are needed to ascertain the absolute sensitivity of LFT as a test of infectiousness in COVID-19 responses. These studies should include sampling for viral cultures and longitudinal series of LFT and PCR, ideally in cohorts sampled from both contacts of cases and the general population.

Irene Petersen, 1 Alexander Crozier, 2 Iain Buchan, 3 Michael J Mina, 4, 5 Jonathan W Bartlett 6 1Department of Primary Care & Population Health, University College London, London, UK; 2Division of Biosciences, University College London, London, UK; 3Institute of Population Health, University of Liverpool, Liverpool, UK; 4Department of Epidemiology, Department of Immunology and Infectious Diseases, Harvard T.H. Chan School of Public Health, Boston, MA, USA; 5Department of Pathology, Clinical Microbiology, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA, USA; 6Department of Mathematical Sciences, University of Bath, Bath, UKCorrespondence: Irene Petersen Email i.petersen@ucl.ac.ukABSTRACT: Testing for SARS-CoV-2 internationally has focused on COVID-19 diagnosis among symptomatic individuals using reverse transcriptase polymerase chain reaction (PCR) tests. Recently, however, SARS-CoV-2 antigen rapid lateral flow tests (LFT) have been rolled out in several countries for testing asymptomatic individuals in public health programmes. Validation studies for LFT have been largely cross-sectional, reporting sensitivity, specificity and predictive values of LFT relative to PCR. However, because PCR detects genetic material left behind for a long period when the individual is no longer infectious, these statistics can under-represent the sensitivity of LFT for detecting infectious individuals, especially when sampling asymptomatic populations. LFTs (intended to detect individuals shedding SARS-CoV-2 antigens) validated against PCR (intended to diagnose infection) are not reporting against a gold standard of equivalent measurements. Instead, these validation studies have reported relative performance statistics that need recalibrating to the purpose for which LFT is being used. We present an approach to this recalibration. We derive a formula for recalibrating relative performance statistics from LFT vs PCR validation studies to give likely absolute sensitivity of LFT for detecting individuals who are shedding shedding SARS-CoV-2 antigens. We contrast widely reported apparent sensitivities of LFT with recalibrated absolute sensitivity for detecting individuals shedding SARS-CoV-2 antigens. After accounting for within-individual viral kinetics and epidemic dynamics within asymptomatic populations we show that a highly performant test for SARS-CoV-2 antigen should show LFT-to-PCR relative sensitivity of less than 50% in conventional validation studies, which after re-calibration would be an absolute sensitivity of more than 80%. Further studies are needed to ascertain the absolute sensitivity of LFT as a test of infectiousness in COVID-19 responses. These studies should include longitudinal series of LFT and PCR, ideally in cohorts sampled from both contacts of cases and the general population.Keywords: rapid test, PCR, validation, recalibration, lateral flow tests

PurposeVolume indices and left ventricular ejection fraction (LVEF) are routinely used to assess cardiac function. Ventricular strain values may provide additional diagnostic information, but their reproducibility is unclear. This study therefore compares the repeatability and reproducibility of volumes, volume fraction, and regional ventricular strains, derived from cardiovascular magnetic resonance (CMR) imaging, across three software packages and between readers.MethodsSeven readers analysed 16 short-axis CMR stacks of a porcine heart. Endocardial contours were manually drawn using OsiriX and Simpleware ScanIP and repeated in both softwares. The images were also contoured automatically in Circle CVI42. Endocardial global, apical, mid-ventricular, and basal circumferential strains, as well as end-diastolic and end-systolic volume and LVEF were compared.ResultsBland-Altman analysis found systematic biases in contour length between software packages. Compared to OsiriX, contour lengths were shorter in both ScanIP (-1.9 cm) and CVI42 (-0.6 cm), causing statistically significant differences in end-diastolic and end-systolic volumes, and apical circumferential strain (all pConclusionOsiriX and CVI42 gave consistent results for all strain and volume metrics, with no statistical differences found between OsiriX and ScanIP for mid-ventricular, global or basal strains, or left ventricular ejection fraction. However, volumes were influenced by the choice of contouring software, suggesting care should be taken when comparing volumes across different software.

