Quality of tuberculosis microscopy in KwaZulu-Natal as determined by laboratory proficiency testing.
MetadataShow full item record
Introduction Sputum smears stained by the Ziehl-Neelsen method are the least expensive tool for diagnosing patients with infectious tuberculosis. However, false positive and false negative results have serious implications for treatment of patients. Therefore, controlling the quality of sputum microscopy services is important to ensure that the laboratory produce results that are accurate, reliable and reproducible. Aim The aim of the study was to determine the quality of tuberculosis smear microscopy in public health laboratories in KwaZulu-Natal between the years 2001 and 2006, and to assess the current knowledge and attitude of laboratory workers and laboratory managers to proficiency testing as a quality assurance technique. Methods A secondary analysis of laboratory proficiency testing results, from the KwaZulu-Natal reference laboratory (2001 to 2004) and from the National Health Laboratory Services reference laboratory (2006), was performed. Key informant interviews were conducted to determine the role proficiency testing played as a quality assurance technique. Results Overall laboratory performance was 93% from 2001 to 2004 and 98% in 2006. High false negative results were the predominant error. Sensitivity and specificity improved from 91% (for both) in 2001 to 2004 to 97% (for both) in 2006. Overall performance of primary, district and tertiary health care levels were 92%, 93% and 73% respectively in the period 2001 to 2004 and 98%, 98% and 94% respectively in 2006.There was significant (p<0.01) improvement in both urban (97%) and rural (98%) laboratory performance in 2006. The overall scores by year ranged from 89% (2002) to 98% (2006), but the annual overall scores (2001 to 2006) only achieved the acceptable level twice. Key informants indicated that proficiency testing was an essential exercise, however, they reported challenges such as inconsistent feedback, high workload and need for training. Discussion Overall performance improved from an unacceptable level of 93% (2001-2004) to a satisfactory level of 98% (2006). Likely reasons include improvement in technical skills of microscopists and improvement in preparation of proficiency testing slides. Proficiency testing is considered an essential exercise to improve laboratory performance, however, participants know that they are being tested and may give 'special attention' to proficiency testing slides resulting in a social desirability bias. Recommendations A blinded rechecking programme should be established in conjunction with the use of a standardised checklist during support visits. Feedback, communication and staff training should be improved while the workload should be evaluated.