|
Activity Number:
|
358
- Contributed Poster Presentations: Section on Statistics in Epidemiology
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, July 30, 2019 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
|
Abstract #306909
|
|
|
Title:
|
Effects of Treatment Classifications in Network Meta-Analysis
|
|
Author(s):
|
Aiwen Xing* and Lifeng Lin
|
|
Companies:
|
and Florida State University
|
|
Keywords:
|
Lumping;
Network meta-analysis;
Node-making;
Systematic review;
Splitting;
Treatment classification
|
|
Abstract:
|
Network meta-analysis (NMA) has been a popular tool to simultaneously compare multiple treatments. But the node-making process was often insufficiently reported. This article empirically examined the impact of different treatment classifications on the results in NMAs. We collected nine published NMAs containing similar treatments that may be lumped. The Bayesian random-effects model was applied to these NMAs before and after lumping the similar treatments. We estimated the odds ratios and their 95% credible intervals in the original and lumped NMAs. We proposed the adjusted DIC to assess the model performance in the lumped NMAs. We also used the ratios of credible interval lengths and ratios of odds ratios to evaluate the estimates’ changes. The point estimates of odds ratios of many treatment comparisons had noticeable changes and many of their precisions were substantially improved. The DIC reduced after lumping similar treatments in seven (78%) NMAs, indicating better model performance. Different ways of classifying treatment nodes may substantially affect NMA results. Researchers should investigate the results’ robustness to different ways of classifying treatments.
|
Authors who are presenting talks have a * after their name.
