Identifying patient subgroups in the heterogeneous chronic pain population using cluster analysis

Area of Science:

• Pain Medicine
• Psychology
• Health Services Research

Background:

• Chronic pain presents complex biopsychosocial challenges, often leading to treatment difficulties.
• Treatment failures may stem from a lack of understanding regarding diverse patient subgroups.
• Tailored interventions require precise patient stratification based on psychological profiles.

Purpose of the Study:

• To identify distinct patient subgroups within chronic pain populations using psychological variables.
• To enable more targeted and effective therapeutic interventions for chronic pain.
• To develop a predictive model for reliable cluster allocation using minimal questions.

Main Methods:

• Retrospective cohort study of 5466 patients from Dutch multidisciplinary pain clinics (2018-2023).
• Unsupervised hierarchical clustering based on anxiety, depression, pain catastrophizing, and kinesiophobia.
• Comparison of sociodemographics, pain characteristics, lifestyle, quality of life, and treatment efficacy across clusters.

Main Results:

• Three distinct chronic pain patient clusters emerged based on psychological profiles.
• Cluster 1 (high psychological burden) exhibited significantly lower treatment efficacy (pain reduction <30%) and poorer quality of life.
• Clusters 2 and 3 (lower/intermediate psychological burden) demonstrated over 50% pain reduction with conventional treatments.
• A 15-item psychometric model accurately predicted cluster allocation.

Conclusions:

• Distinct chronic pain patient subgroups can be identified using psychological assessments, particularly anxiety, depression, pain catastrophizing, and kinesiophobia.
• One identified subgroup with a high psychological burden shows a markedly poorer response to standard pain treatments.
• A predictive model and potential web-based tool can aid clinicians in tailoring therapies to specific patient subgroups for improved outcomes.