The large influx of patients to hos­pi­tal emer­gency room is
con­sid­ered an inter­na­tional prob­lem affect­ing not only the ser­vice providers but
the users them­selves. Over­crowd­ing emer­gency depart­ment is asso­ci­ated with
sev­eral fac­tors such as the reduced access to other emer­gency med­ical ser­vices
or pri­mary care and this has caused sev­eral delays in care for urgent patients
and even increased mor­tal­ity.
The main dif­fi­culty in fore­cast the num­ber of users that arrive to the emer­gency
depart­ment and the large num­ber of inter­ven­ing vari­ables in this com­plex
sys­tem, leads to a dif­fi­cult man­age­ment task. It is com­mon for the man­age­ment
team to use met­rics based on empiric knowl­edge, how­ever, the use of such
method is not at all the most effec­tive. Recent stud­ies report fore­cast­ing
meth­ods based on long time series, as the most accu­rate way to pre­dict the
patients arrival in the short and medium term. This work stud­ies this kind of
meth­ods in order to opti­mize the resources, min­i­miz­ing the costs, to pro­vide
more effi­cient health care.