Audit of the practice of sputum smear examination for patients with suspected pulmonary tuberculosis in Fiji.

BACKGROUND In Fiji, patients with suspected pulmonary tuberculosis (PTB) currently submit three sputum specimens for smear microscopy for acid-fast bacilli, but there is little information about how well this practice is carried out. METHODS A cross-sectional retrospective review was carried out in all four TB diagnostic laboratories in Fiji to determine among new patients presenting with suspected PTB in 2011: the quality of submitted sputum; the number of sputum samples submitted; the relationship between quality and number of submitted samples to smear-positivity; and positive yield from first, second and third samples. RESULTS Of 1940 patients with suspected PTB, 3522 sputum samples were submitted: 997 (51.4%) patients submitted one sample, 304 (15.7%) patients submitted two samples and 639 (32.9%) submitted three samples. Sputum quality was recorded in 2528 (71.8%) of samples, of which 1046 (41.4%) were of poor quality. Poor quality sputum was more frequent in females, inpatients and children (0-14 years). Good quality sputum and a higher number of submitted samples positively correlated with smear-positivity for acid-fast bacilli. There were 122 (6.3%) patients with suspected PTB who were sputum smear positive. Of those, 89 had submitted three sputum samples: 79 (89%) were diagnosed based on the first sputum sample, 6 (7%) on the second sample and 4 (4%) on the third sample. CONCLUSION This study shows that there are deficiencies in the practice of sputum smear examination in Fiji with respect to sputum quality and recommended number of submitted samples, although the results support the continued use of three sputum samples for TB diagnosis. Ways to improve sputum quality and adherence to recommended guidelines are needed.