Clusters of Comorbidities in Multiple Sclerosis and Their Influence on Healthcare Resource Usage

MS patients are at increased risk of comorbidities and use more healthcare resources. Multimorbidity approached as the number of conditions is flawed by classifying patients with different needs as equal. We aimed to explore how comorbidities cluster and their impact on healthcare resource usage.

We used latent-class models of up to 10 clusters in a population-based sample of MS patients. The optimal number of clusters was determined using model metrics and similarity/entropy measures, and cluster stability was assessed by bootstrapping. Sociodemographic characteristics and healthcare-resource usage according to the clusters assigned were compared to each other and to patients without comorbidities using univariable and adjusted linear regression models.

In 5548 MS cases, of which 60% had comorbidities, the optimal number of comorbidity clusters was two, comprising a high frequency of cardiovascular comorbidities and psychiatric disorders. Patients in the cardiovascular cluster were older, and in the psychiatric cluster were more frequently female. After adjusting for sociodemographic variables, healthcare resource usage was higher for patients with comorbidities, particularly for nurse (1.1 more average yearly visit; [95% CI 0.41-1.8]; p = 0.002), primary care (1.8 more visits; [95% CI 1.4-2.1]; p < 0.001), and medication dispensation (336 more dosage units; [95% CI 260-402]; p < 0.001) in the cardiovascular cluster, and annual sick-leave days (3.8 more days; [95% CI 0.25-7.3]; p = 0.036) in the psychiatric cluster.

We observed clustering of comorbidities around cardiovascular comorbidities and mental disorders, which impacted healthcare resource usage differently. Further research is needed to assess the influence of these clusters on the prognosis of MS.