64,442 Hits in 5.7 sec

Treatment Effect Estimation using Invariant Risk Minimization [article]

Abhin Shah, Kartik Ahuja, Karthikeyan Shanmugam, Dennis Wei, Kush Varshney, Amit Dhurandhar
2021 arXiv   pre-print
In this work, we propose a new way to estimate the ITE using the domain generalization framework of invariant risk minimization (IRM).  ...  Inferring causal individual treatment effect (ITE) from observational data is a challenging problem whose difficulty is exacerbated by the presence of treatment assignment bias.  ...  effect (PEHE) which is the mean squared error of the estimated ITE for all the individuals in our data: ǫP EHE = 1 n n i=1 (τ (i) −τ (i) ) 2 (2) Invariant Risk Minimization [31] consider datasets De  ... 
arXiv:2103.07788v1 fatcat:2xqbgpclcbefbcctc6zos6t77y

Invariant Representation Learning for Treatment Effect Estimation [article]

Claudia Shi, Victor Veitch, David Blei
2021 arXiv   pre-print
NICE uses invariant risk minimization (IRM) [Arj19] to learn a representation of the covariates that, under some assumptions, strips out bad controls but preserves sufficient information to adjust for  ...  To address this problem, we develop Nearly Invariant Causal Estimation (NICE).  ...  INVARIANT RISK MINIMIZATION To learn an admissible representation, we use IRM. IRM is a framework for learning predictors that perform well across many environments.  ... 
arXiv:2011.12379v2 fatcat:5xa4etwyyrbwniiaxn5aehob2y

Learning Overlapping Representations for the Estimation of Individualized Treatment Effects [article]

Yao Zhang, Alexis Bellot, Mihaela van der Schaar
2020 arXiv   pre-print
Estimating the likely outcome of alternatives from observational data is a challenging problem as all outcomes are never observed, and selection bias precludes the direct comparison of differently intervened  ...  Despite their empirical success, we show that algorithms that learn domain-invariant representations of inputs (on which to make predictions) are often inappropriate, and develop generalization bounds  ...  This work was supported by GlaxoSmithKline (GSK), the Alan Turing Institute under the EP-SRC grant EP/N510129/1, the US Office of Naval Research (ONR), and the National Science Foundation (NSF): grant  ... 
arXiv:2001.04754v3 fatcat:j2sws7hzxngahaw2khlzq6cnt4

Sensitivity of treatment recommendations to bias in network meta-analysis

David M. Phillippo, Sofia Dias, A. E. Ades, Vanessa Didelez, Nicky J. Welton
2017 Journal of the Royal Statistical Society: Series A (Statistics in Society)  
Network meta-analysis (NMA) pools evidence on multiple treatments to estimate relative treatment effects.  ...  The methods use efficient matrix operations and can be applied to explore the consequences of bias in individual studies or aggregate treatment contrasts, in both fixed and random-effects NMA models.  ...  All studies with plausibly small thresholds should be assessed for risk of bias, e.g. by using GRADE (Guyatt et al., 2011) or the Cochrane risk-of-bias tool , and the thresholds and invariant intervals  ... 
doi:10.1111/rssa.12341 pmid:30449954 pmcid:PMC6221150 fatcat:ef6nieigl5czbn4ezlg27tyfgi

Discussion of a Paper by Professor Miettinen

James M. Robins
2015 Epidemiologic Methods  
AbstractProfessor Miettinen offers a scathing critique of the criteria used by official bodies to decide for whom and how often breast cancer screening should be offered (  ...  Professor Miettinen states that the invariant causal effect that one would ideally wish to estimate is the case fatality rate.  ...  It should be noted that g-methods are capable both of estimating the effect of a screening program (with the resultant treatment viewed as representing standard care), as well as the effect of a combined  ... 
doi:10.1515/em-2015-0026 fatcat:l43iirt3cjeajokvqn7hrnd6hy

Antidepressant Sales and the Risk for Alcohol-Related and Non-Alcohol-Related Suicide in Finland—An Individual-Level Population Study

Heta Moustgaard, Kaisla Joutsenniemi, Mikko Myrskylä, Pekka Martikainen, Kenji Hashimoto
2014 PLoS ONE  
antidepressant users with doses reflecting minimally adequate treatment.  ...  However, one percentage point increase in the proportion of antidepressant users receiving minimally adequate treatment reduced non-alcohol-related male suicide risk by one percent (relative risk 0.987  ...  Current study We use large Finnish register-based data to estimate the association between regional antidepressant sales and suicide risk.  ... 
doi:10.1371/journal.pone.0098405 pmid:24892560 pmcid:PMC4043885 fatcat:5ssvhhrqojc6zehecjkkgul62i

Residual Weighted Learning for Estimating Individualized Treatment Rules

Xin Zhou, Nicole Mayer-Hamblett, Umer Khan, Michael R. Kosorok
2017 Journal of the American Statistical Association  
We further obtain a rate of convergence for the difference between the expected outcome using the estimated ITR and that of the optimal treatment rule.  ...  We show that the resulting estimator of the treatment rule is consistent.  ...  Accordingly, we define the T-risk as and, similarly, the minimal T-risk as and .  ... 
doi:10.1080/01621459.2015.1093947 pmid:28943682 pmcid:PMC5607057 fatcat:l3l2uqyxgzbinbljkqhtoymeiq

Optimization-based Causal Estimation from Heterogenous Environments [article]

