Predictive indices of empirical clinical diagnosis of malaria among under-five febrile children attending paediatric outpatient clinic


Background: Malaria has remained an important public health problem in Nigeria with children under 5 years of age bearing the greatest burden. Accurate and prompt diagnosis of malaria is an important element in the fight against the scourge. Due to the several limitations of microscopy, diagnosis of malaria has continued to be made based on clinical ground against several World Health Organization (WHO) recommendations. Thus, we aim to assess the performance of empirical clinical diagnosis among febrile children under 5 years of age in a busy pediatric outpatient clinic. Materials and Methods: The study was a cross-sectional study. Children aged <5 years with fever or 72 h history of fever were recruited. Children on antimalarial prophylaxis or on treatment for malaria were excluded. Relevant information was obtained from the caregiver and clinical note of the child using interviewer administered questionnaire. Two thick and two thin films were made, stained, and read for each recruited child. Data was analysed using SPSS version 16. Results: Of the 433 children studied, 98 (22.6%) were empirically diagnosed as having malaria and antimalarial drug prescribed. Twenty-three (23.5%) of these children were confirmed by microscopy to have malaria parasitemia, while 75 (76.5%) were negative for malaria parasitemia. Empirical clinical diagnosis show poor predictive indices with sensitivity of 19.2%, specificity of 76.0%, positive predictive value of 23.5% and negative predictive value of 71%. Conclusion and Recommendations: Empirical clinical diagnosis of malaria among the under-five children with symptoms suggestive of acute malaria is highly not reliable and hence the need to strengthen parasitological diagnosis.

Keywords: Empirical clinical diagnosis, malaria, negative predictive value, positive predictive value, sensitivity, specificity, under-five

How to cite this article:
Elechi HA, Rabasa AI, Alhaji MA, Bashir MF, Bukar LM, Askira UM. Predictive indices of empirical clinical diagnosis of malaria among under-five febrile children attending paediatric outpatient clinic. Ann Trop Med Public Health 2015;8:28-33


How to cite this URL:
Elechi HA, Rabasa AI, Alhaji MA, Bashir MF, Bukar LM, Askira UM. Predictive indices of empirical clinical diagnosis of malaria among under-five febrile children attending paediatric outpatient clinic. Ann Trop Med Public Health [serial online] 2015 [cited 2020 Aug 8];8:28-33. Available from:



Malaria is endemic in Nigeria, constituting the country’s most significant public health problem. [1] Children under 5 years of age in areas of high transmission are the most vulnerable age group, with the highest malaria morbidity and mortality. [2] They normally have the highest prevalence of malaria infection of all population groups. [2] Most of them experience their first malaria infection during the first year or two of life, when they have not acquired adequate clinical immunity, thus making these early years particularly dangerous. [3]

A key to effective management of malaria is prompt and accurate diagnosis. Accurate diagnosis of malaria is necessary to prevent morbidity and mortality while avoiding the unnecessary use of Antimalarial agents. Historical strategies to diagnose malaria range from basic empirical clinical diagnostic algorithms to examination of stained blood smears by light microscopy. Diagnosis of malaria based on blood slide microscopy has remained the gold standard for many years. [3],[4] However, the inherent limitations [5],[6],[7] of microscopy have severely hindered its universal routine use particularly in a busy outpatient clinic setting. [8] Malaria rapid diagnostic tests (MRDTs) were introduced in the early 1990s with the potential to overcome the weaknesses associated with microscopy. [9] However, these kits have performed poorly in under-five children with uncomplicated malaria. [10] Empirical clinical diagnosis, however, has remained the most convenient and readily available method and hence, diagnosis is often made on clinical ground. [11] The accuracy of empirical clinical diagnosis is poor, since the symptom complex of malaria overlaps with those of many other tropical diseases, and coinfections can occur. [12],[13],[14]

This study thus aims to determine the predictive indices of Empirical clinical diagnosis of malaria among febrile children under-five in a busy outpatient clinic. This is necessary to evaluate the appropriateness, or otherwise, of the day-to-day prescription of antimalarial treatment based on clinical diagnosis of malaria.

