 Nov 2022

Local file Local file

We employed a generalised estimating equation (GEE) logistic model with an exchangeablewithinpatient correlation structure to account for individualpatients having multiple exacerbations.
Análisis estadístico


Local file Local file

Weobserved that an established machinelearning method (GB) narrowlyoutperformed other prediction algorithmsand resulted in a prediction model with ahigh discrimination power (AUC = 0.82),which also showed robust calibration in thevalidation data.
GB machine learning was the best

In particular, we comparedlogistic regression (LR), random forest (RF),neural network (NN), and gradientboosting (GB) methods (20).


Local file Local fileUntitled2

For model comparison with machine learning–basedclassification, we selected the following classifiers: decisiontrees [15], random forests [16], knearest neighbor clustering[17], linear discriminant analysis, and adaptive boosting [18]

Classification algorithms for this study were selected accordingto previously published studies on COPD such as those of Wanget al [13] and Rahman et al [14].

Wang C, Chen X, Du L, Zhan Q, Yang T, Fang Z. Comparison of machine learning algorithms for the identification of acute exacerbations in chronic obstructive pulmonary disease. Comput Methods Programs Biomed 2020 May;188:105267. [doi: 10.1016/j.cmpb.2019.105267] [Medline: 31841787]

Rahman MJ, Nemati E, Rahman MM, Nathan V, Vatanparvar K, Kuang J. Automated assessment of pulmonary patients using heart rate variability from everyday wearables. Smart Health 2020 Mar;15:100081. [doi: 10.1016/j.smhl.2019.100081]



Local file Local file

Adjusted multiple logistic regression models were alsoperformed, including independent variables associated with exacerbation (P 0.20) in the univariate analysis
Statistical analysis for prediction


Local file Local fileuntitled1

First, we analyzedrisk of having frequent exacerbationsduring the first year of followup usinglogistic regression
statistical analysis for prediction


Local file Local file

(1) baseline differences between patients with and withouthospitalized exacerbation during followup were tested using analysisof variance or Wilcoxon ranksum test for continuous variables, andx2 test for categorical variables; (2) the incidence (first hospitalizedexacerbation during the prospective followup) and recurrence (secondhospitalized exacerbation during the prospective followup) of hospitalized exacerbations was summarized as a rate per person per year(PPPY), using a sum of individual patient’s persontime in the studyand standardized per year, accompanied by 95% CIs; (3) factors associated with first hospitalized (and recurrent) exacerbations during the3year followup, were explored using Cox proportional hazards models,adjusted for a wide range of demographics, and clinical and biologicmarkers.
Statistical analysis


Local file Local file

logisticmultivariate regression tests
Statistical analysis


Local file Local file

Timedependentvariables from hospital discharge were analyzed with Cox logisticregression and KaplanMeier statistics.
Cox logistic regression



The mostcommon statistical method was logistic regression (11 out of 25 different statistical methods analysed)followed by Cox regression (10), and correlation analysis between an index (or a multivariable regressionequation) with the outcome (three). Finally, Poisson regression model, negative binomial regression modeland random forest model were each used once.
Métodos estadísticos. Leer los papers que están siendo estudiados
Bertens (29) Ya lo tienes
Motegi (43) Motegi T, Jones RC, Ishii T, et al. A comparison of three multidimensional indices of COPD severity as predictors of future exacerbations. Int J COPD 2013; 8: 259–271.
Almagro (23) Almagro P, Soriano JB, Cabrera FJ, et al. Short and mediumterm prognosis in patients hospitalized for COPD exacerbation: the CODEX index. Chest 2014; 145: 972–980.
Suetomo (48) Suetomo M, Kawayama T, Kinoshita T, et al. COPD assessment tests scores are associated with exacerbated chronic obstructive pulmonary disease in Japanese patients. Respir Investig 2014; 52: 288–295
Mullerova (45) Müllerova H, Maselli DJ, Locantore N, et al. Hospitalized Exacerbations of COPD. Chest 2015; 147: 999–1007.
Thomsen (50) Thomsen M, Ingebrigtsen TS, Marott JL, et al. Inflammatory biomarkers and exacerbations in chronic obstructive pulmonary disease. JAMA 2013; 309: 2353–2361.
Moberg (42) Moberg M, Vestbo J, Martinez G, et al. Validation of the iBODE index as a predictor of hospitalization and mortality in patients with COPD Participating in pulmonary rehabilitation. COPD 2014; 11: 381–387.
Takahashi (49) Takahashi T, Muro S, Tanabe N, et al. Relationship between periodontitisrelated antibody and frequent exacerbations in chronic obstructive pulmonary disease. PLoS One 2012; 7: e40570.
Faganello (33) Faganello MM, Tanni SE, Sanchez FF, et al. BODE index and GOLD staging as predictors of 1year exacerbation risk in chronic obstructive pulmonary disease. Am J Med Sci 2010; 339: 10–14
GarciaAymerich (34) GarciaAymerich J, Farrero E, Félez MA, et al. Risk factors of readmission to hospital for a COPD exacerbation: a prospective study. Thorax 2003; 58: 100–105.
Ko (39) Ko FW, Tam W, Tung AH, et al. A longitudinal study of serial BODE indices in predicting mortality and readmissions for COPD. Respir Med 2011; 105: 266–273.
Echave (32) EchaveSustaeta J, Comeche Casanova L, Garcia Lujan R, et al. Prognosis following acute exacerbation of COPD treated with noninvasive mechanical ventilation. Arch Bronconeumol 2010; 46: 405–410.
Lee (40) Lee SD, Huang MS, Kang J, et al. The COPD assessment test (CAT) assists prediction of COPD exacerbations in highrisk patients. Respir Med 2014; 108: 600–608.
Moy (44) Moy ML, Teylan M, Danilack VA, et al. An index of daily step count and systemic inflammation predicts clinical outcomes in chronic obstructive pulmonary disease. Ann Am Thorac Soc 2014; 11: 149–157
Hurst (36) Hurst JR, Vestbo J, Anzueto A, et al. Susceptibility to exacerbation in chronic obstructive pulmonary disease. N Engl J Med 2010; 363: 1128–1138
Amalakuhan (28) Amalakuhan B, Kiljanek L, Parvathaneni A, et al. A prediction model for COPD readmissions: catching up, catching our breath, and improving a national problem. J Community Hosp Intern Med Perspect 2012; 2: 9915.


