Segmenting Dorset’s Population for Better Health Planning

6044 Population Segmentation GP Registered Population Dorset

Rhianna Everett, Senior Intelligence Analyst, Dorset Intelligence & Insight Service

​

James Roberts, Principal Research Officer,

Dorset Council

The challenge

Why segment the population?

• Health systems face rising demand and limited resources
• People have very different and complex health needs

​

Segmenting the population helps:

• Target interventions more effectively
• Use resources more efficiently
• Plan for future health needs

Letting the Data Speak: Our Clustering Approach

What we did:

• Combined health records with demographic data
• Used k-means clustering, an unsupervised machine learning method
• Grouped people based on similar health characteristics
• Found 7 distinct population segments

Source: Brilliant.org

What we found

Insights from the Clusters:

• Identified 7 distinct clusters within Dorset’s GP-registered population
• Clusters reflect diverse health profiles
• Generally younger clusters were healthier, but…
• One younger cluster showed higher healthcare activity
• Others were older but relatively low-need
• Highlights the complexity of real-world health needs

From Insight to Action: Why This Work Matters

• Helps us to better understand the population
• Supports population forecasting — planning for future health needs
• Enables targeted health interventions — delivering better outcomes for patients
• Next steps:
• Refine clusters with more data
• Integrate with forecasting models

​

Data clustering insights actions