Linear Hazard Regression Research Articles

The clinical features of alcohol use disorders in biological and step‐fathers that predict risk for alcohol use disorders in offspring

Given that Alcohol Use Disorder (AUD) is clinically heterogeneous, can we, in a large epidemiological sample using public registries, identify clinical features of AUD cases in biological and step-fathers that index, respectively, genetic and familial-environmental risk for AUD in their offspring? From all father-offspring pairs where the father had AUD and the offspring was born 1960-1990, we identified not-lived-with (NLW) biological fathers (n = 38,376) and step-father pairs (n = 9,711). The relationship between clinical and historical features of the father's AUD and risk for AUD in offspring was assessed by linear hazard regression. Age at first registration for AUD and recurrence of AUD registration were significantly stronger predictors of risk for AUD in the offspring of NLW fathers than in step-fathers. By contrast, number of AUD registrations in NLW fathers and step-fathers were equally predictive of risk for AUD in offspring. However, while the number of step-father AUD registrations that occurred when he was living them with significantly predicted risk for AUD in his step-children, the number of registrations that occurred when not residing with his step-children was unassociated with their AUD risk. In an epidemiological sample, we could meaningfully differentiate between features of AUD in fathers that indexed genetic risk which was transmitted to biological offspring (early age at onset and recurrence) versus indexed environmental risk (registrations while rearing) which increased risk in step-children.

Clinical features of registry-ascertained alcohol use disorders that reflect familial risk

BackgroundAlcohol Use Disorder (AUD) is clinically heterogeneous. Using a large epidemiological sample ascertained via public registries, is it possible to identify clinical and historical features of AUD that reflect familial risk? MethodsUsing registration in national medical, legal or pharmacy registries, we identified four kinds of relative pairs (n=683,223) starting with a proband with AUD: cousins, half-siblings, full-siblings and monozygotic cotwins. Using linear hazard regression, we examined the interaction between five clinical/historical features of AUD in the proband and risk for AUD in these relatives. ResultsIncreased risk for AUD in relatives was predicted by the proband’s early age at first registration, total number of registrations, recurrence, history of drug abuse and ascertainment in the medical versus the legal or pharmacy registry. In multivariate models, age at first registration, number of registrations, recurrence and history of drug abuse remained significant and in aggregate strongly predicted the risk for AUD in relatives. The risk for AUD in siblings of AUD probands in the highest decile of genetic risk predicted by these four indices was more than twice as great as that predicted in siblings of probands in the lowest risk decile. ConclusionsIn an epidemiological sample, familial risk for AUD can be assessed by simple clinical and historical variables.

Variable selection in partially linear hazard regression for multivariate failure time data

The aim of this paper is to explore variable selection approaches in the partially linear proportional hazards model for multivariate failure time data. A new penalised pseudo-partial likelihood method is proposed to select important covariates. Under certain regularity conditions, we establish the rate of convergence and asymptotic normality of the resulting estimates. We further show that the proposed procedure can correctly select the true submodel, as if it was known in advance. Both simulated and real data examples are presented to illustrate the proposed methodology.

Clinical features of drug abuse that reflect genetic risk

Drug abuse (DA) is a clinically heterogeneous syndrome. Can we, in a large epidemiological sample, identify clinical features of DA cases that index genetic risk? Using registration in medical, legal or pharmacy records, we identified four kinds of relative pairs (n=935,854) starting with a proband with DA: monozygotic co-twins; full siblings; half-siblings; and cousins. Using linear hazard regression, we examined the interaction between three clinical features of DA in the proband and risk for DA in these four relative pairs, ordered by degree of genetic relationship. Increased risk for DA in relatives was robustly predicted by early age at first registration, total number of registrations, and ascertainment in the criminal versus the medical or pharmacy registry. In multivariate models, all three of these variables remained significant and in aggregate strongly predicted DA risk in relatives. The risk for DA in siblings of DA probands in the highest decile of genetic risk predicted by our three indices was more than twice as great as that predicted in siblings of probands in the lowest decile of risk. In an epidemiological sample, genetic risk for DA can be substantially indexed by simple clinical and historical variables.

Partially Linear Hazard Regression with Varying Coefficients for Multivariate Survival Data

Summary The paper studies estimation of partially linear hazard regression models with varying coefficients for multivariate survival data. A profile pseudo-partial-likelihood estimation method is proposed. The estimation of the parameters of the linear part is accomplished via maximization of the profile pseudo-partial-likelihood, whereas the varying-coefficient functions are considered as nuisance parameters that are profiled out of the likelihood. It is shown that the estimators of the parameters are root n consistent and the estimators of the non-parametric coefficient functions achieve optimal convergence rates. Asymptotic normality is obtained for the estimators of the finite parameters and varying-coefficient functions. Consistent estimators of the asymptotic variances are derived and empirically tested, which facilitate inference for the model. We prove that the varying-coefficient functions can be estimated as well as if the parametric components were known and the failure times within each subject were independent. Simulations are conducted to demonstrate the performance of the estimators proposed. A real data set is analysed to illustrate the methodology proposed.

Partially Linear Hazard Regression for Multivariate Survival Data

This article studies estimation of partially linear hazard regression models for multivariate survival data. A profile pseudo–partial likelihood estimation method is proposed under the marginal hazard model framework. The estimation on the parameters for the linear part is accomplished by maximization of a pseudo–partial likelihood profiled over the nonparametric part. This enables us to obtain -consistent estimators of the parametric component. Asymptotic normality is obtained for the estimates of both the linear and nonlinear parts. The new technical challenge is that the nonparametric component is indirectly estimated through its integrated derivative function from a local polynomial fit. An algorithm of fast implementation of our proposed method is presented. Consistent standard error estimates using sandwich-type ideas are also developed, which facilitates inferences for the model. It is shown that the nonparametric component can be estimated as well as if the parametric components were known and the failure times within each subject were independent. Simulations are conducted to demonstrate the performance of the proposed method. A real dataset is analyzed to illustrate the proposed methodology.

Aalen's linear model for sampled risk set data: a large sample study.

Borgan and Langholz (1997) describe a method for estimating the parameter functions in Aalen's linear hazard regression model from sampled risk set data. Using a counting process formulation and the martingale central limit theorem, we provide a study of the asymptotic distributional properties of the estimator. The results are applied to study the efficiencies of the nested case-control and counter-matched designs relative to a full cohort analysis.