Dermatological disorders, although not
terminal, cannot be easily treated and
are generally a greater problem socially.
Usually the disease is contained rather
than cured and therefore relapse is a
major problem. Treatment of microbial
skin diseases is multi-factorial. The onset
of diseases such as acne is dependent on
many factors, including age, sex, sebum
production and hormonal changes. Anti
androgens can be as effective a
treatment as anti-microbial agents: one
targets the sebum, the other the
organism. It is a well established fact
that with a reduction in the sebum
excretion rate, there is a reduction in
the number of Propionibacterium acnes,
(P. acnes) which correlates with an
improvement in acne. It has been shown
that an increase in the amount, and
changes in the composition of skin
surface lipids appear to be directly
related to the increasing population of
P. acnes around puberty.
Initial research involved both
collection over time and analysis of data
using traditional statistical methods.
Due to the large volume of sparse data it
was time consuming and difficult to
integrate and implement the methods
to support the decision making process.
This paper illustrates the value of Object-
Oriented technology and Artificial Neural
Networks in building clinical decision
support systems to analyse skin surface
lipid data from patients with lipid
dependent microbial skin diseases. The
paper proposes an extension to Blum's
framework for analysing data and
postulates an architecture for clinical
decision support systems.
Get full access to this article
View all access options for this article.
References
1.
Blum AL, Rivest RL1992Training a 3-node neural network is NP-CompleteNeural Network, 5 (1) 117-127
2.
Dreiseitl S., Wang D.1993Automatic generation of C++ code for neural network simulation International Workshop on Artificial Neural Networks ProceedingsSpringer-Verlag, 358-363
3.
El Emam K., Khalafalla F., HoptroffR., Hall T.1991Object oriented neural networks International Conference on Control (Publ No 332) IEELondon, UK, 1007-1010
Graham I.1994Object-oriented methods (2nd Ed) Addison Wesley, UK
8.
Hanley D.1994Object oriented techniques for embedding neural networks in medical decision support systems Dissertation for MSc in Computer Science , South Bank University, London , UK
9.
Hecht-Nielsen R.1987Counterpropagation Networks Proceedings of the IEEE 1 st International Conference on Neural NetworksSOS Printing, San Diego, 2 9-32
Mortensen J.E., Shryock T R, Kapral F A1992Modifications of bactericidal fatty acids by an enzyme of Staphylococcal aureusJ Med Microbiol, 36293-298
13.
Nordstrom Nkm1983Ecology of cutaneous propionibacteria PhD thesis presented to the University of London, UK
14.
Patel SD, Noble WC1993Analyses of skin surface lipid in patients with microbially associated skin disease. Clin Exp Dermatol, 18405-409
15.
Patel SD1991Skin surface hpids in microbial skin dieases. PhD Thesis presented to the University of London, UK
16.
Rumelhart D., McClelland J.PDP research group1986Parallel distributed processingMIT Press, Vol 1
17.
Segev A., Shosham A.1988The representation of a temporal data model in the relational environment In Proceedings of the 4th Confon Statistical and Scientific Database Management, 39-61
18.
Sun Y.1994Extending RDBMS to support temporal dataProceedings of the 1994 IEEE Basque International Workshop on Information TechnologyToulouse, France, 345-355
19.
Vondrák I.1994Object oriented neural networksAI Expert , June
20.
Wasserman PD1989Neural computing theory and practiceVan Nostrand Reinhold, New York
21.
Weitzenfeld A., Arbib MA1991A concurrent object-oriented framework for the simulation of neural networksOOPS Messenger2 (2), 119-24
22.
Whittington G., Spracklen CT1991Object oriented design methodologies for the implementation of artificial neural networksIEE colloquium on 'neural networks design techniques and tools' (Digest No 037) IEE London, UK
23.
Wright S.1991Essential fatty acids and the skinBnt Jour Dermatol, 125503-515
