SciELO - Scientific Electronic Library Online

 
vol.28 issue8Use of medicines by individuals with hypertension and diabetes in municipalities covered by the Pharmacy Network in Minas Gerais State, BrazilMusculoskeletal disorders among healthcare workers in Belo Horizonte, Minas Gerais State, Brazil author indexsubject indexarticles search
Home Page  

Cadernos de Saúde Pública

Print version ISSN 0102-311X

Abstract

PINHEIRO, Rejane Sobrino; ANDRADE, Vanusa de Lemos  and  OLIVEIRA, Gisele Pinto de. Underreporting of tuberculosis in the Information System on Notifiable Diseases (SINAN): primary default and case detection from additional data sources using probabilistic record linkage. Cad. Saúde Pública [online]. 2012, vol.28, n.8, pp. 1559-1568. ISSN 0102-311X.  http://dx.doi.org/10.1590/S0102-311X2012000800014.

This study aimed to analyze underreporting of tuberculosis (TB) cases in the Information System on Notifiable Diseases (SINAN), based on the following data sources: Mortality Information System (SIM), Registry and Follow-up Book for TB Case Treatment (LPATB), and Laboratory Registry Book (LRLAB). Probabilistic record linkage was used between the SIM (2007-2008) and SINAN (2002-2008). A search was conducted in LPATB and LRLAB (2007-2008) for cases not recorded in SINAN. There were 125 deaths, of which 44.8% were not recorded in SINAN. In LPATB, 58 cases (5.1%) were in treatment and were not reported in SINAN. LRLAB showed 32 smear-positive cases not reported to SINAN and without treatment, representing primary default. Addition of the retrieved cases, led to a 14.6% increase in the incidence rate in 2007 and 11.6% in 2008. Underreporting of deaths from or with TB in the Mortality Information System and primary default revealed difficulties in access to adequate and timely treatment, calling for rethinking of strategies to detect cases for timely treatment.

Keywords : Tuberculosis; Disease Notification; Information Systems.

        · abstract in Portuguese     · text in Portuguese     · pdf in Portuguese