Ensuring accurate testing for human immunodeficiency virus in Myanmar

Abstract Problem Until 2005, the quality of rapid diagnostic human immunodeficiency virus (HIV) testing was not monitored and no regular technical support was provided to hospital laboratories in Myanmar. Approach The national reference laboratory introduced a national external quality assessment scheme. The scheme involved (i) training laboratory technicians in HIV testing and in the requirements of the quality assessment system; (ii) implementing a biannual proficiency panel testing programme; (iii) on-site assessments of poorly-performing laboratories to improve testing procedures; and (iv) development of national guidelines. Local setting In 2011, a total of 422 public hospitals in Myanmar had laboratories providing HIV tests. In addition, private laboratories supported by nongovernmental organizations (NGOs) conducted HIV testing. Relevant changes The scheme was started in 65 public laboratories in 2005. In 2012, it had expanded nationwide to 347 laboratories, including 33 NGO laboratories. During the expansion of the scheme, laboratory response rates were greater than 90% and the proportion of laboratories reporting at least one aberrant result improved from 9.2% (6/65) in 2005 to 5.4% (17/316) in 2012. Lessons learnt National testing guidelines and a reference laboratory are needed to successfully implement quality assurance of HIV testing services. On-site assessments are crucial for all participating laboratories and the only source for insight on the causes of aberrant results; lessons that the reference laboratory can share nationally. Proficiency testing helps laboratory technicians to maintain HIV testing skills by ensuring that they regularly encountered HIV-positive samples.


Introduction
Early diagnosis of human immunodeficiency virus (HIV) infection is needed to ensure timely access to care and prevent disease transmission. 1 The emergence of rapid tests that detect HIV antibodies in body fluids has enabled the expansion of HIV diagnosis in resource-poor settings. However, many laboratory services remain inefficient because of a lack of equipment and technicians. 2,3 There are concerns regarding testing accuracy, quality and interpretation of algorithms. 4 Accurate HIV tests are essential for patient care and outcomes. 5,6 Others have described some of the challenges of establishing national quality assessment schemes for HIV testing services. [7][8][9][10][11] We report lessons learnt during eight years of establishing such a scheme in Myanmar.

Approach
The Myanmar Ministry of Health provides laboratory services through the national reference laboratory which began to implement quality assessment in 2005 with technical and financial support from the Japan International Cooperation Agency. The scheme included four parts: (i) training workshops for laboratory personnel; (ii) an external proficiency panel testing programme for participating laboratories; (iii) on-site assessment by national reference laboratory staff; and (iv) development of national guidelines.
From 2005, the national reference laboratory conducted two cycles of training each year. The training was held at the reference laboratory and supported technically by the Japan International Cooperation Agency.
Biannually, the national reference laboratory sent five serum panels to all participating laboratories. The laboratories were expected to return test results within one month of receipt. The panels included strong-positive, weak-positive and negative HIV antibody samples. The number of positive and negative samples differed by panel. To maintain a satisfactory response rate, the national reference laboratory contacted every laboratory twice by phone, before sending the samples and afterwards to confirm receipt of samples. After assessing results from the participating laboratories, the national reference laboratory provided feedback by mail, including the proportion of the laboratories that kept the deadline.
In 2010, the Myanmar Ministry of Health legislated national guidelines for the scheme based on guidelines produced by the Joint United Nations Programme on HIV/AIDS. 12 This was done to standardize the procedures and facilitate the expansion of the scheme. 13 The guidelines describe each step in the scheme and the HIV testing procedures, including the serial testing algorithm for rapid tests. The guidelines propose that Alere Determine (Alere, Waltham, United States of America) be used for screening, Uni-Gold Recombigen® HIV-1/2 (Trinity Biotech, Wicklow, Ireland) for confirmation and HIV 1/2 STAT-PAK® (Chembio, Medord, USA) for second confirmation. 14 Lessons from the field HIV testing in Myanmar Latt Latt Kyaw et al.

Local setting
There are 422 public hospitals in Myanmar, including six teaching hospitals, 28 general hospitals, 19 specialized hospitals, 45 district hospitals, and 324 township hospitals that provide HIV testing. In addition, private laboratories supported by nongovernmental organizations (NGOs) also provide HIV testing services. Between 2005 and 2012, the national reference laboratory selected 314 public hospital laboratories to be part of the scheme. Large hospitals were selected first. The number of HIV tests performed and availability of transportation and communication systems were also considered. Upon request, NGOsupporting laboratories were gradually included.