Randomized trials with continuous outcomes are often analyzed using analysis of covariance (ANCOVA), with adjustment for prognostic baseline covariates. The ANCOVA estimator of the treatment effect is consistent under arbitrary model misspecification. In an article recently published in the journal, Wang et al proved the model-based variance estimator for the treatment effect is also consistent under outcome model misspecification, assuming the probability of randomization to each treatment is 1/2. In this reader reaction, we derive explicit expressions which show that when randomization is unequal, the model-based variance estimator can be biased upwards or downwards. In contrast, robust sandwich variance estimators can provide asymptotically valid inferences under arbitrary misspecification, even when randomization probabilities are not equal.

Multiple imputation has become one of the most popular approaches for handling missing data in statistical analyses. Part of this success is due to Rubin's simple combination rules. These give frequentist valid inferences when the imputation and analysis procedures are so called congenial and the complete data analysis is valid, but otherwise may not. Roughly speaking, congeniality corresponds to whether the imputation and analysis models make different assumptions about the data. In practice imputation and analysis procedures are often not congenial, such that tests may not have the correct size and confidence interval coverage deviates from the advertised level. We examine a number of recent proposals which combine bootstrapping with multiple imputation, and determine which are valid under uncongeniality and model misspecification. Imputation followed by bootstrapping generally does not result in valid variance estimates under uncongeniality or misspecification, whereas bootstrapping followed by imputation does. We recommend a particular computationally efficient variant of bootstrapping followed by imputation., Comment: Updated (fixed) reference based simulation results. Now included tables which were previously not included as they were in supplementary information document. Swapped order of the two simulation studies. Added acknowledgement and funding statements

How to adequately handle nonproportional hazards (NPH) in clinical trials is an important and timely question, particularly given recent advances in immuno-oncology treatments, in which survival cu...

Missing outcomes are a commonly occurring problem for cluster randomised trials, which can lead to biased and inefficient inference if ignored or handled inappropriately. Two approaches for analysing such trials are cluster-level analysis and individual-level analysis. In this study, we assessed the performance of unadjusted cluster-level analysis, baseline covariate-adjusted cluster-level analysis, random effects logistic regression and generalised estimating equations when binary outcomes are missing under a baseline covariate-dependent missingness mechanism. Missing outcomes were handled using complete records analysis and multilevel multiple imputation. We analytically show that cluster-level analyses for estimating risk ratio using complete records are valid if the true data generating model has log link and the intervention groups have the same missingness mechanism and the same covariate effect in the outcome model. We performed a simulation study considering four different scenarios, depending on whether the missingness mechanisms are the same or different between the intervention groups and whether there is an interaction between intervention group and baseline covariate in the outcome model. On the basis of the simulation study and analytical results, we give guidance on the conditions under which each approach is valid. © 2017 The Authors. Statistics in Medicine Published by John Wiley & Sons Ltd.

This paper provides examples of defining estimands in real-world scenarios following ICH E9(R1) guidelines. Detailed discussions on choosing the estimands and estimators can be found in our companion papers. Three scenarios of increasing complexity are illustrated. The first example is a proof-of-concept trial in major depressive disorder where the estimand is chosen to support the sponsor decision on whether to continue development. The second and third examples are confirmatory trials in severe asthma and rheumatoid arthritis respectively. We discuss the intercurrent events expected during each trial and how they can be handled so as to be consistent with the study objectives. The estimands discussed in these examples are not the only acceptable choices for their respective scenarios. The intent is to illustrate the key concepts rather than focus on specific choices. Emphasis is placed on following a study development process where estimands link the study objectives with data collection and analysis in a coherent manner, thereby avoiding disconnect between objectives, estimands, and analyses. ispartof: THERAPEUTIC INNOVATION & REGULATORY SCIENCE vol:54 issue:2 pages:370-384 ispartof: location:Switzerland status: published

