Background
Approximately half of the global population is exposed to malaria [
1]. Early parasitological diagnosis and efficacious treatment in all age groups are recommended by the World Health Organization (WHO) as key strategies to control malaria [
2]. However, next to policies, availability of good quality drugs and parasite based diagnostic is essential [
3]. Other important factors are access to health care services as well as the health seeking behaviour of the population. Policies are principally based on high quality research evidence, but implementation remains challenging on various levels because of some field realities.
In Sub Saharan Africa, which holds the gross of the malaria burden, the proportion of people treated for malaria with a confirmed diagnosis is low compared to other regions of the world [
4]. In addition, the accuracy of malaria diagnosis at primary health care has been reported to be poor in many areas [
5-
8]. This may explain the fact that health practitioners tend to prescribe antimalarial drugs regardless of laboratory results. Furthermore, patients can purchase antimalarial drugs directly on the informal market without any medical prescription [
9,
10].
In 2010, only a few months after its publication, the Congolese National Malaria Control Program (NMCP) adopted the second edition of the WHO malaria treatment guidelines [
2]. The major novelty was the recommendation to give antimalarial treatment upon parasitological evidence of malaria infection even in children under the age of five years. The purpose was to improve targeting of treatment and better quality of care, in particular in an environment of increased malaria control and or declining malaria endemicity [
2]. Indeed non malaria cases would be identified and may benefit from timely treatment thereby reducing morbidity and mortality. In DRC, confirmed uncomplicated malaria cases are recommended to be treated either with artesunate+amodiaquine (ASAQ) or artemether+lumefantrine (AL) as the first line, and quinine+antibiotic having antimalarial activity as second line [
11].
Improving malaria case management requires the involvement of all stakeholders and detection of the operational bottle necks. Inspired by the Piot Model [
12], the health seeking behaviour of the patients and practices of the health care providers were assessed going from the first presumed malaria symptoms until the delivery of care in formal health centres (HC). The objective was to assess health seeking behaviour of patients and practices related to the management of uncomplicated malaria by the health professional at primary health care level (PHC) compared to the policies.
Methods
Study setting and design
This was an observational study embedded in a two-stage cluster randomized survey conducted in Kinshasa, the capital city of the Democratic Republic of Congo (DRC). This city is divided into 35 health zones among them 12 were randomly selected: Bandalungwa, Binza ozone, Bumbu, Kimbanseke, Kingabwa, Kokolo, Limete, Makala, Maluku 1, Masina 1, Police and Selembao. In each health zone, one public or private HC was randomly selected among those reporting to the National Health Information System. The former study consisted of evaluating the accuracy of Optimal-IT®, as well as Paracheck-Pf® malaria rapid diagnostic tests (RDTs) [
5]. RDTs were performed and microscopy slides prepared at recruitment. HC laboratory technicians had to read the slides and were blinded to the RDTs results. The same slides were read by expert microscopists for quality control (QC) and they were blinded to previous results. The QC results constituted the gold standard to assess the accuracy of RDTs and microscopy at PHC level. The sample size of 624 was calculated for the main study based on the accuracy of Optimal-IT® RDT and the same number was involved in these supplementary analyses.
Data were collected from May to June 2011. All the patients attending the health facility and clinically diagnosed with malaria and to whom a blood smear (BS) was requested, were eligible. Clinical diagnosis was established by clinicians based on unspecific known malaria symptoms like fever, history of fever, headache, chills, weakness, muscle and joint pain and anorexia. Data was collected in 2 phases. First, a questionnaire was filled during an interview. Socio-demographic data were recorded and the patient/guardians’ health seeking behaviour was assessed i.e. current illness, action taken and health seeking practices prior to the current visit. In a second phase, the medications (antimalarial and others) prescribed by the clinicians before and after the BS results as well as the laboratory results were also recorded. Clinicians were not informed on the data collection to avoid observer’s bias. Antimalarial drugs taken before seeking care and/or prescribed at the PHC were identified using the list of antimalarial drugs available on the market in Kinshasa (e.g. Artemether+lumefantrine: Coartem®, Coartesiane®, Lonart®, Luther®, etc.).
Laboratory examination
The collected BSs were stained with 10% Giemsa for 10 minutes. Thin BSs were fixed with methanol prior to the staining. Microscopy reading was first performed by the HC laboratory technicians and result was given to clinicians, who were supposed to manage the cases according to parasitological diagnosis and National guidelines and policies [
11]. Slides were subsequently transported to the parasitology unit, Kinshasa University for QC.
Data analysis
Data were double-entered and validated in Epi info version 3.5.1 software and analysed using Stata version 11 (Stata Corp, Lakeway, College Station, Texas, USA). The primary outcome was the proportion of patients to whom a recommended antimalarial drug was prescribed after a positive test and those for whom no prescription was made after a negative test. Secondary outcomes included description of patients’ health seeking behaviour (time for consultation regarding the illness episode onset, medication taken before consultation) and malaria testing outcome compared to patients’ health seeking behaviour and clinicians’ practices. Descriptive statistics were used to get frequencies and percentages. Cross tabulations were performed to check relationship between variables of interest using the Chi-square test of Pearson. Predictors of malaria confirmation were assessed by a logistic regression, referring to experts’ microscopy reading result. The cluster effect was checked and did not influence the results.