Local file Local file

n most cases, if censoring isnegligible and the followup period clearly defined, logistic regression is used; if censoring is significant or time to event is important,then a survival time approach using a Cox proportional Hazardsmodel is preferred. Other more complex model approaches such asmachine learning or competing risk models exist but are beyond thescope of this chapter [18, 19].
Métodos estadísticos para predecir variables binarias

To ensure stability of the model coefficients in logistic and Coxregression, an event frequency of at least 10/events per degree offreedom in the model is advised [13]. For example, in a cohort of1000 patients where 100 outcomes have been observed, the prediction model should include at most 10 variables. Ratios of lessthan 10 events per variable can result in overfitting of the data,leading to poor generalizability in other patient cohorts. All thesegeneral aspects of study and model specification should bedescribed in the methods to allow assessment of internal validity.
Importante



Cox proportional hazards modeling would alsohave been a valid approach for risk estimation, but we choselogistic regression analysis because we considered each exacerbation within our predefined time frame of 2 years to beof equal importance, regardless of whether this exacerbationoccurred early or late in the followup period.
Esto es importante. Comparación entre logística y regresión de cox

We used logistic regression modeling to estimate the riskof occurrence of COPD exacerbations within the proceeding24 months.
Análisis estadístico


Local file Local file

Common classification algorithms for supervisedlearning in the healthcare field include artificialneural networks, 36 decision trees, 37 random forests, 38Bayesian networks, 39 knearest neighbors,40 supportvector machines, 41 linear discriminant analysis, 42 kmeans clustering 43 and logistic regression. 44
Haremos un classification. Evaluar la posibilidad de regresión de Cox.
 Artificial neural networks
 Decision trees
 Random forests
 Bayesian networks
 knearest neighbors
 Support vector machines
 Linear discriminant analysis
 kmeans clustering
 Logistic regression



Five different prediction models for the annual exacerbation rate were estimated using negative binomial regression2
nb regression. Han hecho 5 modelos, pero en el sentido de que han utilizado diferentes variables como predictoras


Local file Local file

These methods included logistic regression with multipleregularization methods (lasso, ridge and elastic net), random forest and gradient boosted trees models (XGBoost).
Regularization methods and XGboosted random and boosted trees

This set of models(along with support vector machines and neural networks,which were not taken into consideration being more challenging to interpret) are considered gold standard formachine learning classification studies done on tabulardata. Resampling was applied during crossvalidation,making sure that only training folds of each crossvalidation iteration are affected, and the effect of resampling istested on the nonresampled test fold in each crossvalidation iteration.
Procesos estadísticos que acompañan a los modelo predictivos.


Local file Local file

Our study assumes that the SVM model can achieve acertain prediction effect in predicting the risk of readmission in COPD patients, and the results have certain reference value. Therefore, it is proposed to use SVM to builda 30day acute exacerbation readmission risk predictionmodel for elderly COPD patients, and evaluate its prediction effect, so as to provide a basis for early identificationof patients with high risk of readmission in the future.
SVM for the prediction analysis



First, although ACCEPTshowed goodtoexcellent discrimination overalland appears superior to exacerbation history alone,improvement in risk prediction was smaller in thosewith previous history of exacerbations than in thosewithout.
Esto también es importante. Es importante porque nosotros deberíamos dividir el desempeño del modelo según varias poblaciones de interés. Por ejemplo, en este caso, han comparado el desempeño del modelo en población con y sin exacerbaciones previas.

Usinga joint survival–logistic model, this risk tool providesan individualised risk estimate for exacerbations in thesubsequent year and their severity, as well as the rate offuture events.
Este documento es un comentario al estudio del ACCEPT. Nosotros podemos hacer lo mismo, pero teniendo en cuenta un mes(?) como referencia.


Local file Local file

Use of the ubiquitous proportional hazardsmodel, with time to first exacerbation as the outcome, is acommon mode of inference in contemporary clinical trialsof COPD. While it is robust in estimating treatment effectin randomized controlled trials, this analytical method fallsshort of providing other features, such as background rateof exacerbations or the shape of the incidence function, toenable predictions about the rate and (absolute or relative)duration of time to future events for a given patient. Asmentioned by Cox et al. (11), making such informative predictions has been hindered by the widespread use of semiparametric proportional hazards models.
desventajas del proportinal hazards

In the present work, we used a joint parametric recurrentevent and logistic regression model toenable full quantification of exacerbation incidence andseverity and their correlation
objetivo del análisis estadístico



We used a joint accelerated failure time and logistic modelto characterise rate and severity of exacerbations. We havepreviously published details of this approach elsewhere.14
análisis estadístico
