JMIR Publications

ABSTRACT

Background:

Degenerative cervical myelopathy (DCM), a progressive spinal cord injury caused by spinal cord compression from degenerative pathology, often presents with neck pain, sensorimotor dysfunction in the upper and/or lower limbs, gait disturbance and bladder or bowel dysfunction. Its symptomatology is very heterogenous, making it hard to detect early and to measure or understand what drives consequences.

Objective:

The objective of this study was to explore whether machine learning can identify clinically meaningful groups of patients based on clinical features alone.

Methods:

A survey was conducted wherein participants were asked to specify the clinical features that they had experienced, their principal presenting complaint, and time to diagnosis, as well as demographic information including disease severity, age, and sex. K-means clustering was used to divide respondents into clusters according to their clinical features using the Euclidean distance measure and the Hartigan-Wong algorithm. The clinical significance of groups was subsequently explored by comparing their time to presentation, time with disease severity, and other demographics.

Results:

After review of both ancillary and cluster data, it was determined by consensus that the optimal number of DCM response groups was 3. There were 40 respondents in Cluster 1, 92 in Cluster 2 and 57 in Cluster 3. The ratio of males to females was 13:21 in Cluster 1, 27:65 in Cluster 2 and 9:48 in Cluster 3. 6 people did not report biological sex in Cluster 1. The mean age for Cluster 1 was 56.2 (SD 10.5), 54.7 (SD 9.63) for Cluster 2 and 51.8 (SD 8.4) for Cluster 3. Patients across clusters significantly differed in total number of clinical features reported, with more clinical features in Cluster 3 and least in Cluster 1 (Kruskal-Wallis rank sum test: χ22 = 159.46, p < 0.001). There was no relationship between pattern of clinical features and severity. There were also no differences between clusters with regards to time since diagnosis and time with DCM

Conclusions:

Using machine learning and patient reported experience, three groups of DCM patients were defined. This differed in the number of clinical features reported but not in the severity of DCM or time with DCM. The significance and generalization of this remains uncertain. Overall, this study confirms a role for machine learning in DCM research, but more pressingly the need to curate the right datasets.