This study took place in 9 neighbourhoods in the 3 boroughs of Parakou city. These were precisely Albarika, Dépôt and Titirou for the 1^st borough, Banikanni, Nima-sokounon and Arafat in the 2^nd borough, then Zongo, Guêma and Tranza for the 3^rd borough.

It was a descriptive cross-sectional analytical survey that took place from April 15 to June 24, 2017.

The target population consisted of all the inhabitants of the nine neighbourhoods of Parakou city mentioned above.

This was a stratified probabilistic sampling. The minimum sample size calculated by the Schwartz formula with a 95% confidence interval and an accuracy of 5% was 237 respondents. They were recruited at least 27 respondents per neighbourhood. Three neighbourhoods were drawn lots in each of the three boroughs, making a total of nine neighbourhoods randomly selected out of 58 in the city of Parakou. Then, at the entrance of each neighbourhood, a pen rotation was used to determine the direction to follow. In this direction, households are systematically recruited. Within each household, simple random sampling was used to select only one respondent from those who met the inclusion criteria.

Data were collected using an individual survey sheet during a direct interview with each respondent. A pre-test questionnaire was carried out on fifteen randomly chosen respondents in a non-selected neighbourhood for the survey.

The dependent variable was anti-malarial self-medication. This consists of treating malaria by taking a drug without any medical prescription and regardless of the supply place. To purchase anti-malarial drugs in pharmacies in Benin there is no need to present a medical prescription, only the purchasing of drugs in pharmacy on the advice of a State-certified pharmacist was considered regulatory and not self-medication.

Independent variables were related to the sociodemographic, economic, clinical, therapeutic data, the drug dosage form, the supply place, the dosage, the distance from home to supply place, the distance from home to the nearest pharmacy and the nearest health facility, the respondents’ perceptions of self-medication, the respondents’ knowledge of self-medication risks.

Data were entered and analysed using Epi-data 3.1 software. Variables frequency tables and graphs were made using Word and Excel version 2013 software. A bivariate analysis helped to examine the association between self-medication and the variables studied. For each variable studied, the proportion of respondents “having practiced anti-malarial self-medication” and “not practicing an anti-malarial self-medication” was calculated. Significant differences in proportions were tested using a Chi2 test or Fisher's exact test. Relative risks and their confidence interval were also calculated. Statistical analyses were performed with a 95% confidence interval. The significance threshold was set at 5%.

This study received the favourable opinion of the Ethics Committee for Biomedical Research of the University of Parakou. The oral consent of each respondent was obtained before using the questionnaire. A population survey authorization was obtained from the administrative authorities. Data collected in this study were treated anonymously and confidentially.

Language barrier was the main difficulty encountered during this survey. This difficulty is linked to the multitude of ethnic groups and dialects, which regularly requires a translator with the respondent’s consent if conditions are met.

A male predominance was noted with a sex-ratio of 1.4. The mean age was 29.69 ± 12.11 years and more than 3/4 of respondents had between 20 to 40 years old. By marital status, 183 were single (54.63%) and 148 were married (44.18%). Average monthly income was less than 70 US dollars in 243 respondents (72.54%). Table 1 presents the sociodemographic characteristics of the respondents.

The main malaria symptoms cited by the respondents were fever (129% or 38.51%), headache (93% or 27.76%), asthenia (58% or 17.31%) and muscular aches and pains (30% or 8.96%). Table 2 presents the malaria symptoms cited by the respondents.

Of the 335 respondents considered in this study, 141 (42.09%) practiced self-medication, including 130 (38.81%) with anti-malarial drugs (Figure 1). For drugs used in self-medication, there were mainly 135 anti-malarial drugs (63.08%), 71 antipyretics/analgesics (33.18%), 5 antibiotics (2.34%). Quinine (60% or 44.45%) and ACTs (46% or 34.07%) were the most used anti-malarial drugs followed by sulfadoxine-pyrimethamine (14% or 10.37%) and chloroquine (3% or 2.22%). Twelve respondents used “Api-palu”, an anti-malarial drug manufactured in Benin. Table 3 shows the drugs used by the self-medication respondents and the supply places. Eighty-seven respondents (66.92%) did not have a good knowledge of self-medication drugs dosage.

The reasons given to justify self-medication were mainly the high cost of consultation fees (99% or 54.10%), good knowledge of one’s illness (53% or 28.96%) and easy access to medicines in pharmacies (20% or 10.93%) (Table 4).

Respondents were unaware of self-medication consequences. Only one respondent knew chemoresistance as a self-medication harmful consequence. Self-medication associated factors were fever (p = 0.04), supply of non-prescription drugs in pharmacies (p < 0.01), supply of drugs in informal market (p < 0.01), previous satisfactory use of the drug (p < 0.01), accessible cost of the drug (p = 0.03) and drug recommendation by a relative or friend (p = 0.02).