Materials and Methods

Study area

The study was carried out at the pediatric general outpatient (PGOP) unit of University of Maiduguri Teaching Hospital (UMTH), Maiduguri, Borno State of Nigeria. Maiduguri, the capital of Borno State is located in the northeastern part of Nigeria. It is a semi-arid zone lying between latitude 11.5΀ N and longitude 13.5΀ E with a sunny weather and a temperature that may be as high as 45΀ C, especially in the hot dry season, and an annual rainfall of 1.14 mm to 771.90 mm. [15] UMTH is a center of excellence for infectious diseases and immunology. It serves as a referral center not only for the six states in the region (Adamawa, Bauchi, Borno, Gombe, Taraba, and Yobe) but also for the neighboring countries of Cameroon, Chad, and Niger. The PGOP unit is a busy clinic with an average population of 100-150 patients per day with no attached side laboratory.

Study design

The study was a hospital based cross-sectional observational study.

Study population and sampling method

Under-five febrile children attending the PGOP unit of UMTH were eligible to participate after meeting the inclusion criteria. Convenient sampling method was employed and patients were recruited consecutively after fulfilling the inclusion criteria. Calculated minimum sample size was 377 using Taylor’s formula [16] and value of p was taken from the study of Ikeh et al.[17] from Jos, Nigeria.

Inclusion criteria

  1. Age of 0-59 months
  2. Fever (axillary temperature >37.5΀C), and/or history of fever in the 72 hours prior to presentation. [12]
  3. Informed parental consent.

Exclusion criteria

  1. Children on antimalarial treatment or prophylaxis prior to presentation.

Ethical considerations

Approval was sought from and granted by the research and ethical committee of UMTH. Signed or thumbprinted informed consent was obtained from each parent/guardian with unlimited liberty to deny consent or opt out of the study at any stage without any negative consequence. Information and results obtained were kept confidential. Results of the tests were disclosed to the guardians and those with positive malaria parasitemia were given antimalarial (Artemether/lumefantrine tablets 20 mg/120 mg) free of charge at the expense of the researchers.

Study procedure

The study was carried out from August 5 to October 20, 2011. On the day of inclusion, demographic and clinical information were obtained from the care giver and the clinical note of the attending physician. These included biodata, parents occupation and educational qualification. Others are presenting complaints, clinical diagnosis and treatments offered. Weight was measured using a digital bathroom weighing machine (Salter Glass Electronic Bathroom Scale) in kilograms to two decimal places and for children who could not stand, the caregiver was weighed alone and then with the child and the difference of the two was taken as the child’s weight. Length was measured using a tape meter on a hard cardboard surface to the nearest centimeters. Axillary temperature was measured using a digital thermometer (JOYCARE; ) in centigrade to one decimal place. Socioeconomic status was determined from parental education and occupations using the model by Ogunlesi et al.[18] A score of 1-5 was awarded for each of education and occupation of both parents separately and the mean of these four scores to the nearest whole number was the socioeconomic status (I, II, II, IV and V) assigned to the child. Classes I and II belong to upper class, while class III and classes IV and V belonged to the middle and lower socioeconomic classes, respectively. Nutritional status was assessed using the Z-score system in accordance with the National Center for Health Statistics (NCHS)/World Health Organization (WHO) reference population, [19] a Z-score of ≥-2 was classified as normal and Z <-2 as under-nutrition for the three anthropometric indices computed. Thick and thin blood smears were prepared from a capillary blood sample. A number was allotted to every participant at the point of entry and was used for identification of slides and questionnaire from the same patient. The thin blood smears were fixed with methanol and the thick smears were left unfixed. Each slide was subsequently stained with 10% Giemsa solution for ten minutes. [20] All blood smears were examined microscopically under x100 oil immersion. The thick smears were used for diagnosis of Plasmodium specie. Smears were considered negative if no parasites were seen in 100 oil-immersion fields. The thin smears were examined to confirm the parasite species for positive samples. All slides were double-read, blinded, by the 6 th author, a qualified and experienced microscopist from the Department of Parasitology at UMTH and the lead author, who was retrained and certified by a parasitologist prior to commencement of the study, with an agreement of >95% between the lead author and the microscopist in slide reading. The slides with discordant findings were resolved through discussion and re-examination of such slide by the both authors at the same time with consensus reached on each case.

