ResearchAI Research

Machine Learning Links Psychosocial Factors to Chronic Kidney Disease

Md. Atik Shams, David Eisenberg, Sumaiya Fatema, Asma Sultana, D. M Hasibul Islam, Junnatul Mawa, Anindita Datta, Nafiya Ahmed, Danastan Tasaouf Mridula, SK. Sazid Mahmud, Simon Bin Akter, Tanjila Helaly, Jorge Fresneda Fernandez, Humayera Islam, Tanmoy Sarkar Pias· August 19, 2026 View original

Key takeaways

  • Machine learning can accurately classify CKD status from self-reported data.
  • Psychosocial factors, including mental health stress, are significant CKD predictors.
  • Regular medical check-ups and blood pressure are also critical indicators.
  • Explainable AI (SHAP) provides interpretable insights into disease drivers.

Who benefits

HealthcarePublic HealthInsurancePharmaHealthTech

Summary

This two-part study uses large-scale telehealth data and machine learning to classify self-reported Chronic Kidney Disease (CKD) status and identify key drivers. A stacked ensemble model achieved high accuracy, and SHAP analysis highlighted critical predictors including regular check-ups, age, blood pressure, and mental health stress indicators.

A comprehensive two-part study has utilized large-scale telehealth data combined with advanced machine learning techniques to investigate Chronic Kidney Disease (CKD). The research aimed both to accurately classify self-reported CKD status and to pinpoint the significant factors contributing to the disease's development. The study leveraged extensive datasets from the Behavioral Risk Factor Surveillance System (BRFSS) and the National Health Interview Survey (NHIS), addressing common data challenges such as missing values through nine state-of-the-art imputation methods and mitigating class imbalance with various sampling strategies. A custom-built stacked ensemble model achieved balanced accuracy between 72.56-76.12% and AUROC scores of 79.59-82.29%. Further analysis using SHapley Additive exPlanations (SHAP) revealed critical predictors of CKD, which were subsequently reviewed by clinicians. These included regular medical check-ups, age, blood pressure, and notably, indicators of mental health stress. The findings establish a robust and interpretable framework for CKD risk stratification and offer actionable insights into its associated psychosocial factors.

Why it matters

For healthcare professionals, public health officials, and insurance providers, this research provides a powerful, interpretable tool for early CKD risk stratification. Identifying psychosocial factors alongside traditional clinical markers can lead to more holistic prevention strategies and improved patient outcomes.

How to implement this in your domain

  1. 1Integrate machine learning models, such as stacked ensembles, into population health initiatives for early disease risk stratification.
  2. 2Utilize explainable AI techniques like SHAP to identify and interpret key predictors of chronic diseases from large datasets.
  3. 3Develop telehealth-based screening programs that incorporate psychosocial factors alongside traditional clinical data for CKD risk assessment.
  4. 4Design public health interventions that address mental health stress and promote regular medical check-ups as part of CKD prevention.
  5. 5Collaborate with data scientists to leverage large-scale survey data (e.g., BRFSS, NHIS) for identifying disease associations.

Original post by Md. Atik Shams, David Eisenberg, Sumaiya Fatema, Asma Sultana, D. M Hasibul Islam, Junnatul Mawa, Anindita Datta, Nafiya Ahmed, Danastan Tasaouf Mridula, SK. Sazid Mahmud, Simon Bin Akter, Tanjila Helaly, Jorge Fresneda Fernandez, Humayera Islam, Tanmoy Sarkar Pias

"arXiv:2608.17174v1 Announce Type: new Abstract: Chronic kidney disease (CKD) progresses silently and severely undermines quality of life, making early detection critical for improving patient outcomes. We present a two-part study that combines large-scale telehealth data with adv…"

View on X

Originally posted by Md. Atik Shams, David Eisenberg, Sumaiya Fatema, Asma Sultana, D. M Hasibul Islam, Junnatul Mawa, Anindita Datta, Nafiya Ahmed, Danastan Tasaouf Mridula, SK. Sazid Mahmud, Simon Bin Akter, Tanjila Helaly, Jorge Fresneda Fernandez, Humayera Islam, Tanmoy Sarkar Pias on X · view source

Want to go deeper?

Turn these trends into skills with Learnijoy's hands-on AI & tech courses.

Explore courses

More in AI Research