Neural networks to estimate multiple sclerosis disability and predict progression using routinely collected...

Area of Science:

• Neurology
• Data Science
• Public Health

Background:

• The Expanded Disability Status Scale (EDSS) is crucial for measuring multiple sclerosis (MS) disability but is often unavailable in administrative datasets.
• This limitation hinders population-level MS research and healthcare planning.
• This study addresses this gap by leveraging routinely collected healthcare data.

Purpose of the Study:

• To develop and validate AI algorithms for estimating MS-related disability using administrative data.
• To predict the progression of MS disability.
• To improve MS monitoring and healthcare planning through advanced data analytics.

Main Methods:

• A population-based study was conducted using administrative data from the Campania Region, Italy (2015-2021).
• Deep learning models were employed to estimate EDSS scores.
• A hybrid model combining survival analysis and neural networks was used to forecast EDSS progression risk.

Main Results:

• The EDSS estimation model achieved an accuracy of 0.68, precision of 0.68, and F1-score of 0.67.
• The hybrid model for predicting progression demonstrated a high predictive performance of 0.92.
• Analysis revealed disability distribution from 2016-2021, with projections indicating significant progression for a majority of individuals with lower EDSS scores.

Conclusions:

• Advanced data analytics on administrative data show promise for improving MS monitoring.
• These methods can significantly aid healthcare planning and decision-making for MS patients.
• The study highlights the potential of AI in transforming population-level health management for chronic conditions.