Ethical approval
The study was approved by the Committee for Medical Ethics of the Antwerp University Hospital, Belgium (approval reference: 14/36/236) and the Ethical Committee of the School of Public Health, Kinshasa University, DRC (approval reference: ESP/CE/082/10). Before inclusion, written informed consent was obtained from all patients or their legal guardians if minors.
Discussion
Results of this survey highlight the alarming situation of malaria management at PHC in the DRC because of a huge discrepancy between policy and field reality. Almost half of the patients (45.7%) were managed in line with NMCP guidelines. However, 7 patients (2.1%) received a prescription for AS+SP which is not officially recommended in DRC but still in line with the WHO recommendations. Besides, a proper prescription is not enough to ensure accurate malaria management. The steps forward are availability of quality-assured antimalarial drugs and the patient/ care giver’s compliance to the treatment [
13,
14]. Nevertheless, this study did not explore those aspects. The lack of compliance with NMCP guidelines has been reported elsewhere in sub-Saharan Africa [
15-
20]. Malaria control programs need to elaborate an efficient strategy for the follow up of adherence to guidelines by practitioners. In addition, a dialogue between policy makers, researchers and program managers is needed to discuss and address the gaps in the implementation of malaria management policies [
21].
The overtreatment reported in this survey could be the result of unsatisfactory diagnostic accuracy at PHC [
5,
22]. This may explain why clinicians seldom relied on the BS results. Another explanation could be related to the mistrust of policies evidence, as reported in other studies [
23,
24]. In Tanzania, Reyburn
et al. [
25] found that almost half of patients with a negative malaria test were treated with an antimalarial drug. The responsibility of the overtreatment is to be shared with the patients who can influence health practitioners’ practice [
26,
27]. In Ghana, participants to a focus group discussion claimed that a test should result in a diagnosis, even following a negative result [
28]. In this aspect, clinicians may choose to prescribe antimalarial drugs, mostly when laboratory diagnoses for other infections are not available.
Clinicians tended to prescribe more antimalarial drugs to children younger than five, even when laboratory results were negative. Clear guidelines for management of non-malaria febrile illnesses as well as appropriate tests and drugs are urgently needed [
29]. By the time this survey was conducted, ASAQ and AL were the recommended first line ACTs. It was noticed that clinicians tended to prescribe more ASAQ when BS result was positive. When prescription was done before BS or in spite of negative BS result, either Quinine or AL was preferred. Assuming that they were not aware about the introduction of AL in the national guidelines some months prior to this assessment, the hypothesis is that they had preference for non-recommended antimalarial drugs in absence of malaria confirmation.
Combination treatment using ACTs is the key strategy to reduce the impact of failing monotherapies [
30,
31]. Unfortunately and against the national guidelines, this survey highlighted a worrying proportion of prescribing monotherapy, including artemisinin monotherapy (AMT), a practice that could increase treatment failures and contribute to the spread of resistance. Despite the WHO’s call to stop AMT production and marketing to refrain the development of drug resistance [
31], the field reality is still different. However, a trend towards their disappearance from the market in Kinshasa is notably observed. In 2009 they constituted 10% and 41% of antimalarial drugs stocked in public/ Not-for-Profit and private sector outlets respectively [
13]. Four years later their availability decreased below 1% in both sectors [
32].
It is worth mentioning the ongoing initiative in the country supported by the Global Fund that aims to supply RDTs as well as recommended medications. The support is presently available in 60% of the health zones in DRC. At the time this survey was conducted, two of the twelve HC involved were supplied, although on an irregular basis.
Apparently, none of the clinicians involved was aware, 6 months after the NMCP guidelines update, that quinine should not be used as a monotherapy. This survey did not focus on the knowledge nor the background of practitioners and further research will be needed to assess explanatory causes and identify the bottlenecks where training or adapted measurements can be taken. The lack of this information is a limitation of the study, but these findings may be reproducible in other settings.
Almost half of the patients (44.2%) undertook a treatment prior to the current consultation and they tended to be more infected with malaria. This means that either the medication taken was not effective or the treatment course was not appropriate. Most of them used antimalarial, antibiotics and antipyretics. Such inappropriate use of antibiotics and antimalarials is dangerous as it may contribute to the emergence and spread of drug resistant bacteria and
Plasmodium [
29,
33]. Self-treatment has been reported to be common, mainly in rural areas of Africa [
9]. People should be encouraged to address to the health facilities as soon as the clinical episode starts, for proper diagnosis and treatment. In the urban settings like Kinshasa, geographical access is not the main obstacle but affordability of the health services can be the main limiting factor. This could explain why few people with history of fever attended the HC the same day. Access to health may be increased by increasing the health insurance coverage either through the Government or the private stakeholders.
Approximately a third (35.9%) of those owning a mosquitonet slept under it the night before consultation. The protective impact on sleeping under mosquitonet was shown by the multivariate analysis. Hence the campains are highly needed to sensitize the mass on the benefits of using the mosquitonet.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
HMM, PL and JPV designed the study protocol. HMM and GI supervised the activities on the field. HMM performed the statistical analysis. RIL, HMM and JPV drafted the manuscript. HMM, JPV, PL, GI, RIL and AK reviewed the paper and all authors approved the final version.