John C. Galati

What Is Meant by “Missing at Random”?

The concept of missing at random is central in the literature on statistical analysis with missing data. In general, inference using incomplete data should be based not only on observed data values but should also take account of the pattern of missing values. However, it is often said that if data are missing at random, valid inference using likelihood approaches (including Bayesian) can be obtained ignoring the missingness mechanism. Unfortunately, the term missing at random has been used inconsistently and not always clearly; there has also been a lack of clarity around the meaning of valid inference using likelihood. These issues have created potential for confusion about the exact conditions under which the missingness mechanism can be ignored, and perhaps fed confusion around the meaning of analysis ignoring the missingness mechanism. Here we provide standardised precise definitions of missing at random and missing completely at random, in order to promote unification of the theory. Using these definitions we clarify the conditions that suffice for valid inference to be obtained under a variety of inferential paradigms.

The burden of rotavirus-related illness among young children on the Australian health care system

Objective: To provide estimates of the annual number and cost of hospital admissions, emergency department (ED) visits and general practitioner (GP) visits for rotavirus (RV) related acute gastroenteritis (AGE) in young children in Australia.

Cohort Profile: The Peri/post-natal Epigenetic Twins Study

Cancer and Disease Epigenetics Group, Murdoch Childrens Research Institute (MCRI), Royal Children’s Hospital, Parkville, Victoria, Australia, Early Life Epigenetics Group, MCRI, Royal Children’s Hospital, Parkville, Victoria, Australia, Clinical Epidemiology and Biostatistics Unit, MCRI, Royal Children’s Hospital, Parkville, Victoria, Australia, Department of Paediatrics, University of Melbourne, Victoria, Australia, The Ritchie Centre, Monash Institute of Medical Research, Victoria, Australia, Department of Obstetrics and Gynaecology, University of Melbourne, Royal Women’s Hospital, Melbourne, Victoria, Australia, Department of Obstetrics and Gynaecology,University of Melbourne, Mercy Hospital for Women, Heidelberg, Victoria, Australia Present address: Department of Public Health, Twin Study, Hjelt Institute, University of Helsinki, Finland

Decreased cortical muscarinic M1 receptors in schizophrenia are associated with changes in gene promoter methylation, mRNA and gene targeting microRNA

Many studies have shown decreased cortical muscarinic M1 receptors (CHRM1) in schizophrenia (Sz), with one study showing Sz can be separated into two populations based on a marked loss of CHRM1 (∼75%) in ∼25% of people (Def-Sz) with the disorder. To better understand the mechanism contributing to the loss of CHRM1 in Def-Sz, we measured specific markers of gene expression in the cortex of people with Sz as a whole, people differentiated into Def-Sz and people with Sz that do not have a deficit in cortical CHRM1 (Non-Def-Sz) and health controls. We now report that cortical CHRM1 gene promoter methylation and CHRM1 mRNA are decrease in Sz, Def-Sz and Non-Def-Sz but levels of the micro RNA (miR)-107, a CHRM1 targeting miR, are increased only in Def-Sz. We also report in vitro data strongly supporting the notion that miR-107 levels regulate CHRM1 expression. These data suggest there is a reversal of the expected inverse relationship between gene promoter methylation and CHRM1 mRNA in people with Sz and that a breakdown in gene promoter methylation control of CHRM1 expression is contributing to the global pathophysiology of the syndrome. In addition, our data argues that increased levels of at least one miR, miR-107, is contributing to the marked loss of cortical CHRM1 in Def-Sz and this may be a differentiating pathophysiology. These latter data continue to support the hypothesis that microRNAs (miRNA) have a role in the underlying neurobiology of Sz but argue they are differentially affected in subsets of people within that syndrome.

Side-to-Side Aorto-GoreTex Central Shunt Warrants Central Shunt Patency and Pulmonary Arteries Growth

BACKGROUNDnCentral shunts may be associated with a high rate of thrombosis and pulmonary artery distortion.nnnMETHODSnBetween January 2000 and April 2010, 68 consecutive patients underwent side-to-side aorto-Gore-Tex central shunts (W.L. Gore & Associates, Flagstaff, AZ). Median age at surgery was 31 days (8 to 122). Cardiac morphologies were tetralogy of Fallot (33), pulmonary atresia with collateral dependent lung circulation (20), and other (15). Shunt sizes ranged from 3 to 6 mm. The procedure was performed on cardiopulmonary bypass in 43 patients (63%).nnnRESULTSnSix patients died during hospital stay (9%) of low output syndrome (3), sepsis (2), and stroke (1). Only one shunt needed early redo. Follow-up was 100% complete. After a median follow-up of 236 days (116 to 340), there were 7 late deaths related to sudden death (3), sepsis (2), reoperation (1), and lack of growth of pulmonary arteries with a patent shunt (1). Repair was completed in 42 patients and still pending in 12. Only one patient needed a late redo shunt (221 days). A larger shunt was performed in 5 patients after a median of 139 days (130 to 258). In the 45 patients who had serial assessment of pulmonary arteries sizes, Nakata index increased from 83 ± 62 mm(2)/m(2) to 153 ± 83 mm(2)/m(2) over a median period of 227 days (146 to 330), with equal growth observed in both pulmonary arteries (p < 0.001).nnnCONCLUSIONSnCentral shunts can be performed in neonates and children with minimal risk of shunt occlusion. Side-to-side aorto-Gore-Tex anastomosis seems to reliably warrant shunt patency and harmonious growth of pulmonary arteries.

