Cluster Analysis Reveals Subgroups with Different Risk Profiles and Sickness Absence Patterns in an Occupational...

Area of Science:

• Occupational health
• Data science in healthcare
• Machine learning applications

Background:

• Sickness absence (SA) poses significant challenges for employee well-being and organizational productivity.
• Understanding the heterogeneity of employee health and its association with SA is crucial for targeted interventions.

Purpose of the Study:

• To identify distinct employee groups using machine learning based on health and workplace factors.
• To analyze the relationship between these identified employee clusters and patterns of sickness absence (SA).

Main Methods:

• Employed unsupervised and supervised machine learning on data from 12,099 Finnish employees (2011-2019).
• Utilized principal component analysis for dimensionality reduction, followed by K-means clustering.
• Assessed associations between clusters and long (>30 days) or repetitive short (1-10 days) SA episodes using logistic regression.

Main Results:

• Six employee clusters were identified, characterized by factors like managerial performance, workplace atmosphere, mood/depression, cardiovascular diseases, sensory symptoms, and work ability.
• Cluster 5 exhibited the highest rate of repetitive short SA, linked to numerous symptoms.
• Cluster 6 showed the highest rate of long SA, associated with work ability deficiencies.

Conclusions:

• Machine learning successfully delineated six clinically meaningful employee clusters.
• These clusters provide insights into characteristic combinations and distinct sickness absence risk profiles.
• Findings support tailored workplace health strategies based on identified employee group characteristics.