Statistical models to predict adverse perioperative outcomes: A case for longer follow up time frames

Large datasets have become available and widely adopted by perioperative medicine researchers in the last decade [1–3]. The use of electronic medical records has facilitated massive data collection not only by single institutions, but also by multicenter initiatives [4–6]. Large datasets enabled the evaluation of infrequent (but important) perioperative outcome with the expectation that interventions could have been implemented to improve the care of patients undergoing surgical procedures.

http://ift.tt/2nSUxAh