A cross-sectional study for algorithm in diagnosing simple uncomplicated malaria in children in health facilities without laboratory backup in Nigeria.

AIMS AND OBJECTIVES

The objective of this study was to determine an algorithm for malaria diagnosis using presenting signs and symptoms of children (aged 0-13 years) with uncomplicated malaria in Gwagwalada Area Council of Abuja, Nigeria.

MATERIALS AND METHODS

A validated questionnaire was used to obtain relevant data from 400 children diagnosed presumptively of simple malaria by clinicians and 400 other children of similar sex and age considered as not having malaria. Giemsa-stained thick blood films were used to determine parasitaemia. Data obtained was analysed using Epi-Info version 3.3.2.

RESULTS

Thirty-eight per cent of children with presumptive diagnosis of malaria had parasitaemia. Fever, rigor, vomiting, jaundice, pallor and spleen enlargement had significant statistical relationship with parasitaemia on bivariate analysis, but only fever (p=0.00), rigor (p=0.00), vomiting (p=0.00), and pallor (p=0.00) maintained the relationship when subjected to logistic regression analysis. But these symptoms individually had low sensitivity and/or specificity. Candidate algorithms (combinations of symptoms) were then successively subjected to bivariate, logistic and validity analyses. Fever with vomiting gave the highest sensitivity (56.2%), specificity (76.4%) and PPV (60.0%) and were therefore adopted as the algorithm of choice.

CONCLUSION AND RECOMMENDATIONS

Children presenting with fever and vomiting without any other obvious cause in health facilities without laboratory support in the research area should receive antimalarial treatment, to help reduce the malaria scourge. This algorithm should be field-tested and if found reliable should be adopted to ease the problem of malaria diagnosis in peripheral health facilities.