ONCOCIN is an expert system, a clinical decision support system that was developed in 1979 at Stanford by Shortliffe's group and used primarily from 1981 to 1987 at the Stanford Oncology Clinic as well as various other locations.1 ONCOCIN, arose from an effort to increase the explanation producing power of an existing expert system.2 This system was the successor of MYCIN and predecessor of the Protégé and Eon systems. ONCOCIN uses artificial intelligence techniques to offer advice to the physician on medicines, dosages, and testing; in this process it integrated medical record keeping with decision support. It can determine these drug doses on the basis of time schedule, toxicity, and blood counts.5 It was designed to aid the physician in decision-making by combining clinical data with chemotherapy protocol guidelines and knowledge provided by expert oncologists. To be clear on terminology used within ONCOCIN's definition, oncology is the branch of medicine dealing with the physical, chemical, and biological properties of tumors, including study of their development, diagnosis, treatment, and prevention; and a protocol in the sense used here is a plan for a course of medical treatment. The typical users of ONCOCIN were residents and clinical assistants rather than certified physicians. When the system was being implemented there were high hopes for its use because in oncology the knowledge was already formally documented. Unfortunately, the system did not live up to expectations after the system was put to use. In some cases the situations at hand did not fit into the rules known by the system, also it took about six weeks to enter the rules for a new protocol and to test them.1 Although the knowledge was documented it was not all-inclusive; new protocols were being found and used all the time, there was no way of ever getting a complete set of the protocols. The advice provided by ONCOCIN was approved by experts in only 79% of the cases.6 ONCOCIN used the same rule-based approach as MYCIN.
It gave rules such as:
To determine the attenuated dose for dugs in MOPP chemotherapy or for all drugs in the PAV chemotherapy:
IF 1. This is the start of the first cycle after a cycle has been aborted, and
2. The blood counts do not warrant dose attenuation
THEN
Conclude that the current attenuated dose is 75% of the previous dose.1
ONCOCIN ran on Xerox personal workstations (See image above), which contained four major components: graphic display, keyboard, mouse, and the processor box.4 Initial versions of the system functioned as an expert system that produced plans that essentially consisted of a set of drugs and dosages. The users of ONCOCIN were annoyed at having to override the system's advice when they did not agree with the generated treatment plan.2 The initial version of ONCOCIN was improved when OPAL transformed the process of inputting information into the system. With the use of OPAL inputting the information now consisted of filling in structured forms and drawing flowcharts on the computer.6 Also, protocols could now be added within two weeks. The physicians now entered the data directly into the database using the mouse to select values from menus instead of manually.
Examples of some ONCOCIN menus:
At this point ONCOCIN was one of the first decision support systems to attempt to model decisions and sequenced actions over time using OPAL as a customized flowchart language.4 The history of prior events was very important to this system. Even with these improvements due to OPAL, physicians were wary to trust the systems advice, and moreover the advice was still only approved by experts in about 80% of the cases. For these reasons, physicians used ONCOCIN as a critique to their own work. ONCOCIN was converted into an embedded critic: rather than use the system primarily to generate treatment plans, doctors were intended to routinely enter their own plans into ONCOCIN and the system offered criticism as a side benefit.2 They would input the information and look at the significant differences between the plan proposed by ONCOCIN and their own. Soon new systems such as Protege and Eon made the use of ONCOCIN obsolete.
References:
1 http://clinicalinformatics.stanford.edu/scci_seminars/slides/SCCImusen.pdf
2 http://argouml.tigris.org/docs/robbins_dissertation/diss3.html
3 http://www.jco.org/cgi/content/abstract/3/10/1409
4 http://www.amia.org/pubs/symposia/D004291.PDF
5 http://www.ncbi.nlm.nih.gov/books/bv.fcgi?rid=cmed.section.42000
6 Wiederhold, G., Shortliffe, E.H., Fagan, L.M., Perreault L.E. Medical Informatics: Computer Applications in Health Care and Biomedicine. New York: Springer, 2001.