<Consulting in Statistics> Speaker：Dr. Jeng-Jung Chen (Hsinchu MacKay Memorial Hospital)


Topic：Applications of Statistics in Medical Research

Speaker：Dr. Jeng-Jung Chen (Hsinchu MacKay Memorial Hospital)

Time：Nov 20 (Friday) , 2026, 13:20-15:10 

Place： 4F-427, Assembly Building I


Abstract
Statistics plays a crucial role in medical research. It provides researchers with the tools to analyze medical data, compare different treatments or clinical conditions, evaluate the strength of evidence, and predict patient outcomes. In this session, three medical studies will be presented as case studies to introduce commonly used statistical methods and their applications in medical research. Through these examples, students will learn how statistical methods can be applied to answer different types of clinical and research questions. 
  • Biostatistics: Comparing Medical Conditions 
   The first study will introduce the basic concepts and applications of biostatistics in medical research. Statistical methods can be used to compare outcomes between two or more medical conditions or patient groups and to determine whether observed differences are statistically significant. This example will illustrate how researchers use descriptive statistics, hypothesis testing, and other commonly used statistical methods to interpret medical data. 
  • Meta-analysis: Combining Evidence from Multiple Studies 
   The second study will focus on meta-analysis, an important method for synthesizing evidence from multiple studies. By combining the results of several randomized controlled trials (RCTs), researchers can obtain a more comprehensive estimate of the effect of a treatment and assess the consistency of the findings across studies. This example will introduce the basic concepts of effect estimates, forest plots, heterogeneity, and the interpretation of pooled treatment effects. 
  • Machine Learning: Predicting Patient Outcomes 
   The third study will introduce the application of machine learning in medical research. Machine learning methods can use clinical characteristics, laboratory results, imaging data, or other medical information to develop models for predicting patient outcomes. This example will demonstrate how predictive models are developed and evaluated, and will introduce concepts such as model performance, discrimination, validation, and the potential applications and limitations of machine learning in clinical practice. 

        Overall, these three examples will demonstrate how statistical methods have evolved from traditional statistical comparisons to evidence synthesis and data-driven prediction. The goal is to help students understand not only the statistical methods themselves, but also how to select and apply appropriate methods to different medical research questions.