Data Documentation and Metadata: Orientation

A dataset without documentation is fragile. Even if the data are accurate, future users may not understand what variables mean, how values were coded, what population is represented, which records were excluded, which dates were derived, or what missing codes mean. Clinical research data are often used long after collection ends. They may support manuscripts, audits, secondary analyses, data sharing, regulatory review, or future pooled analyses. Documentation makes these uses possible.