Statistical Analysis

Sample size calculations and power analysis

120

Sample Size

80%

Statistical Power

r=0.45

Effect Size

96

Expected Completers

Power Curve
Statistical power vs sample size (r=0.45, α=0.05)
20304050607080100120140160180200Sample Size (n)0255075100Power (%)80%n=120
Sample Size Justification

Primary Analysis

With n=120 and expected effect size r=0.45, we achieve >80% power for the primary correlation analysis.

Subgroup Analyses

With ~60 patients per sex and ~60 per intervention, we maintain ~90% power for sex-stratified and intervention-stratified analyses.

Dropout Adjustment

Assuming 2400% dropout, we expect 96 completers from 120 enrolled.

Statistical Parameters
ParameterValueDescription
Sample Size (n)120Total enrolled participants
Effect Size (r)0.45Expected correlation coefficient (conservative estimate)
Alpha (α)0.05Type I error rate (two-sided)
Power (1-β)>80%Probability of detecting true effect
Expected Dropout2400%Anticipated loss to follow-up
Recruitment Rate2.2 pts/mo/centerExpected enrollment rate per center
Primary Endpoint

Correlation between baseline DIT (from LPMT) and % total body weight loss at 12 months

Discussion: Statistics

Sign in to join the discussion

Sign in

No comments yet. Be the first to comment!