We use cookies.

To make your experience the best it can be, we use cookies and similar technologies on our site. We need your permission to allow these technologies, which will maximise browsing experience. For more information on how we use cookies and how to change your cookie settings, please see our cookies and privacy policy.



Please complete this short form to get in touch with a member of our team and we will get back to you as soon as we can.

Header image for the current page Generating actionable insights from a bespoke segmentation model

Generating actionable insights from a bespoke segmentation model

Share this page

When NHS BLMK CCG wanted to better understand the healthcare needs and consumption of the local population, Arden & GEM’s Advanced Analytics Unit developed SMITH, a bespoke segmentation model, to enable person-level analysis which has informed service planning and commissioning.


Understanding the healthcare needs and priorities of the local population is fundamental to service planning within health and care systems.

With a patient population of almost 1 million, NHS Bedfordshire, Luton and Milton Keynes (BLMK) CCG wanted to better understand healthcare consumption, outcomes and inequalities, and how these compared across the CCG’s geography.

Arden & GEM’s Business Intelligence service has been providing data and analytics support to BLMK CCG since June 2021, including generating actionable insight through developing a bespoke analytics platform and adapting proven tools.


The Advanced Analytics Unit, part of our Business Intelligence service, is experienced in designing and using health segmentation approaches, including developing an award-winning in-house high level segmentation model.

The team built upon this expertise and experience to develop the Segmentation Model Individualised Towards Health (SMITH) for BLMK. SMITH is a needs-based model which divides the population into 12 groups based on common characteristics, health status and priorities.

The model, which brings together secondary and primary care data, is informed by medical literature and endorsed by clinicians.

What we did

Working closely with the Information Governance team, we supported the CCG to flow and link their primary care data into the SMITH dataset.

After a calibration process, to check data quality and assumptions, all patients were then assigned to one of the 12 segments. In order to assign people to a segment identifying the correct level of need – high, medium or low – within the model, a clinically validated long term condition matrix was adapted from the British Columbia Clinical Health Matrix. By incorporating individual conditions and interventions alongside comorbidities, needs could be classified based on their impact on quality of life and the severity of treatment.

Analytical outputs

We worked alongside the CCG’s in-house teams responsible for Population Health Management and Public Health to embed the model and enable ownership of the data and insight. Weekly meetings with the client identified a number of target areas for analysis, using SMITH, to highlight opportunities for intervention.

Expenditure profiles
Our health economists were able to show acute spend per capita in each segment for Luton, with Pareto chart visualisations, before making comparisons to the CCG as a whole. Econometric modelling was used to adjust for characteristics such as age and sex to generate evidence statements.

Health equity analysis
By calculating the Slope Index of Inequality, our analysts were able to rigorously measure the relationship between higher levels of deprivation and increased likelihood of diabetes. The same methodology can now be used for different outcomes data such as alcohol related admissions or self-harm, moving from descriptive to prescriptive analysis.

Learning disabilities and mental health
Policy direction was informed by an analysis of mental health need and learning disability (LD) prevalence with a focus on acute spend, sociodemographic characteristics and disease burden. This analysis also provided insights into the impact of mental health and LD on life expectancy.


"BLMK CCG has started to use the SMITH population segmentation model and is interested to further develop the model to understand the health and wellbeing of our population, for population health management planning, for building an ICS system model for the total cost of care, and also potentially in support of outcomes reporting and incentives. The CCG has already made progress by using the model to understand the health and wellbeing of the frail elderly and complex multimorbidity segments in Central Bedfordshire, and also in support of population profiles in BLMK’s four places – Bedford Borough, Central Bedfordshire, Luton and Milton Keynes local authority areas."

Charles Wheatcroft, PHM Programme Director at NHS BLMK CCG

Find out more about our Population Health Management solutions here.