Development and validation of a prediction model for airflow obstruction in older Chinese: Guangzhou Biobank Cohort Study

Jing Pan,Peymane Adab,K.K. Cheng,Chao Qiang Jiang,Wei Sen Zhang,Feng Zhu,Ya Li Jin,G. Neil Thomas,Ewout W. Steyerberg,Tai Hing Lam
DOI: https://doi.org/10.1016/j.rmed.2020.106158
IF: 4.3
2020-11-01
Respiratory Medicine
Abstract:<h3 class="u-h4 u-margin-m-top u-margin-xs-bottom">Objective</h3><p>To develop and validate a prediction model for airflow obstruction (AO) in older Chinese.</p><p>Methods.</p><h3 class="u-h4 u-margin-m-top u-margin-xs-bottom">Design</h3><p>Multivariable logistic regression analysis in large population cohort of Chinese aged ≥50 years.</p><h3 class="u-h4 u-margin-m-top u-margin-xs-bottom">Participants</h3><p>Model development: 8762 Chinese aged ≥50 years were selected from the early phase recruits to the Guangzhou Biobank Cohort Study (GBCS) (recruited from September 2003 to May 2006). Internal validation: 100 bootstrap samples drawn with replacement from the development sample. External validation: 8395 Chinese aged ≥50 years from later phase GBCS (recruited from September 2006 to January 2008).</p><h3 class="u-h4 u-margin-m-top u-margin-xs-bottom">Outcomes</h3><p>AO was defined by a forced expiratory volume in 1 s/forced vital capacity ratio &lt; lower limits of normal.</p><h3 class="u-h4 u-margin-m-top u-margin-xs-bottom">Results</h3><p>839 (9.6%) and 764 (9.1%) individuals had AO in the development and temporal validation samples respectively. The predictors in the prediction model included sex, age, body mass index groups, smoking status, presence of respiratory symptoms, and history of asthma. Model development and validation was stratified by sex. Model performance including calibration (calibration-in-the-large −0.017 vs. −0.157; and calibration slope 0.88 vs. 1.02), discrimination (C-statistic 0.72 vs. 0.63 with 95% confidence interval 0.69–0.75 vs. 0.62–0.73) and clinical usefulness (decision curve analysis) in the external temporal validation sample were more satisfactory in men than that in women. Prediction models with risk thresholds (13% in men and 7% in women) and easy-to-use nomograms were developed to assess the probability of AO.</p><h3 class="u-h4 u-margin-m-top u-margin-xs-bottom">Conclusion</h3><p>The diagnostic models based on readily available epidemiologic and clinical information with satisfactory performance can assist physicians to identify older individuals at high risk of AO and may improve the efficiency of spirometry for active case finding. Further validation beyond the Chinese population is warranted.</p>
cardiac & cardiovascular systems,respiratory system
What problem does this paper attempt to address?
The problem that this paper attempts to solve is to develop and validate a diagnostic model for predicting airflow obstruction (AO) in the Chinese elderly population. Specifically, the study aims to use epidemiological and clinical information from a large - scale population cohort to establish an easy - to - use prediction model, which can help doctors identify high - risk individuals and improve the efficiency of actively detecting undiagnosed chronic obstructive pulmonary disease (COPD) through spirometry. ### Research Background Chronic obstructive pulmonary disease (COPD) is a worldwide public health problem characterized by airflow obstruction, and most cases are undiagnosed. Undiagnosed COPD patients have a higher risk of acute exacerbation, pneumonia and death. In China, according to the data of the China Pulmonary Health (CPH), from 2012 to 2015, the prevalence of COPD (i.e., airflow obstruction) defined by spirometry among 57,779 Chinese people was 8.6%, but only 2.6% of them were aware of their respiratory conditions, which was much lower than that in Western countries (11% - 54%). Therefore, screening for undiagnosed COPD patients has become particularly important. ### Research Objectives 1. **Develop the model**: Based on the large - scale data of the Chinese elderly population in the Guangzhou Biobank Cohort Study (GBCS), develop a diagnostic model for predicting the risk of airflow obstruction. 2. **Validate the model**: Evaluate the performance of the model in internal and external validation samples, including calibration, discrimination and clinical utility. 3. **Provide a practical tool**: Develop an easy - to - use nomogram for practical application. ### Main Results - **Model performance**: - In men, the discrimination (AUC) of the model was 0.72 and the calibration was good. - In women, the discrimination of the model was lower (AUC was 0.63), but it still had certain clinical utility. - **Risk threshold**: - The high - risk threshold for men was 13%. - The high - risk threshold for women was 7%. - **Clinical utility**: - Decision - curve analysis showed that using this prediction model was more effective in detecting airflow obstruction within a wide range of clinical thresholds (about 4% - 35%) than not using the model. ### Conclusion This study successfully developed and validated a diagnostic model based on epidemiological and clinical information, which can help doctors identify high - risk Chinese elderly people, thereby improving the efficiency of spirometry. However, the performance of this model in men was better than that in women, and it needs to be further validated in non - Chinese populations.