Background Recent published studies have shown meaningful discrepancies between local investigator and blinded, independent, central review (BICR) assessed median progression-free survival (PFS). When the local review but not BICR shows progression, generally, no further assessments are carried out and patients are censored in the BICR analysis, leading to violation of the statistical assumptions of independence between censoring and outcome used in survival analysis methods. Methods We carried out a simulation study to assess methodological reasons behind these discrepancies and corroborated our findings in a case study of three BRCA-mutated ovarian cancer trials. We briefly outline possible methodological solutions that may lead to improved estimation of the BICR medians. Results The Kaplan–Meier (KM) curve for the BICR PFS can often be exaggerated. The degree of bias is largest when there is reasonably strong correlation between BICR and local PFS, especially when PFS is long compared with assessment frequency. This can result in an exaggeration of the medians and their difference; however, the hazard ratio (HR) is much less susceptible to bias. Our simulation shows that when the true BICR median PFS was 19 months, and patients assessed every 12 weeks, the estimated KM curves were materially biased whenever the correlation between BICR and local PFS was 0.4 or greater. This was corroborated by case studies where, in the active arm, the BICR median PFS was between 6 and 11 months greater than the local median PFS. Further research is required to find improved methods for estimating BICR survival curves. Conclusions In general, when there is a difference between local and BICR medians, the true BICR KM curve is likely to be exaggerated and its true median will probably lie somewhere between the observed local and BICR medians. Presentation of data should always include both BICR and local results whenever a BICR is carried out.

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 () and the "substantive model compatible" (MI-SMC) method of Bartlett et al. (). We also apply the "MI matched set" approach of Seaman and Keogh () 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.

Studies often follow individuals until they fail from one of a number of competing failure types. One approach to analyzing such competing risks data involves modeling the cause-specific hazards as functions of baseline covariates. A common issue that arises in this context is missing values in covariates. In this setting, we first establish conditions under which complete case analysis (CCA) is valid. We then consider application of multiple imputation to handle missing covariate values, and extend the recently proposed substantive model compatible version of fully conditional specification (SMC-FCS) imputation to the competing risks setting. Through simulations and an illustrative data analysis, we compare CCA, SMC-FCS, and a recent proposal for imputing missing covariates in the competing risks setting.

Analyses of randomised trials are often based on regression models which adjust for baseline covariates, in addition to randomised group. Based on such models, one can obtain estimates of the marginal mean outcome for the population under assignment to each treatment, by averaging the model-based predictions across the empirical distribution of the baseline covariates in the trial. We identify under what conditions such estimates are consistent, and in particular show that for canonical generalised linear models, the resulting estimates are always consistent. We show that a recently proposed variance estimator underestimates the variance of the estimator around the true marginal population mean when the baseline covariates are not fixed in repeated sampling and provide a simple adjustment to remedy this. We also describe an alternative semiparametric estimator, which is consistent even when the outcome regression model used is misspecified. The different estimators are compared through simulations and application to a recently conducted trial in asthma.

Health-care professionals, patients, and families seek as much information as possible about prognosis for patients with Alzheimer's disease (AD); however, we do not yet have a robust understanding of how demographic factors predict prognosis. We evaluated associations between age at presentation, age of onset, and symptom length with cognitive decline as measured using the Mini-Mental State Examination (MMSE) and Clinical Dementia Rating sum-of-boxes (CDR-SOB) in a large dataset of AD patients. Age at presentation was associated with post-presentation decline in MMSE (p < 0.001), with younger patients showing faster decline. There was little evidence of an association with change in CDR-SOB. Symptom length, rather than age, was the strongest predictor of MMSE and CDR-SOB at presentation, with increasing symptom length associated with worse outcomes. The evidence that younger AD patients have a more aggressive disease course implies that early diagnosis is essential.

