Algorithmic approach to finding people with multiple sclerosis using routine healthcare data in Wales

Area of Science:

• Epidemiology
• Health Informatics
• Neurology

Background:

• Identifying multiple sclerosis (MS) cases in routine healthcare data is challenging due to complex diagnosis and infrequent primary hospital admissions.
• Data limitations, such as lack of drug treatment or non-notifiable disease information, further complicate MS case ascertainment.
• Existing methods struggle with the protracted diagnostic process and varied clinical presentations of MS.

Purpose of the Study:

• To develop and validate a novel algorithm for the reliable identification of multiple sclerosis (MS) cases within a national health data bank.
• To enhance the accuracy of MS case detection in large-scale health datasets.
• To provide a robust tool for epidemiological research on multiple sclerosis.

Main Methods:

• A retrospective analysis of the Secure Anonymised Information Linkage (SAIL) databank was conducted.
• A novel algorithm was developed to identify MS cases.
• Algorithm performance was evaluated using two independent datasets: a clinically validated, population-based MS cohort and a self-registered national MS registry.

Main Results:

• The algorithm identified 6194 living MS cases in Wales by December 31, 2020, from 4,757,428 records, yielding a prevalence of 221.65 per 100,000.
• High case-finding sensitivity (96.8%) and specificity (99.9%) were achieved when validated against the population-based cohort.
• Sensitivity of 96.7% was demonstrated using the self-declared national registry data.

Conclusions:

• The developed algorithm successfully identifies multiple sclerosis (MS) cases in the SAIL databank with high accuracy.
• Validation against two independent MS populations confirms the algorithm's reliability and utility.
• This algorithm is a valuable tool for large-scale epidemiological studies of multiple sclerosis.