Sin ttulo de diapositiva

1
IMPROVEMENT OF AN INTERNETS DISCOVERY SUPPORT
SYSTEM OF RECOMBINANT HYPOTHESES WITH MEDICAL
HEALTH UNIFYING PRINCIPLES
Rodolfo J. Stüsser, MD Clinical Research Centre,
Havana University, Havana, Cuba stusser_at_infomed.sl
d.cu
2
INTRODUCTION TO THE PROJECT Since the
18-century, some medical scientists have been
looking for unifying principles to make
experimental and clinical medicine a unified
science, encouraged by the model of the Newtonian
physic.
In the 20-century, the unifying principles of
informatics were discovered making the
development of personal computers and of Internet
possible. However, few investigations seeking the
unifying principles of medical and health
sciences were made, to assist in the hypothesis
formulation beside empirical data. An
important breakthrough on discovery support
systems, has been the Arrowsmith at
http//kiwi.uchicago.edu (Swanson Smalheiser,
1994-5). It helps to evoke (and assess) novel
scientific hypotheses, using the structure of
relationships in the cross-speciality knowledge
of Medline Internet database.
3
HOW DOES ARROWSMITH WORK? Arrowsmith is an
experimental set of interactive software and
database search strategies by Internet, to which
anyone can access to perceive implicit
relationships and connections that may never have
been made explicit in published form. Its
non-automated idea was born in 1986. It extends
the power of PubMed (1966-99-) conventional
searches, based on the premise that results
developed in one area of research can be of value
in another without anyone being aware of the
fact. (It could also be used with Biosis, Embase,
or Scisearch, Internet databases) Arrowsmith
unites the findings across specialities, from
words and phrases of the titles of the PubMed
records (9 million), helping the investigator to
make a heuristic selection from lists of unknown
interactions (B) and afterwards of causes (A),
starting from a given effect (C).
4
HOW DOES ARROWSMITH WORK?
LIST C EFFECTS
LIST B INTERACTIONS
LIST A CAUSES
5
The output list C of a PubMed conventional
search of the effect, is the Arrowsmith input. It
is generally a disease or process not-well known.
By probabilistic correlation and reduction
through the lists B and A, the system should
produce as output a new hypothetical factor of
risk or protection, drug or health measure. The
target-search strategies use the categories or
subheadings of Medline, allowing the finding of
the new cause, cure or prophylaxis. For instance
exogenous factors as deficiency states, dietary
factors, poisons, toxicity, drugs or intrinsic
factors as the genetic one.
Besides the etio-pathogenic, therapeutic and
preventive approaches, other applications of
Arrowsmith include anticipating adverse drug
reactions, identifying mechanisms by which agents
modulate cellular or organismal responses, and
identifying potential animal models for human
conditions.
6
PROBLEM OF THE PROJECT Although Arrowsmith uses
a robust mathematical logic method, its present
target-search strategies are still guided by
empirical and simple concepts developed only with
classification aims, from laboratory and clinical
medicine sciences, as well as from public health
sciences.
Other problem is that these target-search
strategies in use represent, each one in itself,
above all, previous hypotheses about such
categories of hypothetical risk or protective
factors, treatments, or preventive measures,
which should be found for the disease under
investigation. In addition, it is still not clear
if all the types of causes, cures or preventions,
necessarily can be subsumed under a more general
search strategy. Finally, so far these specific
search strategies only allow the discovery of
hypothetical causal correlation of the bivariate
type.
7
HYPOTHESIS OF THE PROJECT The Arrowsmiths
target-search strategies could be improved, if
the unifying principles of the sciences of
laboratory, clinic, and health, could be found.
With them the discovery of multivariate
hypotheses, including acquired and inherited
elements, within a more general search strategy,
would be possible. OBJECTIVE OF THE PROJECT
To search for the unifying principles of the
sciences of laboratory, clinic and health, which
will strengthen the semantics and syntax of the
Arrowsmith's target-search strategies.
8
THE UNIFYING PRINCIPLES WILL IMPROVE ARROWSMITHS
TARGET-SEARCH STRATEGIES
HEALTH SCIENCES
CLINICAL SCIENCES
LABORATORY SCIENCES
9
METHODS OF THE PROJECT The research will have two
main phases and methods.
Phase I. Achieving the principles. Crucial
fragmented information of multiple specialities
will be obtained through Internet facilities and
synthesised by multinational cross-disciplinary
teams of experts. An axiomatic-like research
strategy will be used to obtain the unifying
principles articulated within a general theory,
by integrating clinical, laboratory, and health
theories, through an unconventional induction.
Phase 2. Improving the strategies. The unifying
principles, partial theories, and data, using
non-orthodox ways of deductive-inductive
inference will allow the structuring of more
integral hypotheses. They will guide a better
selection of the categories to be used in the
target-search strategies, discovering
multivariate hypotheses with acquired and
inherited elements, within a more general search
strategy, making also use of the artificial
intelligence tools.
10
EXPECTED RESULTS The main unifying principles of
the medical and health sciences, and the
improvement of Arrowsmith's target-search
strategies of new hypotheses, are expected to be
obtained. They will allow an increase in their
medical specificity, rationality and fortress,
and in the output by the system of multivariate
hypothesis with acquired, genetic and other
elements, within a more general-search strategy.
Arrowsmith could also work in the future in a
more open frame, broadening the PubMed titles
with ad hoc and post hoc abstracts of at least
100 words, as well as extending the database from
1965 to 1900 or even until 1800 completing one
or two centuries of partial scientific ideas and
results of medical and health sciences. It
could also have the possibility in the future, to
work with the discussion fragments of the papers
electronically published by the Project E-biomed
Internet database proposed at http//www.nih.gov/w
elcome/director/ebiomed/ebiomed.htm (Varmus,
Lipman, Brown, et al, 1999).
11
FEASIBILITY DISCUSSION A main Kantian thesis,
used in the physical sciences for three
centuries, attempted the reconciliation of
rationalism and empiricism in the methodology of
science, by saying that while knowledge itself
comes from experience, the mind uses reason to
structure knowledge. If this thesis is true for
all the sciences, then this project will achieve
the improvement of Arrowsmith, once the unifying
principles of laboratory, clinical and health
sciences have been found. REFERENCES -Sadegh-Zad
eh K. Fundamentals of clinical methodology 2.
Etiology. Artif. Intell. Med. 199812227-270. -Sm
alheiser NR, Swanson DR. Using Arrowsmith a
computer-assisted approach to formulating and
assessing scientific hypotheses. Comp. Meth.
Prog. Biomed. 199857149-153.
12
-Stüsser RJ. Emergentist research strategy to
formulate unifying principles, and medical and
health integral hypotheses, to complement the
reductionist research program. Persp. Biol. Med.
(Under review) -Swanson DR, Smalheiser NR. An
interactive system for finding complementary
literatures a stimulus to scientific discovery.
Artif. Intell. Med. 19979183-203. -Swanson DR.
Intervening in the life cycles of scientific
knowledge. Libr. Trends 199341606-631. -Swanson
DR."Complementary Structures in Disjoint Science
Literatures". In Bookstein A, Chiaramella Y,
Salton G, Raghavan VV, editors. SIGIR 91
Proceedings of the 14th Annual International
ACM/SIGIR Conference on Research and Development
in Information Retrieval, Chicago. New York ACM
Press, 1991280-289.
13
CAN HELP
A NEW IDEA