scholarly journals Minimizing bias in massive multi-arm observational studies with BCAUS: balancing covariates automatically using supervision

2021 ◽  
Vol 21 (1) ◽  
Author(s):  
Chinmay Belthangady ◽  
Will Stedden ◽  
Beau Norgeot
Keyword(s):  
Propensity Score ◽  
Causal Inference ◽  
Real World ◽  
Observational Studies ◽  
Treatment Effects ◽  
Infant Health ◽  
Development Program ◽  
Health Condition ◽  
Simultaneous Intervention ◽  
Inference Methods

Abstract Background Observational studies are increasingly being used to provide supplementary evidence in addition to Randomized Control Trials (RCTs) because they provide a scale and diversity of participants and outcomes that would be infeasible in an RCT. Additionally, they more closely reflect the settings in which the studied interventions will be applied in the future. Well-established propensity-score-based methods exist to overcome the challenges of working with observational data to estimate causal effects. These methods also provide quality assurance diagnostics to evaluate the degree to which bias has been removed and the estimates can be trusted. In large medical datasets it is common to find the same underlying health condition being treated with a variety of distinct drugs or drug combinations. Conventional methods require a manual iterative workflow, making them scale poorly to studies with many intervention arms. In such situations, automated causal inference methods that are compatible with traditional propensity-score-based workflows are highly desirable. Methods We introduce an automated causal inference method BCAUS, that features a deep-neural-network-based propensity model that is trained with a loss which penalizes both the incorrect prediction of the assigned treatment as well as the degree of imbalance between the inverse probability weighted covariates. The network is trained end-to-end by dynamically adjusting the loss term for each training batch such that the relative contributions from the two loss components are held fixed. Trained BCAUS models can be used in conjunction with traditional propensity-score-based methods to estimate causal treatment effects. Results We tested BCAUS on the semi-synthetic Infant Health & Development Program dataset with a single intervention arm, and a real-world observational study of diabetes interventions with over 100,000 individuals spread across more than a hundred intervention arms. When compared against other recently proposed automated causal inference methods, BCAUS had competitive accuracy for estimating synthetic treatment effects and provided highly concordant estimates on the real-world dataset but was an order-of-magnitude faster. Conclusions BCAUS is directly compatible with trusted protocols to estimate treatment effects and diagnose the quality of those estimates, while making the established approaches automatically scalable to an arbitrary number of simultaneous intervention arms without any need for manual iteration.

Abstract 154: Estimating Treatment Effects in More Than Two Treatment Groups via Propensity Score Weighting: Practical Guidance and Application from Anticoagulant Therapy in Atrial Fibrillation Study

2016 ◽  
Vol 9 (suppl_2) ◽  
Author(s):  
Gboyega Adeboyeje ◽  
Gosia Sylwestrzak ◽  
John Barron
Keyword(s):  
Atrial Fibrillation ◽  
Propensity Score ◽  
Anticoagulant Therapy ◽  
Observational Studies ◽  
Treatment Effects ◽  
Propensity Scores ◽  
Multinomial Logistic Regression ◽  
Baseline Risk ◽  
Risk Scores ◽  
Treatment Groups

