Main Article Content
Quality assessment of sputum smears microscopy for detection of acid fast bacilli in peripheral health care facilities in Dar es Salaam, Tanzania
Abstract
Background: There is no published information regarding the quality of sputum smear microscopy in Tanzania.
Objective: To evaluate technical quality and results of smear microscopy for acid-fast bacilli (AFB) in peripheral health care facilities in Kinondoni and Ilala Districts in Dar es Salaam, Tanzania.
Design: Cross-sectional study.
Setting: All tuberculosis diagnostic centres in Dar es Salaam, Tanzania.
Results: The proportion of well prepared smears was 86.2% and that of well stained smears was 81.2%. The overall average agreement in reading was (89.2%). The overall sensitivity was 88.5% and specificity was 100%. High false negatives (HFN) were the major errors found in this study and Low false negative (LFN) and quantification errors (QE) were the minor errors found. There were no false positive errors. Minor errors occurred more frequently in hospitals than dispensaries, while major errors occurred more frequently in dispensaries than in hospitals.
Conclusions: The types of errors found in this survey, HFN, LFN and QE, suggest a systematic under-reading of smears in all the surveyed health facilities, probably due to a number of technical factors (quality of smears, poor stains, bad microscopes, or inadequate training) and other factors such as overwork and lack of motivation which need to be addressed.
Recommendations: Regular supervision using the new WHO quality assurance guidelines should be conducted countrywide. We do recommend that blind re-checking as the most efficient means of making the first broad assessment of sputum smear microscopy in Tanzania.
East African Medical Journal Vol. 83(6) 2006: 306-310
Objective: To evaluate technical quality and results of smear microscopy for acid-fast bacilli (AFB) in peripheral health care facilities in Kinondoni and Ilala Districts in Dar es Salaam, Tanzania.
Design: Cross-sectional study.
Setting: All tuberculosis diagnostic centres in Dar es Salaam, Tanzania.
Results: The proportion of well prepared smears was 86.2% and that of well stained smears was 81.2%. The overall average agreement in reading was (89.2%). The overall sensitivity was 88.5% and specificity was 100%. High false negatives (HFN) were the major errors found in this study and Low false negative (LFN) and quantification errors (QE) were the minor errors found. There were no false positive errors. Minor errors occurred more frequently in hospitals than dispensaries, while major errors occurred more frequently in dispensaries than in hospitals.
Conclusions: The types of errors found in this survey, HFN, LFN and QE, suggest a systematic under-reading of smears in all the surveyed health facilities, probably due to a number of technical factors (quality of smears, poor stains, bad microscopes, or inadequate training) and other factors such as overwork and lack of motivation which need to be addressed.
Recommendations: Regular supervision using the new WHO quality assurance guidelines should be conducted countrywide. We do recommend that blind re-checking as the most efficient means of making the first broad assessment of sputum smear microscopy in Tanzania.
East African Medical Journal Vol. 83(6) 2006: 306-310
