Incidence of pulmonary tuberculosis and HIV coinfection in the province of Seville, Spain, 1998.

Using the National Notifiable Disease Surveillance System (NNDSS) and Minimum Basic Data Set (MBDS) for hospital discharges we evaluated pulmonary tuberculosis (PTB) surveillance and estimated real PTB incidence and human immunodeficiency virus (HIV) coinfection in Seville in 1998. In addition, we assessed the sensitivity and the positive predictive value (PPV) of the surveillance system. Real incidence of pulmonary tuberculosis was estimated by the capture-recapture method. In 1998, the province of Seville reported 225 cases of pulmonary tuberculosis to the NNDSS, an incidence of 13.2 per 100,000 population. Of the 225 cases reported, 18.2% presented with HIV coinfection, while a total of 194 had confirmed diagnoses. The MBDS accounted for 106 new cases. Of these, 24.8% presented with HIV coinfection and were 58% less likely to be reported to the NNDSS (25-76%, p = 0.001). Applying the capture-recapture method, 426 cases were estimated, an overall incidence of 25.6 per 100,000 population (21.5-28.8). Completeness for each source was similar (47%), and for both jointly was 72.7%. The NNDSS had a sensitivity of 65.3% and a PPV of 89.3%. In conclusion, the NNDSS underestimates PTB and PTB-HIV coinfection in Seville. The high incidence observed in young adults suggests a high degree of tuberculosis endemicity. Hospital records provide a readily accessible, low-cost means of estimating disease incidence.