Tuberculosis incidence rate among the homeless population: The impact of

socio-demographic andhealth-related variables

Tuberculosis (TB) is an example of a long-standing epidemic:it has evolved over

more than 2000 years and is the ninthleading cause of death worldwide and the
leading cause ofinfectious disease.1Homelessness is widely acknowledgedas a TB
risk factor: in fact, the need to address the mostvulnerable and hard-to-reach
groups --- among which thehomeless are listed --- has been recently highlighted in
anaction framework aiming at the elimination of TB in low-incidence
countries.2In Portugal, data from 2011 shows thatTB incidence among homeless
was five times higher thanwithin the general population.3This study intended to
iden-tify socio-demographic and health-related variables in thegeneral population
and among TB patients that are associ-ated with TB incidence in the homeless
population.

This was a retrospective study focused on a seven-yearperiod (2008---2014) and

on the 18 districts of mainlandPortugal. Information regarding TB incidence and
patients’characteristics, such as homelessness status (defined assomeone living in
the streets with no shelter or lacking afixed address), HIV co-infection, foreign-
born, alcohol abuse(based on subjective information --- CAGE
questionnaire4)and illicit drug abuse (considered if there are withdrawalor
tolerance symptoms, not including occasional consump-tion) was extracted from
the Portuguese TB surveillancesystem (SVIG-TB). Socio-demographic and
health-relateddata, such as total population, population density, work-ing age
population (aged 15---64 years), elderly population(aged ≥65 years), immigrant
population, unemployed pop-ulation, physicians (proportion in the population)
and HIVnotification rate, were collected from Statistics Portugal,Employment and
Vocational Training Institute and NationalHealth Institute Doutor Ricardo
Jorge.The longitudinal effect of the studied variables on theincidence rate of TB
among the homeless was estimated bya mixed-effects linear regression model
with: (1) a randomintercept taking the inter-district variability into account;(2) a
residuals variance function depending exponentiallyon the fitted
values.5Comparison between models wasbased on the likelihood ratio test for
nested models and onthe Akaike Information Criteria (AIC)
otherwise.6Statisticalanalysis was performed with the R language and
softwareenvironment for statistical computation, version 3.3.2.7TB incidence rate
in mainland Portugal has steadilydecreased from 24 cases per 100 000 inhabitants
in 2008 to17 cases per 100 000 inhabitants in 2014. However, the inci-dence rate
of TB among the homeless has not accompaniedthe same decreasing trend: it
remained approximately sta-ble from 1.2 cases per 100 000 inhabitants in 2008 to
0.97cases per 100 000 inhabitants in 2014.Our results indicate that TB incidence
rate among home-less people is associated with HIV co-infection and
alcoholabuse among TB patients and the proportion of elderly people in the
overall population. Fig. 1 describes the lon-gitudinal evolution of homeless TB
patients per 100 000inhabitants per district and the predictors found to be asso-
ciated. Table 1 describes the results obtained from the fittedregression model:
briefly, an increase of 100 HIV co-infectedcases or alcohol abusers among TB
patients is associatedwith an increase on the incidence of TB among the home-less
by an average of 14 or 11 cases per 100 000 population,respectively. The
proportion of elderly people has a smallerimpact: an increase of 1000 elderly
persons in the generalpopulation is associated with an increase on the incidence
ofTB among the homeless by an average of 3 cases per 100 000population.The
relationship between alcohol abuse and TB, aswell as between TB and the elderly,
has been previouslyacknowledged.8,9Moreover, HIV co-infection is
universallyrecognized as the single most important risk factor for TB inregions
with a high TB burden.10Importantly, both alcoholabuse and HIV infection have
been previously associated tohomeless TB patients.3Our study has a number of
strengths that ought to behighlighted: we collected data from the whole of main-
land Portugal for a period of seven years, obtained fromseveral national
institutions. On the other hand, a fewlimitations should also be acknowledged:
although TB notifi-cation is mandatory, reporting coverage may be
particularlylow among the homeless, given that this high-risk group maytend to
avoid primary care facilities, delaying or preventinga timely TB diagnosis; and
finally, as all data was extractedfrom national databases, we were limited by the
degreeof completeness and accuracy, as well as by the variablesconsidered, which
may have left out important indicators.Overall, this study allowed us to conclude
that preventiveinterventions targeting HIV co-infected and alcohol abusers among
TB patients, as well as the elderly in the general pop-ulation, may decrease TB
incidence rate among homelesspeople. In fact, although an early diagnosis and
effectivetreatment remains the cornerstone strategy, complement-ing this
approach with preventive measures targeting socialdeterminants and biological
risk factors, such as the oneshighlighted, may be useful to achieve the target of TB
elim-ination in Portugal and other low-burden countries.