{
  "abstract": "Introduction Manual scoring although time-consuming is the current gold standard when reporting sleep studies. Most sleep analysis software offers auto-analysis programmes and are often marketed as accurate and reliable.Aim The aim of this audit was the evaluate the agreement between automatic analysis and manual scoring in identifying patients with suspected Obstructive Sleep Apnoea (OSA) and its severity and the reliability of automatic analysis for accurate clinical outcomes.Methods A retrospective service audit was performed on sleep study data acquired between March 2024 to September 2024. 44 respiratory polygraphy studies were reviewed comparing automatic and manually scored Apnoea-Hypopnea Index (AHI). A regression analysis and Bland-Altman analysis was carried out to investigate the differences and correlation between the data and considering clinical outcomes into significance.Results 19/44 results (43%) changed scoring severity classification, more so at AHI >5 events/hour and AHI >15 events/hour. Bland-Altman analysis shows that there is a wide variation with a mean difference of +2.578 and the 95% limits (-7.507 and +12.66) ( figure 1: Bland-Altman plot of mean difference and the 95% limits. The red dots result in change in severity category and the purple dots remained in the same severity.). 12/44 patients (27%) would potentially be in a different treatment category based on the automatic scoring analysis (figure 2).Abstract P33 Figure 1Bland-Altman plot of mean difference and the 95% limits. The red dots result in change in severity category and the purple dots remained in the same severityAbstract P33 Figure 2A linear regression analysis of automatic analysis (y-axis) vs. the gold standard manual scoring (x-axis). There are clear differences between the data as they do not all fit on the line of identity (orange dashed line)Conclusion The use of automatic analysis using our current sleep software Stowood (Visi-Download) for respiratory polygraphy studies is limited. It can provide a useful guide on predicting the severity of the AHI however is not reliable in providing accurate diagnosis for patients with suspected Obstructive Sleep Apnoea (OSA).",
  "authors": [
    {
      "affiliations": [
        "Great Western Hospital, Chippenham, UK"
      ],
      "name": "Janina Mallari-Saguyan"
    },
    {
      "affiliations": [
        "Great Western Hospital, Chippenham, UK"
      ],
      "name": "James Pratt"
    },
    {
      "affiliations": [
        "Great Western Hospital, Chippenham, UK"
      ],
      "name": "Jessica O’Keeffe"
    },
    {
      "affiliations": [
        "University of West UK"
      ],
      "name": "Adrian Kendrick"
    }
  ],
  "title": "P33 Automatic analysis vs. manual scoring: comparing the two methods of scoring respiratory polygraphy sleep studies using the visi-download stowood sleep software",
  "uid": "e3a44ce8-1b64-5053-a406-d20e180ea5c2"
}