Use of Improved Panel Randomization as Quality Control in a Large Scale HIV Serology Proficiency Testing Scheme: Best Practices in Kenya
- Geofrey Bundi, National Public Health Laboratory
- Frankline Kitheka, National Public Health Laboratory
- Sophie Mwanyumba, National Public Health Laboratory
- Thomas Gachuki, National Public Health Laboratory
INTRODUCTION
The Ministry of Health, Kenya through the National Public Health Laboratory provides HIV Proficiency Testing (PT) scheme to determine performance and monitor their continuing testing performance in rapid HIV testing. Participants include regional laboratories, antenatal clinics, primary health centers, facility based and stand-alone counseling and testing sites. To address sample homogeneity and stability challenges, the PT provider adopted the use of dried tube specimen (DTS), a cold-chain independent approach with a sample stability of at least a month at temperatures encountered in many parts of Kenya. Initially, the PT scheme was modeled on simultaneous distribution of blinded six test samples to participants biannually for concurrent testing. Summary statistics generated from Data analyses revealed trends of results outcomes among participants within same participant's locality. The PT provider realized this observation and enlisted a change in the PT statistical design to include randomization of PT sampling but maintaining simultaneous distribution within the same frequency.
METHODS
PT samples for two testing cycles were randomized based on a participants’ unique identification using PT
system.
Samples were simultaneously distributed to participants, tested following the National country algorithm for
rapid HIV. Results were submitted on line within two weeks and feedback from PT provider was provided
using the same mechanism within a month.
RESULTS
When the switch was made, 11% of 16,298 HIV Service providers incorrectly identified at least one of the six
panels tested. During the subsequent cycle, 8.7% of 21,370 HIV Service providers got an incorrect result for at
least one of the 6 panels.
CONCLUSION
The new strategy improved the quality in testing. More than one set of reference results can be used to
evaluate participants in a single proficiency testing cycle.