Exercise Physiology: Energy, Nutrition and Human Performance

SNAP-II and SNAPPE-II: Simplified newborn illness severity and mortality risk scores

Accurate information regarding prognosis is fundamental to optimal clinical care. The best approach to assess patient prognosis relies on prediction models that simultaneously consider a number of prognostic factors and provide an estimate of patients' absolute risk of an event. Such prediction models should be characterized by adequately discriminating between patients who will have an event and those who will not and by adequate calibration ensuring accurate prediction of absolute risk. This Users' Guide will help clinicians understand the available metrics for assessing discrimination, calibration, and the relative performance of different prediction models. This article complements existing Users' Guides that address the development and validation of prediction models. Together, these guides will help clinicians to make optimal use of existing prediction models.

https://jamanetwork.com/journals/jama/articlepdf/2656816/jama_alba_2017_ug_170001.pdf

Discrimination and Calibration of Clinical Prediction Models: Users' Guides to the Medical Literature.

Statistical models support medical research by facilitating individualized outcome prognostication conditional on independent variables or by estimating effects of risk factors adjusted for covariates. Theory of statistical models is well-established if the set of independent variables to consider is fixed and small. Hence, we can assume that effect estimates are unbiased and the usual methods for confidence interval estimation are valid. In routine work, however, it is not known a priori which covariates should be included in a model, and often we are confronted with the number of candidate variables in the range 10-30. This number is often too large to be considered in a statistical model. We provide an overview of various available variable selection methods that are based on significance or information criteria, penalized likelihood, the change-in-estimate criterion, background knowledge, or combinations thereof. These methods were usually developed in the context of a linear regression model and then transferred to more generalized linear models or models for censored survival data. Variable selection, in particular if used in explanatory modeling where effect estimates are of central interest, can compromise stability of a final model, unbiasedness of regression coefficients, and validity of p-values or confidence intervals. Therefore, we give pragmatic recommendations for the practicing statistician on application of variable selection methods in general (low-dimensional) modeling problems and on performing stability investigations and inference. We also propose some quantities based on resampling the entire variable selection process to be routinely reported by software packages offering automated variable selection algorithms.

https://onlinelibrary.wiley.com/doi/pdf/10.1002/bimj.201700067

Variable selection - A review and recommendations for the practicing statistician.

Background
Current approaches to predict cardiovascular risk fail to identify many people who would benefit from preventive treatment, while others receive unnecessary intervention. Machine-learning offers opportunity to improve accuracy by exploiting complex interactions between risk factors. We assessed whether machine-learning can improve cardiovascular risk prediction.
Methods
Prospective cohort study using routine clinical data of 378,256 patients from UK family practices, free from cardiovascular disease at outset. Four machine-learning algorithms (random forest, logistic regression, gradient boosting machines, neural networks) were compared to an established algorithm (American College of Cardiology guidelines) to predict first cardiovascular event over 10-years. Predictive accuracy was assessed by area under the ‘receiver operating curve’ (AUC); and sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) to predict 7.5% cardiovascular risk (threshold for initiating statins).
Findings
24,970 incident cardiovascular events (6.6%) occurred. Compared to the established risk prediction algorithm (AUC 0.728, 95% CI 0.723–0.735), machine-learning algorithms improved prediction: random forest +1.7% (AUC 0.745, 95% CI 0.739–0.750), logistic regression +3.2% (AUC 0.760, 95% CI 0.755–0.766), gradient boosting +3.3% (AUC 0.761, 95% CI 0.755–0.766), neural networks +3.6% (AUC 0.764, 95% CI 0.759–0.769). The 78 highest achieving (neural networks) algorithm predicted 4,998/7,404 cases (sensitivity
79 67.5%, PPV 18.4%) and 53,458/75,585 non-cases (specificity 70.7%, NPV 95.7%), correctly predicting 355 (+7.6%) more patients who developed cardiovascular disease compared to the established algorithm.
Conclusions
Machine-learning significantly improves accuracy of cardiovascular risk prediction, increasing the number of patients identified who could benefit from preventive treatment, while avoiding unnecessary treatment of others.

/pdf/can-machine-learning-improve-cardiovascular-risk-prediction-cb0gcjm7s7.pdf

Can machine-learning improve cardiovascular risk prediction using routine clinical data?

Applied Logistic Regression: Hosmer/Applied Logistic Regression

Assessing the Fit of the Model

Introduction to the Logistic Regression Model

BackgroundAlthough some studies have reported an increased incidence of patellar dislocations within active populations, few studies have reported incidence rates and examined risk factors for this injury.PurposeTo examine the incidence of patellar dislocation injuries and the influence of demographic and occupational risk factors associated with injury among active-duty United States (US) service members between 1998 and 2007.Study DesignCohort study; Level of evidence, 3.MethodsUsing the Defense Medical Surveillance System, a search was performed for International Classification of Disease, 9th Revision (ICD-9) code 836.3 among all US service members on active duty during the study period. Multivariable Poisson regression analysis was used to estimate the rate of patellar dislocation per 1000 person-years at risk to injury. Incidence rates (IRs) and incidence rate ratios (IRRs) for patellar dislocation along with 95% confidence intervals (CIs) were estimated by gender, age, race, branch of military serv...

https://brettowensmd.com/assets/patella-dislocations-are-common-in-military-personnel.pdf

Incidence of Acute Traumatic Patellar Dislocation Among Active-Duty United States Military Service Members

Purpose De Quervain's tenosynovitis is thought to occur most frequently in women, with presentation of pain and swelling in the first dorsal extensor sheath. The epidemiology of this extensor tendinitis is not well described. We evaluated the incidence and demographic risk factors for de Quervain's tenosynovitis using a large database of military personnel. Methods The Defense Medical Epidemiology Database (DMED) collects International Classification of Diseases, 9th Revision, and Clinical Modification (ICD-9-CM) coding information for every patient encounter occurring for United States military personnel. We queried the DMED system by race, gender, military service, rank, and age for the years 1998–2006 using the ICD-9 code 727.04, limiting data to first presentations. Multivariate Poisson regression was used to estimate the rate of de Quervain's tenosynovitis per 1000 person-years, as well as incidence rate ratios and 95% confidence intervals. Results There were 11,332 cases of de Quervain's tenosynovitis in the population at risk of 12,117,749 person-years. Women had a significantly higher rate of de Quervain's tenosynovitis at 2.8 cases per 1000 person-years, compared to men at 0.6 per 1000 person-years. Age greater than 40 was also a significant risk factor, with this age category showing a rate of 2.0 per 1000 person-years compared to 0.6 per 1000 in personnel under 20 years. There was also a racial difference, with blacks affected at 1.3 per 1000 person-years compared to whites at 0.8. Conclusions In analysis of a large population, we have described the epidemiology of stenosing tenosynovitis of the first extensor compartment. Risk factors for de Quervain's in our population include female gender, age greater than 40, and black race. Type of study/level of evidence Prognostic II.

Rodney X. Sturdivant

Papers

Applied Logistic Regression: Hosmer/Applied Logistic Regression

Assessing the Fit of the Model

Introduction to the Logistic Regression Model

Incidence of Acute Traumatic Patellar Dislocation Among Active-Duty United States Military Service Members

Incidence of de Quervain's Tenosynovitis in a Young, Active Population