A diagnostic evaluation of single screen testing for malaria in the returning traveler: A large retrospective cohort study

Screening for malaria in the returning traveler has often required repeat testing; however, audit data suggest that patients have not been reattending. We sought to ascertain if this was safe by examining the diagnostic efficacy of a single screen consisting of a rapid diagnostic test (RDT) and a thin film.


| 777
DIAGNOSTIC ACCURACY OF A SINGLE RAPID MALARIA TEST AND A THIN BLOOD FILM the use of serial thick or thin films has been the reference standard for diagnosis. 2 The most recent Hematology Task Force guidelines recommend that when there is a strong clinical suspicion of malaria but the initial films are negative, repeat films should be made and examined after 12 to 24 h and again after an additional 24 h. 3 For patients presenting to the ED who are discharged home, obtaining repeat films would require return visits to ambulatory care or outpatient clinics. This results in a cost and time implication for both the patient and the hospital.
In a Cochrane review, Abba et al. 4 described rapid diagnostic tests (RDTs) by categories according to antigen targets. BinaxNOW is a type 2 RDT, meaning that it detects the presence of histidine-rich protein (HRP-2) and aldolase, allowing the detection of Plasmodium falciparum, or nonfalciparum species or a mixed infection with P. falciparum plus a nonfalciparum species. 4 In their meta-analysis of the use of type 2 RDTs in endemic countries for the detection of P. falciparum they demonstrated a sensitivity of 96.0% (95% confidence interval [CI] = 94.0% to 97.3%). In lower-prevalence settings the diagnostic accuracy appears to persist with stand-alone RDT strategies, with sensitivities ranging from 93.3% to 97.7%. However, the parasitemias of the samples to which these sensitivities apply were not stated. [4][5][6][7] These parasitemias are useful to contextualize the diagnostic accuracy of the RDT. Hypothetically, malaria may be more difficult to detect in low-parasitemia populations.
Since their introduction, RDTs have frequently been combined with microscopy in current clinical practice. This potentially allows an improved diagnostic sensitivity that may obviate the need for repeat testing on every patient, reducing inconvenience for them, improving efficiency for health services, and reducing the associated cost of repeat testing. The safety netting that is a core part of medical practice could be utilized here to direct patients to return if they feel unwell, potentially mitigating the harm from false negatives.
A local audit in a major teaching hospital ED examined the adherence to the malaria diagnostic standard of three serial screens over 72 hours over 10 years. During that time only 5.8% had the recommended three blood tests. These numbers were concerning; patients appeared to be opting not to return. This led us to question whether mandating repeat tests was truly clinically necessary. In this study we aimed to determine the diagnostic efficacy of a combined blood film and simultaneous RDT strategy, for the detection of Plasmodium falciparum and nonfalciparum malaria parasites in the setting of real-world practice of an ED.

Study design and setting
We undertook a retrospective diagnostic accuracy study at

Participants
We included all patients with suspected malaria who attended the adult or pediatric EDs at Manchester University NHS Foundation Trust over a 5-year period (2014-2019). Participants were identified from the laboratory database if they underwent testing for suspected malaria.

Index test
The patients are selected by emergency medicine clinicians who are suspicious of a diagnosis of infection by Plasmodium spp. This is defined by departmental guidelines as a returning traveler from an endemic area, with a fever or who is generally unwell. The clinical diagnostic pathway at the hospital mandated that each malaria test should consist of a thin film and RDT (BinaxNOW).
The Manchester hospitals used BinaxNOW during the period of this study, regularly reviewing the test performance data. The

Reference standard
We retrieved all results of serial testing for malaria from local da- reporting is to Health Protection Scotland. In the event of a negative screen, we examined any serial testing that had been conducted locally to identify in discrepancies.
Due to the high mortality and morbidity associated with P. falciparum an assumption was made that patients who originally tested negative would likely re-present and that the national registry could then be used to identify false negatives. A further assumption was made that patients infected with other species of malaria, but tested negative initially, would represent due to ongoing illness and universal free health care from the United Kingdom's National Health Service.