Background: The methods for estimating and assessing propensity scores in the analysis of treatment effects between two treatment arms in observational studies have been well described in the outcomes research methodology literature. However, in practice, the decision makers may need information on the comparative effectiveness of more than two treatment strategies. There’s little guidance on the estimation of treatment effects using inverse probability of treatment weights (IPTW) in studies where more than two treatment arms are to be compared. Methods: Data from an observational cohort study on anticoagulant therapy in atrial fibrillation is used to illustrate the practical steps involved in estimating the IPTW from multiple propensity scores and assessing the balance achieved under certain assumptions. For all patients in the study, we estimated the propensity score for the treatment each patient received using a multinomial logistic regression. We used the inverse of the propensity scores as weights in Cox proportional hazards to compare study outcomes for each treatment group Results: Before IPTW adjustment, there were large and statistically significant baseline differences between treatment groups in terms of demographic, plan type, and clinical characteristics including validated stroke and bleeding risk scores. After IPTW, there were no significant differences in all measured baseline risk factors. In unadjusted estimates of stroke outcome, there were large differences between dabigatran [Hazard ratio, HR, 0.59 (95% CI: 0.53 - 0.66)], apixaban [HR, 0.69 (CI: 0.57, 0.83)], rivaroxaban [HR, 0.60 (CI: 0.53 0.68)] and warfarin users. After IPTW, estimated stroke risk differences were significantly reduced or eliminated between dabigatran [HR, 0.89 (CI: 0.80, 0.98)], apixaban [HR, 0.92 (0.76, 1.10)], rivaroxaban [HR, 0.84 (CI: 0.75, 0.95)] and warfarin users. Conclusions: Our results showed IPTW methods, correctly employed under certain assumptions, are practical and relatively simple tools to control for selection bias and other baseline differences in observational studies evaluating the comparative treatment effects of more than two treatment arms. When preserving sample size is important and in the presence of time-varying confounders, IPTW methods have distinct advantages over propensity matching or adjustment.

scholarly journals Analysis of cause-effect inference by comparing regression errors

2019 ◽  
Vol 5 ◽  
pp. e169 ◽  
Author(s):  
Patrick Blöbaum ◽  
Dominik Janzing ◽  
Takashi Washio ◽  
Shohei Shimizu ◽  
Bernhard Schölkopf
Keyword(s):  
Causal Inference ◽  
Least Squares ◽  
Real World ◽  
Causal Relation ◽  
Data Sets ◽  
Noise Distribution ◽  
Real World Data ◽  
World Data ◽  
Causal Direction ◽  
Inference Methods

We address the problem of inferring the causal direction between two variables by comparing the least-squares errors of the predictions in both possible directions. Under the assumption of an independence between the function relating cause and effect, the conditional noise distribution, and the distribution of the cause, we show that the errors are smaller in causal direction if both variables are equally scaled and the causal relation is close to deterministic. Based on this, we provide an easily applicable algorithm that only requires a regression in both possible causal directions and a comparison of the errors. The performance of the algorithm is compared with various related causal inference methods in different artificial and real-world data sets.

scholarly journals Real world effects of COPD medications: a cohort study with validation against RCT results

2020 ◽  
pp. 2001586
Author(s):  
Kevin Wing ◽  
Elizabeth Williamson ◽  
James R Carpenter ◽  
Lesley Wise ◽  
Sebastian Schneeweiss ◽  
...  
Keyword(s):  
Cohort Study ◽  
Propensity Score ◽  
Drug Treatment ◽  
Real World ◽  
Treatment Effects ◽  
Control Group ◽  
Real World Data ◽  
Exacerbation Rate ◽  
Active Comparator ◽  
World Data

Real-world data provide the potential for generating evidence on drug treatment effects in groups excluded from trials, but rigorous, validated methodology for doing so is lacking. We investigated whether non-interventional methods applied to real-world data could reproduce results from the landmark TORCH COPD trial.We performed a historical cohort study (2000–2017) of COPD drug treatment effects in the UK Clinical Practice Research Datalink (CPRD). Two control groups were selected from CPRD by applying TORCH inclusion/exclusion criteria and 1:1 matching to TORCH participants: control group 1- people with COPD not prescribed fluticasone propionate-salmeterol (FP-SAL), control group 2- people with COPD prescribed salmeterol (SAL). FP-SAL exposed groups were then selected from CPRD by propensity-score matching to each control group. Outcomes studied were COPD exacerbations, death from any cause and pneumonia.2652 FP-SAL exposed people were propensity-score matched to 2652 FP-SAL unexposed people while 991 FP-SAL exposed people were propensity-score matched to 991 SAL exposed people. Exacerbation rate ratio was comparable to TORCH for FP-SAL versus SAL (0.85, 95% CI 0.74–0.97 versus 0.88, 0.81–0.95) but not for FP-SAL versus no FP-SAL (1.30, 1.19–1.42 versus 0.75, 0.69–0.81). Active comparator results were also consistent with TORCH for mortality (hazard ratio 0.93, 0.65–1.32 versus 0.93, 0.77–1.13) and pneumonia (risk ratio 1.39, 1.04–1.87 versus 1.47, 1.25–1.73).We obtained very similar results to the TORCH trial for active comparator analyses, but were unable to reproduce placebo-controlled results. Application of these validated methods for active comparator analyses to groups excluded from RCTs provides a practical way for contributing to the evidence base and supporting COPD treatment decisions.

