Diagnosis of malaria in children's outpatient departments in Abuja,Nigeria

Introduction

Nigeria is the most populated country in Africa (over 140 million), with approximately one in four of those at risk of malaria in Africa living in Nigeria. The reported incidence of malaria in Nigeria for 2005 was put at 1858/100,000 and it is reported as the cause of death in 25% of children. ¹ The over-diagnosis of malaria has been reported from a number of sites, with a small number from Nigeria suggesting that it might account for over 50% of reported cases. ² The Abuja declaration, from the Roll Back Malaria launch in 2000, highlights strategies for malaria control which include ‘at least 60% of those suffering from malaria to have prompt access to and are able to use correct, affordable and appropriate treatment within 24 hours of the onset of symptoms’. ³

The World Health Organization recommends clinical diagnosis, confirmed by parasitological slides if available, within two hours and treatment with artemisinin-based combination therapy (ACT). ⁴ The current national recommendation for the treatment of uncomplicated malaria is artemether-lumefantrine for both unconfirmed and laboratory confirmed cases. ¹ The practice in many hospitals in sub-Saharan Africa remains that most febrile illnesses should be treated as malaria. This strategy, promoted for areas without diagnostic facilities and largely designed with rural settings where malaria is common in mind, continues in many health facilities which could perform blood microscopy for malaria parasites and in peri-urban settings where malaria is much less common.

The changeover to artemether-lumefantrine in Nigeria comes at a high financial cost both to donors and, in the private sector, to end-users. This is cost-effective in areas of drug resistance so long as the drugs are targeted to those who need treatment; where diagnosis is not correctly targeted the rate of cost-effectiveness falls away rapidly. ⁵ Alternative causes of treatable and potentially serious infections are missed which may lead to a substantial burden of significant morbidity. ⁶ Over-use of antimalarials on a large scale may speed up the rate of development and spread of drug resistance. Given the very large population of Nigeria, the impact of substantial avoidable overprescription of antimalarials would be significant. This study set out to examine what proportion of ACT and other antimalarials were prescribed following a positive malaria test in four typical district hospitals serving Abuja, the capital of Nigeria and site of the Abuja Declaration.

Methods

An observational study was carried out over a period of six weeks at the paediatric outpatient clinics in the capital city of Abuja, Nigeria during July and August 2007 when malaria transmission is at its highest. Abuja is divided into five districts and six suburban districts with an area of 713 km² and a population of 1.4 million. Four government district hospitals located in different districts (Asokoro, Maitama, Nyanya and Wuse) provide health services for all residents at no cost for consultation. Each hospital runs a paediatric outpatient clinic which provides paediatric and child health services; 56,963 malaria cases and 20 deaths were reported between 2004 and June 2005 in Abuja.

Data were collected by interviewer administered questionnaires at these hospitals from mothers or caregivers who brought their children/wards to the outpatient clinics. Group consent was given before the start of the clinic, after a brief presentation about the study, and individual written consent after consulting a doctor. Interviews lasted a maximum of five minutes and were conducted in a language which the mothers/caregivers understood. Recorded diagnoses by doctors were retrieved from case notes after clinics and requested laboratory results recorded from documented laboratory records cross-checked with case notes, and prescriptions were recorded.

The primary outcome was the number of patients seen at the clinic and treated for malaria without a confirmatory positive test. Other outcomes of interest were how many patients who were diagnosed and treated for malaria had positive or negative results, how many results were not available for review by the doctor, and the type of antimalarial used by test result (positive, negative, not requested or not present).

Exclusion criteria were those admitted on the children's ward with severe disease, those seen at the general outpatient clinic and emergencies seen outside the normal clinic times of 0800–1600 h. In order to reduce the Hawthorne effect, the nature of the questionnaires was discussed with the consultant in charge of the clinics but not with the doctors running the clinics. Each morning a clinic was randomly selected for interviews. A target of 1000 patient episodes was set. Data were collected by four interviewers who were trained to record the data identically. The data were entered into EpiData 3.1 and transferred to STATA 9.0 for analysis. Simple proportions were calculated and compared where appropriate using chi-squared.

