How do you optimize a slow-running SQL query that processes millions of rows of patient records at Medpace?
