|Year : 2015 | Volume
| Issue : 4 | Page : 572-575
How better is random blinded re-checking results in revised national TB Control Programme, India?
BN Sharath1, AMV Kumar2, R Ranjini3, S Anand4, H Sundaram5, SK Singh4, P Kumar5
1 Office of World Health Organization Representative to India, New Delhi and Employee's State Insurance Corporation Medical College and Post Graduate Institute of Medical Science and Research, Bangalore, Karnataka, India
2 International Union Against TB and Lung Disease, South East Asia Office, Qutub Institutional Area, New Delhi, India
3 Office of World Health Organization Representative to India, New Delhi, India
4 Central TB Division, Nirman Bhavan, Ministry of Health and Family Welfare, New Delhi, India
5 National TB Institute, Bangalore, Karnataka, India
|Date of Submission||09-Jun-2014|
|Date of Acceptance||23-Feb-2015|
|Date of Web Publication||16-Oct-2015|
B N Sharath
Office of World Health Organization Representative to India, New Delhi and Employee's State Insurance Corporation Medical College and Post Graduate Institute of Medical Science and Research, Bangalore, Karnataka
Source of Support: None, Conflict of Interest: None
Background: The Revised National Tuberculosis Control Programme (RNTCP) is implementing the External Quality assurance (EQA) and Random blinded re-checking (RBRC) as one of its important component. This nationwide study was conducted to determine (1) the number and types of RBRC errors and (2) the sensitivity and specificity among rechecked slides. Materials and Methods: The study was based on the monthly RBRC reports submitted by ~13,000 designated microscopy centres (DMCs) across the country under routine programmatic settings in 2010. The DMCs reports were compiled at district, state and national level. Results: A total of 11, 89,564 slides were rechecked from 11,039 DMCs. Of which 99.5% of rechecked slides did not have any errors. The sensitivity and specificity of the rechecked slides had 98% sensitivity and 100% specificity. Conclusion: RBRC is the crucial component of EQA and the results from the programme are found to be satisfactory. Based on the study findings, the earlier value of 80% sensitivity used for calculation of annual sample size for RBRC has been increased to 90% sensitivity. The annual RBRC sample size for DMCs has been increased by 1.5–2 folds.
Keywords: External quality assurance, random blinded re-checking, Revised National Tuberculosis Control Programme
|How to cite this article:|
Sharath B N, Kumar A, Ranjini R, Anand S, Sundaram H, Singh S K, Kumar P. How better is random blinded re-checking results in revised national TB Control Programme, India?. Indian J Med Microbiol 2015;33:572-5
|How to cite this URL:|
Sharath B N, Kumar A, Ranjini R, Anand S, Sundaram H, Singh S K, Kumar P. How better is random blinded re-checking results in revised national TB Control Programme, India?. Indian J Med Microbiol [serial online] 2015 [cited 2019 Dec 8];33:572-5. Available from: http://www.ijmm.org/text.asp?2015/33/4/572/167318
| ~ Introduction|| |
In India, the Revised National Tuberculosis Control Programme (RNTCP) is being implemented adopting the World Health Organization Directly Observed Treatment, Short-Course (WHO DOTS) strategy since 1997. Sputum smear microscopy is the basis for diagnosis of tuberculosis under the programme. For diagnosis of TB, sputum microscopy is performed by trained laboratory technicians (LT) at designated microscopy centres (DMCs) using Ziehl-Neelsen staining technique.
In 2010, there were nearly 13,000 DMCs (one DMC for every 100,000 population) in the country. Due emphasis is provided on well-functioning laboratory network that provides high quality smear microscopy services which is implementing the revised WHO External Quality Assurance (EQA) programme of Lot Quality Assurance System (LQAS) since 2005. Poor quality microscopy services have serious implications on patients and programme which includes failure to detect patients with infectious TB who continue to spread infection in the community, or leading to unnecessary treatment for 'non-cases'. Errors in the reading of follow-up smears may result in patients being placed on prolonged treatment, or treatment being discontinued prematurely. The success and sustainability of the TB programme depends on the effective monitoring of the laboratory network for quality implementation.
To ensure good quality laboratory services, different quality assurance mechanisms for laboratory network are in place. Essentially, it includes EQA which is a process to assess laboratory performance and includes on-site evaluation of the laboratory to review quality control (QC) and evaluation of entire process of smear microscopy, and random blinded re-checking of routine smears. EQA is also termed 'proficiency testing' as described by International Union Against Tuberculosis and Lung Disease (IUATLD). The EQA activity is supervised and conducted by a programme staff specially recruited for this activity called senior TB laboratory supervisor (STLS); they are responsible for supervision and monitoring of EQA activity in a population of 500,000 under the supervision of district tuberculosis officer.
Since, there was no information available on the performance of random blinded re-checking (RBRC) implementation in RNTCP. The study was undertaken to determine (1) the number and types of RBRC errors (2) the sensitivity and specificity among rechecked slides.
| ~ Materials and Methods|| |
It is a retrospective study based on review of EQA records and reports in RNTCP.
The study is based on the RBRC reports of 11,039 DMCs spread over 27 states in the country. The network of DMCs is supervised by larger regional laboratories (Intermediate Reference Laboratories or IRLs), and National TB Reference Laboratories (NRLs). The NRLs work closely with the IRLs, monitor and supervise the IRL's activities.