Missing data are a commonly occurring threat to the validity and efficiency of epidemiologic studies. Perhaps the most common approach to handling missing data is to simply drop those records with 1 or more missing values, in so-called “complete records” or “complete case” analysis. In this paper, we bring together earlier-derived yet perhaps now somewhat neglected results which show that a logistic regression complete records analysis can provide asymptotically unbiased estimates of the association of an exposure of interest with an outcome, adjusted for a number of confounders, under a surprisingly wide range of missing-data assumptions. We give detailed guidance describing how the observed data can be used to judge the plausibility of these assumptions. The results mean that in large epidemiologic studies which are affected by missing data and analyzed by logistic regression, exposure associations may be estimated without bias in a number of settings where researchers might otherwise assume that bias would occur.

Multiple imputation is a practical, principled approach to handling missing data. When used to impute missing values in covariates of regression models, imputation models may be misspecified if they are not compatible with the substantive model of interest for the outcome. In this article, we introduce the smcfcs command, which imputes covariates by substantive-model compatible fully conditional specification. This modifies the popular fully conditional specification or chained-equations approach to multiple imputation by imputing each covariate compatibly with a user-specified substantive model. We compare the smcfcs command with standard fully conditional specification imputation using mi impute chained in a simulation study and illustrative analysis of data from a study investigating time to tumor recurrence in breast cancer.

Background Early growth of HIV-exposed, uninfected (HEU) children is poorer than that of their HIV-unexposed, uninfected (HUU) counterparts but there is little longitudinal or longer term information about the growth effects of early HIV exposure. Methods We performed a longitudinal analysis to compare growth of HEU and HUU infants and children using data from two cohort studies in Lusaka, Zambia. Initially 207 HUU and 200 HEU infants from the Breastfeeding and Postpartum Health (BFPH) study and 580 HUU and 165 HEU from the Chilenje Infant Growth, Nutrition and Infection Study (CIGNIS) had anthropometric measurements taken during infancy and again when school-aged, at which time 66 BFPH children and 326 CIGNIS children were available. We analysed the data from the two cohorts separately using linear mixed models. Linear regression models were used as a secondary analysis at the later time points, adjusting for breastfeeding duration. We explored when the main group differences in growth emerged in order to estimate the largest ‘effect periods’. Results After adjusting for socioeconomic status and maternal education, HEU children had lower weight-for-age, length-for-age and BMI-for-age Z-scores during early growth and these differences still existed when children were school-aged. Exposure group differences changed most between 1 and 6 weeks and between 18 months and ~7.5 years. Conclusions HEU children have poorer early growth than HUU children which persists into later growth. Interventions to improve growth of HEU children need to target pregnant women and infants. Electronic supplementary material The online version of this article (doi:10.1186/s12887-017-0828-6) contains supplementary material, which is available to authorized users.

Objectives: We hypothesized that CSF neurofilament light (NFL) levels would be elevated in dementias with subcortical involvement, including vascular dementia (VaD), but less elevated in dementias primarily affecting gray matter structures, such as Alzheimer disease (AD), and that elevated CSF NFL would correlate with disease severity and shorter survival time irrespective of clinical diagnosis. Methods: We included 3,356 individuals with dementia who had CSF NFL analyzed in our laboratory between 2005 and 2012. Clinical diagnoses and Mini-Mental State Examination (MMSE) scores were obtained from the Swedish Dementia Registry, and in selected cases (n = 478), date of death from the Swedish Mortality Registry. Results: CSF NFL differed among clinical diagnoses, with the highest levels seen in frontotemporal dementia, VaD, and mixed AD and VaD. Early-onset AD (younger than 65 years) had the lowest levels. High CSF NFL correlated with low MMSE score and short survival time irrespective of diagnosis, and was also particularly evident in AD. Conclusions: CSF NFL differs among different neurodegenerative diseases and is especially high in dementias engaging subcortical brain regions, such as VaD and mixed AD and VaD, but also in frontotemporal dementia. The association of high CSF NFL levels with disease severity and short survival supports the notion that high CSF NFL levels indicate more aggressive disease processes.

