Globally gridded climate products are often used as substitutes for ground-based station data in health studies. However, each product has unique strengths and limitations that require careful evaluation before use in health applications. Given its high spatiotemporal variability, this is especially true for precipitation, which is often a key driver of disease burden. It is therefore important to systematically assess global gridded precipitation products (GGPPs) for their utility in models of health impacts. Here, we compared and evaluated six GGPPs with different characteristics and spatial resolutions, including reanalysis (ERA5/-Land), satellite-based (CHIRPS, PERSIANN-CDR), and interpolated gauge-based products (CRUTS, GPCC). We provide the first comprehensive analysis specifically evaluating the appropriateness of different precipitation datasets for health research across South America. Our analysis focuses on Brazil and Colombia, two diverse countries differing e.g., in orography, climate, and size. We explored how spatial deviations and uncertainties in these GGPPs affect area-level precipitation estimates typically used in health studies, where epidemiological data are usually attributed to administrative units. Emphasis was placed on analyzing seasonal patterns and deriving extreme precipitation indices, along with disease-related bioclimatic variables. Each dataset was evaluated against national weather station data. The findings revealed substantial variation in the accuracy of local and aggregated precipitation estimates across GGPPs and indicates that estimating tropical precipitation was especially challenging for reanalysis products. Conversely, CHIRPS consistently demonstrated the best overall performance. These results underscore the importance of careful dataset selection for health impact studies.