“Api-palu” = antimalarial drug made in Benin. ACT = artemisinin-based combination therapy.

This study found 42.09% of adults living in Parakou routinely use self-medication when malaria is suspected. Houngnihin and al. noted that 88% of mothers first resorted to self-medication for fever in their children [9] . This confirms the entrenchment of such practice in Beninese population. Self-medication practice is very real in many other African countries. Its frequency was 91.4% among students in southwestern Nigeria in 2012 [20] , 81% in the northern district of Dakar in Senegal in 2006 [21] , 65% in Libreville, Gabon [22] , 64.5% in Ethiopia in 2003 [3] and 55.7% in Cameroon in 2011 [5] .

Quinine (60% or 44.45%) and ACTs (46% or 34.07%) were the most used anti-malarial drug for self-medication among respondents. There is, therefore, a higher tendency to use quinine than CTAs. This could simply due to the lower cost of generic quinine compared to the cost of ACTs. Quinine is also a historically well-known medicine and remains widely used by the majority of Beninese population despite its side effects. Chloroquine has been used in self-medication although it is no longer recommended for preventive or curative treatment in Benin due to chemoresistance. Unfortunately, it continues to be sold in the informal market. Similarly, sulfadoxine-pyrimethamine, found in 10.37% of cases, is currently recommended in Benin only as intermittent preventive treatment in pregnant women. For this reason, pregnant women were excluded from this survey to avoid confusion. But perhaps they advise this medicine to other members of their family as a cure. It is also a cheaper drug and therefore easy to access which can justify its use. In all self-medication studies, there are almost the same drugs [3] [5] [23] .

Seventy-one respondents (33.18%) associated an antipyretic (paracetamol) with their medication. Malaria is a feverish disease, so this result is not surprising. Some patients often take only paracetamol delaying the use of care and promoting the disease evolution to complications. This finding was made by Diop and al. in Dakar in 2012 [24] and Metta and al. in Tanzania in 2014 [18] .

In this study, antibiotics were also used as an anti-malarial self-medication drug. These are amoxicillin, ciprofloxacin and cotrimoxazole used to treat malaria. These medicines remain cheaper and more accessible in Benin regardless of the supply source. They can also be bought in pharmacies without any prescription. This facilitates their excessive and inappropriate use increasing thus the risk of resistance to these antibiotics. Many other studies mentioned antibiotics use in self-medication for treating malaria [3] [18] [19] [21] [22] .

Self-medication drugs dosage was inadequate in 87 respondents or 66.92% of cases. None of the respondents knew the normal duration of the treatment used in self-medication. There was a tendency to reduce the treatment dosage and duration. Not only this increases the resistance risk to anti-malarial drugs but also their sickness worsening risk in case of malaria. Yet 88.65% of self-medication respondents in this study were in school. Self-medication drugs misuse is well documented in the literature. Most of the time, the treatment respects neither the duration nor the recommended dose [3] [19] [21] .

Respondents obtained the drugs at three locations. Pharmacies came first, followed by the informal “street drugs” market. This is due to the non-prescription drugs dispensing in pharmacies and drugs sale in the informal sector despite numerous sensitizations to the population [8] . Indeed, a study conducted in Benin in 2015 noted that direct requests were so far the preferred way to buy medicines in pharmacies. While formal prescriptions were submitted by some clients, the majority of them were simple sheet of papers on which only the products’ names were indicated. Others are served directly without any paperwork simply by explaining their symptoms to pharmacy staff who act as consultants and prescribers at pharmacies counter while they are not qualified for [8] . The sale of anti-malarial drugs in pharmacies without a medical prescription has been reported by several authors [8] [10] [11] [18] [19] . Fight against self-medication must include the regulation of prescription drugs sale in pharmacies on the one hand, and the fight against “false drugs” on the other hand.

Factors associated with self-medication were fever (p = 0.04), the supply of non-prescription drugs in pharmacies (p < 0.01), the drugs supply in the neighbourhood (p < 0.01), the satisfaction history use of the drug (p < 0.01), accessible cost of the drug (p = 0.03) and drug recommendation by a relative or friend (p = 0.02).the distance between the respondent’s home and the nearest health centre was not a factor associated with self-medication in this study as was the case in a Ugandan study [6] . And only a minority of self-medicated respondents (4.26%) lived more than 2 km from the nearest health centre.

These factors could be the consequence of the respondents’ low purchasing power. The main reason given by respondents for self-medication was the high cost of consultation and medical care in health facilities. Therefore, to get around this difficulty they go directly to pharmacies or market to access drugs. Establishing a universal health insurance system could improve attendance to health facilities and significantly reduce the use of self-medication among population.