Hippocampal volumetric measures may be useful for Alzheimer's disease (AD) diagnosis and disease tracking; however, manual segmentation of the hippocampus is labour-intensive. Therefore, automated techniques are necessary for large studies and to make hippocampal measures feasible for clinical use. As large studies and clinical centres are moving from using 1.5 Tesla (T) scanners to higher field strengths it is important to assess whether specific image processing techniques can be used at these field strengths. This study investigated whether an automated hippocampal segmentation technique (HMAPS: hippocampal multi-atlas propagation and segmentation) and volume change measures (BSI: boundary shift integral) were as accurate at 3T as at 1.5T. Eighteen Alzheimer's disease patients and 18 controls with 1.5T and 3T scans at baseline and 12-month follow-up were used from the Alzheimer's Disease Neuroimaging Initiative cohort. Baseline scans were segmented manually and using HMAPS and their similarity was measured by the Jaccard index. BSIs were calculated for serial image pairs. We calculated pair-wise differences between manual and HMAPS rates at 1.5T and 3T and compared the SD of these differences at each field strength. The difference in mean Jaccards (manual and HMAPS) between 1.5T and 3T was small with narrow confidence intervals (CIs) and did not appear to be segmentor dependent. The SDs of the difference between volumes from manual and automated segmentations were similar at 1.5T and 3T, with a relatively narrow CI for their ratios. The SDs of the difference between BSIs from manual and automated segmentations were also similar at 1.5T and 3T but with a wider CI for their ratios. This study supports the use of our automated hippocampal voluming methods, developed using 1.5T images, with 3T images.

A variable is 'systematically missing' if it is missing for all individuals within particular studies in an individual participant data meta-analysis. When a systematically missing variable is a potential confounder in observational epidemiology, standard methods either fail to adjust the exposure-disease association for the potential confounder or exclude studies where it is missing. We propose a new approach to adjust for systematically missing confounders based on multiple imputation by chained equations. Systematically missing data are imputed via multilevel regression models that allow for heterogeneity between studies. A simulation study compares various choices of imputation model. An illustration is given using data from eight studies estimating the association between carotid intima media thickness and subsequent risk of cardiovascular events. Results are compared with standard methods and also with an extension of a published method that exploits the relationship between fully adjusted and partially adjusted estimated effects through a multivariate random effects meta-analysis model. We conclude that multiple imputation provides a practicable approach that can handle arbitrary patterns of systematic missingness. Bias is reduced by including sufficient between-study random effects in the imputation model.

BACKGROUND: Ethnic minorities/immigrants have differential health as compared with natives. The epidemic in child overweight/obesity (OW/OB) in Sweden is leveling oft but lower socioeconomic groups and immigrants/ethnic minorities may not have benefited equally from this trend. We investigated whether nonethnic Swedish children are at increased risk for being OW/OB and whether these associations are mediated by parental socioeconomic position (SEP) and/or early-life factors such as birth weight, maternal smoking, BMI, and breastfeeding. METHODS: Data on 10,628 singleton children (51% boys, mean age: 4.8 y, born during the period 2000-2004) residing in Uppsala were analyzed. OW/OB was computed using the International Obesity Task Force's sex- and age-specific cutoffs. The mother's nativity was used as proxy for ethnicity. Logistic regression was used to analyze ethnicity-OW/OB associations. RESULTS: Children of North African, Iranian, South American, and Turkish ethnicity had increased odds for being overweight/obese as compared with children of Swedish ethnicity (adjusted odds ratio (OR): 2.60 (95% confidence interval (CI): 1.57-4.27), 1.67 (1.03-2.72), 3.00 (1.86-4.80), and 2.90 (1.73-4.88), respectively). Finnish children had decreased odds for being overweight/obese (adjusted OR: 0.53 (0.32-0.90)). CONCLUSION: Ethnic differences in a child's risk for OW/OB exist in Sweden that cannot be explained by SEP or maternal or birth factors. As OW/OB often tracks into adulthood, more effective public health policies that intervene at an early age are needed.

