Random blinded rechecking of AFB smears in a pilot project at an intermediate reference laboratory.

BACKGROUND: In most developing countries, sputum smear microscopy for acid-fast bacilli remains the front line and often the only diagnostic tool for the diagnosis of tuberculosis (TB), making quality assurance of smear microscopy an important activity. OBJECTIVE: To evaluate the results of a pilot study, where the random blinded rechecking for the entire state of Delhi was conducted at a reference laboratory. METHODOLOGY: Slides from 25 Revised National Tuberculosis Control Programme designated districts (200 peripheral microscopy centres) in Delhi were re-read after proper coding by all the Senior Tuberculosis Laboratory Supervisors (STLS) at an intermediate reference laboratory under proper supervision. RESULTS: Of 12,162 re-read slides, 204 discrepant results were found. Of these, 150 (73.5%) errors were attributed to the peripheral microscopy centres and 54 (26.5%) to STLS. High false-positive errors were observed at a frequency of 12/150 (8%), and high false-negative errors at a frequency of 38/150 (25%). Minor errors, i.e., low false-negative, low false-positive and quantification errors, were observed at frequencies of respectively 68/150 (45.3%), 17/150 (11.3%) and 15/150 (10.0%). CONCLUSION: Greater stringency in the supervision of random blinded rechecking at the district level is essential to make smear rechecking more efficient and effective.