Strategy for data synthesis
When possible, data extracted for sensitivity and specificity will be analysed using random effects meta-analysis using hierarchical summary receiver operating characteristic (ROC) curve analysis in case of three or more studies for a given combination of index test and food.
The GRADE approach will be used to assess heterogeneity13,14, and will not solely rely on statistical methods such as the I2 statistic. Depending on the included studies, the following reasons for heterogeneity will be discussed:
The GRADE approach will be used to evaluate the certainty of the body of evidence.