For observational studies the data has often already been collected. Annually survey a few tens of thousands of people, anonymize that, store it... in 10 or 20

For observational studies the data has often already been collected. Annually survey a few tens of thousands of people, anonymize that, store it... in 10 or 20 years you go back and see how those people did. Induct a new cohort each year. The same base data can be used for an infinite number of studies.