Meta-analytic methods for pooling rates when follow-up duration varies: a case study

Area of Science:

• Biostatistics
• Epidemiology
• Health Services Research

Background:

• Meta-analysis is crucial for synthesizing evidence but faces challenges with count data when study follow-up durations differ.
• Accurate pooling of rate measures requires appropriate statistical methods to handle variable follow-up times.
• Clinical interpretability of meta-analysis results is vital for effective healthcare decision-making.

Purpose of the Study:

• To compare statistical approaches for pooling count data with varying follow-up durations.
• To evaluate differences in effect estimates, precision, and clinical interpretability across methods.
• To identify optimal methods for meta-analysis of rate data in pediatric asthma self-management studies.

Main Methods:

• Utilized data from a Cochrane Review on pediatric asthma self-management education.
• Analyzed school absences and emergency room (ER) visits using standardized weighted mean differences (SMD), incidence rate differences (IRD), and incidence rate ratios (IRR).
• Employed Poisson regression models to assess the impact of clustering by study and adjust for heterogeneity.

Main Results:

• All methods indicated a positive effect of the intervention for both school absences and ER visits.
• Incidence rate methods (IRD and IRR) provided more clinically interpretable results than SMD.
• Poisson regression with adjustment for clustering reduced precision; failure to account for study indicators altered effect estimates for ER visits.

Conclusions:

• The choice of statistical method impacts clinical interpretability more than overall inference.
• Incidence rate methods (IRD, IRR) are recommended for enhanced clinical interpretability in meta-analyses of count data.
• Poisson regression offers advantages for adjusting for study-level heterogeneity in such analyses.