Estimate of the real incidence of tuberculosis in the Leon Health Area: application of the capture-recapture method to compare two information sources]

The actual incidence of tuberculosis is probably higher than that previously published in national and international records. Under-reporting is estimated to fluctuate between 7% and 27%, according to studies.To estimate the incidence rate of tuberculosis in the area of León for 2008 and 2009 using the capture-recapture method in order to compare two sources of information: prescribed tuberculostatic drugs (combination of rifampicin-isoniazid) and the regional epidemiological surveillance system register (SIVE).Retrospective descriptive study in an area of 351,086 inhabitants of tuberculosis cases using as sources: (i), information on prescribed tuberculostatic drugs, and (ii), the SIVE register. We calculated incidence rates for each source by the capture-recapture method. We analyzed epidemiological and demographic data, symptoms, diagnosis, treatment and follow-up.The incidence based on the SIVE data for 2008 was 18.80/100,000 inhabitants and according to the pharmacy register, the rate was 26.77. The estimated value for 2009 based on the SIVE data was 18.23/100,000 inhabitants, and according to the pharmacy register, it was 22.50. After applying the capture-recapture method, the annual incidence for 2008 was 44.14/100,000 (95% CI; 37.88-50.41) and for 2009, it was 34.17/100,000 (95% CI; 30.19-38.17). In the study of all these years we have found that the number of cases were higher in the pharmacy register than the SIVE one.The SIVE data on the incidence of tuberculosis in our study area underestimates the actual incidence rate. The source of information that involves case record of tuberculosis in the community is under-used. The capture-recapture method is a good alternative to measure the incidence of tuberculosis, and to check the surveillance systems.