{
  "abstract": "Background Type 2 diabetes mellitus (T2DM) is estimated to be among the top causes of disability and years of life lost and mortality worldwide, with a high burden estimated for low- and middle-income countries. Early diagnosis and management is beneficial, but it is not clear how high-risk groups for screening should be identified in low-resource settings. This review assesses performance of population-based risk prediction scores to identify T2DM risk in adults in low- and middle-income countries.Methods A systematic search was conducted in PubMed, OVID Embase, and Cochrane Library in June 2024. After the systematic search, additional hand searches of reference lists and Google Scholar were performed. No restrictions were used for publication dates and languages. We included studies that developed or validated T2DM risk prediction scores applicable in general populations and based on adults without known T2DM from low—and middle-income countries. Results from the systematic search were combined and duplicates were removed. Two reviewers independently screened search results using Covidence according to predetermined inclusion and exclusion criteria with a third reviewer used in case of disagreement. Data extraction and risk of bias assessment were based on critical appraisal and data extraction for systematic reviews of prediction modelling studies (CHARMS) and prediction model risk of bias assessment tool (PROBAST), respectively. This systematic review is registered with PROSPERO (CRD42024584501).Results We identified 7979 articles. After removing duplicates and retracted articles, we screened 6787 articles based on title and abstracts, followed by a full-text assessment for eligibility of 87 articles. A total of 46 articles were included in the review. There were 48 diagnostic and 23 prognostic population-based T2DM risk prediction scores developed and/or validated in 18 countries with variable performance. In some populations, Finnish and Australian (FINDRISC and AUSDRISK) models performed well for T2DM diagnosis (area under receiver operating characteristic curve > 0.80), however, none of the prognostic scores performed well. Only two articles were considered to have a low risk of bias: one which validated FINDRISC and its modifications for diagnosis and one which validated the FINDRISC, AUDRISK, and American Diabetes Association (ADA) risk scores for diagnosis and for 5-year prognosis.Conclusion Some T2DM risk prediction scores have demonstrated broad applicability across diverse populations. Further robust validation of population-based T2DM risk prediction scores in low- and middle-income countries are needed.",
  "authors": [
    {
      "affiliations": [
        "Usher Institute, University of Edinburgh, Edinburgh, UK",
        "Faculty of Public Health, Universitas Indonesia, Depok, Indonesia"
      ],
      "name": "Fadila Wirawan"
    },
    {
      "affiliations": [
        "Usher Institute, University of Edinburgh, Edinburgh, UK",
        "Department of Emergency Medical Services, King Saud Bin Abdulaziz University for Health Sciences, Jeddah, Saudi Arabia"
      ],
      "name": "Areej Almutairi"
    },
    {
      "affiliations": [
        "Usher Institute, University of Edinburgh, Edinburgh, UK"
      ],
      "name": "Regina Prigge"
    },
    {
      "affiliations": [
        "Usher Institute, University of Edinburgh, Edinburgh, UK"
      ],
      "name": "Sarah Wild"
    }
  ],
  "title": "P28 Population-based risk prediction scores for type 2 diabetes mellitus in low- and middle-income countries: a systematic review",
  "uid": "a6ed7de5-5a2e-5464-aa63-3a91596da8f8"
}