scholarly journals Robust Causal Estimation from Observational Studies Using Penalized Spline of Propensity Score for Treatment Comparison

Stats ◽  
2021 ◽  
Vol 4 (2) ◽  
pp. 529-549
Author(s):  
Tingting Zhou ◽  
Michael R. Elliott ◽  
Roderick J. A. Little
Keyword(s):  
Propensity Score ◽  
Observational Studies ◽  
Treatment Effects ◽  
Single Measure ◽  
Propensity Score Model ◽  
Penalized Spline ◽  
Treatment Comparison ◽  
Control Subjects ◽  
Score Model ◽  
The Impact

Without randomization of treatments, valid inference of treatment effects from observational studies requires controlling for all confounders because the treated subjects generally differ systematically from the control subjects. Confounding control is commonly achieved using the propensity score, defined as the conditional probability of assignment to a treatment given the observed covariates. The propensity score collapses all the observed covariates into a single measure and serves as a balancing score such that the treated and control subjects with similar propensity scores can be directly compared. Common propensity score-based methods include regression adjustment and inverse probability of treatment weighting using the propensity score. We recently proposed a robust multiple imputation-based method, penalized spline of propensity for treatment comparisons (PENCOMP), that includes a penalized spline of the assignment propensity as a predictor. Under the Rubin causal model assumptions that there is no interference across units, that each unit has a non-zero probability of being assigned to either treatment group, and there are no unmeasured confounders, PENCOMP has a double robustness property for estimating treatment effects. In this study, we examine the impact of using variable selection techniques that restrict predictors in the propensity score model to true confounders of the treatment-outcome relationship on PENCOMP. We also propose a variant of PENCOMP and compare alternative approaches to standard error estimation for PENCOMP. Compared to the weighted estimators, PENCOMP is less affected by inclusion of non-confounding variables in the propensity score model. We illustrate the use of PENCOMP and competing methods in estimating the impact of antiretroviral treatments on CD4 counts in HIV+ patients.

scholarly journals Estimates of Causality with Medical Image in Oncology

ANALES RANM ◽  
2021 ◽  
Vol 138 (138(01)) ◽  
pp. 16-23
Author(s):  
Luis Martí-Bonmatí
Keyword(s):  
Causal Inference ◽  
Real World ◽  
Observational Studies ◽  
Clinical History ◽  
Response To Treatment ◽  
Control Analysis ◽  
Real World Data ◽  
World Data ◽  
New Research ◽  
Case Control Analysis

This work defines a research on data strategy focused on medical imaging and derived image biomarkers to critically assess the concept of causal inference and uncertainties. Computational observational studies will be valued to generate casual inference from real world data. Our main goal is to propose a scientific methodology that allows to estimate causalities from observational studies through quality control of large databases, definition of plausible hypotheses, using computational estimated models and artificial intelligence tools. The computational approach of radiology to precision medicine by using epidemiological strategies is based on causal inference studies relies on real-world data observational, longitudinal, case-control analysis designed (being case the presence, and control the absence of the event to be estimated). In this new research setting, we consider disease in classical epidemiology as phenotyping, response to treatment and final prognosis; and exposure equals to the presence of a radiomic, dynamic image biomarker or AI modeling solution. Research with data on which causality is to be inferred must control for recruitment of closed cases, in which the researcher does not intervene in the patient’s clinical history but works on databases, collecting data to be secondary used in generating consistent causalities.

Sign in / Sign up

Export Citation Format

Share Document