Comparison of methods for imputing ordinal data using multivariate normal imputation: a case study of non-linear effects in a large cohort study.

BACKGROUNDnMultiple imputation is becoming increasingly popular for handling missing data, with Markov chain Monte Carlo assuming multivariate normality (MVN) a commonly used approach. Imputing categorical variables (which are clearly non-normal) using MVN imputation is challenging, and several approaches have been suggested. However, it remains unclear which approach should be preferred.nnnMETHODSnWe explore methods for imputing ordinal variables using MVN imputation, including imputing as a continuous variable and as a set of indicators, and various methods for assigning imputed values to the possible categories (rounding), for estimating a non-linear association between an ordinal exposure and binary outcome. We introduce a new approach where we impute as continuous and assign imputed values into categories based on the mean indicators imputed in a separate round of imputation. We compare these approaches in a simple setting where we make 50% of data in an ordinal exposure missing completely at random, within an otherwise complete real dataset.nnnRESULTSnMethods that impute the ordinal exposure as continuous distorted the non-linear exposure-outcome association by biasing the relationship towards linearity irrespective of the rounding method. In contrast, imputing using indicators preserved the non-linear association but not the marginal distribution of the ordinal variable.nnnCONCLUSIONSnImputing ordinal variables as continuous can bias the estimation of the exposure-outcome association in the presence of non-linear relationships. Further work is needed to develop optimal methods for handling ordinal (and nominal) variables when using MVN imputation.

Association of in vitro fertilization with global and IGF2 / H19 methylation variation in newborn twins

In vitro fertilization (IVF) and its subset intracytoplasmic sperm injection (ICSI), are widely used medical treatments for conception. There has been controversy over whether IVF is associated with adverse short- and long-term health outcomes of offspring. As with other prenatal factors, epigenetic change is thought to be a molecular mediator of any in utero programming effects. Most studies focused on DNA methylation at gene-specific and genomic level, with only a few on associations between DNA methylation and IVF. Using buccal epithelium from 208 twin pairs from the Peri/Postnatal Epigenetic Twin Study (PETS), we investigated associations between IVF and DNA methylation on a global level, using the proxies of Alu and LINE-1 interspersed repeats in addition to two locus-specific regulatory regions within IGF2/H19, controlling for 13 potentially confounding factors. Using multiple correction testing, we found strong evidence that IVF-conceived twins have lower DNA methylation in Alu, and weak evidence of lower methylation in one of the two IGF2/H19 regulatory regions and LINE-1, compared with naturally conceived twins. Weak evidence of a relationship between ICSI and DNA methylation within IGF2/H19 regulatory region was found, suggesting that one or more of the processes associated with IVF/ICSI may contribute to these methylation differences. Lower within- and between-pair DNA methylation variation was also found in IVF-conceived twins for LINE-1, Alu and one IGF2/H19 regulatory region. Although larger sample sizes are needed, our results provide additional insight to the possible influence of IVF and ICSI on DNA methylation. To our knowledge, this is the largest study to date investigating the association of IVF and DNA methylation.

Probabilities of heart donors arising within specified times for child recipients

Aim:u2003 To determine the availability of donor hearts for children of different blood group and weight needing urgent heart transplantation.

Favourable anatomy after end-to-side repair of interrupted aortic arch.

OBJECTIVEnTo evaluate cardiovascular outcomes in patients with aortic arch repair and their possible correlation with arch geometry.nnnMETHODSnTen patients who underwent end-to-side repair for aortic arch interruption (IAA), older than 10 years were compared to a cohort of 10 post coarctation (CoA) repair patients matched for age, sex and age at repair. Mean age at operation was 9.7±6.5 days. Patients underwent a resting and 24 h blood pressure measurements, exercise study, MRI, transthoracic echocardiography and vascular studies.nnnRESULTSnSeven patients developed hypertension, two from IAA group and five from CoA group. Nine patients (45%) had gothic arch geometry, three from IAA group and six from CoA group. Despite differences in arch geometry, both groups had normal LV mass, LV function and vascular function.nnnCONCLUSIONnNo differences in functional or morphologic outcomes could be demonstrated between the end-to-side repair of the arch by sternotomy and the conventional coarctation repair by thoracotomy. A favourable arch geometry can be achieved after the end-to-side repair of the aortic arch. In the present study, we could not correlate adverse arch geometry with any adverse cardio-vascular outcomes. After neonatal arch repair, the contributive role of aortic arch geometry to late hypertension remains uncertain.