<s>
DXplain	B-Application
is	O
a	O
Clinical	B-Application
decision	I-Application
support	I-Application
system	I-Application
(	O
CDSS	O
)	O
available	O
through	O
the	O
World	O
Wide	O
Web	O
that	O
assists	O
clinicians	O
by	O
generating	O
stratified	O
diagnoses	O
based	O
on	O
user	O
input	O
of	O
patient	O
signs	O
and	O
symptoms	O
,	O
laboratory	O
results	O
,	O
and	O
other	O
clinical	O
findings	O
.	O
</s>
<s>
Designed	O
by	O
the	O
Laboratory	O
of	O
Computer	O
Science	O
at	O
the	O
Massachusetts	O
General	O
Hospital	O
,	O
work	O
on	O
DXplain	B-Application
began	O
in	O
1984	O
with	O
a	O
first	O
version	O
being	O
released	O
in	O
1986	O
.	O
</s>
<s>
Use	O
of	O
DXplain	B-Application
as	O
a	O
tool	O
for	O
medical	O
consultation	O
has	O
been	O
common	O
to	O
some	O
institutions	O
since	O
it	O
fills	O
a	O
gap	O
,	O
particularly	O
for	O
medical	O
students	O
in	O
clinical	O
rotations	O
,	O
that	O
is	O
not	O
adequately	O
covered	O
by	O
textbook	O
literature	O
.	O
</s>
<s>
The	O
system	O
's	O
large	O
knowledge	O
base	O
combined	O
with	O
its	O
ability	O
to	O
formulate	O
diagnostic	O
hypotheses	O
have	O
made	O
it	O
a	O
popular	O
education	O
tool	O
for	O
US-based	O
medical	O
schools	O
;	O
by	O
2005	O
,	O
DXplain	B-Application
was	O
supporting	O
more	O
than	O
33,189	O
total	O
users	O
.	O
</s>
<s>
DXplain	B-Application
generates	O
ranked	O
differential	O
diagnoses	O
using	O
a	O
pseudo-probabilistic	O
algorithm	O
.	O
</s>
<s>
Each	O
clinical	O
finding	O
entered	O
into	O
DXplain	B-Application
is	O
assessed	O
by	O
determining	O
the	O
importance	O
of	O
the	O
finding	O
and	O
how	O
strongly	O
the	O
finding	O
supports	O
a	O
given	O
diagnosis	O
for	O
each	O
disease	O
in	O
the	O
knowledge	O
base	O
.	O
</s>
<s>
Using	O
this	O
criterion	O
,	O
DXplain	B-Application
generates	O
ranked	O
differential	O
diagnoses	O
with	O
the	O
most	O
likely	O
diseases	O
yielding	O
the	O
lowest	O
rank	O
.	O
</s>
<s>
Analysis	O
of	O
accuracy	O
has	O
shown	O
promise	O
in	O
DXplain	B-Application
and	O
similar	O
clinical	B-Application
decision	I-Application
support	I-Application
systems	I-Application
.	O
</s>
<s>
In	O
a	O
preliminary	O
trial	O
investigation	O
of	O
46	O
benchmark	O
cases	O
with	O
a	O
variety	O
of	O
diseases	O
and	O
clinical	O
manifestations	O
,	O
the	O
ranked	O
differential	O
diagnoses	O
generated	O
by	O
DXplain	B-Application
were	O
shown	O
to	O
be	O
in	O
alignment	O
with	O
a	O
panel	O
of	O
five	O
board-certified	O
physicians	O
.	O
</s>
<s>
In	O
another	O
study	O
investigating	O
how	O
well	O
decision	O
support	O
systems	O
work	O
at	O
responding	O
to	O
a	O
bioterrorism	O
event	O
,	O
an	O
evaluation	O
of	O
103	O
consecutive	O
internal	O
medicine	O
cases	O
showed	O
that	O
Dxplain	B-Application
correctly	O
identified	O
the	O
diagnosis	O
in	O
73%	O
of	O
cases	O
,	O
with	O
the	O
correct	O
diagnosis	O
averaging	O
a	O
rank	O
of	O
10.7	O
.	O
</s>
<s>
Despite	O
its	O
usage	O
in	O
clinician	O
training	O
,	O
similar	O
to	O
other	O
clinical	B-Application
decision	I-Application
support	I-Application
systems	I-Application
,	O
DXplain	B-Application
has	O
not	O
expanded	O
beyond	O
the	O
research	O
laboratory	O
or	O
medical	O
training	O
setting	O
,	O
due	O
in	O
part	O
to	O
a	O
lack	O
of	O
support	O
by	O
clinicians	O
in	O
real-world	O
settings	O
.	O
</s>
