Parasite-based Diagnosis for Malaria in Uganda: Feasibility and Cost-Effectiveness

The development of drug resistance in malaria parasites lead the Uganda Ministry of Health (MoH) to change the first-line anti-malarial treatment from the cheap Chloroquine/Sulfadoxine-Pyrimethamine combination to a more expensive Artemether-Lumefantrine. The MoH recommends treatment of all fever cases as malaria within 24 hours of illness with first-line drug. Under this policy, patients who do not have malaria but present with febrile illness will receive Coartem® resulting in significant drug wastage. With the increased cost of first-line drugs, this wastage places a substantial and potentially remediable burden on the health budget. As such, there is a need to gauge whether more accurate malaria diagnosis through microscopy and/or rapid diagnostic tests, might improve the cost-effectiveness of the new treatment regimes.

Specific objectives

1. To assess the feasibility of rapid test and microscopy in diagnosis of malaria at health centre III.
2. To compare the cost-effectiveness of treating malaria with Artemether- Lumefantrine based on microscopy, rapid test and presumptive diagnosis in different transmission intensities
3. To assess whether introduction of malaria parasite-based diagnosis (rapid test or microscopy) at government Health Centre III improves the overall cost-effectiveness of outpatient management of febrile illness

Sample size determination The sample size was determined using standard formula. Estimating the sensitivity of either test to be 90%, Zα =1.96 and after stratifying for age (\<5 years and ≥5 years), the calculated sample is 272 per health centre. Therefore, the total number of patients for 6 health centres in two districts = 6 x 272 = 1632.

Baseline: Baseline data was collected on the current malaria treatment practices; clinicians' view of malaria diagnosis; concerns towards the new treatment; and health centre staffing. Geographical locations of all visited government health centres was recorded using Germin etrex global positioning system (Germin International Inc., Olathe, USA). At six selected health centres, social and demographic data was collected from 613 patients.

Implementation of intervention: The main intervention is comprised of three malaria diagnostic approaches: presumptive diagnosis, field microscopy and rapid test (Paracheck Pf® device - Orchid Biomedical Systems, Goa, India). Each of these approaches was randomly allocated to a health centre. The first-line drug used is Artemether/Lumefantrine (20mg/120mg) (Novartis, Switzerland).

Consenting subjects are consecutively enrolled at the point when the attending clinician suspects that they have uncomplicated malaria. Where microscopy is the main diagnostic method, thick and thin blood smears are prepared per patient enrolled. Where rapid test is the main method, all patients are tested. One hundred patients per health centre are randomly selected to provide blood specimens for validation using expert microscopy and PCR as "gold standard." After enrolment, patients are systematically tracked until departure from the health centre. Patients are followed up on the seventh day of treatment to assess their clinical improvement. Those who fail to return on the scheduled date are traced from their homes on the eighth day.

Follow-up activities include: documentation of drugs prescribed and or dispensed; clinically assess the patient's status in comparison to Day 0; perform further tests including PCR if the patient does not show improvement; pill counting of drugs remaining - if the patient has not completed the dose; documentation of reasons for not completing dose; and documentation if the patient bought the prescribed drugs that were out of stock on Day 0.

Effectiveness is measured as the number of patients commencing treatment with Artemether/Lumefantrine. However, patients are followed up and effectiveness also measured on the 7th day of treatment indicated by their clinical improvement. The feasibility of the diagnostic methods is ongoing from the March 2010 to date.