Data Synthesis
Data were summarized narratively and presented in tabular format. A
meta-analysis was done with the meta package in the R program (Vienna,
Austria). A forest plot was used to visualize the proportion of
gastrointestinal manifestations in each study and the overall pooled
prevalence estimates. The fixed effect model will be used to derive the
pooled prevalence estimates, assuming that all effect estimates in each
study estimate the same underlying effect and avoiding giving greater
weighting to studies with smaller sample sizes. The heterogeneity was
measured using the I2. Subgroup analyses were done to
compare studies based on the study location (African vs. non-African
studies) and age groups (Adult only, Mixed, Pediatrics only). Test for
subgroup differences was performed, and a p-value for interaction of
<0.1 indicates a statistically significant difference between
subgroups. The publication bias was qualitatively assessed by
visualizing the studies with a funnel plot and quantitatively using an
Egger’s test. A p-value of less than 0.05 suggests the presence of
publication bias.