Data collection
The local clinical database was used to extract the malaria screen results, regional center verification, demographics, physiologic parameters, and full blood counts. The data were then collated in a central database for analysis.

Data analysis
The sample size was determined by the availability of relevant data.
We planned to conduct measures of diagnostic accuracy including sensitivity, specificity, and positive and negative predictive values. A total of 1331 patients had a complete first screen of both RDT and blood film. The blood film and RDT strategy identified all P. falciparum infections (see Table 2). This gave sensitivity and specificity of 100% (95% CI = 95.0% to 100%) and 99.4% (95% CI = 98.9% to 99.8%), respectively.

RE SULTS
The blood film and RDT strategy did not identify all Plasmodium spp. infections; there was one false negative. The false-negative case was admitted and subsequently found to be positive on a serial screen conducted 14 hours later. The second screen also had a negative RDT but the film detected P. ovale. This gave sensitivity and specificity of 99.0% (95% CI = 94.8% to 100%) and 99.5% (95% CI = 98.9% to 99.8%), respectively.

DISCUSS ION
Our findings suggest that negative results for the RDT and blood film, following a single blood test in the ED, excludes malaria sufficiently well to enable a safe diagnostic pathway when combined with safety netting. With the use of appropriate safety netting, advising patients strongly to return for follow-up testing if their symptoms persist, routine serial sampling for 3 days on every patient would not be required. In our study population this does not appear to increase the risk of a missed diagnosis and has the advantage of reducing inconvenience for patients and improving efficiency of health services. This study showed that there was already a reluctance of patients to reattend, and within the study period we did not find an increased risk for patients after a single screen with RDT and thin smear. We demonstrated an excellent sensitivity for P. falciparum of 1.00 (95% CI = 0.95 to 1.00) in our cohort of patients, with a mean parasitemia of 33,500 parasites/µL. Crucially, it was not sufficient to exclude all species of malaria, but the sensitivity was still high 99.0 (95% CI = 94.8 to 100). Given the far greater severity of P. falciparum versus P. vivax, P. ovale, and P. malariae (though not P. knowlesi), this is F I G U R E 1 STARD diagram of study participants for P. falciparum. The national opt-out registry is cross-referenced when retrospective research is conducted NHS databases; any participants who have preemptively indicated that they do not wish to participate in research are excluded [Color figure can be viewed at wileyonlinelibrary.com] an advantageous diagnostic profile. Furthermore, the false-negative case was already admitted on clinical grounds and therefore came to no harm.

TA B L E 1 Baseline characteristics
For those sent home from EDs, appropriate safety netting must be provided. Preferably this would be in a written form (see Data Supplement S1, available as supporting information in the online version of this paper, which is available at http://onlin elibr ary.wiley. com/doi/10.1111/acem.14216/ full), to ensure that those patients with a negative initial malaria screen whose illness continues in the absence of a viable alternative diagnosis present expediently for reevaluation, including another malaria test.
The benefit of early exclusion of malaria is not only one of patient safety and satisfaction. There is a strong economic argument for where it is clinically safe to do so, in reducing the number of follow-up appointments. An outpatient visit has an economic cost to the hospital, £120, 9 and to the patient's economic activity, £59. 10  There is concern that HRP-2-based RDTs, such as the one used in this study, may fail to detect clinical cases of P. falciparum due to to that of nested PCR. 19 Unlike malaria microscopy, its performance does not require lengthy training and LAMP has the potential to replace microscopy as the primary malaria screen in settings like the one described in this study, contributing greater confidence to the efficacy of a single visit strategy.
TA B L E 2 Diagnostic accuracy statistics of a blood film and RDT strategy for the diagnosis of P. falciparum and any species important caveats. The sensitivity of this screen and the detection