scholarly journals Diving Deeper into Models

2018 ◽  
pp. 121-133
Author(s):  
Alberto Traverso ◽  
Frank J. W. M. Dankers ◽  
Biche Osong ◽  
Leonard Wee ◽  
Sander M. J. van Kuijk
Keyword(s):  
Machine Learning ◽  
Logistic Regression ◽  
Prediction Model ◽  
Unsupervised Learning ◽  
Learning Objectives ◽  
Clinical Prediction ◽  
Clinical Prediction Model ◽  
Supervised And Unsupervised Learning

AbstractPre-requisites to better understand the chapter: knowledge of the major steps and procedures of developing a clinical prediction model.Logical position of the chapter with respect to the previous chapter: in the last chapters, you have learned how to develop and validate a clinical prediction model. You have been learning logistic regression as main algorithm to build the model. However, several different more complex algorithms can be used to build a clinical prediction model. In this chapter, the main machine learning based algorithms will be presented to you.Learning objectives: you will be presented with the definitions of: machine learning, supervised and unsupervised learning. The major algorithms for the last two categories will be introduced.

scholarly journals Prediction of Unsuccessful Endometrial Ablation: Random Forest vs Logistic Regression

2020 ◽  
Author(s):  
Kelly Yvonne Roger Stevens ◽  
Liesbet Lagaert ◽  
Tom Bakkes ◽  
Malou Evi Gelderblom ◽  
Saskia Houterman ◽  
...  
Keyword(s):  
Machine Learning ◽  
Logistic Regression ◽  
Prediction Model ◽  
Treatment Options ◽  
Premenopausal Women ◽  
Endometrial Ablation ◽  
Machine Learning Techniques ◽  
Multivariate Logistic Regression Model ◽  
Clinical Prediction ◽  
Study Objective

Abstract Background: Five percent of premenopausal women experience abnormal uterine bleeding. Endometrial ablation (EA) is one of the treatment options for this common problem. However, this technique shows a decrease in patient satisfaction and treatment efficacy on the long term Study objective: To develop a prediction model to predict surgical re-intervention (for example re-ablation or hysterectomy) within two years after EA by using Machine Learning (ML). The performance of the developed prediction model was compared with a previously published multivariate logistic regression model (LR). Design: This retrospective cohort study, with a minimal follow up time of two years, included 446 pre-menopausal women (18+) that underwent an EA for complaints of heavy menstrual bleeding. The performance of the ML- and the LR model was compared using the area under the Receiving Operating Characteristic (ROC) curve. Results: We found out that the ML model (AUC of 0.65 (95% CI 0.56-0.74)) is not superior compared to the LR model (AUC of 0.71 (95% CI 0.64-0.78)) in predicting the outcome of surgical re-intervention within two years after EA. Conclusion: Although Machine Learning techniques are gaining popularity in development of clinical prediction tools, this study shows that ML is not necessarily superior to the traditional statistical LR techniques. The performance of a prediction model is influenced by the sample size, the number of features of a dataset, hyperparameter tuning and the linearity of associations. Both techniques should be considered when developing a clinical prediction model.

scholarly journals Development of a multivariate clinical prediction model for the diagnosis of mild stroke/TIA in physician first-contact patient settings

2016 ◽  
Author(s):  
Maximilian B Bibok ◽  
Andrew M Penn ◽  
Mary L Lesperance ◽  
Kristine Votova ◽  
Robert Balshaw
Keyword(s):  
Logistic Regression ◽  
Prediction Model ◽  
Clinical Symptoms ◽  
Diagnostic Model ◽  
Clinical Prediction ◽  
Clinical Prediction Model ◽  
Abcd2 Score ◽  
Mild Stroke ◽  
Sensitivity Specificity ◽  
First Contact

