Text this: Overcoming challenges in prevalence meta-analysis: the case for the Freeman-Tukey transform