Relevant changes
The scheme was started in 65 laboratories in 2005 and gradually expanded to include 347 participating laboratories in 2012 (Fig. 1), which included almost all 330 townships in Myanmar. Of these laboratories, 33 were supported by NGOs. During the expansion of the scheme, laboratories' response rates continued to be over 90% despite the inclusion of laboratories in remote areas with communication difficulties. The proportion of laboratories reporting at least one aberrant test result was improved from 9.2% (6/65) in 2005 to 5.4% (17/316) in 2012 ( Fig. 1).
On-site assessment of poorly performing laboratories revealed some misunderstandings regarding HIV testing procedures, such as wrong incubation time or inadequate amount of sample being used (Box 1). We often observed that the test result was read immediately after the appearance of the control band, without allowing sufficient incubation time for a weak-positive sample to test positive. Most of the aberrant results (166/263) were false negatives. These results most frequently occurred in weak-positive samples that were diluted by the reference laboratory to cause a weak reaction. Thus, the weak-positive panel appeared to be effective for revealing misunderstandings relating to testing procedures. Some aberrant results were attributed to the lack of necessary equipment in laboratories, such as timers and micropipettes.
The scheme found that nine types of test kit were used in the different laboratories and around 40% of the laboratories initially used a kit that had not been recommended in the guidelines. Most laboratories have now begun to use the recommended test kits, due to clearer guidance and improved procurement of HIV test kits resulting from increased funding.

Discussion
A strong commitment by the national reference laboratory is crucial for any quality assurance programme. 4 The Myanmar national reference laboratory has the mandate to promote and maintain the quality of all laboratory services and took full responsibility for implementing the scheme. However, we found that maintaining the scheme in a resource-poor setting required intensive efforts. For example, one of the most difficult tasks was to sustain a high response rate from the participating laboratories in remote areas. While expanding the scheme, smaller less-

Background factors causing false-positive results
• Cross-contamination during the process of testing equipped laboratories located in remote areas needed technical assistance and arrangements to solve limitations in communication and postal infrastructure. The national guidelines helped the national reference laboratory to implement the scheme by allocating responsibility and authority. 13 The guidelines clearly describe the role of the national reference laboratory in the programme and require all public laboratories to participate in the scheme.
Maintaining the quality of HIV testing in the national reference laboratory itself was also important for the quality of the scheme. Therefore the laboratory was certified by an international external quality assessment scheme through the National Serology Reference Laboratory, Australia. 15

Proficiency testing
Proficiency testing programmes are an effective tool for improving the quality of laboratory services. 16 Although they have some limitations, (such as incomplete assessment of the whole testing process and testing materials being treated differently to patient materials), these programmes remain useful. Compared with retests of samples from a participating laboratory performed by a reference laboratory, proficiency panel testing is a simpler and more feasible method to monitor quality in a resourcepoor setting.
Acquiring and maintaining skills for HIV testing can be challenging for technicians in countries with a low HIV prevalence, such as Myanmar, because positive specimens are rarely encountered during everyday work. Proficiency testing helps technician to maintain skills by regularly identifying positive samples.

On-site assessment
On-site assessments encourage problem solving and motivate staff. They are useful for improving the performance of health-care services. 17 However, the assessments increase the workload of supervisors and supervisees, since they require time and resources. Thus, they were not suitable for all participating laboratories, but were an important component of the scheme. Supervisors from the national reference laboratory provided on-site training using proficiency samples. The findings from these on-site assessments generated practical information regarding the most probable causes of aberrant results. These findings were shared with all participating laboratories and were used to improve the content of training materials.
We also found that laboratories at local hospitals sometimes faced difficulties due to lack of commitment by the supervisors. We tried to solve this during on-site assessments. Sometimes we also provided simple equipment if it was the major cause of aberrant results.

Running costs
The running costs and work burden associated with distributing proficiency sample panels often hinder the establishment of external quality assessment. 18 In Myanmar, the sustain-ability of funding and the stretching of human resources at the national reference laboratory to maintain the scheme were a challenge. The annual cost of running the scheme in 350 laboratories was approximately 11 700 United States dollars (US$) in 2012, which included training (US$ 3000), panel preparation (US$ 4200), postage and communication (US$ 1000), report publications (US$ 500) and on-site assessment (US$ 3000).

Conclusion
To ensure reliable HIV testing services in this context, assessment is needed. The on-site component is particularly important, as it is the only way to find the causes of aberrant results; information that can then be used to improve the performance of all laboratories in the country (Box 2). ■ Funding: This study was funded by a grant from the National Center for Global Health and Medicine, Japan (25-8).
Competing interests: None declared.