To evaluate the causes of non-malarial febrile illness in children living in an area of perennial malaria transmission and to determine if these children who test negative for malaria by rapid diagnostic test receive any benefit from antimalarial therapy.

Until recently, national and global malaria control authorities recommended clinical diagnosis-based solely on the presence or history of fever-for most malaria treatment settings in sub-Saharan Africa where malaria transmission is sustained and intense. To some extent this recommendation was based on the fact that conventional antimalarial treatments like chloroquine or sulfadoxine-pyrimethamine (SP) were relatively affordable and safe and that microscopic diagnosis was complex and difficult to maintain in remote rural settings. It was economically advantageous and logistically more feasible to treat all potential cases as malaria than to extend microscopic diagnosis to every level of the health system. This approach has resulted in extensive over-treatment, particularly among older children and adults, and may have contributed to the rapid development of antimalarial drug resistance.

Although much has been written recently on the cost-effectiveness of expanding malaria diagnosis, available information is scarce on a number of other important reasons why clinical diagnosis has been recommended for so long, especially among children living in high transmission settings. First, uncomplicated malaria can progress to severe or fatal illness within 24 to 48 hours of onset. Numerous care-seeking studies have demonstrated that caretakers seldom arrive at formal health facilities within 24 or 48 hours after the onset of uncomplicated febrile illness. If a diagnostic test imposes additional barriers-such as cost, time delay, or referral-requiring a positive parasitological diagnosis could put children whose cause of fever is malaria infection at greater risk of progressing to severe or fatal illness. Second, although the current approach based on clinical diagnosis appears to result in substantial over-treatment, it is still possible to demonstrate that children living in malaria transmission areas benefit from additional scheduled doses of antimalarial treatment, even when they are not ill. For example, a meta-analysis of six trials of sulfadoxine-pyrimethamine (SP) given to children at routine immunization visits demonstrated an average decrease of 30% in episodes of clinical malaria, 15% in anemia, and 24% in all-cause hospital admissions among children receiving SP compared to children who did not receive the drug at these visits. Finally, providers and clients may be inclined to disregard a negative blood slide or RDT, especially in situations where they have not identified an additional treatable cause of illness. Withholding antimalarial treatments from such children might adversely affect provider and client satisfaction and poor client satisfaction may reduce subsequent health facility utilization. It might also encourage disappointed clients to seek treatment in the private sector where a broad range of antimalarial drugs-most of them single drug treatments that contribute to the development of resistance and which are not recommended in the national treatment policy-can be obtained without diagnostic confirmation.

We propose a longitudinal cohort study to evaluate the identifiable causes of treatable fever among 1000 malaria-negative children presenting to outpatient health clinics in Miono, Bagamoyo District, Tanzania using a variety of clinical, microbiological and serologic methods. In addition we intend to follow these 1000 malaria-negative children for up to 91 days or until their next malaria infection to assess their clinical progress and need for further malaria treatment. To compare the relative benefit of providing antimalarial treatment even to malaria-negative children, half of the participants will be randomized to receive first-line treatment for malaria as currently recommended; the other half will receive treatment only for other identified illnesses.

Alternative Hypothesis: Febrile, parasite-negative children treated for malaria have better clinical and longitudinal outcomes (as measured by prevalence of anemia at the end of the follow up period, reticulocyte count repeat visits to the health facility, hospitalization, and time to next infection with malaria parasites) than febrile, parasite-negative children not treated for malaria in areas of high transmission.

Inclusion Criteria:

• Age 6 to 59 months.
• Present to health facility with fever (oral or rectal temperature ≥38°C or axillary temperature ≥37.5°C) or history of fever in the past 48 hours.
• Have negative rapid diagnostic test for malaria.
• Live within the boundaries of the officially recognized catchment area of Miono Health Center (within approximately 10 km of the health facility).

Exclusion Criteria:

• Plan to travel or leave the area within the next 3 months.
• Have been treated for malaria in the 2 weeks prior to enrollment.
• Have clinical evidence or history of danger signs: convulsions, lethargy, loss of consciousness, unable to eat or drink, vomiting everything.
• Have severe, life-threatening anemia: hemoglobin ≤5g/ dL.
• Have very low weight for age, severe pneumonia, or very severe disease as defined in the Integrated Management of Childhood Illness algorithms.
• Have a history of sensitivity to artemisinin derivatives or Artemether-Lumefantrine.
• Have previously been enrolled in this study or another ongoing cohort study of malaria treatment options at these health facilities
• Chronic disease requiring ongoing medical care (i.e HIV on cotrimoxazole).