Overall trends in anti-malarial prescriptions
A large proportion of patients were treated with anti-malarials, although malaria was not confirmed parasitologically in most cases. This was due to the policy regarding management of patients in malaria endemic regions, before the introduction, in 2007, of rapid diagnostic tools to confirm malaria diagnosis. Until 2003, management of the patients consisted in early diagnosis and treatment at the health care facility or home-based treatment of fevers with CQ. In health centres, suspected cases of uncomplicated malaria were treated with CQ whereas quinine had to be saved for severe cases. Actually, health agents did not always comply with official guidelines and anti-malarials were extensively used. Frequently, patients suffering from another disease received anti-malarial drugs as well (19.8% in this study), usually CQ, as a "precautionary principle". Besides, quinine was not used for severe cases only.
Even if anti-malarials were prescribed throughout the year for a great number of patients, their use was mainly concentrated during the rainy season. Seasonality of prescriptions followed malaria morbidity patterns in this mesoendemic area, where malaria transmission occurs principally in a four-month period, leading to a rise in malaria cases between August and December, with a peak in the month of October. Presumptive malaria cases represented approximately 40% of total attendance through the observed period.
After a period of increasing clinical malaria cases and CQ prescriptions from 1992 to 1996, a sudden drop of prescriptions was observed in 1997, following the fall in consultations the same year in the health care facilities (Figure
2). Two explanations are proposed. First, rainfall was lower in 1997 than in the previous years (419 mm precipitations
vs a yearly average of 522 mm during 1992–1996, as recorded in the Niakhar station by our own rain gauges), which may have reduced malaria transmission and, therefore, the morbidity attributable to malaria. Second, the immunization trials conducted until 1997 may have played a motivating role on the sanitary activities during the preceding years [
21,
25] inducing a high level of prescriptions. Other drug prescriptions also decreased by 10–20% between 1996 and 1997, associated to a decline in medical consultations.
Until the late 1990s, health workers did not change their practices because either they have not realized there was a rise in CQ resistance, as CQ decreasing effectiveness was not obvious, or they have lacked other therapeutic options. Besides, an additional study (Munier et al, in press) showed an absence of significant increase in patients' returns to the health centre after anti-malarial treatment over the period of study, which is not in favour of a perception of CQ decreasing effectiveness at the dispensary level.
In the early 2000s, the slight decrease in CQ prescription was concomitant with the persistence of the use of quinine at a high level and the rise of SP prescription. It could be attributed to a better awareness of the growing resistance of P. falciparum, which may have become clinically apparent, prompting health agents to opt for alternative drugs (SP or quinine). Another hypothesis is that health care providers anticipated the implementation of the new policy. Being aware of discussions held at national and local levels about the change of anti-malarial strategy, they may have started to change their practices before the new policy was adopted. Finally, patients' complaints, such as itching and eye allergies, intensified after the National Pharmacy liberalized the CQ sources of supply, possibly due to the presence of different excipients in these products that would be less tolerated by patients. This may have incited prescribers to use alternative drugs. At that time, they could also order cheaper generic SP at the health district (Maloxine®), compared to the Fansidar® proprietary drug.
Furthermore, contrary to CQ, interest in quinine remained stable in this area. As it is known to be very effective, patients often request injections and they even seek them from other providers, when they cannot get them at the dispensary [
20]. Thus health agents made an excess use of quinine.
As the delivery of drugs to the facility was dependant on the drug policy, the new directive was followed by the abrupt drop of CQ prescription in 2003–2004 (Figures
2 and
3) and the use of AQ and SP prescribed in association.
Therapeutic practices according to the health care facility
Beyond the general tendency, use of anti-malarials was highly variable according to the health centre, indicating distinct therapeutic practices, especially in 2004 after the change in national anti-malarial policy.
After the policy change, the National malaria control programme recalled the remaining CQ stocks from public health centres. However, some facilities finished their own CQ stock before ordering the new anti-malarial drugs. In any case, CQ was not supposed to be sold anymore at the district pharmacy's level. While in Toucar the national recommendations were followed thoroughly and CQ delivery was stopped in 2004, the delay was higher in the two other centres. In Ngayokhem, it seems that there were remaining stocks of CQ, which were further used during the dry season. The new efficacious drugs started to be prescribed in the rainy season, when a high increase of malaria morbidity was expected. Finally, in privately-run Diohine, CQ was still prescribed in 2004 all year long, although quinine prescriptions rose above CQ during the rainy season. It may be explained by the continuing supply of CQ by a catholic organization, which had not terminated its stocks. Only three patients received AQ/SP bi-therapy. In Toucar, the attendance of the same experienced nurse since 1996 is most probably a key factor in the respect of treatment guidelines.
Other practices outside official health care facilities, such as self-medication and purchase of medicines via the parallel drug market may have played a role. Previous socio-anthropological studies conducted in this rural zone [
26‐
28] have shown that these practices are associated with complex behaviour resulting in a high proportion of patients not attending modern health facilities. Even though self-medication is generally the first response in case of a child's fever, populations preferably use symptomatic than etiological treatment [
26]. They also considered that CQ was more efficient in preventing than treating malaria. A population-based survey conducted in the same area in 2001 [
28] showed that anti-malarials represented only 18.2% of total self-medication (versus 64.7% for antipyretics). Furthermore, CQ was seldom available on markets in the years preceding the policy change, even though it was a cheap and affordable drug. Actually, customers rarely asked about it, and for the retailer, CQ was not profitable enough, being useful in a three-month period only. To cure malaria, they rather sold paracetamol or aspirin, which would treat the fever symptom (Le Hesran and Baxerres, personal communication).
This health-seeking behaviour was present during the entire period of study and probably does not interfere with overall trends in anti-malarial prescriptions in health care facilities, which represent the main anti-malarials providers in this area.
Neither AQ nor SP was found on the parallel market until 2003. Shortly after the policy change, it was noticed that some health workers in a neighbouring region [
20] had not been properly trained for the new policy and did not always follow guidelines, prescribing AQ and SP in monotherapy, or quinine injections.