Automated history-taking

What is it about?

Medical histories are the key to good clinical results. This will remain true for the foreseeable future despite recent advances in such disciplines as genomics. History-taking in everyday practice is limited because the process is knowledge- and time-intensive. History-taking is based on rules, however; so the process can be automated to alleviate limits imposed by the limited cognition and time of humans. This paper reviews the history of software development to collect medical histories by direct interactions between patients and computers and the current state of the art.

Featured Image

Why is it important?

Increasingly, physicians rely on molecular measurements as diagnostic and prognostic tools. There is little evidence that these have improved either task in a significant way but considerable evidence that molecular level data, as currently conceived, does not report on the integrity/disarray of complex systems that account for good health and occurrence of disease. The best current method for collecting data on the function of these systems is the patient's medical history. These data are critical too, but generally unavailable, for correlating molecular level data with normal and abnormal function. And absence of detailed, standardized histories can explain why the majority of clinical research cannot be replicated, i.e., there is insufficient history data to replicate study cohorts. Automated history-taking hence is important technology not only for good clinical outcomes but also for advancing life science research.