- Segmentation Home >
- Part 2: Evaluation of the Bridges to Health Segmentation Model for Outcome Measurement
Publication Date: Nov 2016Download PDF
Part two takes a more detailed look at the ‘Bridges to Health’ model, and describes a methodology developed by OBH, for the practical implementation of this segmentation model, specifically for the purpose of outcomes measurement.
A. Evaluation of the Bridges to Health Model
Bridges to Health Summary
The ‘Bridges to Health’ model is fundamentally a person-focused segmentation approach, with the principal goal of ‘pursuing the health of each population segment’. This model does not continue to treat each individual health condition as a separate segment, in the way that some segmentation and clinical approaches do. In this respect, the ‘Bridges to Health’ model appears very well suited to whole-population outcomes based approaches. Being care setting agnostic, the model lends itself well to meaningful population outcomes measurement, reporting and incentivisation. There is emerging national and international evidence of use of the model, or close variants (such as that used by North West London, London Health Commission), for the purposes of population health management approaches.
‘Bridges to Health’ proposes ‘segmenting’ the entire population into eight core groups, shown in Table 1. The proposal of the eight groups was shaped around three considerations1:
- The set of population segments must be limited if the health care system is to offer a sensible array of integrated services for each segment.
- The set of population segments should include everyone; that is, at every point in their life, every person should fit into one of these categories.
- The people in each population segment must have sufficiently similar health care needs, but each segment must be different enough to justify separate consideration.
Throughout the paper, a variety of descriptive terms were used by the authors to describe each of the population groups, to help the reader better understand the eight populations, and to enable flexibility in implementation approaches depending on the purpose of use of the model.
Table 1: Bridges to Health Population Segments. Source: Using population segmentation to provide better health care for all: The “Bridges to Health” Model.
|Segment 2:||Maternal and Infant Health|
|Segment 3:||Acutely ill|
|Segment 4:||Chronic conditions, normal function|
|Segment 5:||Stable but serious disability|
|Segment 6:||Short period of decline before dying (mostly cancer)|
|Segment 7:||Limited reserve and exacerbations (organ failure)|
|Segment 8:||Frailty with or without dementia|
As outlined by Vuik et al (2016), there are important trade-offs between simplicity and precision, which apply to any segmentation model.2 The ‘Bridges to Health’ model strikes a good balance between the two, defining segments and movements between segments in a sufficiently simple, and clinically relevant manner. At the same time, it allows for precise definitions of each to be constructed for population-level outcomes based commissioning, and outcomes measurement purposes.
The key dimensions of the ‘Bridges to Health’ Model have been evaluated below:
Purpose:the model seeks to build a framework that can shape resource planning, care arrangement and service delivery at a ‘macro-’ level. Thus ensuring that each person’s health needs can be met effectively and efficiently.
Method: whole-population model, ensuring that every individual is accounted for and “assigned” to one (or more) segments at any point in time, while allowing for movement between segments. How this movement happens in light of precise segment definitions is discussed below.
Defining Variables:health prospects and priorities. Segments are divided according to four main goals for health – staying healthy, getting well, living with illness or disability, and coping with illness at the end of life – as well as eight distinct health priority concerns. These variables describe segmentation characteristics that are relatively stable over a person’s life course. More importantly, from an outcomes based perspective, they allow for the identification of similar health needs between each of the eight core groups. This enables a robust, whole-population outcomes framework to be established, monitored and incentivised.
In summary, the ‘Bridges to Health’ model:
- Remains sufficiently high-level to be usable/helpful for organising and planning clinical care, with sufficiently distinct segments to design care pathways around;
- Contains segments which are homogenous enough for the purpose of outcomes based commissioning, despite challenges in dealing with segment overlaps, and understanding how people flow between segments;
- Is sufficiently detailed to accommodate precise and meaningful outcomes measurement for sub-populations, subject to any data constraints; and
- Provides sufficient guidance to create granular enough views of financial information for specific sub-populations, in order to devise a capitated budget, including payments contingent on outcomes.
While there are a number of useful risk stratification models, incorporating a wide range of defining variables, the target variable is frequently risk of care activity and/or cost. This renders it generally unsuitable for person-centred outcomes measurement, on its own. The ‘Bridges to Health’ model (or very close variants) appear to be one of the most suitable, and increasingly widely adopted models internationally, for wholepopulation, outcomes based commissioning.3
Segmentation around population characteristics, rather than provider characteristics, can be challenging. Health systems have historically been organised largely around provider characteristics (for example, clinical specialties like cardiology), rather than population characteristics (for example, people living with frailty). This is increasingly unsustainable both financially and in terms of improving people’s outcomes. Care systems organised around people (rather than providers), offers potentially the only sustainable long term solution. Getting population segmentation right is key to this.