External quality assessment of AFB smear microscopy performances and its associated factors in selected private health facilities in Addis Ababa, Ethiopia

Tuberculosis (TB) is still a public health problem in sub Saharan African countries. In resource-limited settings, TB diagnosis relies on sputum smear microscopy, with low and variable sensitivities, especially in paucibacillary pediatric and HIV-associated TB patients. Tuberculosis microscopy centers have several weaknesses like overworking, insufficiently trained personnel, inconsistent reagent supplies, and poorly maintained equipments; thus, there is a critical need for investments in laboratory infrastructure, capacity building, and quality assurance schemes. The performance of TB microscopy centers in the private health facilities in Addis Ababa is not known so far. The main objective of the study was to assess laboratory performance of acid fast bacilli (AFB) smear microscopy and its associated factors in selected private health facilities in Addis Ababa, Ethiopia. A cross-sectional study was conducted in 33 selected private health facilities of Addis Ababa, Ethiopia comprising 7 hospitals, 2 NGO health centers, 23 higher clinics and 1 diagnostic laboratory that provide AFB smear microscopy services. The study was conducted from January to April 2014. A total of 283 stained sputum smears were randomly collected from participant laboratories for blinded rechecking, 320 panel slides were sent to 32 microscopy centers to evaluate their performance on AFB reading, staining and reporting. Checklists were used to assess quality issues of laboratories. Data were captured, cleaned, and analyzed using SPSS version 16.0; χ2 tests, kappa statistics were used for comparison purpose. P value < 0.05 considered statistically significant. Among the 32 participant laboratories, 2-scored 100%, 15 scored 80-95% & the remaining 15 scored 50-75% for overall proficiency test performance. There were 10 (3.15%) major errors and 121 (37.8%) minor errors. The sensitivity, specificity, PPV and NPV of panel reading by microscopy centers were 89%, 96%, 96%, and 90% respectively. Out of 283 randomly selected slides for blind rechecking, 11 (3.9%) slides interpreted falsely for AFB, with overall agreement of 97.5%, sensitivity of 88.4% and specificity of 99.3%. In terms of slide quality assessment, 71.6% of AFB slides were graded as good for evenness, cleanness, thickness, size, staining and labeling. The performance score for AFB slide evenness was 56.9% (161 slides) and for labeling quality was 90.8% (257 slides); having significant difference in slide quality (p value < 0.05). On-site evaluation indicated problems in terms of infrastructure, standard operating procedure, reagent quality; equipment maintenance, data management and training issues. Most of the health facilities had poor maintenance scheme for microscope (53.5%) and poor inventory management (25.0%) system. Microscopy centers that scored a proficiency of 75.5%; which is below the acceptable minimum score of 80% and an overall error rate of 3.9% for blinded rechecking needs attention. Moreover, there are gaps identified through on site assessment including poor SOP, reagent quality, equipment maintenance, data management & lack of updated training on AFB microscopy techniques, requiring a concerted effort to alleviate the bottle neck problems and strengthening the public private partnership to control TB.


