A crossroads in predictive analytics monitoring for clinical medicine.

A new goal for medical informatics is to develop robust tools that integrate clinical data on a patient in order to estimate the risk of imminent adverse events. This new field of predictive analytics monitoring is growing very quickly. Its claims, however, can be vulnerable when clinicians fail to use the best mathematical and statistical tools, when quantitative scientists fail to grasp the nuances of clinical medicine, and when either fails to incorporate knowledge of physiology. Its potential, though is clear: we can provide more effective clinical decision support and make better predictive analytics monitoring tools if we apply principles learned from physiology and mathematics to the right problems in clinical medicine.