Performance Evaluation of Standard Health System Microscopy and Rapid Diagnostic Test for the Focused Screening and Treatment of Malaria in South-eastern Tanzania

Background: The Focused screening and treatment (FSAT) has been identied as one of the key approaches for reducing the malaria burden in Tanzania. However, the diagnostic performance of the local standard health system microscopy and rapid diagnostic tests (RDT) for the malaria infections is yet to be established. Thus, we aimed to evaluate the performance of the local standard health system microscopy and RDTs in comparison to that of the gold standard modality through the re-examination of blood slide microscopy ndings. Methods: We used the paired RDTs and standard health system microscopy results from the participants screened in the FSAT during 2015–2016. With the gold standard modality as reference, the results of local standard health system microscopy and RDTs were evaluated according to their sensitivity, specicity and reliability. Results: A total of 1,497 paired standard health system microscopy and RDTs results were analysed. Of these, 679 (45.4%) samples and 818 (54.6%) samples were from the high transmission areas (HTA) and low transmission areas (LTA) in the Ruji District of south eastern Tanzania, respectively. With the gold standard as the reference, the sensitivity of RDTs were higher than those of standard health system microscopy (87.10% vs 76.88%) and (86.99% vs 65.04%) in the HTA and LTA, respectively. Further, the false-negative rates of RDTs were lower than those of standard microscopy in both HTA (12.90% vs 13.01%) and LTA (23.12% vs 34.96%). Conclusions: RDTs have better performance than the standard health system microscopy in terms of sensitivity and specicity for FSAT in study areas in Tanzania.


Background
Malaria remains one of the most serious vector-borne diseases impairing human health worldwide [1,2]. More than 228 million cases of malaria were reported in 2018, of which 93% were in Africa. In Tanzania, as in many other countries of sub-Saharan Africa, the national malaria prevalence has decreased to 9.3% in 2017 from 18.1% in 2008 [3]. The "China-UK-Tanzania Pilot Project on Malaria Control" (hereinafter "Pilot Project") was implemented in the South-eastern Tanzania to explore a new model for reducing malaria burden and possibly scaling-out the approach into other malaria-endemic countries, a collaboration among the National Institute of Parasitic Diseases (NIPD), Chinese Center for Disease Control and Prevention, and the Ifakara Health Institute (IHI), was initiated in 2015 in the Ru ji District in South-eastern Tanzania. Baseline epidemiological data were collected, and microscopy and rapid diagnostic tests (RDTs) modalities were used to identify the malaria hotspots through the focused screening and treatment (FSAT) strategy.
Given the heterogeneity of the malaria transmission intensity, prioritizing areas with the highest numbers of symptomatic cases is a rational initial approach.
In this context, FSAT has been established as a key strategy for reducing the high malaria burden immediately. For example, during the campaign to contain artemisinin resistance in the Cambodia-Thailand border, 49% of cases were from 3 of the 109 villages, and 77% of cases were from 10 of the 109 villages.
Microscopy is a simple, convenient, and direct diagnostic tool for the parasitological detection, including of Plasmodium species (Plasmodium sp.), and is thus the gold standard modality for the parasitological con rmation. However, the accuracy of microscopic diagnosis is affected by the skill of the technician.
In general, the blood testing can detect only approximately 50-100/µl Plasmodium sp., and it has a low detection rate in asymptomatic Plasmodium sp. carriers with low-density infection [4]. Further, for a patient who used anti-malarial drugs with the consequent reduction of Plasmodium sp. density and morphological changes, challenges with parasite detection and differentiation increase, leading to further errors in microscopy [5].
RDTs are immunological methods for detecting speci c antigens of Plasmodium spp.. from the peripheral blood. It has the advantages of having high sensitivity, simple operation, and swift obtainment of results. Further, only short-term training is required for the operation of local healthcare staff. Therefore, well as local logistical convenience, the Ru ji District was selected. The selected study area covered four wards. Of the four wards, two were assigned to the intervention arm and the other two were controls.

Sample selection
FSAT was carried out from 1 November 2015 to 20 November 2016 in the Ru ji District to screen for symptomatic and asymptomatic local residents. In a preliminary investigation on the Ru ji pilot project by Tanzania local slide examiners, 1497 blood slides randomly selected from seven villages in Ikwiriri, a low transmission area (LTA), (n = 679, 45.36%) and from seven villages in Muhoro, a high transmission area (HTA), (n = 818, 54.64%) were tested. All the blood slides were stored at 4°C for the next re-examination to evaluate the microscopy performance of the local slide examiner, to ultimately compare the accuracy of microscopy and RDTs in the study areas (Fig. 1). The production and microscopy of these blood samples were completed by local health facility staff.

Standard health system microscopy
The thin and thick slides were made at the focal points. Peripheral blood slides were created immediately after sample collection on a clean, grease-free microscope slide and allowed to air dry. The lms were stained with 10% Giemsa solution for 10 mins. The slides were allowed to the air dry and subsequently examined with light microscopy using an oil immersion objective lens. A slide was declared negative only after 100 microscopic elds with no parasites were observed. For each specimen, the thick lms were only examined for malaria parasite detection after the parasite speciation was identi ed through thin lms. The result was recorded as positive when both microscopists recorded a positive result and the same species. Discrepancies were resolved by involving a third microscopist [11].

