Logistic Model in Clinical and Health Research—the Elephant and Blind Men
Open Access
- 15 September 2025
- journal article
- Published by Wolters Kluwer Health in American Journal of Clinical Oncology
- Vol. 48 (11) , 572-577
Abstract
Logistic models are everywhere in scientific research, including clinical and health research. However, there is no practical report on research results that will utilize many available statistical tools but mostly 1 or 2 only, for a logistic model. We introduce 6 powerful statistical tools, forest plot, AUC curve, nomogram, and decision curve analysis, as well as bootstrapping sampling and cross validation, by applying head and neck cancer data to a logistic model. We hope that these tools will make the logistic model, and accordingly our research, more comprehensive and clinically more applicable. In the 6 parts of the article, we introduce each of the 6 statistical tools (methods) with relevant figures and interpretations on key statistical concepts to show how we can improve our understanding on the logistic model and the clinical prospects behind the data. The statistical tools we present in the current special communication for reporting research or clinical trials in a logistic model, if popularized among researchers and clinicians, will make a research conclusion more comprehensive, valid and clinically applicable to other cases.Keywords
This publication has 4 references indexed in Scilit:
- Sarcopenia and Treatment Toxicity in Older Adults Undergoing Chemoradiation for Head and Neck Cancer: Identifying Factors to Predict FrailtyCancers, 2022
- A simple, step-by-step guide to interpreting decision curve analysisDiagnostic and Prognostic Research, 2019
- Nomograms in oncology: more than meets the eyeThe Lancet Oncology, 2015
- Decision Curve Analysis: A Novel Method for Evaluating Prediction ModelsMedical Decision Making, 2006