Comparison of Efficacy of Serological and Photogrammetric Testing Algorithms for HIV Screening in Nonlaboratory Settings

Abstract

Proficiency testing for Human Immunodeficiency Virus offers a platform for institutions and individuals to assess their technical competencies. It boosts morale and confidence among the laboratory operators and provides opportunities for improvement among participating institutions. Post-market field validation of HIV rapid diagnostic test is important because it increases confidence in the quality of testing services. Moreover, it confirms that the kit prequalified by WHO, is of superior quality (and vice versa) in terms of sensitivity and turnaround time. However, a shortcoming in test validation is that once kits are procured, rarely does field validation take place. As of 2008, no study in East Africa in general and Kenya, in particular, had looked at the field validation of photogrammetric testing. This current study is thus aimed at comparing the efficacy of serological and photogrammetric testing for HIV screening in non-laboratory resource-limited settings within Nairobi County. Besides, the study explores a new cost-effective method which can complement routine proficiency testing as a training tool for field-based validation. A longitudinal study was conducted using three rounds of proficiency and photogrammetric testing respectively. A total of 234 experienced and non-experienced operators were recruited using purposive sampling. A total of 702 dried tube specimen panel samples were tested using the Determine algorithm, while 702 photos were visually interpreted. The study revealed that the validity of serological testing was 98.07% and 96.21% for sensitivity and positive predictive values respectively and 70.37% and 82.61% for specificity and negative predictive values respectively. The validity of photogrammetric testing was 96.43% and 98.63% for sensitivity and positive predictive values respectively, and 70% and 46.7% for specificity and negative predictive value respectively. Lastly, the overall accuracy was 94.5% and 95.30% for proficiency and photogrammetric testing respectively, calculated as the percentage of true positives and true negatives on overall results. This study concludes that Determine algorithm is still sensitive and specific as such, it can still be used for proficiency testing of HIV panels. Besides, it was established that photogrammetric testing could be interpreted with higher accuracy compared to proficiency testing. The study recommendes that a higher accuracy rate of interpretation of HIV photographed tests can be used to assess proficiency levels of operators, while photogrammetric testing is viable as a complementary tool for identifying and monitoring operators testing competencies in resource-limited settings.