Data obtained were entered into a computer to generate a data base. Analysis was done using SPSS version 16.0 (SPSS, Chicago, ILL, USA). Baseline characteristics (demographic, clinical, and parasitological) were analyzed using descriptive statistics; mean, mode, medians, and standard deviation were used as appropriate. Results were presented in tables. Frequencies and proportions were compared using Chi-square (x 2 ), strength of association were tested using the contingency coefficient. A 95% confidence interval (95% CI) and a P-value of <0.05 were considered significant.


Sociodemographic feature of the study population

A total of 433 children aged less than 60 months were studied. Out of these, there were 238 (55%) males and 195 (45%) females (M:F ratio 1.2:1). The mean age of the study population was 19.2 ± 14.3 months. Approximately half of the children studied, 203 (46.9%), were aged 12 months and younger. The least frequency, 18 (4.2%), was observed among the 49 months and above age category [Table 1].

Table 1: Age and sex distribution of the study population

Click here to view

The majority, 377 (77.8%), of the studied population were of low socioeconomic status. The remaining 77 (17.8%) and 19 (4.4%) of them belonged to the middle and upper socioeconomic classes respectively. This is due to the fact that majority of the parents did not go beyond secondary education (62.2% of the fathers and 81.1% of the mothers) and therefore are low-income earners.

Greater than a third, 164 (37.9%), were underweight while 123 (28.4%) of them were stunted. Among the under-nourished children, moderate under-nutrition (-2 > weight for age z-score [WAZ], weight for height z-score [WHZ], or height for age z-score [HAZ] ≥ -3) was more frequent than severe undernutrition (WAZ, WHZ, or HAZ < -3) accounting for 60%, 64% and 55% of under-nutrition for WAZ, WHZ, and HAZ, respectively.

Clinical feature of the study population

Looking at clinical features at presentation, 168 (38.8%) of the study population had fever at presentation with axillary temperature ranging between 37.6 and 40.1΀ C, while 264 (61.2%) had history of fever within the preceding 72 h. The mean, median, and mode of the axillary temperature of the studied population were 37.2΀ C, 37΀ C, and 38΀ C, respectively.

Two hundred and sixty (60.1%) of the patients presented with cough, while 100 (23.1%), 82 (18.9%), and 65 (15%) presented with diarrhea, vomiting, and nasal discharge, respectively.

Eighty-five (19.6%) of them presented with other symptoms such as: Fast breathing, abdominal pain, dysuria, ear discharge, and septic wounds. Acute respiratory tract infection was the commonest diagnosis, 278 (64.2%), while acute diarrheal disease and malaria were experienced in 100 (23.1%) and 98 (22.6%), respectively. Other clinical diagnoses accounted for 27 (6.2%) and included skin sepsis, meningitis, osteomyelitis, drug allergy, acute poisoning, and so on [Table 2].

Table 2: Frequency of empirical diagnoses among under-five febrile children in Maiduguri

Click here to view

Prevalence of malaria parasitemia

The prevalence of malaria parasitemia in this study was 27.7%. The effect of different variables on the prevalence of malaria parasitemia in the study population is given in [Table 3]. The age-group specific prevalences for malaria parasitemia were 26.6, 27.0, 27.3, 27.8 and 44.4% for 0-12, 13-24, 25-36, 37-48, and 49-59 months, respectively. Although, slight differences were observed in the age group specific prevalences of malaria parasitemia in this study, this difference was not statistically significant (x2 = 2.680, P = 0.611). There was slightly higher preponderance of malaria parasitemia among males (29.8%) compared to female (25.1%), this difference was, however, not statistically significant (x2 = 1.184, P = 0.277). The lower socioeconomic (SEC) class recorded the highest prevalence of 41%. The middle SEC had the least prevalence of 19.5% while the upper SEC recorded 36.8%. However, this difference was not statistically significant (X2 = 1.417, P = 0.234). The prevalence of malaria parasitaemia was higher among the under-nourished children across the three measured anthropometric indices [Table 3]. However, these differences were not statistically significant (x2 = 1.014, 2.597 and 0.868, for WAZ, WHZ, and HAZ respectively, P > 0.05).

