Causal Inference

Objective: Questions of causality are fundamental inspirations behind innovations in philosophy, business, and science, including biomedicine. Answering causal questions leads to better predictive and prescriptive modeling. It is also the key to the correct identification of unknown effects as well as the latent factors that influence outcomes, and to produce hypotheses, validation, and proof.

Large-scale observational data offers a new window for verifying our existing causal understandings and for inferring new causal relationships at a fast pace. We are working to rethink the existing frameworks of causal inference in health sciences by introducing ideas from other scientific disciplines and by inventing new concepts and analytical methods.

• If you are working in causal inference add your project in the following. Also, you can find projects to collaborate or join to stay informed.

• If you have an idea on causal inference that you want to test it or even develop it, you probably will find collaborators here

Project Leads:

* Abbas Shojaee * Yalda Aryan

Ongoing Projects

Causal Inference Using Composition of Transactions (CICT)

CICT is a novel computational method that uses large-scale health data to predict potential causal relationships between clinical conditions. A pipeline for epidemiological etiological inference is also developed to validate the results of CICT. CICT and the validation pipeline have been used by different teams to identify latent risk factors and unknown effects of clinical conditions or procedures, which resulted in reporting of multiple novel findings during 2018-2019. Two large-scale population-level claims datasets from HCUP California and Florida have been used for discovery and validation. OHDSI will be used to empower the engine for the new phase.

CICT pipeline can produce highly accurate novel hypotheses and analyses results at a fast pace. Accordingly, we are now open to collaboration requests from clinical researchers who are seeking novel hypotheses from data and/or epidemiological evidence from population-level data. We are also eager to start collaborations with individuals or organizations that can provide us access to large-scale claims/EMR datasets in OHDSI, based on their existing access.

* Publications and conference papers: *

Viral Pneumonia is Associated with Increased Risk and Earlier Development of Post-Inflammatory Pulmonary Fibrosis

Robust discovery of gene regulatory networks from single-cell gene expression data by Causal Inference Using Composition of Transactions

Asthma-Neoplasms Relationships: New Insights Using Machine Inference, Epidemiological Reasoning, And Big Data

The Associations of Invasive Procedures and Subsequent Psychiatric Diagnoses

Obstructive Sleep Apnea Increases the Risk of Diastolic Heart Failure: Results of a Large Population Study

Viral Pneumonia Is an Independent Risk Factor for Pulmonary Fibrosis: Results of Large-Scale Longitudinal Population-Level Data|

Platelet Transfusion Is an Independent Risk Factor for Idiopathic Pulmonary Fibrosis: Results of Large-Scale Longitudinal Population-Level Data

Repository:

TBD

News Link

TBD

Literature Repository

TBD