Chapter 10--Hypothesis Testing--Categorical Data

The patients are followed for 5 years with survival

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: ions for paired data. Example: Two chemotherapy regimens for breast cancer after mastectomy are to be compared. A matched study is set up such that patients are assigned to pairs matched on age (within 5 years) and clinical conditions. A random member of each matched pair gets treatment A (chemotherapy) perioperatively (within 1 week after mastectomy) and for an additional six months, whereas the other member gets treatment B chemotherapy only perioperatively. The patients are followed for 5 years with survival as the outcome variable. The rate of 5 year survivals were 526/621 and 515/621 for treatments A and B, respectively. Chapter 10: Hypothesis Testing: Categorical Data Stat 491: Biostatistics Introduction Two-Sample Test for Binomial Proportions McNemar’s Test Estimation of Sample Size and Power R × C Contingency Tables Chi-Square Goodness-of-Fit Test The Kappa Statistic McNemar’s Test: Introduction Cont’d... Assuming independent samples (WRONG assumption), prop.test(x=c(526,515),n=c(621,621)) gives a p-value= 0.441 which implies no significant difference between the two treatments. A more appropriate analysis of the data is using matched pairs as sampling units. Outcome for Treatment B Outcome for Treatment A survive Die Total Survive 510 16 526 Die 5 90 95 Total 515 106 621 Chapter 10: Hypothesis Testing: Categorical Data Stat 491: Biostatistics Introduction Two-Sample Test for Binomial Proportions McNemar’s Test Estimation of Sample Size and Power R × C Contingency Tables Chi-Square Goodness-of-Fit Test The Kappa Statistic McNemar’s Test: Concordant and Discordant Pairs A concordant pair (discordant pair) is a matched pair in which the outcome is the same (different) for each member of the pair. A type A discordant pair is a matched pair in which the treatment A member of the pair has the event but treatment B member does not. A type B discordant pair is a matched pair in which the treatment B member of the pair has the event but treatment A member does not. The concordant pairs provide no information about differenc...
View Full Document

Ask a homework question - tutors are online