[UCR] Decision Support Use Case

[UCR] Decision Support Use Case

 

Approval of Section 2.2 Decision Support:  No.

 

I think this text is not okay unless changes are made.  A general
problem is that it's not possible to tell where and how the RIF is
functioning to assist decisions in this use case.  The description of
MEDIC suggests an application that seems to involve much more than
exchanging rules, though rule sharing might be a pre-requisite (see
emphasis below):  

 

"The rules in MEDIC have been automatically generated [how?] using the
rules [and presumably something more, to derive new information - dln]
for product and medical event compatibility from other knowledge bases
(e.g, more general knowledge about effects of drugs on kidneys, drugs'
clearance time, etc.)."

 

We need to understand how the RIF is used in the creation, maintenance,
and/or use of the MEDIC system.   

 

It's not clear that the RIF is used at all by the other two
decision-support systems mentioned in the use case, viz., the expert
system that evaluates/diagnoses the patient's brain scan ("an expert
system that helps interpret brain scans"), and the e-learning software
used by residents.

 

In addition, it would be helpful to have an expanded introductory
paragraph describing different general types of decision support and
ways in which the RIF could facilitate it.  One aid to decision support
is already covered by the first use case (2.1), Information
Integration.  Use cases 2.3 and 2.4, and elements of others, also
describe uses of the RIF in applications that help people make better
decisions.  Perhaps the following are some ways to characterize
distinctive uses of the RIF in decision support, which might be
exemplified in the use case.  A Rule Interchange Format may aid
decision support applications by:

 

            1. Improving situation assessment - e.g., diagnosing and
prescribing for Bob.  Another place where RIF might be useful in this
example would be if the modifications of Bob's drug regime had been
suggested to his doctor by a diagnostic program that compared Bob's lab
results to guidelines for the use of pharmaceuticals.  If the drug
manufacturers encoded their guidelines, contraindications, clearance
times, etc., in RIF or a format that was mapped to RIF, then that
information could be made available both to a diagnostic program that
helps Bob's physician update the patient's medications, and to an
application like MEDIC which relates drug information to events such as
MRIs or surgery.  

            2. Improving situation monitoring - e.g., enriched handling
of ICU patients' real-time data.

            3. Improving medical event planning - e.g., scheduling
Bob's MRI.  This could be viewed more generally as plan assessment.
(It would be more useful if the hospital's system automatically checked
physician-prescribed courses of action against something like MEDIC,
rather than relying on the technician to check.)

            4. Assisting professional/technical education - e.g.,
e-learning, and also pushing information to physicians and technicians
as it becomes available from manufacturers of drugs and equipment, from
journal articles and from evaluations and/or policies.  In the case of
e-learning, it might be emphasized that data and rules describing a
learner's profile might interact with rules about the pedagogical
presentation of information in a domain to customize the learner's
experience.

 

It would also be good to close the decision-support use case with a
paragraph that summed up the uses of RIF in the example and generalized
to analogous non-medical cases.  It could also mention additional types
of cases not detailed in the example.