There is interest in using cerebrospinal fluid (CSF) biomarkers to enrich mild cognitive impairment (MCI) clinical trials; however, there are conflicting views on their utility. We identified MCI subjects with CSF from the Alzheimer's Disease Neuroimaging Initiative. We measured brain and hippocampal atrophy rates, ventricular expansion rates, and decline in ADAS-Cog 13 and MMSE over one year. We examined proportionate and absolute reductions in mean outcome measures. Using a CSF Aβ1-42 cut-off (192 pg/ml), we estimated sample size ratios for clinical trials based on these outcomes for targeted (i.e., low-Aβ-MCI subjects only) compared with unrestricted recruitment. We further examined sample size ratios assuming only low-Aβ-MCIs would benefit from treatment. We found that for proportionate reductions in mean outcomes targeted recruitment led to significantly smaller sample sizes for all outcomes apart from ADAS-Cog. No sample size reduction was demonstrated for outcomes based on absolute reductions. Excluding subjects who would not benefit from treatment always reduces sample sizes. Using CSF biomarkers as inclusion criteria for AD trials increases the number of subjects needing to be screened but may reduce required sample sizes and reduce the risk of exposing patients without amyloid pathology to treatment side-effects. Whether targeted recruitment reduces required sample sizes depends critically on assumptions regarding treatment effects.

Medial temporal lobe atrophy (MTA) and cerebrospinal fluid (CSF) markers of Alzheimer's disease (AD) pathology may aid the early detection of AD in mild cognitive impairment (MCI). However, the relationship between structural and pathological markers is not well understood. Furthermore, while posterior atrophy (PA) is well recognized in AD, its value in predicting conversion from late-onset amnestic MCI to AD is unclear. In this study we used visual ratings of MTA and PA to assess their value in predicting conversion to AD in 394 MCI patients. The relationship of atrophy patterns with CSF Aβ1-42, tau, and p-tau(181) was further investigated in 114 controls, 192 MCI, and 99 AD patients. There was a strong association of MTA ratings with conversion to AD (p < 0.001), with a weaker association for PA ratings (p = 0.047). Specific associations between visual ratings and CSF biomarkers were found; MTA was associated with lower levels of Aβ1-42 in MCI, while PA was associated with elevated levels of tau in MCI and AD, which may reflect widespread neuronal loss including posterior regions. These findings suggest both that posterior atrophy may predict conversion to AD in late-onset MCI, and that there may be differential relationships between CSF biomarkers and regional atrophy patterns.

Bayesian approaches for handling covariate measurement error are well established and yet arguably are still relatively little used by researchers. For some this is likely due to unfamiliarity or disagreement with the Bayesian inferential paradigm. For others a contributory factor is the inability of standard statistical packages to perform such Bayesian analyses. In this paper, we first give an overview of the Bayesian approach to handling covariate measurement error, and contrast it with regression calibration, arguably the most commonly adopted approach. We then argue why the Bayesian approach has a number of statistical advantages compared to regression calibration and demonstrate that implementing the Bayesian approach is usually quite feasible for the analyst. Next, we describe the closely related maximum likelihood and multiple imputation approaches and explain why we believe the Bayesian approach to generally be preferable. We then empirically compare the frequentist properties of regression calibration and the Bayesian approach through simulation studies. The flexibility of the Bayesian approach to handle both measurement error and missing data is then illustrated through an analysis of data from the Third National Health and Nutrition Examination Survey.

