Dr David Lowe is a Consultant in Emergency Medicine and Co Director of EmQuire the QEUH ED Research Group. He leads a portfolio of projects centred on unscheduled care and specifically data, devices and decisions. Collaborations include Glasgow, Strathclyde University, Glasgow School of Art, private and third sector, seeking to create innovative solutions that can create impactful and sustainable change. His research focuses on decision-making within acute care and the interface between data, the clinician and the patient. He is co-developing novel app based solutions in range of conditions including trauma, COPD and chest pain to support and enhance clinical decision-making and operational efficiency. He is PI on innovatateUK grants to develop digital service model for COPD that leverages machine learning, wearables, and patient report outcomes to risk stratify patients in the community surfacing those at risk of exacerbation, treatment failure and need for home NIV. The other grant focus on developing a clinical decision support app for major trauma within the emergency department enabling granular data collection and promoting team situational awareness with the aim to align team members and optimise resuscitation care. He is a co-applicant on the CAIPE bid to create a platform for digital imaging AI leading on an exemplar using CXR.
13:00 - 14:00
- Appar / Innovativ/forskning / Personcentrering, samverkan över organisationsgränser
- Beslutstöd sepsis/diab/kol
- Patientorganisationer/Brukarorganisationer / Vårdpersonal
Optimizing COPD patient care includes anticipating and preventing exacerbations of COPD for patients in the community in an effort to proactively mitigate their need for acute care services and resultant hospital admission. Doing so enables the ambulatory management of COPD patients in their community, strengthening ties with their primary care providers and empowering individuals with COPD to remain in their homes. Such an approach also promotes “equal care" as all COPD patients, regardless of income, geography, or other factors can benefit from this service model. We present our solution of proactive community management of patients with a long-term condition with a nested clinical trial designed to demonstrate clinical and patient outcomes.
The core of this solution is aggregation and analysis of data from multiple disparate sources:
• Non-Invasive ventilation
• Patient Reported outcomes
• Electronic Health Record (EHR)
Combining these data sources provides the opportunity to develop machine-learning based algorithms to stratify, predict and highlight patient-level risk. The aim of the project is to develop and validate a continuous preventative service model in the largest health board in Europe, NHS Greater Glasgow & Clyde (GGC). Greater Glasgow & Clyde will integrate digital health into COPD care from the hospital to the community, leveraging risk prediction and patient-generated health data to support proactive, participatory and preventative models of care.
The innovative project co-developed within a UK government-funded framework, will allow NHS GGC to deploy a holistic digital health solution for patient engagement, enhanced connectivity and improved integration of data across a range of chronic conditions and scale beyond Scotland. Clinical teams will be provided with real-time clinical decision support to remotely manage and optimise care for this high utilization group. A patient facing application will enable patient-reported outcomes to be collected and synthesised with physiology from wearable and connected non-invasive ventilation devices to surface ML-derived risk prediction scores. A clinical cockpit with embedded decision support and asynchronous messaging facility will enable respiratory services based on data-driven insights optimising care delivery and initiating anticipatory care planning.
The nested research project will target two patient cohorts at highest risk of further COPD exacerbations:
• 100 remote-managed home NIV patients with severe COPD
• 300 patients presenting to hospital with an acute exacerbation of COPD
All enrolled patients will have a wearable, only the 100 remote-managed patients will have NIV.
Novel features of this work include:
• Exploiting the patient reported and continuous physiology dataset acquired (via project connected hardware and patient wearables), combined with available historical cohort data to train machine-learning algorithms to provide risk predictive scores for COPD outcomes.
• Integrating and presenting these risk-predictive scores on a clinician-facing platform to provide the insights necessary to inform clinical decision making and individualized quality care (a major step towards personalized medicine).
• Providing outputs including COPD admission reduction and service blueprint to demonstrate the utility and value of the collaborators' innovations.
During the presentation the team will outline:
• Prior work between the partners in building and evaluating machine learning models for COPD patients in the acute care setting (A&E admission, length of inpatient stay).
• The enablers for successful integration of ML insights into service delivery,
• The process of creating a patient facing application that promotes patient engagement,
• The value of physiology data aggregated from wearables and home ventilation systems to risk prediction and ML models, and
• Model for solution scaling and the potential impact of this work to other chronic condition patient cohorts.