Home - Knowledge
Center - Healthcare Technologies
- Computer Software
Decision Support Systems
Decision support systems (DSS) have been around since the
beginning of the era of distributed computing. The first decision support system made its
appearance in the mid to late 1960s and can now be found in almost all
industries where information systems are used.
Decision support systems are increasingly being used in healthcare, where doctors,
for use during their consultations, design some while others are aimed at the
wider industry for not only doctors, but also other healthcare professionals
and patients.
advertisement
These decision support systems generally provide two types of support:
-
Diagnostic Support: - Here systems provide support concerning diagnosis or
prognosis. They provide outcomes that reduce the uncertainty concerning the
patient’s current or future situation.
-
Management Support: - Systems provide support by providing suggestions on how
best to manage a patient’s condition. Some of the suggestions might involve
tests that have to be carried out, what medication or treatment should be
considered, sometimes with financial and ethical considerations taken into
account.
Decision support systems aid clinicians in applying new information to patient care
through the analysis of various patient specific data and enhance diagnostic
and management outcomes.
Decision support systems operate in three modes – active (systems triggered
automatically and make decisions without any intervention), semi active (raise
reminders and alarms according to the users input) and passive (where the user
must make an explicit request to the system in order to gain advice).
Whatever mode the decision support system operate at, it must provide accurate and
reliable data, which is retrieved from a knowledge base. The knowledge base is
made of several sources of information from various medical disciplines, which
might include patient observations, medical books and journals, and the medical
experience of several physicians.
The outcomes that are derived from the knowledge base are
usually modelled on the following:
- Mathematical models: are used to describe complex
biological or physiological systems
- Statistical models: are mainly based on
multi-dimensional classification systems. Some of the statistical tools they
make use of are multiple regression and discriminate analysis.
- Bayesian networks: are probability-based models and
essentially use Baye’s Theorem, which provides a mathematical model to update
probabilities on the basis of new information.
- Artificial Intelligence (AI): has been used to develop
expert systems, which are used mainly in specialized domains to provide
functions that would have normally been done by a human expert. The terms
“expert systems” and “decision support systems” have often been used
interchangeably.
- Neural networks: are made of rules that are linked by
arcs, mimicking the structure and operation of the human brain, hence its name.
Information is processed by being propagated from an inner layer, through
intermediate layers to an outer layer.
Sample Decision Support Systems
-
DXplain –
Decision support system that uses a set of clinical findings to produce a
ranked list, which might help explain or be associated with the clinical
manifestations. DXplain is owned by Massachusetts General Hospital.
-
QMR (Quick Medical
Reference) – Developed by the University of Pittsburgh, Pennsylvania
and First DataBank, Inc., QMR is a diagnostics decision support system with a
knowledge base of diseases, diagnosis, findings, disease associations and
laboratory information.
-
PRODIGY – Developed at
the Sowerby Centre for Health Informatics Newcastle (SCHIN) and funded by the
National Health Service (NHS) in the United Kingdom. PRODIGY is a computerised
prescribing decision support system for General Practice.
-
HELP system – HELP stands for Health Evaluation through Logical Processing. It
was developed at the Latter Days Saint (LDS) Hospital in Salt Lake City, Utah,
and is integrated into a hospital information system. It takes laboratory and
dosage data and generate alerts and warnings, usually when a patient’s records
are updated.
|
|
|