Prepare report on study design #28
Description
IRBs will approve only minimal datasets meeting the requirements of a study. My impression is that name + DOB + ZIP-5 isn't scalable for deduplicating large datasets and would not provide linkage quality sufficient for record merging. It would be great to provide researchers with a summary and language for their application along the lines of, "We truly require all these identifier fields. Here's the evidence that having fewer fields will cause the analysis to fail, and this is how we will ensure patient privacy."