0000000000021535
AUTHOR
Michal Abrahamowicz
Empirical study of the dependence of the results of multivariable flexible survival analyses on model selection strategy
Flexible survival models, which avoid assumptions about hazards proportionality (PH) or linearity of continuous covariates effects, bring the issues of model selection to a new level of complexity. Each ‘candidate covariate’ requires inter-dependent decisions regarding (i) its inclusion in the model, and representation of its effects on the log hazard as (ii) either constant over time or time-dependent (TD) and, for continuous covariates, (iii) either loglinear or non-loglinear (NL). Moreover, ‘optimal’ decisions for one covariate depend on the decisions regarding others. Thus, some efficient model-building strategy is necessary. We carried out an empirical study of the impact of the model …
Advantages and limitations of using national administrative data on obstetric blood transfusions to estimate the frequency of obstetric hemorrhages
International audience; BACKGROUND: Obstetric hemorrhages are a frequent cause of maternal death all over the world, but are not routinely monitored. Health systems administrative databases could be used for this purpose, but data quality needs to be assessed. OBJECTIVES: Using blood transfusion data recorded in administrative databases to estimate the frequency of obstetric hemorrhages. Research design A population-based study. Subjects Validation sub-sample: all mothers who gave birth in a French region in 2006-07 (35 123 pregnancies). Main study: all mothers who gave birth in France in 2006-07 (1 629 537 pregnancies). METHOD: Linkage and comparison of administrative data on blood transfu…
Multi-state relative survival modelling of colorectal cancer progression and mortality.
Abstract Accurate identification of factors associated with progression of colorectal cancer remains a challenge. In particular, it is unclear which statistical methods are most suitable to separate the effects of putative prognostic factors on cancer progression vs cancer-specific and other cause mortality. To address these challenges, we analyzed 10 year follow-up data for patients who underwent curative surgery for colorectal cancer in 1985–2000. Separate analyses were performed in two French cancer registries. Results of three multivariable models were compared: Cox model with recurrence as a time-dependent variable, and two multi-state models, which separated prognostic factor effects …
Flexible statistical models provided new insights into the role of quantitative prognostic factors for mortality in gastric cancer.
Abstract Objectives To reassess the effects of prognostic factors on mortality in gastric cancers, and to illustrate the advantages of flexible modeling. Study Design and Setting A prospective population-based cohort of persons diagnosed with gastric cancers in 1976 to 1995 in Burgundy, France, was followed for 5 years since diagnosis. Multivariable survival analyses, stratified by cancer stage, involved both conventional Cox's model and its flexible generalization, which permitted testing the underlying assumptions and accounting for changes over time in the effects of prognostic factors. Results Conventional assumptions of proportional hazards (PH) (P = 0.003) and linear increase in risk …
Estimation of National Colorectal-Cancer Incidence Using Claims Databases
Background.The aim of the study was to assess the accuracy of the colorectal-cancer incidence estimated from administrative data.Methods.We selected potential incident colorectal-cancer cases in 2004-2005 French administrative data, using two alternative algorithms. The first was based only on diagnostic and procedure codes, whereas the second considered the past history of the patient. Results of both methods were assessed against two corresponding local cancer registries, acting as “gold standards.” We then constructed a multivariable regression model to estimate the corrected total number of incident colorectal-cancer cases from the whole national administrative database.Results.The firs…
Ophthalmic Outcomes of Congenital Toxoplasmosis Followed Until Adolescence
BACKGROUND: Congenital toxoplasmosis (CT) can elicit severe damage to several organs, especially the eye, and may be manifested at birth or later. We assessed the long-term ocular prognosis in a cohort of congenitally infected children treated according to a standardized protocol and monitored for up to 22 years. METHODS: This prospective study included confirmed cases of CT, which were identified by obligatory antenatal screening at the Lyon (France) reference center between 1987 and 2008. Data obtained through ocular examinations were recorded on a standardized form and confirmed by an independent external committee. Risk factors for retinochoroiditis were identified by using a multivari…
Congenital \textittoxoplasma infection: monthly prenatal screening decreases transmission rate and improves clinical outcome at age 3 years
Background. Toxoplasma infection during pregnancy exposes the fetus to risks of congenital infection and sequelae that depend heavily on gestational age (GA) at time of infection. Accurate risk estimates by GA are necessary to counsel parents and improve clinical decisions. Methods. We analyzed data from pregnant women diagnosed with acute Toxoplasma infection in Lyon (France) from 1987 to 2008 and assessed how the risks of congenital toxoplasmosis and of clinical signs at age 3 years vary depending on GA at the time of maternal infection. Results. Among 2048 mother-infant pairs, 93.2% of mothers received prenatal treatment and 513 (24.7%) fetuses were infected. Because of a significant red…
Comparison of the ability of alternative birthweight and fetal weight standards to identify preterm newborns at increased risk of perinatal death
Objective To compare prediction of perinatal deaths among preterm infants based on fetal weight standards versus a new subpopulation-based birthweight standard. Design Population-based cohort study. Setting France. Population A total of 9100 preterm singletons, born between 24 and 36 weeks of gestation in 2000–09, in Burgundy (France). Methods We first classified all newborns as either small for gestational age (SGA) or not, based on alternative fetal weight or birthweight standards, including a new birthweight standard that excludes infants born to mothers with disease related to the weight of a fetus. Based on discrepancies between the different classifications, we then divided the newbor…