OBJECTIVE: To develop a clinical prediction model for diagnosing mild stroke/transient ischemic attack (TIA) in first-contact patient settings. DESIGN: Retrospective study design utilizing logistic regression modeling of patient clinical symptoms collected from patient chart histories and referral data. SETTING: Regional fast-track TIA clinic on Vancouver Island, Canada, accepting referrals from emergency departments (ED) and general practice (GP). PARTICIPANTS: Model development: 4187 ED and GP referred patients from 2008-2011 who were assessed at the TIA clinic. Temporal hold-out validation: 1953 ED and GP referred patients from 2012-2013 assessed at the same clinic. OUTCOMES: Diagnosis of mild stroke/TIA by clinic neurologists. RESULTS: 123 candidate predictors were assessed using univariate feature selection for inclusion in the model, and culminated in the selection of 50 clinical features. Post-hoc investigation of the selected predictors revealed 12 clinically relevant interaction terms. Model performance on the temporal hold-out validation set achieved a sensitivity/specificity of 71.8% / 72.8% using the ROC01 cutpoint (≥ 0.662), and an AUC of 79.9% (95% CI, 77.9%-81.9%). In comparison, the ABCD2 score (≥ 4) achieved a sensitivity/specificity of 70.4% / 54.5% and an AUC of 67.5% (95% CI, 65.2%-69.9%). The logistic regression model demonstrated good calibration on the hold-out set (β0 = -0.257); βlinear = 1.047). CONCLUSIONS: The developed diagnostic model performs better than the ABCD2 score at diagnosing mild stroke/TIA on the basis of clinical symptoms. The model has the potential to replace the use of the prognostic ABCD2 score in diagnostic medical contexts in which the ABCD2 score is currently used, such as patient triage.

scholarly journals Prediction of unsuccessful endometrial ablation: random forest vs logistic regression

2021 ◽  
Vol 18 (1) ◽  
Author(s):  
Kelly Yvonne Roger Stevens ◽  
Liesbet Lagaert ◽  
Tom Bakkes ◽  
Malou Evi Gelderblom ◽  
Saskia Houterman ◽  
...  
Keyword(s):  
Machine Learning ◽  
Logistic Regression ◽  
Prediction Model ◽  
Endometrial Ablation ◽  
Machine Learning Techniques ◽  
Multivariate Logistic Regression Model ◽  
Clinical Prediction ◽  
Clinical Predictors ◽  
Study Objective ◽  
Menopausal Women

Abstract Background Five percent of pre-menopausal women experience abnormal uterine bleeding. Endometrial ablation (EA) is one of the treatment options for this common problem. However, this technique shows a decrease in patient satisfaction and treatment efficacy on the long term. Study objective To develop a prediction model to predict surgical re-intervention (for example re-ablation or hysterectomy) within 2 years after endometrial ablation (EA) by using machine learning (ML). The performance of the developed prediction model was compared with a previously published multivariate logistic regression model (LR). Design This retrospective cohort study, with a minimal follow-up time of 2 years, included 446 pre-menopausal women (18+) that underwent an EA for complaints of heavy menstrual bleeding. The performance of the ML and the LR model was compared using the area under the receiving operating characteristic (ROC) curve. Results We found out that the ML model (AUC of 0.65 (95% CI 0.56–0.74)) is not superior compared to the LR model (AUC of 0.71 (95% CI 0.64–0.78)) in predicting the outcome of surgical re-intervention within 2 years after EA. Based on the ML model, dysmenorrhea and duration of menstruation have the highest impact on the surgical re-intervention rate. Conclusion Although machine learning techniques are gaining popularity in development of clinical prediction tools, this study shows that ML is not necessarily superior to the traditional statistical LR techniques. Both techniques should be considered when developing a clinical prediction model. Both models can identify the clinical predictors to surgical re-intervention and contribute to the shared decision-making process in the clinical practice.

scholarly journals Prediction of Postoperative Delirium in Geriatric Hip Fracture Patients: A Clinical Prediction Model Using Machine Learning Algorithms

2021 ◽  
Vol 12 ◽  
pp. 215145932110622
Author(s):  
Jacobien H. F. Oosterhoff ◽  
Aditya V. Karhade ◽  
Tarandeep Oberai ◽  
Esteban Franco-Garcia ◽  
Job N. Doornberg ◽  
...  
Keyword(s):  
Machine Learning ◽  
Hip Fracture ◽  
Prediction Model ◽  
Postoperative Delirium ◽  
Machine Learning Algorithms ◽  
Brier Score ◽  
Clinical Prediction ◽  
Clinical Prediction Model ◽  
Preoperative Prediction ◽  
Geriatric Hip Fracture

