Evaluation of PROforma as a language for implementing medical guidelines in a practical context

From Clinfowiki
Jump to: navigation, search

Sutten et al was looking at a system to assist clinicians with the management of patients with hypertension. They tried to address this through improving connectivity between patients and physicians by extending the role of the community pharmacist. This study specifically looked at patients with hypertension.


There were six community pharmacies that participated in the study. The pharmacist then initially identified 250 patients that were suitable for the study and invited them to participate in the study. This project builds upon prior work evaluating computer-based decision support. In the project the pharmacies had access to computerized clinical guidelines for patient medication management. The system worked by identifying if a patients medication need to be altered and then passed that information onto the physician who has the authorization to change the prescription.


Prior reviews have shown improved quality of care by the use of clinical guidelines. However, healthcare providers have complained that hypertension guidelines are cumbersome and too many for compliance to be realistic to use. There have been a number of languages developed for representing knowledge and they are: Arden Syntax, Asbru, EON, GLIF, GUIDE, PRODIGY, and PROforma. Arden Syntax is a rule-based language for Medical Logic Modules and encodes logic as necessary for medical decision making. The rest of the languages use “Task Network Models”. This represents interaction medical decision and action in sequence or parallel over time. All of these represent “plans” which contain decisions, actions, and next sub-plans. All of the six contain expression language whose criteria influence decision and control the execution.


The creation and evaluation of PROforma was viewed as having three but overlapping phages which are: knowledge acquisition, system design, and implementation. The British Hypertension Society was the basis for the guidelines used in the computer based system for the knowledge acquisition portion. During the design phase a web-based tool lead the way for use. The requirements of the system were established through Use Case. Several things had to be decided including how frequently to monitor, when to refer to the clinicians, and when referrals were of an emergent matter. What was need to be decided were blood pressure target, deviations, and cause for immediate concern. Them system provided less information to the clinical and left the decision up to them on how to treat when compared to other systems. The web interface that was used by the pharmacist was built on Java Servelets and simple HTML pages. An SQL database and interface was used between the web pages and the PROforma engine.


The noteworthy points are considered under each of the following headings logical adequacy, heuristic power, notational convenience, and explanation support. Under logical adequacy it was noted that PROforma meet most of the requirements of the application in the different phases. The deficiencies were in implementation and concern for the lack of support for information hiding and data abstraction. Heuristic power is the ability to solve problems and draw inferences within the application. PROforma proved to be a language to encode the reasoning required. An issue that comes into play is deciding whether or not to incorporate the knowledge or not is the organization, legal, and ethical factors.


The lack of a graphical notation of the PROforma language was found to be a noteworthy weakness in terms of notational convenience. It was stated that they semantics of PROforma were better defined then UML.


PROforma was successful as part of a web-based tool and integrated with other software, most of all a patient information database was also used that was held secure. This study showed a number of limitations in logical adequacy. It was noted to be of great value with notational convenience in the choosing of the guideline which it was written. It is essential that the languages allow creation of a complete and unambiguous graphical representation of the guidelines.


Krystal Lloyd