1. Electronic health records (EHRs) are increasingly being used for biomedical research, but data quality issues can lead to misleading findings.
2. Validation of EHR data is costly and time-consuming, so researchers often validate subsamples of patient records instead.
3. Multiwave sampling strategies have been proposed to optimize resources in practice by nearly matching the efficiency of optimal sampling designs.
The article is generally trustworthy and reliable, as it provides a comprehensive overview of the current state of EHR validation sampling strategies and their potential applications in biomedical research. The authors provide evidence from previous studies to support their claims, and they discuss potential biases that could arise from using EHR data for research purposes. They also note the importance of multiwave sampling strategies for optimizing resources in practice, which is an important point to consider when designing validation studies with limited funds.
However, there are some points that could be further explored in the article. For example, the authors do not discuss any potential risks associated with multiwave sampling strategies or how these risks can be mitigated. Additionally, they do not provide any examples of how multiwave sampling has been implemented in practice or what results have been obtained from such implementations. Finally, while the authors mention functional principal components analyses (FPCAs) as a way to estimate maternal weight gain during pregnancy, they do not provide any details on how this method works or why it is preferable over other methods for this purpose.