Introduction
Tuberculosis (TB) is an infectious and transmissible disease caused by M. tuberculosis and occasionally byM.bovis; the main pathogenic species within the M. tuberculosis complex. It remains a major cause of human morbidity and mortality, with an estimated 9 million new cases and 1.3 million deaths per year, with most cases occurring in low-income countries [1]. In resource-limited countries, sputum smear microscopy is the routine and available diagnostic method. The technique is simple, inexpensive, and efficient in detecting those cases of pulmonary tuberculosis that are most infectious. Since its yield is highly dependent on execution of the laboratory personnel or even auxiliary, personnel performing smear.
Smears may be negative or may require even more careful screening to identify low numbers of AFB [2]. If the laboratory results of AFB are unreliable, then patients with infectious TB may not be diagnosed, resulting in ongoing transmission of disease in the community and more severe disease in the individual.
Alternatively, patients without TB may be treated unnecessarily.
Therefore, quality assurance of AFB smear microscopy is essential [3]. Accuracy and reliability of laboratory testing are critical to the success of TB control programs. It must be monitored to ensure the quality of the overall process, to detect and reduce errors, and to improve consistency between testing sites. The minimum number of AFB necessary to produce a positive smear result has been estimated to be 5000-10,000/ml of sputum. If AFB concentration is below 1000 bacilli per ml of sputum, the chance of observing the bacilli in a smear is less than 10% [4].
Tuberculosis control strategy needs to be integrated with clinicians, laboratorians, and TB-control officials and effective public-private partnerships that require prompt, complete, and accurate communication among the laboratory systems [5]. Quality assurance of sputum microscopy includes laboratory arrangement & administration, equipment, specimens and request forms, reagents, stains, staining, smear examination and reporting. Sputum microscopy is recommended in current TB control strategies due to its attractive technology for public-health programs, provides visual evidence of TB's bacterial burden, is specific enough that no confirmatory testing is needed, only tiny amounts of material are examined, as little as 0.2 ml, hence bacteria must be presented in high concentrations to be visible, typically over 10,000 AFB per ml, takes about 3-5 minutes [6]. Onsite evaluation allows observation of worker performance under actual working environment, including condition of equipment, laboratory safety, and adequacy of supplies, specimen collection, and the process for smearing, staining, reading, recording, and reporting of stained smears. When problems are detected, solutions can be suggested and potentially implemented immediately [7]. Previous work in Ethiopia focusing in public laboratories showed that , the performance of laboratories were poor in terms of reading slides and unsatisfactory reporting (below acceptable score < 80%) and major & minor errors were identified [8]. Information concerning the performance of AFB microscopy centers from the private health facilities in Addis Ababa is very scarce; hence we planned to determine the performance of the selected private laboratories using the three types of external AFB quality assessment schemes.

Methods
Study design and setting: Cross-sectional study was conducted in Addis Ababa private health facilities that provide AFB smear microscopy diagnostic services.
Sample size determination: According to health and healthrelated indicators released from federal ministry of health there are 35 private hospitals in Addis Ababa, among these 11 are on direct observed treatment (DOT) program and a report from Addis Ababa health bureau, 25 of them provide anti-TB drugs officially. We used the standard sample size determination method with the formula n = z (a/2) 2 p (1-p)/ d 2 and n was 384, and from this the finite population correction factor was used to reach on the final sample size. Since the calculated sample size from the above formula was 384, the total population was 36 (11+25); therefore, using the finite population correction factor to estimate final sample size (nf) from target population (N), the sample size was reduced to (Nf=

Participant microscopy centers
In this study, 33 private health facilities were included comprising 7 (21.2%) general hospital, 2 (6.06%) NGO health centers, 23 (69.7%) were higher clinics and 1 (3.03%) diagnostic laboratory (Table 1). All laboratories were willing to participate in the study except one higher clinic and three hospitals that refused to admit the research project for unknown reason and later substituted by higher clinics.

Panel testing
A total of 330 AFB smears of various grades of slides were used.
The response rates of participant laboratories were 97% as one health institution refused to read the panel slides after distribution.
The results were given back within one day of panel distribution.

On-site evaluation
A standardized checklist was used to assess the status of infrastructure, SOP, reagent and equipment, maintenance of microscope, biosafety and waste disposal, training related to AFB quality assessment, data and supply management. Nineteen consistent as overall proficiency result shows 75.6% whose reading agreement was very good with kappa value of 0.87 (Table 2).  EQA is a method performed to assess the performance of laboratory personnel through the proficiency testing, random blinded re checking and onsite evaluation checklist.

What this study adds
 Generated data on quality performance of TB laboratory in private health facilities which didn't addressed previously in the capital of Ethiopia.
 It identified gaps of microscopy centers concerning EQA from the private health facilities point of view that further facilitate accreditation process.
 Determined overall situation of the AFB microscopy centers.