The development of Primary Care Networks (PCNs) is a key part of delivering the NHS Long Term Plan. With all general practices required to be in a network by June 2019, the vision is to build on the core of current primary care services and enable greater provision of proactive, personalised, and integrated health and social care.
To assist PCNs across the North Norfolk health economy achieve this vision, Arden & GEM’s advanced analytics team has developed a population health management (PHM) approach that identifies new models of care that will improve the health and wellbeing of the registered population.
This unique analytic approach incorporates data from over 470,000 local records including: mental health, acute and community data flows. By creating a high-level segmentation model, adapted from the British Columbia’s Health System Matrix specification, a detailed view of service utilisation across the care continuum has been delivered. This allows service redesign to focus on the needs of specific segments, prioritise the most effective interventions and therefore deliver the most efficient improvements in patient outcomes.
The recent Royal College of General Practitioners ‘Fit for the future’ vision emphasised the need for PCNs to explore how data can be used to inform decisions on the transformation of services and move towards a population health approach.
The formation of PCNs in North Norfolk has created an opportunity to reshape services in order to improve integrated working, generate efficiencies and improve the health of the local population. As a recently formed collaboration, a PHM approach was created by applying econometric modelling to segmented data.
This revealed opportunities to improve patient health by delivering proactive care as well as highlighting system efficiencies.
Transitioning to a PHM model has universal challenges ranging from accessing and sharing data to supporting the cultural acceleration needed to deliver system change. At a local level, our challenge was to identify areas of variation that could directly translate into practical improvement projects and create a robust case for change.
Drawing upon in-house health analysts, econometricians and mathematicians, our business intelligence team wanted to build a model that would highlight the best opportunities for North Norfolk PCNs.
A key priority was also to create a segmentation and profiling approach that was replicable in different localities with minimal adaptation.
Adapted from the British Columbia’s Health System Matrix specification, and sharing similarities with the Bridges to Health model, a High-Level Segmentation Model (HLSM) was created.
This was driven by linked datasets from acute, mental health and inpatient community services and provided detailed analysis in four key areas:
- Health system matrix modelling created 14 segments of similar needs with an underlying health record for each patient
- Unwarranted variation was assessed, by theme, accounting for demography, comorbidity profile and expenditure
- Economic modelling generated insights at PCN, CCG and system level, this included calculation of the probability of outcomes for comprehensive options appraisal
- Marginal analysis examined the additional benefits compared to the additional costs to further enhance decision making.
Clinical validation was also completed to review the diagnostic and procedure codes within the model. This work reviewed the Long Term Conditions (LTC) complexity matrix and checked the allocations against real life clinical knowledge. This validation resulted in refining of the clinical aspect of the model in areas including patients living with IBS and Crohn's disease.
Completing this modelling has created a detailed value proposition for North Norfolk PCNs which will help them to understand the most effective interventions, determine the value that will be created from new models of care and provide a robust case for change. This has proved invaluable insight to the PCNs as they did not have the in-house expertise required to build an impactability matrix to identify opportunity areas.
Assessing variation by theme has also highlighted projects where primary care can make the greatest impact. For example, a case for change to proactively target preventative care for individuals with LTCs who have not accessed mental health support but are at risk of developing co-morbid issues. The model demonstrates that improving patient’s resilience and wellbeing in the community could reduce their associated acute costs by up to 42%.
Arden & GEM regularly presents findings and recommendations to the local population health steering group. We will continue to support with the implementation and evaluation of new care models throughout the PCNs transformational journey.
As an iterative model, the more data that gets fed into the analysis the more accurate it becomes, so we continue to work collaboratively with the PCN to analyse findings and refine recommended actions. We are also working across the patch to install a robust information governance process that will enhance access to primary care data.
The model is flexible and can be adapted to fit any population, with work already underway to ensure it is utilised across wider geographies and health economies to support their individual PHM approach.
We see Population Health Management as a methodology rather than a product. Working closely with colleagues at Arden & GEM has given us some really important intelligence from the various sources of data. We asked the CSU to highlight an area to focus on to help improve health outcomes and through their modelling they have given us the first of our projects to look at.
I’m hopeful that this approach will enable us to encourage clinicians to start asking questions, knowing we’ll be able to work with our CSU colleagues to provide the information required to form the answers.
Wayne Bolt, Practice Manager and Director of North Norfolk PCN