Implementing the Mantel-Haenszel estimator in a Meta-analysis
The M-H estimator can be used in situations with less than 10 studies in a Meta-analysis. We will use a similar Python implementation as before, but this time, we will use a specific argument to change the estimator. If you change the setting of the algorithm to MH, we can implement the MH method which is more appropriate for this meta-analysis (having four studies):
# Define settings
settings = {
"datatype": "CATE", # 'CATE' for binary data or 'CONT' for continuous data
"models": "Random", # 'Fixed' or 'Random'
"algorithm": "MH", # Mantel–Haenszel method
"effect": "RR" # RR/OR/RD for binary data; SMD/MD for continuous data
}
# Perform the analysis...
