Page 237 - Contributed Paper Session (CPS) - Volume 3

CPS2003 Bruno de S. et al.
lungs, and can be transmitted through the air when the bacteria is expelled by
coughing, sneezing or speaking.
From all notified cases in the WHO European Region in 2017, about 80%
had pulmonary localization (PTB) [1,2], a fact also verified in Portugal, with
73.5% of the cases in our database being PTB. An earlier study conducted in
Portugal in 2011 aimed at identifying critical areas for the joint occurrence of
PTB and HIV/AIDS (Acquired Immune Deficiency Syndrome). The study, based
on spatiotemporal clustering analyses, identified the Oporto and Lisbon
Metropolitan Areas as critical areas for both diseases, either independently or
jointly occurring [3].
Research on spatial and temporal correlations among PTB incidence rates
together with disease factors are of the utmost importance from a Public
Health perspective. This study will focus on analyzing through STAR
(Structured Additive Regression) modeling temporal trends and geographic
patterns of PTB incidence rates associated with notified PTB cases in
Continental Portugal (278 municipalities) from 2000 to 2010.

2. Methodology
2.1 The data
This study was entirely based on data from registers with the permission
of the National Program for Tuberculosis Control. The data was extracted from
SVIG-TB (Sistema de Vigilância da TB em Portugal) database of the National
Program for Tuberculosis Control and included information from all confirmed
TB cases, whose notification is mandatory in Continental Portugal (henceforth
referred to as Portugal) between 2000 and 2010. Ethics committee approval
and informed consent were not required, as data was based on an Official
National Surveillance System, provided by the General Directorate of Health,
and was previously anonymized.
A total of 25,279 new cases with PTB were used, together with the
information regarding municipality of residence, age, sex and disease risk
factors, such as alcohol dependence, intravenous drug dependence (IV Drugs),
other drug dependence, being an inmate, homeless, an immigrant and co-
infected with HIV. This study considered a new case as one defined by WHO
[1], that is, a patient with PTB disease involving lung parenchyma who has
never received a treatment or who has been taking anti-TB drugs for less than
one month. Yearly population data (global and per municipality) were taken
from Statistics Portugal.

2.2 The model
Structured Additive Regression Models (STAR) enable the placement
within the same framework of nonlinear effects of continuous covariates,
spatial effects, time trends and the usual linear or fixed effects in regression

226 | I S I W S C 2 0 1 9