Table 3: Prevalence of malaria parasitemia by various variables in the study population

Click here to view

Predictive indices of empirical diagnosis

Empirical diagnosis of malaria was made in 98 (22.6%) patients in this study. Of these, 23 (23.5%) were confirmed microscopically, while the remaining 75 (76.5%) were negative for malaria parasitaemia [Table 4]. Empirical diagnosis of malaria showed poor predictive indices with sensitivity of 19.2% (95% CI: 12.2-26.2) and specificity of 76.0% (95% CI: 71.3-80.7). The positive and negative predictive values were 23.5% and 71%, respectively.

Table 4: Empirical clinical diagnosis in the diagnosis of malaria in the under-five febrile children in Maiduguri

Click here to view



The accuracy of empirical clinical diagnosis is poor, even in countries of endemicity with high malaria incidence rates. The symptom complex of malaria overlaps with those of many other tropical diseases, and coinfections can occur. [12],[13],[14] In high transmission areas of sub-Saharan Africa, the distinction between the clinical disease of malaria and malaria parasitemia is especially difficult. Persons may present with a wide variety of other fever-inducing diseases accompanied by a parasitemia that is not related to the presenting symptoms. [21] In such settings, no testing paradigm currently exists to actually confirm that a given illness is caused by malaria parasitemia.

The sensitivity of empirical diagnosis of malaria in this study of 19.2% was very low. This implies that greater than 80% of the children with malaria would have been missed if diagnosis and treatment were based on empirical criteria alone. However, this assumption must be considered with caution as parasitaemia is not synonymous to clinical malaria in malaria endemic area. [21] Ben-Edet et al.[22] from Lagos reported similar low sensitivity while other authors have found very high sensitivities. [23],[24],[25],[26],[27] The low sensitivity observed may be attributed to the low prevalence of malaria in this study. Several studies have demonstrated the direct relationship between prevalence of malaria and sensitivity of empirical diagnosis. [23],[24],[25] Again, lack of standardized criteria for clinical diagnosis of malaria in this study may explain the low sensitivity in comparison to other studies with standardized criteria and very high sensitivity.

The high specificity of 76% found in this work for empirical malaria implies that a quarter of children who had no malaria parasitaemia were erroneously diagnosed clinically as malaria. This would have led to failure to look for alternative causes of illness and wastage of antimalarial drugs in 24% of those with empirical malaria if management decision were based solely on empirical diagnosis. Other studies have found comparable specificities. [28],[29] However, most studies have found low specificities with some recording specificity of 0%. [23],[26],[27] The high specificity recorded in this study could be attributed to the non-specific criteria defining clinical malaria. This is supported by the study of Bojang et al.[29] in the Gambia where a hot body alone and normal chest on examination had specificity of 20 and 17%, while pooling together several predictors in the same study resulted in overall specificity of 61%.

The positive predictive value (PPV) of 23.5% implies that the probability of empirically diagnosed malaria being truly malaria is only 0.235. This makes it unreasonable to rely on clinically diagnosed malaria in making treatment decision. Some studies have reported equally low PPV, [26],[27],[30] while others had higher PPV. [31],[32],[33] The low PPV in this study may be due to low prevalence of malaria in this study.

The negative predictive value on the other hand was moderate at 71%. This implies that the probability that a negative case of empirically diagnosed malaria being truly negative is 0.71. This level of probability emphasizes the need to further exclude malaria by other means when a child is negative for malaria empirically. Similar or comparable NPV have been reported by other workers, [25],[31],[32] while others have reported higher NPV. [26],[27],[30] This level of NPV in this study could be due to the low prevalence recorded in this work.

