{
  "abstract": "Aim This paper demonstrates how temporal graph analysis, combined with clustering techniques, can provide simple and insightful visualisations. Applied to routinely collected Electronic Patient Record (EPR) data, these visualisations can highlight potential resourcing issues within a hospital ward, a task that is inherently complex and challenging.Method This study focused on patients admitted to Bear Ward in June 2024. Bear Ward specialises in cardiac care, treating patients at various stages of their treatment at GOSH. Temporal graphs were generated for patients, healthcare professionals (HCPs), and the ward. The periods for these graphs varied: 24 hours for patients and HCPs, and eight hours for the ward. To define co-occurrence, a time window of 60 minutes was used for patients and HCPs, while a 15-minute window was applied to the ward. Clustering was performed using the k-means algorithm, with five clusters determined by a silhouette coefficient analysis.Results In June 2024, 101 patients attended Bear Ward. Care was provided by 803 HCPs, who performed over 500,000 patient activities, there were significant differences in patterns of HCP activity across days and between patients. A total of 10,963 temporal graphs were produced. Individual patient progress on the ward was depicted by the intensity of care, visualised with a colour scale from green (least intense) to red (highest intensity). Individual HCP activity was categorized by professional group (Nursing, Medical, Allied Health Professional, and Othe). Ward visualisation divided days into 8-hour periods, comparing HCP, action, interaction, and patient activities.Conclusion This paper has investigated the categorisation of patient care intensity through the clustering of temporal graph metrics. By expanding this analysis to include not only patients but also HCPs and wards as graph entities, the findings offer a comprehensive visualisation of ward resourcing. This extended analysis provides a deeper understanding of the patient care environment.Acknowledgements for Funding or Support John Booth is Funded by CIRP via GOSH Children’s Charity Award VS0618. The study and researchers were part funded through the NIHR GOSH Biomedical Research Centre.",
  "authors": [
    {
      "affiliations": [
        "DRIVE, Great Ormond Street Hospital for Children NHS Foundation Trust, UK"
      ],
      "name": "John Booth"
    },
    {
      "affiliations": [
        "DRIVE, Great Ormond Street Hospital for Children NHS Foundation Trust, UK"
      ],
      "name": "William Bryant"
    },
    {
      "affiliations": [
        "UCL Institute of Health Informatics, UK"
      ],
      "name": "Spiros Denaxas"
    },
    {
      "affiliations": [
        "UCL GOS ICH, UK"
      ],
      "name": "Maria Eriksson"
    },
    {
      "affiliations": [
        "NIHR Great Ormond Street Hospital Biomedical Research Centre, Great Ormond Street Hospital for Children NHS Foundation Trust, UK"
      ],
      "name": "Stephen Marks"
    },
    {
      "affiliations": [
        "UCL GOS ICH, UK"
      ],
      "name": "Rebecca Pope"
    },
    {
      "affiliations": [
        "UCL GOS ICH, UK"
      ],
      "name": "Joshua Spear"
    },
    {
      "affiliations": [
        "Great Ormond Street Hospital for Children NHS Foundation Trust, UK"
      ],
      "name": "Neil Sebire"
    }
  ],
  "title": "82 A hospital ward activity analysis using routinely collected Electronic Patient Record data and temporal graph analytics. A cohort study",
  "uid": "57b10778-0284-5900-854c-ac572209809e"
}