Treatment effects for assessing the efficacy of a novel therapy are typically defined as measures comparing the marginal outcome distributions observed in two or more study arms. Although one can estimate such effects from the observed outcome distributions obtained from proper randomization, covariate adjustment is recommended to increase precision in randomized clinical trials. For important treatment effects, such as odds or hazard ratios, conditioning on covariates in binary logistic or proportional hazards models changes the interpretation of the treatment effect under noncollapsibility and conditioning on different sets of covariates renders the resulting effect estimates incomparable.
We propose a novel nonparanormal model formulation for adjusted marginal inference allowing the estimation of the joint distribution of outcome and covariates featuring the intended marginally defined treatment effect parameter – including marginal log-odds ratios or log-hazard ratios. Marginal distributions are modelled by transformation models allowing broad applicability to diverse outcome types. Joint maximum likelihood estimation of all model parameters is performed. From the parameters not only the marginal treatment effect of interest can be identified but also an overall coefficient of determination and covariate-specific measures of prognostic strength can be derived. A free reference implementation of this novel method is available in R add-on package tram.
For the special case of Cohen’s standardized mean difference d, we theoretically show that adjusting for an informative prognostic variable improves the precision of this marginal, noncollapsible effect. Empirical results confirm this not only for Cohen’s d but also for log-odds ratios and log-hazard ratios in simulations and four applications.
Software: tram package
]]>Over the last five decades, we have seen strong methodological advances in survival analysis, using parametric methods and, more prominently, methods based on non-/semi-parametric estimation. As the methodological landscape continues to evolve, the task of navigating through the multitude of methods and identifying available software resources is becoming increasingly challenging – especially in more complex scenarios, such as when dealing with interval-censored or clustered survival data, non-proportional hazards, or dependent censoring.
This tutorial explores the potential of using the framework of smooth transformation models for survival analysis in the R system for statistical computing. This framework provides a unified maximum-likelihood approach that covers a wide range of survival models, including well-established ones such as the Weibull model and a fully parametric version of the famous Cox proportional hazards model, and various extensions for more complex scenarios. We explore models for non-proportional/crossing hazards, dependent censoring, clustered observations and extensions towards personalised medicine within this framework.
Using survival data from a two-arm randomised controlled trial on rectal cancer therapy, we demonstrate how survival analysis tasks can be seamlessly navigated in R within this framework using the implementation provided by the tram package, and few related packages.
Software: tram package
]]>Researchers have always been responsible for creating and maintaining their own research environment tailored to the specific needs and challenges in their field of research. With the rise of open research practices, it becomes a more and more common task to demonstrate, to peers and the general public, that these research environments are in fact suitable in the sense that research findings produced under the conditions of these environments are methodologically sound and, as an ultimate goal, reproducible.
In empirical research, computer systems play a pivotal role in research environments, as they are used to collect, to manage, to analyse, and to publish data from relevant experiments. In an ideal world, all cascades of this pipeline are under full control of the investigators ensuring that published results have been obtained from a well-run experiment and a well-defined data analysis procedure.
By definition, closed source software cannot be part of such a system, simply because it is impossible to understand, validate, criticise, or improve the way such software products work. Since the beginning of the computer age in the 1950ies, researchers have invested considerable time and energy to develop and maintain a free research software universe giving all of its users the right to understand, validate, criticise, and improve programs relevant for their research tasks.
In recent months, the concept of Digital/IT Sovereignty received increasing attention. Concerns regarding data privacy and security risks associated with strong dependencies on foreign cloud infrastructure have not only been raised by researchers but in the general public. Protecting research integrity by building resilience against compromised, or simply quiescent, network services is therefore on everybody’s agenda.
The aim of this workshop is to introduce researchers and students to the free research software universe and help them to navigate this world. Seminar participants will install a free operating system on computers and will learn how to set-up and maintain a computing environment covering all aspects of empirical research, most importantly the management, analysis, and reporting of data.
The seminar will focus on the following topics:
Workshop participants will be provided with a laptop. We meet Wednesday September 10 from 13:00 to 17:00 and Thursday September 11 from 9:00 to 15:00 in Hirschengraben 82 HIT-E03. Please register by email to Torsten Hothorn, Torsten.Hothorn@uzh.ch.