Conclusion and Recommendations

Empirical clinical diagnosis of malaria among children under-five with symptoms of acute malaria is highly unreliable. It has enormous implications of missing children with malaria, as well as over-diagnosis with attendant unnecessary antimalarial prescription and failure to search for alternative causes of fever. Hence, parasitological diagnosis of malaria is recommended in children with features of acute malaria as a requisite for antimalarial prescription.



National Population Commission [Nigeria] and ICF Macro. Nigeria Demographic and Health Survey 2008. Abuja: National Population Commission and ICF Macro; 2009. p. 187-96.
WHO. The Role of Laboratory Diagnosis to Support Malaria Disease Management: Focus on the use of Rapid Diagnostic Tests in Area of High Transmission – Report of WHO Technical Consultation, 25-26 October 2004. Geneva: World Health Organization; 2006. p. 4-28.
McGregor IA, Wilson RJ. Specific immunity acquired in man. In: Wernsdorfer WH, McGregor IA, editors. Malaria: The Principles and Practice of Malariology. Vol. 1. Edinburgh: Churchill Livingstone; 1988. p. 735-51.
Murray CK, Bell D, Gasser RA, Wongsrichanlai C. Rapid diagnostic testing for malaria. Trop Med Int Health 2003;8:876-83.
Coleman RE, Sattabongkot J, Promstaporm S, Maneechai N, Tippayachai B, Kengluecha A, et al. Comparison of PCR and microscopy for the detection of asymptomatic malaria in a Plasmodium falciparum/vivax endemic area in Thailand. Malar J 2006;5:121.
Kain KC, Harrington MA, Tennyson S, Keystone JS. Imported malaria: Prospective analysis of problems in diagnosis and management. Clin Infect Dis 1998;27:142-9.
Zurovac D, Midia B, Ochola SA, English M, Snow RW. Microscopy and outpatient malaria case management among older children and adults in Kenya. Trop Med Int Health 2006;11:432-40.
Iwelunor J, Airhihenbuwa CO, King G, Adedokun A. Contextualizing child malaria diagnosis and treatment practices at an outpatient clinic in Southwest Nigeria: A qualitative study. ISRN Infect Dis 2013, Article ID 101423.
Murray CK, Jasin WB. Rapid diagnosis of malaria. Interdiscip Perspect Infect Dis 2009;2019:415953.
Ratsimbasoa A, Ravony H, Vonimpaisomihanta JA, Raherinjafy R, Jahevitra M, Rapelanoro R, et al. Management of uncomplicated malaria in febrile under five-year-old children by community health workers in Madagascar: Reliability of malaria rapid diagnostic tests. Malar J 2012;11:85.
Manirakiza A, Njuimo SP, Le Faou A, Malvy D, Millet P. Availability of antimalarial drugs and evaluation of the attitude and practices for the treatment of uncomplicated malaria in Bangui, Central African Republic. J Trop Med 2010, Article ID 510834.
Chandramohan D, Jaffar S, Greenwood B. Use of clinical algorithms for diagnosing malaria. Trop Med Int Health 2002;7:45-52.
Berkley JA, Maitland K, Mwangi I, Ngetsa C, Mwarumba S, Lowe BS, et al. Use of clinical syndromes to target antibiotic prescribing in seriously ill children in malaria endemic area: Observational study. BMJ 2005;330:995.
Hamer DH, Ndhlovu M, Zurovac D, Fox M, Yeboah-Antwi K, Chanda P, et al. Improved diagnostic testing and malaria treatment practices in Zambia. JAMA 2007;297:2227-31.
Office of the Zonal Meteorological Inspector, Meteorological Agency, Federal Ministry of Transport and Aviation, Maiduguri, Borno State.
Hassard TH. Understanding Biostatistics. Saint Louis, MO: Mosby Year Book; 1991. p. 167-81.
Ikeh EI, Teclaire NN. Prevalence of malaria parasitaemia and associated factors in febrile under-5 children seen in primary health care centres in Jos, North Central Nigeria. Niger Postgrad Med J 2008;15:65-9.