Introduction Postoperative delirium in geriatric hip fracture patients adversely affects clinical and functional outcomes and increases costs. A preoperative prediction tool to identify high-risk patients may facilitate optimal use of preventive interventions. The purpose of this study was to develop a clinical prediction model using machine learning algorithms for preoperative prediction of postoperative delirium in geriatric hip fracture patients. Materials & Methods Geriatric patients undergoing operative hip fracture fixation were queried in the American College of Surgeons National Surgical Quality Improvement Program database (ACS NSQIP) from 2016 through 2019. A total of 28 207 patients were included, of which 8030 (28.5%) developed a postoperative delirium. First, the dataset was randomly split 80:20 into a training and testing subset. Then, a random forest (RF) algorithm was used to identify the variables predictive for a postoperative delirium. The machine learning-model was developed on the training set and the performance was assessed in the testing set. Performance was assessed by discrimination (c-statistic), calibration (slope and intercept), overall performance (Brier-score), and decision curve analysis. Results The included variables identified using RF algorithms were (1) age, (2) ASA class, (3) functional status, (4) preoperative dementia, (5) preoperative delirium, and (6) preoperative need for mobility-aid. The clinical prediction model reached good discrimination (c-statistic = .79), almost perfect calibration (intercept = −.01, slope = 1.02), and excellent overall model performance (Brier score = .15). The clinical prediction model was deployed as an open-access web-application: https://sorg-apps.shinyapps.io/hipfxdelirium/ . Discussion & Conclusions We developed a clinical prediction model that shows promise in estimating the risk of postoperative delirium in geriatric hip fracture patients. The clinical prediction model can play a beneficial role in decision-making for preventative measures for patients at risk of developing a delirium. If found to be externally valid, clinicians might use the available web-based application to help incorporate the model into clinical practice to aid decision-making and optimize preoperative prevention efforts.

scholarly journals Development and validation of a clinical prediction model for endocervical curettage decision-making in cervical lesions

BMC Cancer ◽  
2021 ◽  
Vol 21 (1) ◽  
Author(s):  
Yuanxing Li ◽  
Haixia Luo ◽  
Xiu Zhang ◽  
Jingjing Chang ◽  
Yueyang Zhao ◽  
...  
Keyword(s):  
Logistic Regression ◽  
Prediction Model ◽  
Squamous Epithelium ◽  
Characteristic Curve ◽  
Multivariable Logistic Regression Analysis ◽  
Clinical Prediction ◽  
Cervical Lesions ◽  
Clinical Prediction Model ◽  
Multivariable Logistic Regression ◽  
Hpv Status

Abstract Background In the absence of practical and reliable predictors for whether the endocervical curettage (ECC) procedure should be performed, decisions regarding patient selection are usually based on the colposcopists’ clinical judgment instead of evidence. We aimed to develop and validate a practical prediction model that uses available information to reliably estimate the need to perform ECC in patients suspected of having cervical lesions. Methods In this retrospective study, 2088 patients who underwent colposcopy, colposcopically directed biopsy (CDB) and ECC procedures between September 2019 and September 2020 at the Second Hospital of Shanxi Medical University were included. The data were analyzed with univariate and multivariable logistic regression. Least absolute shrinkage and selection operator (LASSO) was used to select predictors for ECC positivity. The ECC prediction model was presented as a nomogram and evaluated in terms of discrimination and calibration. Furthermore, this model was validated internally with cross-validation and bootstrapping. Results Significant trends were found for ECC positivity with increasing age (P = 0.001), menopause (P = 0.003), Human papillomavirus (HPV) status (P < 0.001), severity of ThinPrep Cytological Test (TCT) (P < 0.001), original squamous epithelium ectopia (P = 0.037) and colposcopy impression (P < 0.001) by multivariable logistic regression analysis. The ECC prediction model was developed based on the following predictors: age, menopause, symptom of contact bleeding, severity of TCT, HPV status, cervix visibility, original squamous epithelium ectopia, acetowhite changes and colposcopic impression. This model had satisfactory calibration and good discrimination, with an area under the receiver operator characteristic curve (AUC) of 0.869 (95% confidence interval 0.849 to 0.889). Conclusions A readily applicable clinical prediction model was constructed to reliably estimate the probability of ECC positivity in patients suspicious of having cervical lesions, which may help clinicians make decisions regarding the ECC procedure and possibly prevent adverse effects.

scholarly journals Minimum sample size for external validation of a clinical prediction model with a binary outcome

2021 ◽  
Author(s):  
Richard D. Riley ◽  
Thomas P. A. Debray ◽  
Gary S. Collins ◽  
Lucinda Archer ◽  
Joie Ensor ◽  
...  
Keyword(s):  
Prediction Model ◽  
Sample Size ◽  
External Validation ◽  
Minimum Sample Size ◽  
Binary Outcome ◽  
Clinical Prediction ◽  
Clinical Prediction Model ◽  
Minimum Sample

Develop a Clinical Prediction Model for Postoperative Cognitive Dysfunction after Major Noncardiac Surgery in Elderly Patients: A Protocol for a Prospective Observational Study