]]>The estimation of heterogeneous treatment effects (HTEs) has attracted considerable interest in many disciplines, most prominently in medicine and economics. Contemporary research has so far primarily focused on continuous and binary responses where HTEs are traditionally estimated by a linear model, which allows the estimation of constant or heterogeneous effects even under certain model misspecifications. More complex models for survival, count, or ordinal outcomes require stricter assumptions to reliably estimate the treatment effect. Most importantly, the noncollapsibility issue necessitates the joint estimation of treatment and prognostic effects. Model-based forests allow simultaneous estimation of covariate-dependent treatment and prognostic effects, but only for randomized trials. In this paper, we propose modifications to model-based forests to address the confounding issue in observational data. In particular, we evaluate an orthogonalization strategy originally proposed by Robinson (1988, Econometrica) in the context of model-based forests targeting HTE estimation in generalized linear models and transformation models. We found that this strategy reduces confounding effects in a simulated study with various outcome distributions. We demonstrate the practical aspects of HTE estimation for survival and ordinal outcomes by an assessment of the potentially heterogeneous effect of Riluzole on the progress of Amyotrophic Lateral Sclerosis.
Software: model4you package, htesim package.
]]>Insects have a pivotal role in ecosystem function, thus the decline of more than 75% in insect biomass in protected areas over recent decades in Central Europe and elsewhere has alarmed the public, pushed decision-makers4 and stimulated research on insect population trends. However, the drivers of this decline are still not well understood. Here, we reanalysed 27 years of insect biomass data from Hallmann et al., using sample-specific information on weather conditions during sampling and weather anomalies during the insect life cycle. This model explained variation in temporal decline in insect biomass, including an observed increase in biomass in recent years, solely on the basis of these weather variables. Our finding that terrestrial insect biomass is largely driven by complex weather conditions challenges previous assumptions that climate change is more critical in the tropics or that negative consequences in the temperate zone might only occur in the future. Despite the recent observed increase in biomass, new combinations of unfavourable multi-annual weather conditions might be expected to further threaten insect populations under continuing climate change. Our findings also highlight the need for more climate change research on physiological mechanisms affected by annual weather conditions and anomalies.
]]>Estimation of heterogeneous treatment effects (HTE) is of prime importance in many disciplines, ranging from personalized medicine to economics among many others. Random forests have been shown to be a flexible and powerful approach to HTE estimation in both randomized trials and observational studies. In particular “causal forests”, introduced by Athey, Tibshirani, and Wager (2019), along with the R implementation in package grf were rapidly adopted. A related approach, called “model-based forests”, that is geared towards randomized trials and simultaneously captures effects of both prognostic and predictive variables, was introduced by Seibold, Zeileis, and Hothorn (2018) along with a modular implementation in the R package model4you.
Neither procedure is directly applicable to the estimation of individualized predictions of excess postpartum blood loss caused by a cesarean section in comparison to vaginal delivery. Clearly, randomization is hardly possible in this setup and thus model-based forests lack clinical trial data to address this question. On the other hand, the skewed and interval-censored postpartum blood loss observations violate assumptions made by causal forests. Here, we present a tailored model-based forest for skewed and intervalcensored data to infer possible predictive prepartum characteristics and their impact on excess postpartum blood loss caused by a cesarean section.
As a methodological basis, we propose a unifying view on causal and model-based forests that goes beyond the theoretical motivations and investigates which computational elements make causal forests so successful and how these can be blended with the strengths of model-based forests. To do so, we show that both methods can be understood in terms of the same parameters and model assumptions for an additive model under L2 loss. This theoretical insight allows us to implement several flavors of “model-based causal forests” and dissect their different elements in silico.
The original causal forests and model-based forests are compared with the new blended versions in a benchmark study exploring both randomized trials and observational settings. In the randomized setting, both approaches performed akin. If confounding was present in the data generating process, we found local centering of the treatment indicator with the corresponding propensities to be the main driver for good performance. Local centering of the outcome was less important, and might be replaced or enhanced by simultaneous split selection with respect to both prognostic and predictive effects. This lays the foundation for future research combining random forests for HTE estimation with other types of models.
Software: grf package, model4you package, htesim package.
]]>