Clinical Prediction Models

The second edition of this volume provides insight and practical illustrations on how modern statistical concepts and regression methods can be applied in medical prediction problems, including diagnostic and prognostic outcomes. Many advances have been made in statistical approaches towards outcome prediction, but a sensible strategy is needed for model development, validation, and updating, such that prediction models can better support medical practice.

There is an increasing need for personalized evidence-based medicine that uses an individualized approach to medical decision-making. In this Big Data era, there is expanded access to large volumes of routinely collected data and an increased number of applications for prediction models, such as targeted early detection of disease and individualized approaches to diagnostic testing and treatment. Clinical Prediction Models presents a practical checklist that needs to be considered for development of avalid prediction model. Steps include preliminary considerations such as dealing with missing values; coding of predictors; selection of main effects and interactions for a multivariable model; estimation of model parameters with shrinkage methods and incorporation of external data; evaluation of performance and usefulness; internal validation; and presentation formatting. The text also addresses common issues that make prediction models suboptimal, such as small sample sizes, exaggerated claims, and poor generalizability.

The text is primarily intended for clinical epidemiologists and biostatisticians. Including many case studies and publicly available R code and data sets, the book is also appropriate as a textbook for a graduate course on predictive modeling in diagnosis and prognosis. While practical in nature, the book also provides a philosophical perspective on data analysis in medicine that goes beyond predictive modeling.

Updates to this new and expanded edition include:

. A discussion of Big Data and its implications for the design of prediction models

. Machine learning issues

. More simulations with missing 'y' values

. Extended discussion on between-cohort heterogeneity

. Description of ShinyApp

. Updated LASSO illustration

. New case studies

Autorentext
Ewout Steyerberg worked for 25 years at Erasmus Medical Center in Rotterdam before moving to Leiden where he is now Professor of Clinical Biostatistics and Medical Decision Making and chair of the Department of Biomedical Data Sciences at Leiden University Medical Center. His research has covered a broad range of methodological and medical topics, which is reflected in hundreds of peer-reviewed methodological and applied publications. His methodological expertise is in the design and analysis of randomized controlled trials, cost-effectiveness analysis, and decision analysis. His methodological research focuses on the development, validation and updating of prediction models, as reflected in a textbook (Springer, 2009). His medical fields of application include oncology, cardiovascular disease, internal medicine, pediatrics, infectious diseases, neurology, surgery and traumatic brain injury.

Inhalt
PrefaceviiAcknowledgementsxiChapter 1Introduction11.1Diagnosis, prognosis and therapy choice in medicine11.1.1Predictions for personalized evidence-based medicine11.2Statistical modeling for prediction51.2.1Model assumptions51.2.2Reliability of predictions: aleatory and epistemic uncertainty61.2.3Sample size61.3Structure of the book81.3.1Part I: Prediction models in medicine81.3.2Part II: Developing internally valid prediction models81.3.3Part III: Generalizability of prediction models91.3.4Part IV: Applications9Part I: Prediction models in medicine11Chapter 2Applications of prediction models132.1Applications: medical practice and research132.2Prediction models for Public Health142.2.1Targeting of preventive interventions14*2.2.2Example: prediction for breast cancer142.3Prediction models for clinical practice172.3.1Decision support on test ordering17*2.3.2Example: predicting renal artery stenosis172.3.3Starting treatment: the treatment threshold20*2.3.4Example: probability of deep venous thrombosis202.3.5Intensity of treatment21*2.3.6Example: defining a poor prognosis subgroup in cancer222.3.7Cost-effectiveness of treatment232.3.8Delaying treatment23*2.3.9Example: spontaneous pregnancy chances242.3.10Surgical decision-making26*2.3.11Example: replacement of risky heart valves272.4Prediction models for medical research282.4.1Inclusion and stratification in a RCT28*2.4.2Example: selection for TBI trials292.4.3Covariate adjustment in a RCT302.4.4Gain in power by covariate adjustment31*2.4.5Example: analysis of the GUSTO-III trial322.4.6Prediction models and observational studies322.4.7Propensity scores33*2.4.8Example: statin treatment effects342.4.9Provider comparisons35*2.4.10Example: ranking cardiac outcome352.5Concluding remarks35Chapter 3Study design for prediction modeling373.1Studies for prognosis373.1.1Retrospective designs37*3.1.2Example: predicting early mortality in esophageal cancer373.1.3Prospective designs38*3.1.4Example: predicting long-term mortality in esophageal cancer393.1.5Registry data39*3.1.6Example: surgical mortality in esophageal cancer393.1.7Nested case-control studies40*3.1.8Example: perioperative mortality in major vascular surgery403.2Studies for diagnosis413.2.1Cross-sectional study design and multivariable modeling41*3.2.2Example: diagnosing renal artery stenosis413.2.3Case-control studies41*3.2.4Example: diagnosing acute appendicitis423.3Predictors and outcome423.3.1Strength of predictors423.3.2Categories of predictors423.3.3Costs of predictors433.3.4Determinants of prognosis443.3.5Prognosis in oncology443.4Reliability of predictors453.4.1Observer variability45*3.4.2Example: histology in Barrett's esophagus453.4.3Biological variability463.4.4Regression dilution bias46*3.4.5Example: simulation study on reliability of a binary predictor463.4.6Choice of predictors473.5Outcome473.5.1Types of outcome473.5.2Survival endpoints48*3.5.3Examples: 5-year relative survival in cancer registries483.5.4Composite endpoints49*3.5.5Example: composite endpoints in cardiology493.5.6Choice of prognostic outcome493.5.7Diagnostic endpoints49*3.5.8Example: PET scans in esophageal cancer503.6Phases of biomarker development503.7Statistical power and reliab...