New Technology But Old Fashioned Nursing: The Key Findings in the EV-WUDS Study

The Expansion of Virtual Wards Using Data Science (EV-WUDS) project was awarded a grant by UEA Health and Social Care Partners in 2022 to establish Trusted Research Environments (TRE) computing environment which multiple partners can use to perform data science activities and methods; to prove the value of joined up data sets across multiple organisations; to demonstrate exemplar data science approaches to virtual ward evaluation; to validate predictive models as they relate to virtual ward activity.

The project included aspects of research data science applied to virtual wards and its patients, as well as an innovation aspect due to interest from industry partners such as small to medium enterprises (SMEs) supplying the wearable technology used, and large industry partners who are interested in developing virtual ward data products and remote sensing technology.

The project had seven phases of work, which they are happy to share their progress on.

Phase 1: Establishing data infrastructure

TRE computing environment to be established, cleared for use from IG, data ingest, and user access facilitated.

TRE computing environment is live with ability to add users, secure data usage, dual factor authentication, and cleared for use from Norfolk and Norwich University Hospitals NHS Foundation Trust (NNUH) IG.

Phase 2: Data sources and information governance approvals

Data approvals for use and linkage via IG colleagues.

NNUH ++ data sources were sourced and approved (community, primary care, public health, and social care).

Phase 3: Data acquisition and data recovery

Digital health, ICT, analyst support at relevant organisations to provide pseudonymised data for the TRE.

Data from various sources were collated and pseudonymised. These data include patient flow, comorbidities, administrative data, cost data, economic data sets.

Phase 4: Analysis, hypothesis development

Statistics, data science, epidemiology expertise applied to the analysis and outputs.

Initial discovery and hypothesis development was completed on patient flow, comorbidity, cost and length of stay of case and control groups of patients from the acute setting.

Phase 5: Data science model design

Data science skills applied to model design, data engineering, feature selection.

Data science approaches and models were selected, data engineering took place, feature selection on all of the data sets.

Phase 6: Model training, testing, validation

Feature optimisation, model testing and validation.

Model training, testing and validation took place, followed by refinement and validation.

Phase 7: Reporting results

All domains to support analysis and reporting of results.

Reporting of results took place at the UEA HSCP Citizens Academy, Norfolk and Waveney ICS, NHS England National Virtual Ward Working Group and NNUH Digital Health.

Summary

The initial objectives of the work were successfully achieved and the results of the work contributed to the successful business case by Norfolk and Waveney to extend the Virtual Ward across the ICS and it’s partners.

NHS England commented the Norfolk and Waveney Virtual Ward Final Report was the best in the country.

At the time of writing, final stages of preparation for a high-impact journal manuscript is in the final stages of production.

Mike Shemko and the EV-WUDS team also presented their findings at our Collaborate to Innovate Conference in 2023.