Kan Chen

Kan Chen

Postdoctoral Research Fellow at Harvard and Columbia/PhD from UPenn

Peer-Reviewed Publications and Preprint

Causal Inference

This body of work advances statistical methodologies for causal inference using observational data, with a focus on partial identification, sensitivity analysis, mediation analysis, and case-control designs to address challenges such as unmeasured confounding bias, selection bias, and residual bias. These methods are applied to real-world problems in social policy, clinical research, and oncology studies.

  1. Kan Chen, Ruoyu Wang, Zhonghua Liu, Xihong Lin. “The Blessing of Multiple Mediators: Removing Unmeasured Confounding Bias via Factor Analysis.” Journal of the American Statistical Association. In preparation.
  2. Kan Chen, Ting Ye, Greg Ridgeway, Dylan S. Small. “Evaluating ‘Stand Your Ground’ Laws: A Bracketing Approach for Partial Identification in Difference-in-Difference Analysis with Staggered Adoption.” Annals of Applied Statistics. In preparation.
  3. Kan Chen, Jeffrey Zhang, Bingkai Wang, Dylan S. Small. “A Differential Effect Approach to Partial Identification of Treatment Effects.” Biometrika. Major revision.
    (Recipient of the 2024 IMS New Researcher Travel Award)
  4. Kan Chen, Ting Ye, Dylan S. Small. “Sensitivity Analysis for Attributable Effects in Case2 Studies.” Biometrics.
  5. Kan Chen, Jing Cheng, M. Elizabeth Halloran, Dylan S. Small. “Using Case Description Information to Reduce Sensitivity to Bias for the Attributable Fraction Among the Exposed.” Journal of the Royal Statistical Society: Series A.
    (Recipient of the 2023 ASA Statistics in Epidemiology Young Investigator Award)
  6. Ye Ting, Kan Chen, Dylan S. Small. “Combining Broad and Narrow Case Definitions in Matched Case-Control Studies: Firearms in the Home and Suicide Risk.” Journal of the American Statistical Association.
  7. Kan Chen, Siyu Heng, Qi Long, Bo Zhang. “Testing Biased Randomization Assumptions and Quantifying Imperfect Matching and Residual Confounding in Matched Observational Studies.” Journal of Computational and Graphical Statistics.

Causal Machine Learning

This research develops and applies machine learning methods for causal inference and treatment effect estimation, including interpretable deep learning models, privacy-preserving algorithms, and large-scale emulated trials. It spans a broad range of topics, from federated instrumental variable estimation for distributed healthcare data to optimization theory for high-dimensional statistics and overparameterized neural networks, with the goal of integrating modern machine learning tools into rigorous and scalable causal analysis frameworks.

  1. Kan Chen, Inyoung Choi, Ravi Parikh, Qi Long. “CROME: Causal Representation Learning for Composite Outcomes via Multi-Task Estimation in Electronic Health Records.” . In preparation.
  2. Orcutt Xavier, Kan Chen, Ronac Mamtani, Qi Long, Ravi Parikh. “Evaluating Generalizability of Results from Landmark Randomized Controlled Trials in Oncology Using Machine Learning-Based Emulated Trials.” Nature Medicine.
  3. Yinjun Wu, Neelay Velingker, Kan Chen, et al. “DISCRET: Synthesizing Faithful Explanations for Treatment Effect Estimation.” ICML 2024.
    (Spotlight presentation, acceptance rate 3.5%)
  4. Kan Chen, Qishuo Yin, Qi Long. “Covariate-Balancing-Aware Interpretable Deep Learning Models for Treatment Effect Estimation.” Statistics in Biosciences.
    (Recipient of the 2022 ICSA Jiann-Ping Hsu Student Paper Award and NSF Travel Award)
  5. Qiyiwen Zhang, Zhiqi Bu, Kan Chen, Qi Long. “Differentially Private Bayesian Neural Networks on Accuracy, Privacy and Reliability.” ECML PKDD 2022.
  6. Kan Chen, Shiyun Xu, Zhiqi Bu. “Asymptotic Analysis of Sparse Group LASSO via Approximate Message Passing Algorithm.” ECML PKDD 2021.
  7. Zhiqi Bu, Shiyun Xu, Kan Chen. “A Dynamical View on Optimization Algorithms of Overparameterized Neural Networks.” AISTATS 2021.
  8. Kan Chen, Qi Long. “Distributed Gaussian Differential Privacy Via Shuffling.” ICLR 2021 Workshop on Distributed and Private Machine Learning.

Scientific Collaborations

These collaborative studies apply rigorous statistical and machine learning techniques to pressing biomedical and health policy questions. Topics include disparities in cancer care, clinical trial design, palliative care, and the integration of patient-reported outcomes in predictive modeling.

  1. Sophia Shi, Kan Chen, Qi Long, Ravi Parikh. “Effect of Incorporating Symptom Burden With Mortality as a Composite Outcome on Accuracy and Bias in Palliative Care Identification Algorithms in Oncology.” Journal of Clinical Oncology.
  2. Yehoda M. Martei, Kan Chen, Ronac Mamtani, Lawrence N. Shulman, Rebecca A. Hubbard. “Racial disparities in utilization of first-line targeted therapies for metastatic breast cancer.” JCO Oncology Practice. Submitted.
  3. Mia Djulbegovic, Meredith Doherty, Katie Fanslau, Kan Chen, et al. “A Randomized Controlled Trial of a Financial Navigation Program for Patients with Multiple Myeloma.” JCO Oncology Practice. Submitted.
  4. Christopher R. Manz, Yichen Zhang, Kan Chen, et al. “Machine Learning-Triggered Behavioral Nudges on Serious Illness Communication and End-of-Life Outcomes Among Patients With Cancer: A Randomized Clinical Trial.” JAMA Oncology.
    (Selected Best Papers for the 2024 IMIA Yearbook)
  5. Stephen Bagley, Jacob Shabason, Kan Chen, et al. “CTIM-35. A Phase II Study of GITR Agonist INCAGN01876 and PD-1 Inhibitor Retifanlimab in Combination With Stereotactic Radiotherapy in Patients With Recurrent Glioblastoma.” Neuro-Oncology.
  6. Stephen J. Bagley, Divij Mathew, Kan Chen, et al. “PD1 Inhibition and GITR Agonism in Combination With Fractionated Stereotactic Radiotherapy for the Treatment of Recurrent Glioblastoma: A Phase 2, Multi-Arm Study.” Journal of Clinical Oncology.