[New cohort analysis system in new tuberculosis surveillance system in Japan]

To know factors to influence treatment outcome of new cohort analysis method in revised TB surveillance system and important points for quality improvement of the system using hospital based real data of TB patients.

To analyze treatment outcome of new sputum smear positive TB patients hospitalized to Fukujuji Hospital during 2004 year by new cohort analysis method.

One hundred and ninety-four TB patients were hospitalized. Out of them, 166 were new cases. Cohort analysis showed 104 treatment success cases (62.7%), 27 died cases (16.3%), 2 failure cases (1.2%), 9 defaulter cases (5.4 %), 15 transfer-out cases (9.0%), 7 cases with treatment longer than 1 year (4.2%), and 2 other cases (1.2%). Among 27 died cases, 18 cases were due to TB death. Out of other 9 cases, 4 were due to malignancy, 3 due to pneumonia, and 2 other causes. Out of 9 defaulter cases, 6 were self-interruption, 2 were due to medical doctor's decision to resolve side effects. Out of 7 cases with treatment longer than 1 year, half were due to drug resistance and another half were due to side effects. Twenty-eight retreatment cases showed 15 treatment success, 4 failure cases, 5 transferred-out, 2 cases with longer treatment than 1 year, and 2 other cases.

To evaluate TB treatment outcome, died cases should be categorized into TB death and non-TB death. Defaulter cases and cases with treatment longer than 1 year should be categorized by causes into drug resistant cases and cases with interruption by side effects. At national level, data collection of drug sensitivity test results and development of cohort analysis method for drug resistant cases, especially multi-drug resistant cases, are needed to make new cohort analysis method more relevant to TB treatment outcome.