Rapid diagnostic test
The RDTs used in this study speci cally detects P. falciparum and P. vivax and were purchased from Standard Diagnostics (SD BiolineP.f/P.v Biostandard Diagnostics, Gurgaon, Korea). Plasmodium falciparum and P. vivax were detected using histidine rich protein-2 and vivax-speci c lactate dehydrogenase, respectively. A total of 5 µl of whole blood were added to the card pad, and three drops of speci c lying agent were added. The RDT result was read after 15 min according to the manufacturer's instructions and then recorded [11].

Gold standard modality
The re-examination result of microscopy of blood slides was considered the gold standard in this study. The selected slides were re-examined by two NIPD experts with WHO's Level 1 quali cation for malaria microscopy following the WHO guidelines for malaria microscopy. Two microscopists from the NIPD independently re-examined each blood sample under double-blind conditions and recorded the microscopy results. Inconsistencies were resolved by involving a third microscopist.

Statistical analysis
The chi-square test (χ 2 test) was adopted for pairwise comparison of diagnostic accuracy and coincidence rate among different methods, with α = 0.05 taken as the signi cant level. With the gold standard modality as reference, the results of standard health system microscopy and RDTs were evaluated according to their accuracy, reliability, and applicability [12] [10]. Evaluation indexes included sensitivity, speci city, and kappa value. All statistical analyses were performed using SAS 9.3 software.

Results
A total of 1,497 pairs of standard health system microscopy results and RDTs test results were collected in the Ru ji pilot area. Of these, 679 (40.48%) and 818 (59.52%) of the samples were from HTAs and LTAs, respectively. Of the 1,497 slides, 244 (16.30%), 382 (25.45%), and 309 (20.64%) slides were detected to be positive on standard health system microscopy, RDTs, and the gold standard modality ( Table 1). The positive rate of RDTs was higher than that of the gold standard and standard health system microscopy (Table 1).

Evaluation Of Reliability
In the HTAs and LTAs, with respect to consistency, there was a signi cant difference between the standard health system microscopy and the gold standard in both the HTAs and LTAs (u = 21.03138 vs u = 20.34338, p < 0.05). The difference was also signi cant between RDTs and the gold standard (u = 18.56158 vs u

Discussion
The diagnostic performance of RDTs and local standard health system microscopy for malaria in comparison to the gold standard modality is yet to be clari ed. This study showed that RDTs performed better than microscopy in the local settings in Tanzania. RDTs were more sensitive, although microscopy results were more speci c.
In general, RDTs are more suitable for FSAT because microscopy training and equipment are often challenging in the low-income countries. Moreover, there are di culties in maintaining the local staff capacity for diagnosing malaria [11]. RDTs are widely distributed in Tanzania with reliable quality control and are easy to use and interpret. In this study, RDTs positively identi ed P. falciparum in 25.45% of the slides, whereas the positive rate of the local standard health system microscopy was only 16.30%. These results were consistent with the ndings of Shakeley et al. [13] who reported the accuracy and usefulness of RDTs for case detection in Zanzibar. Further, our ndings showed that RDTs have higher sensitivity than the local standard health system microscopy in both HTAs and LTAs, indicating that RDTs are more reliable than local standard health system for case detection.
In addition, the difference in sensitivity was more signi cant in LTAs than that in HTAs. The false negative rates of RDTs were also lower than those of standard microscopy in both HTAs and LTAs, indicating that RDTs more accurately identi ed individuals infected with malaria. For speci city, RDTs showed lower rates than standard health system microscopy in HTAs and LTAs. This indicated that compared with RDTs, the local standard health system microscopy could more correctly identify individuals without malaria. However, it should be noted that at 7-days posttreatment, the microscopy results can be negative, while RDTs may remain positive due to delayed clearance of the HRP2 antigen. In our study, all positive RDT results became negative 14 days after treatment, suggesting that the effective detection time of RDTs is longer. In a previous case report, a positive RDT result was obtained, while a negative microscopic result was not a false positive case, as there might be some residual antigens in the patient [14].
In general, the blood detection of Plasmodium sp. requires at least 50-100/µl of the pathogen in the blood, and it is di cult to detect the asymptomatic Plasmodium sp. carriers with low-density infection. Compared with patients in HTAs, the protozoa density is normally lower in patients in LTAs [4]. In this study, the area under the curve of RDTs was signi cantly higher than that of microscopy in LTAs (P < 0.01), whereas there was no signi cant difference between them in HTAs (P > 0.05). These results demonstrated that RDTs had similar diagnostic accuracy to the gold standard modality in HTAs. In LTAs, RDTs addressed the limitations of microscopic examination for malaria diagnosis [15] [16]. RDTs can be useful as an alternative to standard microscopy, especially in the HTAs. However, the medical institutions with microscopy capabilities (such as those at the district level) should continue to reinforce training on microscopy, external evaluation, and quality control. Improvements in FSAT may lead to the need for various detection methods, particularly those that are more sensitive screening tools (e.g. PCR), to improve the accuracy of detection.
This study has some limitations. The samples in this study were not subjected to PCR evaluation. Hence, some patients with low-density parasitemia, particularly those with asymptomatic infections, may have been missed.

Conclusions
RDTs are more accurate than the local health system standard microscopy for malaria case screening in the pilot district of Ru ji in Tanzania. RDTs are appropriate for the case identi cation. The combination of RDTs and microscopy can effectively improve the detection rate of Plasmodium sp, lowering the false positive rate.