Random blinded re-checking of routine slides
Blinded rechecking is a process of rereading a statistically valid sample of slides from a laboratory to assess whether that laboratory has an acceptable level of performance. A random and representative number of slides from every laboratory are rechecked at a higher level by a controller who does not know the original results of the laboratory. Discordant smears, (positive at the laboratory and negative by the controller, or the reverse) are rechecked by a second controller who serves as the gold standard. The errors are classified based on the variance in agreement of the results of DMC LT and results of the controller. The errors are classified as high false positive (HFP), high false negative (HFN), low false positive (LFP), low false negative (LFN) and quantification errors (QE). The [Table 1] shows classification of errors; the errors like low false positive, low false negative and quantification errors are considered to be errors of milder nature when compared to high false positive and high false negative errors.
The sample size for each laboratory is based on a modified statistical sampling method called Lot Quality Assurance Sampling (LQAS), this acts like a filter to separate acceptable laboratory standards from unacceptable quality. LQAS uses the smallest required numbers of negative smear or rechecking to indicate that a selected parameter (namely the number of false negatives) has not been exceeded in the sample with 95% confidence limit. The Sample size determination is based on the annual negative slide volume, sensitivity, specificity and critical value. Under the programme, an overall sensitivity of 80%, specificity of 100% and an acceptance number of zero has been selected [Table 2].
|Table 2: Recommended Annual Sample size with 80% sensitivity and 100% specificity and ‘0’ acceptance number|
Click here to view
The variables studied included annual slide volume, annual positive slides, number of slides rechecked, number and type of errors in slides rechecked that is, HFP, HFN, LFP, LFN and QE were collected.
The data was entered in the excel sheet and was analyzed for percentages and proportions. The sensitivity and specificity of the slides examined relative to the controller are calculated. Sensitivity is the proportion of 'true positive' slides; while specificity is the proportion of 'true negative' slides detected by the laboratory technician relative to the controller. The new annual sample size is calculated based on the standard reference tables for the required sensitivity value and acceptance number.
Sources of data
The monthly RBRC report from each DMC for a year (2010) was consolidated at districts; and all the district reports were compiled at State/IRL. The data from 27 IRLs were compiled at central TB Division.
| ~ Results|| |
The analysis of the reported data revealed that as part of the EQA mechanism 11, 89,564 slides were rechecked in 11,039 DMCs across the country [Figure 1]. Among the slides re-checked there were 610 HFP errors, 1508 HFN errors, 713 LFP errors, 1131 LFN errors and 2274 QE. The sensitivity and specificity for the above data was calculated to be 98% and 100% respectively.
|Figure 1: Number of slides and errors detected in routine RNTCP EQA implementation, 2010|
Click here to view
| ~ Discussions|| |
This is the first study to analyze the national programmatic data for random blinded rechecking; with 98% sensitivity and 100% specificity the results of EQA are found to be satisfactory. Nevertheless, it has following programmatic implications.
First, only 0.5% of the slides rechecked (6236/11, 89,564) had any errors which signifies that the EQA is being effectively implemented under the programme. Nearly 36% of the errors reported were quantification errors which are milder errors that are usually expected routinely under programmatic conditions. There were approximately 34% high false errors which would have necessitated the programme to build the capacity of the laboratory technicians by subjecting them to re-training exercises.
Second, the 98% sensitivity among the slides rechecked has impact on the annual sample size for RBRC. According to RNTCP EQA guidelines (2005), the sample size drawn for RBRC is based on LQAS sampling technique with 80% sensitivity, 100% specificity and '0' acceptance number. It was also stated that acceptable sensitivity for the subsequent years of EQA implementation will be determined by the RNTCP based on the results from field. With 98% sensitivity among the rechecked samples now it is deemed appropriate for the programme to increase the sensitivity to 90% with other parameters remaining the same to calculate the newer annual sample size for RBRC. The newer sample size is expected to increase the DMC monthly RBRC sample size by 1.5–2 folds [Table 3].
|Table 3: Newer annual sample size based on 90% sensitivity, 100% specificity and ‘0’ acceptance rate|
Click here to view
Third, the operational realities which determine the RBRC activity cannot be ignored (a) The senior TB laboratory supervisors (STLS) and 20–50% of laboratory technicians under RNTCP are contractual positions and one of the criteria for their yearly contract renewal is based on RBRC performance. Hence, there may be a possibility of nexus among the contractual staffs to bury the true findings (b) Blinding of RBRC slides which is the responsibility of the district programme managers may not be strictly followed at all District TB centres (c) Sub-optimal supervision and monitoring of RBRC activities by the IRLs/NRLs.
The above study findings were presented to the 20th National laboratory committee (2011), Central TB Division, Ministry of Health and Family Welfare, Government of India whose mandate is to review and guide the programme to take appropriate policy decisions. The committee gave its consent and approved the programme to adopt newer annual sample size table to calculate the number of slides to be re-checked in a DMC for routine RBRC implementation.
| ~ Conclusion|| |
EQA is crucial for the programme and RBRC is a powerful tool to measure the quality of EQA implementation. Online reporting mechanisms and newer strategies to encourage prompt reporting are to be adopted. Further studies with different newer blinding procedures have to be conducted to maximize the effectiveness of EQA activity.
| ~ References|| |
Central TB Division, DGHS, MoHFW. RNTCP Laboratory Network; 2005.
World Health Organization. External Quality Assessment for AFB smear microscopy; 1999.
Minutes of 20th
National Laboratory Committee meeting. Available from: [www.tbcindia.nic.in
]. [Last accessed on 2014 Feb 04].
[Table 1], [Table 2], [Table 3]