PRIMARY BILIARY-CIRRHOSIS - AN ELECTRONIC DIAGNOSTIC-TOOL TRIAL(C) BASED ON SYMPTOMS, (PAST) HISTORY AND SIGNS ONLY, USING THE EUROPEAN DATABASE EURICTERUS
Y. Reisman et al., PRIMARY BILIARY-CIRRHOSIS - AN ELECTRONIC DIAGNOSTIC-TOOL TRIAL(C) BASED ON SYMPTOMS, (PAST) HISTORY AND SIGNS ONLY, USING THE EUROPEAN DATABASE EURICTERUS, Hepato-gastroenterology, 44(16), 1997, pp. 1104-1109
Background/Aims: Primary Biliary Cirrhosis (PBC) is a relatively rare
chronic progressive disease in which a working diagnosis of PBC easily
leads to a final diagnosis by testing for anti-mitochondrial antibodi
es. Liver transplantation is the only effective treatment. The aim of
this study was to test an electronic diagnostic tool (tool) for it's a
bility to include PBC in the working differential diagnosis. Methodolo
gy: In the European Union Euricterus project a large number of (sub)ic
teric patients in 17 discrete disease categories, PBC being one of the
m, were gathered prospectively. A tool was developed using Bayes (B) a
nd Trial Algorithm (TA) pattern-recognition and based on items related
to the history, symptoms and signs of all Euricterus patients. We hav
e tested the diagnostic tool on 143 PBC Euricterus patients. Results:
PBC was mentioned by the tool in 86% (B) and 91% (TA) of the 143 patie
nts. These figures were higher for patients under 60 and (TA only) fem
ales. Females under 60 (n = 89) scored 92% B and 96% TA. A sole diagno
sis of PBC was made in 31% (B) and 66% (TA). In the other patients wit
h a PBC probability, 7 other (first) diagnoses were presented by the t
ool of which non-alcoholic active liver disease and pancreatic or bili
ary carcinoma were the leaders. These 7 diseases appeared evenly distr
ibuted along the percentual probabilities of PBC given by the tool (B)
and also along Pugh and Mayo scores (B and TA). PBC was mentioned by
the tool in all patients with a Pugh score 10 or higher (advanced dise
ase, class C). In the patients in whom the tool did not mention PBC, t
he primary diagnoses came from 9 other disease categories. Conclusion:
This electronic tool has been able to identify PBC as one of the diff
erential diagnostic modalities in the large majority of a preset popul
ation of PBC patients.