Mingzhang Yin, Yixin Wang, David M. Blei
2021 arXiv   pre-print
This paper presents a new optimization approach to causal estimation.  ...  We describe the theoretical foundations of this approach and demonstrate its effectiveness on simulated and real datasets.  ...  For example, we are often interested in estimating treatment effect, and the treatment variable is a non-descendant of the outcome.  ... 
arXiv:2109.11990v1 fatcat:qj4nfslodre23psfsveigaxyty

An application of propensity score weighting to quantify the causal effect of rectal sexually transmitted infections on incident HIV among men who have sex with men

Adam S Vaughan, Colleen F Kelley, Nicole Luisi, Carlos del Rio, Patrick S Sullivan, Eli S Rosenberg
2015 BMC Medical Research Methodology  
To adjust for behavioral confounding, while accounting for limited HIV infections, we used an inverse probability of treatment weighted (IPTW) Cox proportional hazards (PH) model for incident HIV.  ...  We hypothesized that, after controlling for potentially confounding behavioral and demographic factors, the significant STI-HIV association would attenuate, but yield an estimate of the causal effect.  ...  For time-toevent analyses, application of propensity scores using IPTW (rather than matching, stratification, or adjustment) produces effect estimates with minimal bias [21] .  ... 
doi:10.1186/s12874-015-0017-y pmid:25888416 pmcid:PMC4369368 fatcat:bxdpug5j2bda7ccjdmshth4hd4

Association between statins and infections among patients with diabetes: a cohort and prescription sequence symmetry analysis

Koen B. Pouwels, Niken N. Widyakusuma, Jens H. J. Bos, Eelko Hak
2016 Pharmacoepidemiology and Drug Safety  
However, animal experiments indicate that statins may be more effective in reducing the risk and/or the severity of infection among patients with diabetes.  ...  Hence, we evaluated the effect of statins on antibiotic prescriptions (a proxy for infections) among patients with drug-treated type 2 diabetes using two confounding-reducing observational designs.  ...  to evaluate the risk of receiving antibiotic treatment during periods that statins are used compared with non-use periods.  ... 
doi:10.1002/pds.4052 pmid:27365184 pmcid:PMC5129506 fatcat:sik6qlqt7basjo35g4tpf6hmtu

On Two Approaches to Weighting in Causal Inference

David A. Hirshberg, José R. Zubizarreta
2017 Epidemiology  
In this issue of the Journal, Setodji et al. 1 offer an interesting comparison of two weighting methods for estimating average treatment effects with time-invariant binary treatments: the use of generalized  ...  To get to the heart of the issue, we look at the error in estimating the average treatment effect on the treated, θ, when treatment assignment is unconfounded.  ...  In this issue of the Journal, Setodji et al. 1 offer an interesting comparison of two weighting methods for estimating average treatment effects with time-invariant binary treatments: the use of generalized  ... 
doi:10.1097/ede.0000000000000735 pmid:28817467 fatcat:qd25p7vlozgothchalxg3lvieq

Optimal management of a biochemical incomplete response to therapy in differentiated thyroid cancer: aggressive treatment or cautious observation?

R. Michael Tuttle
2014 Endocrine (Basingstoke)  
These initial risk estimates are used to guide early management decisions and plan follow-up studies that allow us to define the response to initial therapy for each patient.  ...  These observations must make us re-evaluate our current aggressive management approach to the detection and treatment of minimal residual disease.  ... 
doi:10.1007/s12020-014-0213-2 pmid:24615658 fatcat:2fz6rxfs4zhnbcrxh46sqafmta

Allocation Rules for Sequential Clinical Trials [chapter]

D. Siegmund
1983 Mathematical Learning Models — Theory and Algorithms  
A and nk of treatment I In the k stratum, the maximum likelihood estimator of the treatment effect 8 is (in th, obvious notation) (18) -K-; 4 , (xkmk-yk. nk) k-l k+n Let z(z,.n) denote the numerator  ...  treatment allocation rule is used.  ...  L9 particular problem should probably be made before seriously contemplating use of a sequential allocation scheme. Remarks (i).  ... 
doi:10.1007/978-1-4612-5612-0_22 fatcat:nb75uqh4mbgqjhfgb54gdh5qby

Future Cases as Present Controls to Adjust for Exposure Trend Bias in Case-only Studies

Shirley Wang, Crystal Linkletter, Malcolm Maclure, David Dore, Vincent Mor, Stephen Buka, Gregory A. Wellenius
2011 Epidemiology  
However, when the exposure of interest has a transient effect on risk for an abrupt onset outcome, the solution suggested by researchers such as Maclure (case-crossover), 1  ...  Simulation studies show that the case-case-time control can adjust for exposure trends while controlling for time-invariant confounders.  ...  The case-case-time-control design additionally minimizes the risk of introducing selection bias from use of a control group whose study base does not match that of the cases.  ... 
doi:10.1097/ede.0b013e31821d09cd pmid:21577117 pmcid:PMC3110688 fatcat:htuqidfrkvcyrcssk6tuqdhuo4

Minimax designs for causal effects in temporal experiments with treatment habituation [article]

Guillaume Basse, Yi Ding, Panos Toulis
2020 arXiv   pre-print
Randomized experiments are the gold standard for estimating the causal effects of an intervention.  ...  This paper proposes randomized designs for estimating causal effects in temporal experiments when habituation is present.  ...  Acknowledgments We would like thank organizers and participants at the Google Market Workshop, and at the LinkedIn Seminar for useful comments and feedback.  ... 
arXiv:1908.03531v2 fatcat:42j2f54edncm7f2m5p5ady4ziy
« Previous Showing results 1 — 15 out of 64,442 results