{
  "abstract": "Background Early warning systems (EWS) used across the world typically assign a fixed number of points to patients receiving supplemental oxygen, regardless of amount. This ordinal binary approach may fail to recognise deteriorating patients who have an increasing oxygen requirement with otherwise stable observations. It is unclear whether weighting oxygen beyond binary scoring improves recognition of deterioration.Aims We aimed to describe all general adult EWS that grade oxygen beyond binary scoring (part 1). Where reported, we summarised the performance of graded oxygen EWS in comparison to binary scoring (part 2).Methods We systematically reviewed the literature, searching Embase, MEDLINE, CINAHL, Cochrane Central and Web of Science. We included studies of vital-sign-only EWS, for adult inpatients, which included grades of oxygen therapy above binary weighting (‘graded oxygen weighting’). We summarised methods of including graded oxygen therapy. We performed a random-effects meta-analysis of the effects of graded oxygen weighting inclusion in comparison to binary weighting. Risk of bias was assessed using the Prediction model Risk Of Bias ASsessment Tool.Results 15 studies reported the development of 16 EWS with graded oxygen weighting, classified by flow rate, delivery device and/or fraction of inspired oxygen. Four studies compared graded oxygen EWS to binary oxygen EWS. Meta-analysis showed a significant improvement in the performance of graded oxygen EWS over binary oxygen EWS (logit(AUROC)=0.19; 95% CI 0.094 to 0.285; p=0.002). 15/16 models were at high risk of bias.Conclusions 16 EWS with graded oxygen weighting were identified. Graded oxygen models had improved recognition of deterioration. Future work should explore the optimal method of oxygen classification and how this could be integrated into future EWS.PROSPERO registration number CRD42024443362.",
  "authors": [
    {
      "affiliations": [
        "Critical Care Research Group, Nuffield Department of Clinical Neurosciences, University of Oxford, Oxford, UK"
      ],
      "name": "Charlotte H Harrison"
    },
    {
      "affiliations": [
        "Critical Care Research Group, Nuffield Department of Clinical Neurosciences, University of Oxford, Oxford, UK"
      ],
      "name": "Phoebe Tupper"
    },
    {
      "affiliations": [
        "Centre for Statistics in Medicine, Nuffield Department of Orthopaedics, Rheumatology and Musculoskeletal Sciences, University of Oxford, Oxford, UK"
      ],
      "name": "Stephen Gerry"
    },
    {
      "affiliations": [
        "Critical Care Research Group, Nuffield Department of Clinical Neurosciences, University of Oxford, Oxford, UK"
      ],
      "name": "Verena Michael"
    },
    {
      "affiliations": [
        "Critical Care Research Group, Nuffield Department of Clinical Neurosciences, University of Oxford, Oxford, UK",
        "Department of Critical Care, Milton Keynes University Hospital NHS Foundation Trust, Milton Keynes, UK"
      ],
      "name": "Jonathan P Bedford"
    },
    {
      "affiliations": [
        "Bodleian Health Care Libraries, University of Oxford, Oxford, UK"
      ],
      "name": "Carolyn Smith"
    },
    {
      "affiliations": [
        "Ysbyty Gwynedd, Bangor, Gwynedd, UK",
        "School of Medical Sciences, Bangor University, Bangor, UK"
      ],
      "name": "Chris Subbe"
    },
    {
      "affiliations": [
        "Critical Care Research Group, Nuffield Department of Clinical Neurosciences, University of Oxford, Oxford, UK"
      ],
      "name": "Oliver Redfern"
    },
    {
      "affiliations": [
        "Critical Care Research Group, Nuffield Department of Clinical Neurosciences, University of Oxford, Oxford, UK",
        "Oxford Critical Care, Oxford University Hospitals NHS Foundation Trust, Oxford, UK",
        "Oxford Biomedical Research Centre, National Institute for Health and Care Research (NIHR), Oxford, Oxfordshire, UK"
      ],
      "name": "Peter J Watkinson"
    }
  ],
  "title": "Oxygen therapy in early warning scores: a systematic review and meta-analysis",
  "uid": "54d4abd6-f0c2-53a0-bfed-81399ecb994c"
}