Literature DB >> 33716168

Using computable knowledge mined from the literature to elucidate confounders for EHR-based pharmacovigilance.

Scott A Malec1, Peng Wei2, Elmer V Bernstam3, Richard D Boyce4, Trevor Cohen5.   

Abstract

INTRODUCTION: Drug safety research asks causal questions but relies on observational data. Confounding bias threatens the reliability of studies using such data. The successful control of confounding requires knowledge of variables called confounders affecting both the exposure and outcome of interest. However, causal knowledge of dynamic biological systems is complex and challenging. Fortunately, computable knowledge mined from the literature may hold clues about confounders. In this paper, we tested the hypothesis that incorporating literature-derived confounders can improve causal inference from observational data.
METHODS: We introduce two methods (semantic vector-based and string-based confounder search) that query literature-derived information for confounder candidates to control, using SemMedDB, a database of computable knowledge mined from the biomedical literature. These methods search SemMedDB for confounders by applying semantic constraint search for indications treated by the drug (exposure) and that are also known to cause the adverse event (outcome). We then include the literature-derived confounder candidates in statistical and causal models derived from free-text clinical notes. For evaluation, we use a reference dataset widely used in drug safety containing labeled pairwise relationships between drugs and adverse events and attempt to rediscover these relationships from a corpus of 2.2 M NLP-processed free-text clinical notes. We employ standard adjustment and causal inference procedures to predict and estimate causal effects by informing the models with varying numbers of literature-derived confounders and instantiating the exposure, outcome, and confounder variables in the models with dichotomous EHR-derived data. Finally, we compare the results from applying these procedures with naive measures of association (χ2 and reporting odds ratio) and with each other. RESULTS AND
CONCLUSIONS: We found semantic vector-based search to be superior to string-based search at reducing confounding bias. However, the effect of including more rather than fewer literature-derived confounders was inconclusive. We recommend using targeted learning estimation methods that can address treatment-confounder feedback, where confounders also behave as intermediate variables, and engaging subject-matter experts to adjudicate the handling of problematic covariates. Published by Elsevier Inc.

Entities:  

Keywords:  Causal inference; Confounder selection; Confounding bias; Electronic health records; Pharmacovigilance

Mesh:

Year:  2021        PMID: 33716168      PMCID: PMC8559730          DOI: 10.1016/j.jbi.2021.103719

Source DB:  PubMed          Journal:  J Biomed Inform        ISSN: 1532-0464            Impact factor:   6.317


  86 in total

Review 1.  Frequency of adverse drug reactions in hospitalized patients: a systematic review and meta-analysis.

Authors:  Ana Miguel; Luís Filipe Azevedo; Manuela Araújo; Altamiro Costa Pereira
Journal:  Pharmacoepidemiol Drug Saf       Date:  2012-07-04       Impact factor: 2.890

2.  Exploiting semantic relations for literature-based discovery.

Authors:  Dimitar Hristovski; Carol Friedman; Thomas C Rindflesch; Borut Peterlin
Journal:  AMIA Annu Symp Proc       Date:  2006

3.  Embedding of semantic predications.

Authors:  Trevor Cohen; Dominic Widdows
Journal:  J Biomed Inform       Date:  2017-03-08       Impact factor: 6.317

4.  Overall prescription medication use among adults: findings from the Survey of the Health of Wisconsin.

Authors:  Jianhong Che; Kristen C Malecki; Matthew C Walsh; Andrew J Bersch; Vincent Chan; Christine A McWilliams; F Javier Nieto
Journal:  WMJ       Date:  2014-12

5.  Cost of Prescription Drug-Related Morbidity and Mortality.

Authors:  Jonathan H Watanabe; Terry McInnis; Jan D Hirsch
Journal:  Ann Pharmacother       Date:  2018-03-26       Impact factor: 3.154

6.  From Big Data to Smart Data for Pharmacovigilance: The Role of Healthcare Databases and Other Emerging Sources.

Authors:  Gianluca Trifirò; Janet Sultana; Andrew Bate
Journal:  Drug Saf       Date:  2018-02       Impact factor: 5.606

7.  The Mechanics of Omitted Variable Bias: Bias Amplification and Cancellation of Offsetting Biases.

Authors:  Peter M Steiner; Yongnam Kim
Journal:  J Causal Inference       Date:  2016-11-08

8.  Discovering discovery patterns with Predication-based Semantic Indexing.

Authors:  Trevor Cohen; Dominic Widdows; Roger W Schvaneveldt; Peter Davies; Thomas C Rindflesch
Journal:  J Biomed Inform       Date:  2012-07-26       Impact factor: 6.317

9.  Ensembles of natural language processing systems for portable phenotyping solutions.

Authors:  Cong Liu; Casey N Ta; James R Rogers; Ziran Li; Junghwan Lee; Alex M Butler; Ning Shang; Fabricio Sampaio Peres Kury; Liwei Wang; Feichen Shen; Hongfang Liu; Lyudmila Ena; Carol Friedman; Chunhua Weng
Journal:  J Biomed Inform       Date:  2019-10-23       Impact factor: 6.317

10.  Principles of confounder selection.

Authors:  Tyler J VanderWeele
Journal:  Eur J Epidemiol       Date:  2019-03-06       Impact factor: 8.082

View more
  2 in total

1.  Improving Pharmacovigilance Signal Detection from Clinical Notes with Locality Sensitive Neural Concept Embeddings.

Authors:  Justin Mower; Elmer Bernstam; Hua Xu; Sahiti Myneni; Devika Subramanian; Trevor Cohen
Journal:  AMIA Annu Symp Proc       Date:  2022-05-23

Review 2.  Machine Learning in Causal Inference: Application in Pharmacovigilance.

Authors:  Yiqing Zhao; Yue Yu; Hanyin Wang; Yikuan Li; Yu Deng; Guoqian Jiang; Yuan Luo
Journal:  Drug Saf       Date:  2022-05-17       Impact factor: 5.228

  2 in total

北京卡尤迪生物科技股份有限公司 © 2022-2023.