Risk Prediction of a Multiple Sclerosis Diagnosis

TL;DR

This paper develops a risk prediction model for multiple sclerosis using electronic medical records, achieving reasonable accuracy and demonstrating generalizability across healthcare systems to aid early diagnosis.

Contribution

It introduces a novel EMR-based risk prediction model for MS that performs well with limited data and is adaptable across different healthcare systems.

Findings

Achieved AUC of 0.724 with limited patient data

Model generalizes well across different healthcare systems

Focuses on early risk prediction rather than post-diagnosis progression

Abstract

Multiple sclerosis (MS) is a chronic autoimmune disease that affects the central nervous system. The progression and severity of MS varies by individual, but it is generally a disabling disease. Although medications have been developed to slow the disease progression and help manage symptoms, MS research has yet to result in a cure. Early diagnosis and treatment of the disease have been shown to be effective at slowing the development of disabilities. However, early MS diagnosis is difficult because symptoms are intermittent and shared with other diseases. Thus most previous works have focused on uncovering the risk factors associated with MS and predicting the progression of disease after a diagnosis rather than disease prediction. This paper investigates the use of data available in electronic medical records (EMRs) to create a risk prediction model; thereby helping clinicians perform the difficult task of diagnosing an MS patient. Our results demonstrate that even given a limited time window of patient data, one can achieve reasonable classification with an area under the receiver operating characteristic curve of 0.724. By restricting our features to common EMR components, the developed models also generalize to other healthcare systems.