Ogunlesi TA, Dedeke IO, Kuponiyi OT. Socio-economic classification of children attending specialist paediatric clinic in Ogun State, Nigeria. Niger Med Pract 2008;54:21-5.
De Onis M, Blossner M. WHO Glibal Database on Child Growth and Malnutrition. Geneva: World Health Organization (WHO); 1997. p. 45.
Marianne L. Giemsa Staining of thick or thin blood film. In: Kristen M, Inger L, Heduig P, Artur S, Mats W, editors. Methods in Malaria Research. 5 th ed. Paris: BioMalPar; 2008. p. 17.
Taylor TE, Fu WJ, Carr RA, Whitten RO, Mueller JS, Fosiko NG, et al. Differentiating the pathologies of cerebral malaria by postmortem parasite counts. Nat Med 2004;10:143-5.
Ben-Edet AE, Lesi FE, Mafe AG, Grang AO. Diagnosis of plasmodiam falciparum malaria in children using the immuno-chromatographic diagnostic technique. Niger J Paediatr 2004;31:71-8.
Oliver M, Develoux M, Chegou Abari A, Loutan L. Presumptive diagnosis of malaria results in a significant risk of mistreatment of children in urban Sahel. Trans R Soc Trop Med Hyg 1991;85:729-30.
Olaleye B, Williams LA, D’Alessandro U, Weber MM, Mulholland K, Okorie C, et al. Clinical predictors of malaria in Gambian children with fever or a history of fever. Trans R Soc Trop Med Hyg 1998;92:300-4.
Redd SC, Kazembe PN, Luby SP, Nwanyanwu O, Hightower AW, Ziba C, et al. Clinical algorithm for treatment of Plasmodium falciparum malaria in children. Lancet 1998;347:223-7.
Chandramohan D, Carneiro I, Kavishwar A, Brugha R, Desai V, Greenwood B. A clinical algorithm for the diagnosis of malaria: Results of an evaluation in an area of low endemicity. Trop Med Int Health 2001;6:505-10.
Weber MW, Mulholland EK, Jaffar S, Troedsson H, Gove S, Greenwood BM. Evaluation of an algorithm for the integrated management of childhood illness in an area with seasonal malaria in the Gambia. Bull World Health Organ 1997;75(Suppl 1):25-32.
Luxemburger C, Nosten F, Kyle DE, Kiricharoen L, Chongsuphajaisiddhi T, White NJ. Clinical features cannot predict a diagnosis of malaria or differentiate the infecting species in children living in an area of low transmission. Trans R Soc Trop Med Hyg 1998;92:45-9.
Bojang KA, Obaro S, Morison LA, Greenwood BM. A prospective evaluation of a clinical algorithm for the diagnosis of malaria in Gambian children. Trop Med Int Health 2000;5:231-6.
Muhe L, Oljira B, Degefu H, Enquesellassie F, Weber MW. Clinical algorithm for malaria during low and high transmission seasons. Arch Dis Child 1999;81:216-20.
Rooth I, Björkman A. Fever episodes in a holoendemic malaria area of Tanzania: Parasitological and clinical findings and diagnostic aspects related to malaria. Trans R Soc Trop Med Hyg 1992;86:479-82.
Gomes M, Espino FE, Abaquin J, Realon C, Salazar NP. Symptomatic identification of malaria in the home and in the primary health care clinic. Bull World Health Organ 1994;72:383-90.
Genton B, Smith T, Baea K, Narara A, al-Yaman F, Beck HP, et al. Malaria: How useful are clinical criteria for improving the diagnosis in a highly endemic area? Trans R Soc Trop Med Hyg 1994;88:537-41.

Source of Support: None, Conflict of Interest: None


DOI: 10.4103/1755-6783.157275


[Table 1], [Table 2], [Table 3], [Table 4]

Paul Mies has now been involved with test reports and comparing products for a decade. He is a highly sought-after specialist in these areas as well as in general health and nutrition advice. With this expertise and the team behind, they test, compare and report on all sought-after products on the Internet around the topics of health, slimming, beauty and more. The results are ultimately summarized and disclosed to readers.


Please enter your comment!
Please enter your name here