Gerontology ◽  
2021 ◽  
pp. 1-8
Author(s):  
Yang Shen ◽  
Xianchen Li ◽  
Junyan Yao
Keyword(s):  
Observational Study ◽  
Prediction Model ◽  
Cognitive Decline ◽  
Cognitive Dysfunction ◽  
Prospective Observational Study ◽  
Postoperative Cognitive Dysfunction ◽  
Noncardiac Surgery ◽  
Clinical Prediction ◽  
Clinical Prediction Model ◽  
Postoperative Cognitive Decline

Perioperative neurocognitive disorders (PNDs) refer to cognitive decline identified in the preoperative or postoperative period. It has been reported that the incidence of postoperative neurocognitive impairment after noncardiac surgery in patients older than 65 at 1 week was 25.8∼41.4%, and at 3 months 9.9∼12.7%. PNDs will last months or even develop to permanent dementia, leading to prolonged hospital stays, reduced quality of life, and increased mortality within 1 year. Despite the high incidence and poor prognosis of PNDs in the aged population, no effective clinical prediction model has been established to predict postoperative cognitive decline preoperatively. To develop a clinical prediction model for postoperative neurocognitive dysfunction, a prospective observational study (Clinical trial registration number: ChiCTR2000036304) will be performed in the Shanghai General Hospital during January 2021 to October 2022. A sample size of 675 patients aged &#x3e;65 years old, male or female, and scheduled for elective major noncardiac surgery will be recruited. A battery of neuropsychological tests will be used to test the cognitive function of patients at 1 week, 1 month, and 3 months postoperatively. We will evaluate the associations of PNDs with a bunch of candidate predictors including general characteristics of patients, blood biomarkers, indices associated with anesthesia and surgery, retinal nerve-fiber layer thickness, and frailty index to develop the clinical prediction model by using multiple logistic regression analysis and least absolute shrinkage and the selection operator (LASSO) method. The <i>k</i>-fold cross-validation method will be utilized to validate the clinical prediction model. In conclusion, this study was aimed to develop a clinical prediction model for postoperative cognitive dysfunction of old patients. It is anticipated that the knowledge gained from this study will facilitate clinical decision-making for anesthetists and surgeons managing the aged patients undergoing noncardiac surgery.

scholarly journals Predicting severe pneumonia in the emergency department: a global study of the Pediatric Emergency Research Networks (PERN)—study protocol

BMJ Open ◽  
2020 ◽  
Vol 10 (12) ◽  
pp. e041093
Author(s):  
Todd Adam Florin ◽  
Daniel Joseph Tancredi ◽  
Lilliam Ambroggio ◽  
Franz E Babl ◽  
Stuart R Dalziel ◽  
...  
Keyword(s):  
Emergency Department ◽  
Prediction Model ◽  
Severe Disease ◽  
Pediatric Emergency ◽  
Global Network ◽  
Clinical Prediction ◽  
Emergency Research ◽  
Research Networks ◽  
Clinical Prediction Model ◽  
Severity Scores

IntroductionPneumonia is a frequent and costly cause of emergency department (ED) visits and hospitalisations in children. There are no evidence-based, validated tools to assist physicians in management and disposition decisions for children presenting to the ED with community-acquired pneumonia (CAP). The objective of this study is to develop a clinical prediction model to accurately stratify children with CAP who are at risk for low, moderate and severe disease across a global network of EDs.Methods and analysisThis study is a prospective cohort study enrolling up to 4700 children with CAP at EDs at ~80 member sites of the Pediatric Emergency Research Networks (PERN; https://pern-global.com/). We will include children aged 3 months to <14 years with a clinical diagnosis of CAP. We will exclude children with hospital admissions within 7 days prior to the study visit, hospital-acquired pneumonias or chronic complex conditions. Clinical, laboratory and imaging data from the ED visit and hospitalisations within 7 days will be collected. A follow-up telephone or text survey will be completed 7–14 days after the visit. The primary outcome is a three-tier composite of disease severity. Ordinal logistic regression, assuming a partial proportional odds specification, and recursive partitioning will be used to develop the risk stratification models.Ethics and disseminationThis study will result in a clinical prediction model to accurately identify risk of severe disease on presentation to the ED. Ethics approval was obtained for all sites included in the study. Cincinnati Children’s Hospital Institutional Review Board (IRB) serves as the central IRB for most US sites. Informed consent will be obtained from all participants. Results will be disseminated through international conferences and peer-reviewed publications. This study overcomes limitations of prior pneumonia severity scores by allowing for broad generalisability of findings, which can be actively implemented after model development and validation.

Sign in / Sign up

Export Citation Format

Share Document