1) To compare in a setting where microscopy for malaria is available whether introducing rapid diagnostic tests (RDTs) improves targetting of antimalarial drugs and antibiotics (RDT v microscopy). 2) To compare whether, in a setting where microscopy for malaria is not available, introducing rapid diagnostic tests (RDTs) improves targetting of antimalarial drugs and antibiotics (RDT v clinical diagnosis).

BACKGROUND There is good evidence from multiple sites, including in Ghana, that malaria is massively over-diagnosed. It could be argued that where microscopy is not available it can be introduced, and elsewhere it can be improved, but high-quality microscopy is not easy to sustain. If substantial over-diagnosis continues in Ghana in an era where an ACT is now the first-line treatment, it will lead to the intervention being substantially more expensive due to over-prescription than it should be, potentially rendering it unsustainable.

The introduction of rapid diagnostic tests (RDTs) has the potential to provide a way of more accurately directing ACTs to those that need them and may also encourage clinicians to consider alternative diagnoses in test-negative cases, reducing the risk of missing treatable, and potentially fatal, alternative causes of febrile illness. RDTs to direct ACT use also have the possibility to be cost-effective, but only if clinicians prescribe logically on the basis of test results.

Initial data from Tanzania suggests that providing RDTs in the context of formal healthcare settings may have little impact on clinician behaviour, but the health system in Ghana is very different, and both clinician and patient beliefs about malaria are likely to be different.

Additionally, this has not been properly tested in areas with little or no access to microscopy, nor where ACTs are currently available, which may influence clinician behaviour. Many believe this is the most useful setting for RDTs, and may limit over-prescription of anti-malarials but there are no data to support this belief, nor are there data on the cost-effectiveness of this approach. This trial aims to test the impact of RDT use on clinician behaviour directly by means of a randomised trial.

OBJECTIVES Principal Objective To determine by means of a randomized trial the impact of the introduction of Rapid Diagnostic Tests (RDTs) on the appropriate prescription of anti-malarials in the two public healthcare settings found in Ghana.

Specifically,

1. To determine by means of a randomized trial the impact of introducing RDTs on the appropriate prescription of anti-malarials in a setting where diagnosis is currently purely clinical.
2. To determine by means of a randomized trial the impact of introducing RDTs on the appropriate prescription of anti-malarials in a setting where microscopy is available.
3. To determine the sensitivity and specificity of RDTs in the diagnosis of malaria.
4. To explore clinician and patient perceptions on the use of RDTs versus clinical diagnosis in the management of malaria.
5. To explore clinician and patient perceptions on the use of RDTs versus microscopy in the management of malaria.
6. To determine the cost-effectiveness of RDTs for diagnosis of malaria in both settings.

In both cases RDT is being compared to the standard of care in the health centre.

METHODS The study will be carried out in the Dangme West District in the southern part of Ghana. It will be an individually randomized controlled trial

1. Baseline data will initially be collected for a period of one month to document the normal pattern of diagnosing malaria by clinicians in each health facility. At the health centre, all patients for whom microscopy is requested will have a research slide taken at the same time and the laboratory results of the clinic slide documented.
2. Exit interviews with patients and record reviews will be conducted to find out whether patients had been prescribed an anti-malarial and confirm whether this was done on clinical grounds or on the basis of microscopy. Their prescriptions will be documented. At the community clinics as well, a research slide will be taken for all patients with diagnosis of malaria or who are prescribed an anti-malarial. Record review of the outpatient department (OPD) cards of all patients with a diagnosis of malaria or who receive an anti-malarial be carried out to also document presenting symptoms.
3. For the main trial, all patients visiting the health facility and who meet the inclusion criteria will be eligible for enrolment into the study. Allocations to either microscopy or RDT at the health centre and to either clinical diagnosis or RDT at the community clinics will be computer generated. The allocations will be placed in sequentially numbered sealed opaque envelopes which will be prior-labelled with unique study ID numbers. Patients will be identified and consent sought on exit from the consulting room. If the clinician requests a laboratory test, the patient will be sent to the laboratory where a research assistant will open the sealed envelope in the presence of the patient to find out their allocation. The allocated test will be carried out and a research slide taken at the same time. The lab results will be written out as usual in the case of microscopy and the dipstick results will be recorded and sent to the clinician for reading as well.
4. At the community clinics where there is no laboratory, the process of identification of eligible patients, seeking of consent and allocation of study arm will be similar to that of the health centre. However, in this setting, depending on the allocation in the sealed envelope, a rapid diagnostic test will either be carried out by a research assistant or a printed card with "clinical diagnosis" written on it will be given to the patient. In the case of the RDT, the results will be recorded before the dipstick is sent to the clinician for clinician-read diagnosis and subsequent treatment. Whether the patient is assigned to clinical diagnosis or RDT, a research assistant will prepare a research slide per patient and air dry them for later batch reading.
5. The presenting symptoms, temperature at presentation, laboratory results, clinician's diagnoses and treatment of all patients recruited into the study will be documented. Prescriptions given to them will be documented on exit. In addition, their contact addresses will be collected on exit and consent sought to facilitate their easy location and recruitment for focus group discussions later.
6. The gold standard for whether a child or adult has malaria will be slide-proven malaria parasitaemia measured by a double-read research slide. Research slides will be Giemsa stained and read by two independent microscopists who would be blind to the study allocations and test results. They will not be used by clinicians in their treatment decisions.
7. At the end of the trial, individual in-depth interviews will be conducted with the main clinicians and focus group discussions held with all the other clinicians of the participating health facilities. The aim will be to explore their perceptions with regard to the use of clinical diagnosis, microscopy and RDT in the diagnosis of malaria and their practices with regards to this.
8. A random selection of patients from the three health facilities will be contacted to participate in focus group discussions at the end of the trial. There will be at least two focus group discussions comprising 8-12 discussants each from each health facility. These discussions will be carried out by means of interview guides. The discussions will be conducted in the local language and recorded on a tape recorder.

In both settings data will be collected for the cost-effectiveness studies.

Inclusion Criteria:

• Clinician decision to test for malaria

Exclusion Criteria:

• Pregnant women
• Severe malaria requiring referral
• Patient refuses consent
• Clinician specifically requests for microscopy