Biostatistics in Clinical Decision Making What can We Get from a 2×2 Contingency Table

Columbian College of Arts and Science, the George Washington University, Washington D.C., 20052, United States

^* Corresponding author: *shanshan789789@gwu.edu

Abstract

Biostatistics is an essential part when making clinical decisions. Applications of 2×2 contingency tables playing a key role in conducting analysis involving binary variables. When it comes to analysis based on 2×2 contingency tables, most people are familiar with the concept of sensitivity and specificity for evaluating a new test, but predictive values and receiver operating characteristic (ROC) curves would also provide information. Besides, Odds Ratio (OR), Risk Ratio (RR), and Chi-square test are measures based on 2×2 tables and commonly applied in retrospective and prospective studies. This article will first review the two kinds of application of 2×2 contingency tables, evaluating a new test compared with a reference standard, and exploring the relationship of exposures and outcomes in retrospective or prospective studies. Two clinical examples are presented to demonstrate these basic biostatistical concepts: diagnostic accuracy of 64-slice multidetector computed tomography (64-MDCT) to identify periampullary duodenal diverticula, and a randomized clinical trial (RCT) to examine the effectiveness of Dexmedetomidine for prevention of delirium in elderly patients after non-cardiac surgery. Correctly understanding these concepts will assist clinicians and medical researchers to analyze the data and interpret the results, and therefore make accurate decisions in clinical practice.