PIP

From Clinfowiki
Jump to: navigation, search

The Present Illness Program (PIP) system, developed in 1976, was an early diagnostic tool designed to emulate clinicians in the evaluation of patients with edema. [1] It merged facts about the patient with knowledge from a database to develop a hypothesis about what was afflicting the patient. [1]


Introduction

The system had four major components:

  • a set of patient data
  • a long-term memory, the knowledge repository
  • a short-term memory, the intersection of patient data and the knowledge repository
  • a supervisor program to filter knowledge and act on patient input.

A clinician would enter patient facts into the system. The supervisor program would then pull relevant facts into the short term memory and set them as active. Facts related to those just pulled in were marked as semi-active. Eventually the system would finish aggregating facts it would then advice the clinician in one of three ways. It would suggest more questions to help the clinician focus in on the disease in question. It would prompt the user to validate potentially spurious data. It would provide an alert of potentially spurious data as interpreted by the patient. One it has exhausted all possible actions it would then generate a hypothesis about what was afflicting the patient.

Hypothesis are generated by looking at all of the facts, brought in from the knowledge repository, best fit the case in question. Each fact is coupled with a series of rules which allow the system to determine if it is a good candidate or not. The likelihood is computed over the possible candidates and the best scoring hypothesis is reported.

While PIP may not be in high commercial use, certain aspects of its function can be seen in current daya clinical systems. Specifically it resolved around the concept of automatically using clinical knowledge repositories to infer information about the patient and take a proper course of action. Similarly, current day clinical systems can use information on patient allergens and medications to issue warnings of issuing a hazardous prescription or to generate reminders to issue useful tests that check for disease. Given this, PIP could be seen as one of the systems helping to develop the foundation of current clinical decision support systems.

References

  1. 1.0 1.1 Pauker SG, Gorry GA, Kassirer JP, Schwartz WB. Towards the simulation of clinical cognition: taking a present illness by computer. Am J Med 1976; 60:981-96. http://www.ncbi.nlm.nih.gov/pubmed/779466