Population-Level Prediction of Type 2 Diabetes From Claims Data and Analysis of Risk Factors

By Narges Razavian, Saul Blecker, A. Schmidt, Aaron Smith-McLallen, S. Nigam, David A. Sontag. 2015

212 citations recorded by the source corpus.

Abstract We present a new approach to population health, in which data-driven predictive models are learned for outcomes such as type 2 diabetes. Our approach enables risk assessment from readily available electronic claims data on large populations, without additional screening cost. Proposed model uncovers early and late-stage risk factors. Using administrative claims, pharmacy records, healthcare utilization, and laboratory results of 4.1 million individuals between 2005 and 2009, an initial set of 42,000 variables were derived that together describe the full health status and history of every individual. Machine learning was then used to methodically enhance predictive variable set and fit models predicting onset of type 2 diabetes in 2009–2011, 2010–2012, and 2011–2013. We compared the enhanced model with a parsimonious model consisting of known diabetes risk factors in a real-world environment, where missing values are common and prevalent. Furthermore, we analyzed novel and known risk factors emerging from the model at different age groups at different stages before the onset. Parsimonious model using 21 classic diabetes risk factors resulted in area under ROC curve (AUC) of 0.75 for diabetes prediction within a 2-year window following the baseline. The enhanced model increased the AUC to 0.80, with about 900 variables selected as predictive (p < 0.0001 for differences between AUCs). Similar improvements were observed for models predicting diabetes onset 1–3 years and 2–4 years after baseline. The enhanced model improved positive predictive value by at least 50% and identified novel surrogate risk factors for type 2 diabetes, such as chronic liver disease (odds ratio [OR] 3.71), high alanine aminotransferase (OR 2.26), esophageal reflux (OR 1.85), and history of acute bronchitis (OR 1.45). Liver risk factors emerge later in the process of diabetes development compared with obesity-related factors such as hypertension and high hemoglobin A1c. In conclusion, population-level risk prediction for type 2 diabetes using readily available administrative data is feasible and has better prediction performance than classical diabetes risk prediction algorithms on very large populations with missing data. The new model enables intervention allocation at national scale quickly and accurately and recovers potentially novel risk factors at different stages before the disease onset.

Subfields: Artificial Intelligence, Endocrinology, Diabetes and Metabolism, Health Information Management

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Endocrinology, Diabetes and Metabolism

In Health Information Management

Citations in the corpus

Why Are We Using Black Box Models in AI When We Don’t Need To? A Lesson From An Explainable AI Competition (2019, background): “…domains and across many other high-stakes machine learning applications where life-altering decisions are being made (e.g., Caruana et al., 2015; Razavian et al., 2015; Rudin & Ustun, 2018, who all show models with interpretability constraints that perform just as well as unconstrained models).”
The Secrets of Machine Learning: Ten Things You Wish You Had Known Earlier to be More Effective at Data Analysis (2019, background): “Surprisingly, there appears to be no practical significant difference in performance between machine learning methods for a huge number of data science problems [45, 10, 62, 2].”
A study in Rashomon curves and volumes: A new perspective on generalization and model simplicity in machine learning (2019, background): “, [47, 29]), diabetes prediction [36], and pneumonia risk and readmission prediction [10] are in a similar state.”
Predicting Diabetes Mellitus With Machine Learning Techniques (2018, background): “In order to deal with the high dimensional datasets, Razavian et al. (2015) built prediction models based on logistic regression for diﬀerent onsets of type 2 diabetes prediction.”
IDF Diabetes Atlas: Global estimates for the prevalence of diabetes for 2015 and 2040. (2011, background): “[38] Razavian N, Blecker S, Schmidt AM, Smith-McLallen A, Nigam S, Sontag D.”
Comparing deep learning and concept extraction based methods for patient phenotyping from clinical narratives (2018, background): “Accurate patient phenotyping is required for secondary analysis of EHRs to correctly identify the patient cohort and to better identify the clinical context [36, 37].”
Machine Learning and Data Mining Methods in Diabetes Research (2017, background): “Prediction of DM The second category deals with disease prediction a nd diagnosis [40-76].”
Learning a Health Knowledge Graph from Electronic Medical Records (2017, methodology): “L1 regularization was used for logistic regression both to prevent overfitting and to encourage sparsity, which was desirable as we expect most diseases to cause only a small number of symptoms 35 .”
Comparing Rule-Based and Deep Learning Models for Patient Phenotyping (2017, methodology): “efficiently used to detect patients if it does not matter that the model overlooks most of the positives, for example to detect a small at-risk population for interventions.(73)”