Missing outcomes are a commonly occurring problem for cluster randomised trials, which can lead to biased and inefficient inference if ignored or handled inappropriately. Two approaches for analysing such trials are cluster‐level analysis and individual‐level analysis. In this study, we assessed the performance of unadjusted cluster‐level analysis, baseline covariate‐adjusted cluster‐level analysis, random effects logistic regression and generalised estimating equations when binary outcomes are missing under a baseline covariate‐dependent missingness mechanism. Missing outcomes were handled using complete records analysis and multilevel multiple imputation. We analytically show that cluster‐level analyses for estimating risk ratio using complete records are valid if the true data generating model has log link and the intervention groups have the same missingness mechanism and the same covariate effect in the outcome model. We performed a simulation study considering four different scenarios, depending on whether the missingness mechanisms are the same or different between the intervention groups and whether there is an interaction between intervention group and baseline covariate in the outcome model. On the basis of the simulation study and analytical results, we give guidance on the conditions under which each approach is valid. © 2017 The Authors. Statistics in Medicine Published by John Wiley & Sons Ltd.

Attrition is a common occurrence in cluster randomised trials (CRTs) which leads to missing outcome data. Two approaches for analysing such trials are cluster-level analysis and individual-level analysis. This paper compares the performance of unadjusted cluster-level analysis, baseline covariate adjusted cluster-level analysis and linear mixed model (LMM) analysis, under baseline covariate dependent missingness (CDM) in continuous outcomes, in terms of bias, average estimated standard error and coverage probability. The methods of complete case analysis (CCA) and multiple imputation (MI) are used to handle the missing outcome data. Four possible scenarios are considered depending on whether the missingness mechanisms and covariate effects on outcome are the same or different in the two intervention groups. We show that both unadjusted cluster-level analysis and baseline covariate adjusted cluster-level analysis give unbiased estimates of the intervention effect only if both intervention groups have the same missingness mechanisms and the same covariate effects, which is arguably unlikely to hold in practice. LMM and MI give unbiased estimates under all four considered scenarios, provided that an interaction of intervention indicator and covariate is included in the model when the covariate effects are different in the two intervention groups. MI gives slightly overestimation of average standard error, which leads to a decrease in power., 25 pages and 5 Tables

Global gray matter (GM) atrophy rates were quantified from magnetic resonance imaging (MRI) over 6- and 12-month intervals in 37 patients with Alzheimer's disease (AD) and 19 controls using: (1) nonlinear registration and integration of Jacobian values, and (2) segmentation and subtraction of serial GM volumes. Sample sizes required to power treatment trials using global GM atrophy rate as an outcome measure were estimated and compared between the 2 techniques, and to global brain atrophy measures quantified using the boundary shift integral (brain boundary shift integral; BBSI) and structural image evaluation, using normalization, of atrophy (SIENA). Increased GM atrophy rates (approximately 2% per year) were observed in patients compared with controls. Although mean atrophy rates provided by Jacobian integration were smaller than those from segmentation and subtraction of GM volumes, measurement variance was reduced. The number of patients required per treatment arm to detect a 20% reduction in GM atrophy rate over a 12-month follow-up (90% power) was 202 (95% confidence interval [CI], 118–423) using Jacobian integration and 2047 (95% CI 271 to > 10 000) using segmentation and subtraction. Comparable sample sizes for whole brain atrophy were 240 (95% CI, 142–469) using the BBSI and 196 (95% CI, 110–425) using SIENA. Jacobian integration could be useful for measuring GM atrophy rate in Alzheimer's disease as a marker of disease progression and treatment efficacy.

We explore several approaches for imputing partially observed covariates when the outcome of interest is a censored event time and when there is an underlying subset of the population that will never experience the event of interest. We call these subjects 'cured', and we consider the case where the data are modeled using a Cox proportional hazards (CPH) mixture cure model. We study covariate imputation approaches using fully conditional specification. We derive the exact conditional distribution and suggest a sampling scheme for imputing partially observed covariates in the CPH cure model setting. We also propose several approximations to the exact distribution that are simpler and more convenient to use for imputation. A simulation study demonstrates that the proposed imputation approaches outperform existing imputation approaches for survival data without a cure fraction in terms of bias in estimating CPH cure model parameters. We apply our multiple imputation techniques to a study of patients with head and neck cancer. Copyright © 2016 John WileySons, Ltd.

