The study evaluated the performance of the ICT rapid diagnostic test relative to microscopy in an operational setting. The results of the study demonstrate that ICT test is an accurate and reliable method of diagnosing P. falciparum malaria in the context of microscopy. The very high sensitivity of 98% implies that the majority of malaria case among this population would be accurately diagnosed, giving the clinician the confidence that true cases of malaria are rarely missed. Most false negatives are linked to low parasitaemias which are unlikely to be the primary cause of illness. There were three false negatives results, two with parasite density <1,000/μl and one with parasite density > 5,000/μl. The last result where the parasite density was high, has been described by other investigators as a very rare event which may be caused by a prozone effect or by the presence of a mutation or deletion within the HRP2 gene . Two studies have reported cases of false negatives at very high parasitaemia with both a pLDH-based test and a HRP2 test with repeated testing [24, 25], but again in very low numbers – 1 case out of 58 and 2 cases out of 196 patients respectively.
Related to the high sensitivity, IC T demonstrated a very high NPV in this study population that allows a clinician to confidently diagnose negative-test patients as non malaria patients and look for other causes of fevers. The averted non-malarial treatments would lead to cost saving in the health sector and improved quality of care [3, 26]. The results showed a relatively low specificity especially in the under fives where it was 54%. Having a lower specificity, which leads to over diagnosis and treatment, was considered to result into less serious consequences than a low sensitivity in the context of clinical diagnosis. It is worth noting however, that the PPV in the under 5s was 78% which might be perceived as surprisingly high with such a low specificity and is linked to the high prevalence of malaria in this population. It is very likely that the low specificity has resulted from persistent antigenicity in patients who had recently been successfully treated for malaria.
ICT was found to be a reliable test with a high inter-reader and intra-test reliability Only 8 (2%) discordant records were assigned different interpretations as "faint positives" versus negative results. There was a high level of result concordance between ICT and Paracheck tests during longitudinal component. The two HRP2 tests were of comparable accuracy and performance in an operational setting.
This implies that given minimum training, it is feasible for field workers to interpret and report consistent results.
The study also described the duration of ICT false positivity due to persistent antigenicity and its relationship to pre-treatment P. falciparum parasite density in Ugandan children. Overall, persistence of HRP2 antigenicity was detectable in the majority of children at three weeks following successful treatment. The mean duration of 32 days is particularly worrying because HRP2 persistent antigenicity reduces the accuracy of the RDT test. This is more of concern in hyperendemic areas, where frequent malaria infections mean that children are likely to have the antigen in their blood even if their fever is not caused by malaria. There is a danger of over-diagnosis of malaria based on the interpretation of HRP2 test results. Misdiagnosis results in poor quality of care as the appropriate illness may not be addressed in time. It may also result in misuse of anti-malarials, poor resource allocation and, clinicians doubting the efficacy of the medicine used. Clinical skills in taking history and eliciting signs are essential when interpreting RDT results in an effort limit misuse of anti-malarials. Misuse of anti-malarial contributes to drug pressure and spreading of drug resistant P. falciparum , resulting in increased malaria-specific mortality, especially in children.
Previous studies from Asian countries have also reported persistent antigenicity after malaria clinical cure and the levels reported have varied widely, with ranges of 29–65% on day-7 [7, 8, 29–31]. In this study, longer durations of persistent antigenicity were reported and were shown to be associated with high parasite density. In general parasite densities are significantly higher in Africa than in Asia and may explain why similar studies in Asia have not found such a high level of prolonged false positivity. These results show that the high parasitaemia is directly proportional to persistent antigenicity and support what Swarthout and others found in the DR Congo, another hyperendemic region in Africa.
The causes of persistent antigenicity after malaria therapy are still debatable. Previous studies have listed causes of persistent antigenicity as parasite density levels below detection threshold for microscopy, delayed clearance of circulating antigens and, presence of gametocytes that also produce the antigen . The most plausible explanation in this study is the high pre-treatment parasitaemia. Patients with high parasite densities have large amounts of HRP2 antigen secreted in proportion to the parasite numbers  and to their stage of development . Elevated concentrations of HRP2 evidently take longer to eliminate  which seems to be supported by the findings of this study. The observed 28% recurrence of parasitaemia during follow-up may be a reflection of resistant parasites that persisted as sub-patent parasitaemia (<50 parasites/μl) or new infections. This raises the question of whether the positive results with the RDTs were actually false positives or the RDT was correctly detecting these sub-patent infections. However, a recent study in Uganda found 29% risk of recurrent parasitaemia of which 16% was due to recrudescence after treatment with artemether-lumefantrine (AL) over a 42 day follow-up period in a similar setting . Since this current study also reported a similar proportion of recurrence of parasitaemia in a hyper-endemic area, the high proportions reported here as result of persistent antigenicity are possible and not a mere reflection of recrudescence or re-infection.
The study participants were drawn from a pool of all persons suspected of having malaria, regardless of history, clinical status, demographic characteristics or other factors that may affect sensitivity and specificity in a routine health clinic. Inclusion of a large number of individuals improved the precision of study estimates. Therefore, the results obtained here reflect the performance of ICT in the operational setting.
There were some limitations in the conduct of the study. The evaluation was performed using microscopy as a gold standard, fully aware of its limitations in detecting low parasite densities . To improve accuracy of readings, two microscopists read all slides, blinded to each other's reading. A third reading was done to determine a final reading for all discordant field results. It is likely that some participants were included in the cross-sectional component of the study with circulating persistent antigenemia. This may have led to a reduced specificity because HRP2 antigens were detected for patients who may have successfully eliminated their parasites. This effect could have been overcome by using PCR to detect very low parasite densities that may have been missed by microscopy. However, this study was designed to evaluate operational accuracy of ICT where such cases are impossible to distinguish from current infections when using RDTs alone. It is assumed that there was no loss of quality in the RDTs as result of temperature fluctuations during transportation and storage. The storage temperature ranged between 22–29°C during the study period, albeit the single spike to 38°C recorded during transportation. These readings are within the safe temperatures for HRP2-based kits as reported in two previous studies [5, 36].
The study team neither observed therapy nor assessed for adherence to medication for the children during the longitudinal component. Poor adherence to treatment may have led to recurrent parasitaemia as a result of treatment failure.
RDTs devices are new technologies that resemble HIV test kits formats previously used in this setting. During the follow-up, a significant number of patients that had consented to participate failed to honour all the scheduled appointments. Only three patients formally withdrew consent mentioning that their families were suspicious that these were actually HIV studies. It is likely that others may have kept away from the study with similar misinformation. All patients' data collected during the evaluation of persistent antigenicity was utilized to mitigate any effect (bias) caused by loss to follow-up.