Scholarlink

Background and Objective <p>The reliability of a decision model to guide decision making depends on its ability to accurately predict patient outcomes. We present results of an external validation of the MicroSimulation Core Obesity Model (MS-COM) that was developed to compare the cost effectiveness of obesity management interventions in adults.</p> Methods <p>We updated a 2018 systematic literature review of economic models in overweight and obesity and conducted additional targeted searches to identify suitable sources and outcomes to validate against MS-COM in people with overweight or obesity with or without type 2 diabetes. We extracted baseline characteristics and cardiovascular and mortality outcomes, where these were closely matched with MS-COM, and incidence of type 2 diabetes. We performed external-dependent (sources used in MS-COM) and external-independent (sources not used in MS-COM) validation. The extent of concordance between predicted and observed outcomes was assessed using the coefficient of determination (<i>R</i><sup>2</sup>), ordinary least-squares linear regression line (OLS LRL), mean absolute percentage error, root mean square percentage error and mean squared log of accuracy ratio.</p> Results <p>Ninety-nine potential independent validation sources were identified from 6381 screened records, of which nine studies reported cardiovascular and mortality outcomes that were closely matched with MS-COM, along with two studies that reported type 2 diabetes incidence (number of endpoints&#xa0;=&#xa0;106). The dependent validation of cardiovascular and mortality outcomes (<i>N</i>&#xa0;=&#xa0;18), based on the QRisk3 risk equation (normoglycaemia/prediabetes population) and UKPDS 82 (type 2 diabetes population), showed a good linear correlation with observed outcomes (<i>R</i><sup>2</sup>&#xa0;=&#xa0;0.99 and 0.98, respectively). There was some slight overprediction of QRisk3 (OLS LRL slope&#xa0;=&#xa0;1.11) and underprediction of UKPDS 82 (OLS LRL slope&#xa0;=&#xa0;0.97). The independent validation of cardiovascular and mortality outcomes also showed a good linear correlation with observed outcomes, particularly in adults with normoglycaemia/prediabetes (<i>R</i><sup>2</sup>&#xa0;=&#xa0;0.90; OLS LRL slope&#xa0;=&#xa0;0.86); however, an independent validation of type 2 diabetes incidence showed a poorer fit with some degree of underprediction (<i>R</i><sup>2</sup>&#xa0;=&#xa0;0.74; OLS LRL slope&#xa0;=&#xa0;0.66). Mean error estimates were lower in the dependent validation, showing good concordance between predicted and observed values.</p> Conclusions <p>External validation of MS-COM showed good concordance with dependent and independent sources, suggesting the model accurately predicts obesity-related complications in an overweight/obese population with normoglycaemia/prediabetes and type 2 diabetes.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

External Validation of the MicroSimulation Core Obesity Model (MS-COM) to Predict Cardiovascular Outcomes, Mortality and Type 2 Diabetes Mellitus Incidence and Assess Cost Effectiveness

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年11月6日

查看全文
  • Christopher G. Fawsitt,
  • Howard Thom,
  • David Aceituno,
  • Alexander Jarde,
  • Sara Larsen,
  • Christopher Lübker,
  • Edward Kayongo,
  • Edna Keeney,
  • Volker Foos

摘要

Background and Objective

The reliability of a decision model to guide decision making depends on its ability to accurately predict patient outcomes. We present results of an external validation of the MicroSimulation Core Obesity Model (MS-COM) that was developed to compare the cost effectiveness of obesity management interventions in adults.

Methods

We updated a 2018 systematic literature review of economic models in overweight and obesity and conducted additional targeted searches to identify suitable sources and outcomes to validate against MS-COM in people with overweight or obesity with or without type 2 diabetes. We extracted baseline characteristics and cardiovascular and mortality outcomes, where these were closely matched with MS-COM, and incidence of type 2 diabetes. We performed external-dependent (sources used in MS-COM) and external-independent (sources not used in MS-COM) validation. The extent of concordance between predicted and observed outcomes was assessed using the coefficient of determination (R2), ordinary least-squares linear regression line (OLS LRL), mean absolute percentage error, root mean square percentage error and mean squared log of accuracy ratio.

Results

Ninety-nine potential independent validation sources were identified from 6381 screened records, of which nine studies reported cardiovascular and mortality outcomes that were closely matched with MS-COM, along with two studies that reported type 2 diabetes incidence (number of endpoints = 106). The dependent validation of cardiovascular and mortality outcomes (N = 18), based on the QRisk3 risk equation (normoglycaemia/prediabetes population) and UKPDS 82 (type 2 diabetes population), showed a good linear correlation with observed outcomes (R2 = 0.99 and 0.98, respectively). There was some slight overprediction of QRisk3 (OLS LRL slope = 1.11) and underprediction of UKPDS 82 (OLS LRL slope = 0.97). The independent validation of cardiovascular and mortality outcomes also showed a good linear correlation with observed outcomes, particularly in adults with normoglycaemia/prediabetes (R2 = 0.90; OLS LRL slope = 0.86); however, an independent validation of type 2 diabetes incidence showed a poorer fit with some degree of underprediction (R2 = 0.74; OLS LRL slope = 0.66). Mean error estimates were lower in the dependent validation, showing good concordance between predicted and observed values.

Conclusions

External validation of MS-COM showed good concordance with dependent and independent sources, suggesting the model accurately predicts obesity-related complications in an overweight/obese population with normoglycaemia/prediabetes and type 2 diabetes.