Population Health Risk Stratification Solutions

Within a fully integrated health and social care system, Graphnet occupies a unique position to support organisations to undertake risk stratification, impactibility monitoring and utilise real time intelligence to support patient care.

Benefits of risk stratification tools in healthcare

Our solutions bring together a large variety of information into a centralised location enable effective and automatic modelling. Benefits include

From risk stratification to prevention: Our full end-to-end solution enables the identification, classification and prevention of care deficits. Risk stratification engines can be integrated to enable the early detection of patients at risk or those who have negative health trajectories. Automated ‘what-if’ models can give clinicians lists of preventive and reactive models driven by data and evidence.
Cohort mapping engine build localised registries: The cohort mapping engine delivers standardised diseases and wellness registries. The engine also gives users the ability to include their own mappings which enables localisation.
Frailty: The electronic frailty index (eFI) delivers the ability to stratify the entire elderly population. Frailty deficits are calculated and a full history for each patient can be analysed. Graphnet is working with a leading university on the development of advanced artificial intelligence to predict the future frailty state for a population.
Data and model APIs: Our APIs allow the integration of our business intelligence and data science packages with third party tools, and vice versa. Following the REST API structure means that our APIs enable easy to use industry standard integration with third party applications.

Risk stratification case study: Diagnosing diabetes/pre-diabetes for earlier intervention and proactive care

The Diabetes Diagnosis dashboard uses the shared care record to identify patients whose latest HbA1c and/or blood glucose results indicate that they may have diabetes, prediabetes or are at emerging risk but are not coded as being on the QOF diabetic register.

This dashboard is part of a suite of dashboards which identify coding quality/possible missed diagnosis, including hypertensive patients with high BMI/blood pressure and also COPD exacerbation admissions without a primary care diagnosis.

Benefits to care professionals and patients include:

• Identifying ‘missing diabetics’ those who have not been formally diagnosed or recorded as diabetic

• Tracking the progress of individual care plans against national targets

• Identifying and following up on patients who may have disengaged

• Planning interventions to improve patients’ outcomes

• Tracking performance at GP, community and population level and identifying if additional resources are required to manage caseloads

I had a 45 year old male of BAME background with underlying diabetes whose son developed Covid and, a few days later, my patient developed symptoms and tested positive as well...

We enrolled him in the pule oximetry programme and after day 11 his symptoms suddenly deteriorated from 95% saturations to 80% saturations. He was taken straight into hospital and he had a Covid-related pulmonary embolism.

Without the pulse oximetry programme, we wouldn’t have known he had deteriorated and this has saved the patient’s life.

Dr Priya Kumar, Connected Care, Frimley ICS and Berkshire West ICP

Risk Stratification case study: Covid Oximetry @Home / Virtual Ward

Using the Enhanced Case Finding tool, our integrated care partners can interactively filter patient lists to identify and reach out to those patients most at risk from the complications from Covid, ensuring care is targeted to where it is of greatest benefit.

Under the Covid Oximetry @Home / Virtual Ward programme, patients can record their pulse oximetry readings and symptoms electronically using our Personal Health Record (PHR) app. Care professionals (e.g. GPs, acute clinicians, Out of Hours) across integrated care systems have access to the latest patient recordings through our shared record. This means that should their patient’s condition suddenly deteriorate; they can escalate the level of care. Being able to identify and recruit patients using the information routinely held in the shared care record allows for a more targeted and fuller response to a patient’s individual wishes, needs and risks.

This same model could be applied to support many other care pathways such as higher-risk pregnancies, frailty or diabetes.