Predictive Variables for Failure in Administration of Intrapleural Tissue Plasminogen Activator/Deoxyribonuclease in Patients With Complicated Parapneumonic Effusions/Empyema

Khemasuwan, Danai; Sorensen, Jeffrey; Griffin, David C.

doi:10.1016/j.chest.2018.01.037

Cited by 34 publications

(30 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the study by Khemasuwan et al, 3 the XGBoost ROC AUC was 0.953, whereas the logistic model AUC was 0.891. The Hosmer-Lemeshow test failed to demonstrate poor calibration, but as noted previously, this does not prove that the model is well calibrated.…”

mentioning

confidence: 91%

“…1,2 A common clinical question is whether surgical intervention should be delayed while intrapleural agents are given time to work. In this issue of CHEST, Khemasuwan et al 3 address this by developing a clinical prediction model to identify patients with complicated parapneumonic effusions that are likely to fail intrapleural tPA/DNase therapy and require videoassisted thoracoscopic surgery. The investigators developed both a logistic regression model and a machine-learning model (XGBoost).…”

mentioning

confidence: 99%

“…This study merits attention because it illustrates key concepts related to clinical prediction models. 3 Prediction models in medicine are not new; there are models for ICU mortality (Acute Physiology and Chronic Health Evaluation), pulmonary nodule diagnosis, and spread of lung cancer to lymph nodes, to name but a few. [4][5][6] Although statistical models are often useful for hypothesis testing, and can be useful for predicting outcomes of groups, clinical prediction models are different in that they take this one step further and try to provide absolute risk estimates for individuals.…”

mentioning

confidence: 99%

“…This can potentially improve model performance. The study by Khemasuwan et al 3 serves to highlight some of the promises and pitfalls associated with machine learning and underscores the importance of assessing discrimination, calibration, and validation when evaluating clinical prediction models. 7,10 Discrimination is the ability to distinguish between patients that have the event vs those that do not.…”

mentioning

confidence: 99%

“…13 Practicing physicians do not need to know the details of machine learning, but they do need to understand the fundamental concepts of discrimination, calibration, and validation to decide whether a clinical prediction model is suitable for patients. The study by Khemasuwan et al 3 highlights this; the model is promising, but we will have to wait for external validation to see if the promise of machine learning is fulfilled.…”

mentioning

confidence: 99%

See 4 more Smart Citations

Machine Learning for Treating Complicated Parapneumonic Effusions

Ost

2018

Chest

View full text Add to dashboard Cite

mentioning

confidence: 91%

mentioning

confidence: 99%

mentioning

confidence: 99%

mentioning

confidence: 99%

mentioning

confidence: 99%

See 3 more Smart Citations

Machine Learning for Treating Complicated Parapneumonic Effusions

Ost

2018

Chest

View full text Add to dashboard Cite

Machine learning risk score for prediction of gestational diabetes in early pregnancy in Tianjin, China

Liu

Leng

et al. 2020

Diabetes Metabolism Res

View full text Add to dashboard Cite

Aims This study aimed to develop a machine learning–based prediction model for gestational diabetes mellitus (GDM) in early pregnancy in Chinese women. Materials and methods We used an established population‐based prospective cohort of 19,331 pregnant women registered as pregnant before the 15th gestational week in Tianjin, China, from October 2010 to August 2012. The dataset was randomly divided into a training set (70%) and a test set (30%). Risk factors collected at registration were examined and used to construct the prediction model in the training dataset. Machine learning, that is, the extreme gradient boosting (XGBoost) method, was employed to develop the model, while a traditional logistic model was also developed for comparison purposes. In the test dataset, the performance of the developed prediction model was assessed by calibration plots for calibration and area under the receiver operating characteristic curve (AUR) for discrimination. Results In total, 1484 (7.6%) women developed GDM. Pre‐pregnancy body mass index, maternal age, fasting plasma glucose at registration, and alanine aminotransferase were selected as risk factors. The machine learning XGBoost model‐predicted probability of GDM was similar to the observed probability in the test data set, while the logistic model tended to overestimate the risk at the highest risk level (Hosmer–Lemeshow test p value: 0.243 vs. 0.099). The XGBoost model achieved a higher AUR than the logistic model (0.742 vs. 0.663, p < 0.001). This XGBoost model was deployed through a free, publicly available software interface (https://liuhongwei.shinyapps.io/gdm_risk_calculator/). Conclusion The XGBoost model achieved better performance than the logistic model.

show abstract