From: Sample size and power considerations in network meta-analysis
Approach | Strengths | Limitations |
---|---|---|
Effective number of trials | 1. Easy and fast to calculate | 1. Only valid to the extent trial sample sizes are equal and heterogeneity is absent |
2. Lacks flexibility for approximate trial count ratios | ||
Effective sample size (ignoring heterogeneity) | 1. Easy and fast to calculate | 1. Does not account for heterogeneity |
2. Exact calculations for all trial count ratios | 2. Assumes equal meta-analysis population variances across comparisons | |
3. Sample size (no. of patients) resonates well with clinicians | ||
Effective sample size (correcting for heterogeneity) | 1. Exact calculations for all precision ratios | 1. Assumes equal meta-analysis population variances across comparisons |
2. Accounts for heterogeneity | 2. Depends on precise heterogeneity estimation | |
3. Easy to calculate | ||
4. Sample size (no. of patients) resonates well with clinicians | 1. Statistical information does not resonate well with a clinical audience | |
Effective statistical information | 1. Theoretically statistically exact | |
2. Not straight forward to calculate | ||
3. Depends on precise heterogeneity variance estimation or requires elicitation of Bayesian variance priors |