Comparison of visual and automated Deki Reader interpretation of malaria rapid diagnostic tests in rural Tanzanian military health facilities.

BACKGROUND: Although microscopy is a standard diagnostic tool for malaria and the gold standard, it is infrequently used because of unavailability of laboratory facilities and the absence of skilled readers in poor resource settings. Malaria rapid diagnostic tests (RDT) are currently used instead of or as an adjunct to microscopy. However, at very low parasitaemia (usually < 100 asexual parasites/µl), the test line on malaria rapid diagnostic tests can be faint and consequently hard to visualize and this may potentially affect the interpretation of the test results. Fio Corporation (Canada), developed an automated RDT reader named Deki Reader™ for automatic analysis and interpretation of rapid diagnostic tests. This study aimed to compare visual assessment and automated Deki Reader evaluations to interpret malaria rapid diagnostic tests against microscopy. Unlike in the previous studies where expert laboratory technicians interpreted the test results visually and operated the device, in this study low cadre health care workers who have not attended any formal professional training in laboratory sciences were employed. METHODS: Finger prick blood from 1293 outpatients with fever was tested for malaria using RDT and Giemsa-stained microscopy for thick and thin blood smears. Blood samples for RDTs were processed according to manufacturers' instructions automated in the Deki Reader. Results of malaria diagnoses were compared between visual and the automated devise reading of RDT and microscopy. RESULTS: The sensitivity of malaria rapid diagnostic test results interpreted by the Deki Reader was 94.1% and that of visual interpretation was 93.9%. The specificity of malaria rapid diagnostic test results was 71.8% and that of human interpretation was 72.0%. The positive predictive value of malaria RDT results by the Deki Reader and visual interpretation was 75.8 and 75.4%, respectively, while the negative predictive values were 92.8 and 92.4%, respectively. The accuracy of RDT as interpreted by DR and visually was 82.6 and 82.1%, respectively. CONCLUSION: There was no significant difference in performance of RDTs interpreted by either automated DR or visually by unskilled health workers. However, despite the similarities in performance parameters, the device has proven useful because it provides stepwise guidance on processing RDT, data transfer and reporting.

The use of Fionet technology for external quality control of malaria rapid diagnostic tests and monitoring health workers' performance in rural military health facilities in Tanzania.

INTRODUCTION: Internal and external quality control (QC) of rapid diagnostic tests (RDTs) is important to increase reliability of RDTs currently used to diagnose malaria. However, cross-checking of used RDTs as part of quality assurance can rarely be done by off-site personnel because there is no guarantee of retaining visible test lines after manufacturers' recommended reading time. Therefore, this study examined the potential of using Fionet™ technology for remote RDT quality monitoring at seven clinics, identifying reasons for making RDT processing and interpretation errors, and taking corrective actions for improvement of diagnosis and consequently improved management of febrile patients. METHODS: The study was conducted at seven military health facilities in Mainland Tanzania and utilized RDTs capable of detecting Plasmodium falciparum specific Histidine-rich protein 2 (Pf-HRP2) and the genus specific Plasmodium lactate dehydrogenase (pLDH) for other species of plasmodium (P. vivax, P. malariae or P. ovale; pan-pLDH). Patients' data and images of processed RDTs from seven clinics were uploaded on a Fionet web portal and reviewed regularly to monitor preparation procedures and visual interpretation of test results compared to automated analysis using the Deki reader of RDT. Problems detected were rapidly communicated to remote laboratory personnel at the clinic for corrective action and follow-up of patients who were falsely diagnosed as negative and missed treatment. Factors contributing to making errors in visual interpretation of RDT results were analyzed during visits to the health facilities. RESULTS: A total of 1,367 (1.6%) out of 83,294 RDT test images uploaded to the Fionet portal had discordant test results of which 822 (60.1%) and 545 (39.9%) were falsely reported as negative and positive, respectively. False negative and false positive test results were common for a single test line in 515 (62.7%) and 741 (54.2%) tests, respectively. Out of 1,367 RDT images assessed, 98 (7.2%) had quality problems related to preparation procedures of which 95(96.9%) errors were due to putting too much blood on the sample well or insufficient buffer in the respective wells. The reasons for discrepant results included, false reporting of none existent lines in 526 (38.5%) tests, missing a faint positive line in 493 (36.1%), missing a strong positive line in 248(18.1%) and errors caused by poorly processed RDTs in 96 (7.2%) tests. Among the false negative tests (n = 822), 669 (48.9%) patients were eligible for follow-up and only 339 (48.5%) were reached and 291 (85.8%) received appropriate anti-malaria therapy. CONCLUSION: Fionet technology enabled remote monitoring of RDT quality issues, identifying reasons contributing to laboratory personnel making errors and provided a rapid method to implement corrective actions at remote sites to improve malaria diagnosis and consequently improved health care management of febrile patients infected with malaria.

The role of herpes simplex virus type 2 and other genital infections in the acquisition of HIV-1 among high-risk women in northern Tanzania.

BACKGROUND: We examined the role of herpes simplex virus type 2 (HSV-2) and other genital infections on human immunodeficiency virus type 1 (HIV-1) incidence in a cohort study conducted between 2002 and 2005 among female bar/hotel workers in Moshi, Tanzania. METHODS: At baseline and every 3 months thereafter, participants were interviewed, and blood and genital samples were collected. Predictors of HIV-1 incidence were evaluated using a Cox proportional hazards regression model. RESULTS: Of 845 women who were HIV-1 seronegative at baseline, 689 (81.5%) were monitored in the study for a total of 698.6 person-years at risk (PYARs). The overall HIV-1 incidence was 4.6/100 PYARs (95% confidence interval [CI], 3.0-6.2/100 PYARs), and condom use was very low. After adjustment for other risk factors, the risk of HIV-1 was increased among women with HSV-2 at baseline (hazard ratio [HR], 4.3 [95% CI, 1.5-12.4]) and in those who acquired HSV-2 during the study period (HR, 5.5 [95% CI, 1.2-25.4]). Other independent predictors of HIV-1 were baseline chlamydial infection (HR, 5.2), bacterial vaginosis (HR, 2.1), and the occurrence of genital ulcers (HR, 2.7). CONCLUSION: HSV-2 and other genital infections were the most important risk factors for HIV-1. Control of these infections could help to reduce HIV-1 incidence in this population.