Voice processing in neurodegenerative disease is poorly understood. Here we undertook a systematic investigation of voice processing in a cohort of patients with clinical diagnoses representing two canonical dementia syndromes: temporal variant frontotemporal lobar degeneration (n = 14) and Alzheimer’s disease (n = 22). Patient performance was compared with a healthy matched control group (n = 35). All subjects had a comprehensive neuropsychological assessment including measures of voice perception (vocal size, gender, speaker discrimination) and voice recognition (familiarity, identification, naming and cross-modal matching) and equivalent measures of face and name processing. Neuroanatomical associations of voice processing performance were assessed using voxel-based morphometry. Both disease groups showed deficits on all aspects of voice recognition and impairment was more severe in the temporal variant frontotemporal lobar degeneration group than the Alzheimer’s disease group. Face and name recognition were also impaired in both disease groups and name recognition was significantly more impaired than other modalities in the temporal variant frontotemporal lobar degeneration group. The Alzheimer’s disease group showed additional deficits of vocal gender perception and voice discrimination. The neuroanatomical analysis across both disease groups revealed common grey matter associations of familiarity, identification and cross-modal recognition in all modalities in the right temporal pole and anterior fusiform gyrus; while in the Alzheimer’s disease group, voice discrimination was associated with grey matter in the right inferior parietal lobe. The findings suggest that impairments of voice recognition are significant in both these canonical dementia syndromes but particularly severe in temporal variant frontotemporal lobar degeneration, whereas impairments of voice perception may show relative specificity for Alzheimer’s disease. The right anterior temporal lobe is likely to have a critical role in the recognition of voices and other modalities of person knowledge.

Despite growing clinical and neurobiological interest in the brain mechanisms that process emotion in music, these mechanisms remain incompletely understood. Patients with frontotemporal lobar degeneration (FTLD) frequently exhibit clinical syndromes that illustrate the effects of breakdown in emotional and social functioning. Here we investigated the neuroanatomical substrate for recognition of musical emotion in a cohort of 26 patients with FTLD (16 with behavioural variant frontotemporal dementia, bvFTD, 10 with semantic dementia, SemD) using voxel-based morphometry. On neuropsychological evaluation, patients with FTLD showed deficient recognition of canonical emotions (happiness, sadness, anger and fear) from music as well as faces and voices compared with healthy control subjects. Impaired recognition of emotions from music was specifically associated with grey matter loss in a distributed cerebral network including insula, orbitofrontal cortex, anterior cingulate and medial prefrontal cortex, anterior temporal and more posterior temporal and parietal cortices, amygdala and the subcortical mesolimbic system. This network constitutes an essential brain substrate for recognition of musical emotion that overlaps with brain regions previously implicated in coding emotional value, behavioural context, conceptual knowledge and theory of mind. Musical emotion recognition may probe the interface of these processes, delineating a profile of brain damage that is essential for the abstraction of complex social emotions., Research highlights ► Emotion recognition from music is impaired in frontotemporal lobar degeneration. ► This deficit is associated with atrophy in a distributed cerebral network. ► This network includes cortical and mesolimbic areas likely to code social emotions.

e14522Background: Recent published studies have shown large discrepancies between locally evaluated (LE) median PFS times and those assessed by BICR. We show that, in many cases, this can be attrib...

Objective: To identify cognitively normal individuals at risk of Alzheimer disease (AD) based on cerebrospinal fluid (CSF) A beta 1-42, and to determine rates of cerebral atrophy.Methods: Control subjects from the Alzheimer's Disease Neuroimaging Initiative with CSF and serial magnetic resonance imaging (MRI) were dichotomized on CSF A beta 1-42 (normal control [NC]-high >192pg/ml; NC-low