Ethical approval was given by the ethics committee of the Biomedical and Research Unit of the Federal Capital Territory Administration which manages the hospitals and the ethical committee of the London School of Hygiene and Tropical Medicine. A separate ethical approval was given by the ethics committee of the Asokoro hospital which had its own ethics committee.

Results

Data were collected from 1000 children – 115 (11.5%) from Asokoro, 233 (23.3%) Maitama, 286 (28.6%) Nyanya and 366 (36.6%) from Wuse – who presented to the paediatric outpatient department between 9 June and 3 August 2007. Girls accounted for 46% and age range was 0–13 years with a median of one year and 86.5% of all patients aged ≤ five years. A fever or history of fever was seen in 669 (66.9%) patients. Over half of the patients (57.7%) had taken some form of medication before presenting to the hospital – 12.1% had taken chloroquine and only 4.1% had received ACTs.

Of the 669 patients presenting with a fever or a history of fever, 616 (92%) had a diagnosis of malaria documented in the case notes, only 24 (3.8%) of whom where confirmed by positive slides in the hospital laboratories. This was similar across the hospitals (Table 1). Of the children with fever or a history of fever who had been prescribed antimalarials, 23 (3.8%) had negative slides and 53.4% of all patients had a diagnosis of malaria without a laboratory test being requested or available. Of the 82 malaria tests requested, 32 (40%) were not available at the time clinicians issued a prescription. It took a median of 1.5 days for slide results to be made available. Of the 256 children prescribed an ACT, 11 (4.3%) were test negative, eight (3.1%) test positive, in 11 the test was not available (4.3%), the test was not requested for 219 (86%) and seven (3%) were diagnosed as ‘not having malaria’ despite having received a prescription for an ACT. The proportion of available negative slide patients treated with an antimalarial was 23/26 (88%). This was similar to the 22/24 (91%) treated with an antimalarial who were test positive. Testing, when available, therefore made almost no impact on ACT prescribing. The lack of impact of testing on prescribing was seen for all other antimalarials and antibiotics (Table 2).

Discussion

In the capital of Nigeria, the most populous country in Africa and the home to the Abuja declaration, the overwhelming majority of those prescribed ACTs and other antimalarials had received no testing. When testing was requested it was not available for 40% and was largely ignored when it was available. These figures support data from in Lagos in 2007 and may well be representative of many other centres in Nigeria. ⁷

The great majority of the diagnoses were made with no attempt to obtain testing. The considerable delay between the tests being requested and their becoming available is likely to have played at least some part in the clinician's decision not to request them, but the fact that even where the tests were available they were ignored (88% of negative tests being prescribed and antimalarial) suggests that simply speeding up the testing is unlikely to fix this problem. Rapid diagnostic tests (RDTs) for malaria are potentially useful to get around the problems of delay as they are available immediately and can be seen by both clinicians and patients' families. However, if they are deployed under routine conditions there is no good evidence that they have an impact over and above that seen by microscopy, with trials demonstrating limited or no impact. ⁸ RDTs can be cost-effective, but only if the clinicians' prescribing practices are influenced by the results. Deploying them without some method of effectively changing the current practice is likely simply to add to the costs without a resulting benefit. ⁹

A high proportion of the children had taken some medication at home before presenting to hospital. That 12.1% proportion taking chloroquine at home and 3.1% of all hospital prescriptions were for chloroquine, despite its documented local resistance, is indicative of the lack of knowledge and education of the informal and formal prescribers; chloroquine is easily available and is very affordable. The issue of availability and affordability remains a problem in Nigeria, despite efforts by the government drug regulatory agency's efforts in preventing further manufacture and importation of chloroquine. The government of Nigeria provides free ACTs to the under fives in government hospitals and perhaps this is the reason for the documented high level of prescribing ACTs. While this is encouraging, it may not be translatable into rural settings and privately owned clinics throughout Nigeria. The prescription rate for ACTs is higher than the 26.2% seen in Lagos. ⁷

The empirical treatment of most febrile illness in children in Nigeria remains the prevailing practice with 92% treated as malaria. Only a small proportion of malaria is treated following the results of a test; often tests are not available and when they are the results are largely ignored. Improving the targeting of antimalarials is essential when using the more effective but more expensive ACTs. Working out how to do this will be challenging, but should now be a priority in Nigeria.