Socio-Economic and Health Determinants of Rural Households Consent to Prepay for Their Health Care in N’Dali (North of Benin)

In order to know the importance of the pre-finance of the health care in a family, this article aims to analyze the real possibilities of households partici-pating in the pre-financing of their health care. The data used for the analysis were collected from one hundred and twenty (120) households in N’Dali lo-cated in northern Benin. The results obtained show that more than half of rural households can agree to pay 7000 FCFA/individual/year to pre-finance their health care. This willingness to pay households is influenced both by the characteristics specific to the respondents (age, educational level, number of children in the household, income level, etc.) as well as those specific to the local health services (prompt reception of patients, cleanliness of the health center and actual or periodic presence of a doctor in the health service).


Introduction
In the aftermath of their independence, African countries declared health as a right and introduced free healthcare for all. The health systems inherited from the colonial period were no longer adapted to the demographic distribution and the basic health needs of these countries. The 1980s therefore saw a profound reorganization of health systems in Africa, with strong decentralization and a strategy focused on prevention and primary health care (PHC). It was in Alma Ata in 1978 that the member countries of the World Health Organization officially gave the starting point for this new strategy. However, the latter quickly came up against the question of its funding. Over-indebted African countries consumer of goods and services. It is based on the neoclassical individualist paradigm, which postulates that the economic agent always seeks to maximize utility in his purchasing decisions, therefore in the consumption of goods and services. According to this approach, consumer behavior is linked to two determinants: the preferences of individuals on the one hand, and the constraint on their budgets on the other. Studying consumer behavior involves three main steps: identifying individual preferences, that is, understanding how and why agents prefer one good over another; take into account the budgetary constraint weighing on individuals, the combination of preferences and budget constraints then determining consumption choices; identify the combination of goods that agents will choose to maximize their usefulness. Three basic assumptions underlie this theory: individual preferences are complete, that is, they exist for all goods. They are transitive (if A is preferred to B and B to C, then A is always preferred to C).
The third hypothesis is that of unsatiety, that is to say that agents always prefer to have more than less, whatever the good.
According to consumer theory, willingness to pay is what agents are willing to give up in terms of other consumption opportunities, in order to obtain a combination of goods capable of maximizing their utility taking into account their budgetary constraints. The notion of consent to pay (CAP) is very old and has been developed in several contexts. It can be defined in three main situations: Public investment will reduce the journey time between two cities.
For each user of the infrastructure, we can express the amount of money he is willing to sacrifice to benefit from the (certain) reduction in travel time. It should be noted immediately that the CAP is akin to the very old notion of "marginal rate of substitution".
This definition of C.AP has been used very often in the assessment of public investment in infrastructure and the environment. We note two essential characteristics: 1) the usefulness of the interviewee is "two-dimensional": it depends on the journey time and the wealth available for consumption and 2) the context is a situation of certainty.
• A willingness to pay can be expressed in a risk context for one-dimensional utility functions.
In this case the CAP is similar to another familiar concept: that of "risk premium". Consider the following lottery: where p0 is the initial probability of seeing a loss equal to 10,000 F.
In this context, the CAP is defined as the sum of money that someone is ready to invest in order to reduce the initial probability of loss and reduce it from p0 to If we remember that the risk premium is equal to the amount that the individual is ready to pay in order to eliminate the risk (and replace it with his mathematical expectation), we immediately see the link between the two concepts: then that the CAP deals with a marginal change in risk, the risk premium evokes an overall change in risk (in fact its elimination). With regard to lotteries like those described here, the CAP has two important characteristics: 1) the usefulness of the interviewee is "one-dimensional", it depends solely on financial wealth and 2) the context is a risk situation.
• In the domain of health The CAP is defined in a framework, which brings together the two previous ones. Indeed: -the utility function is multi-dimensional: it reflects not only the financial wealth of the individual but also characteristics of his state of health, for example the duration or the quality of his survival; -the context is that of risk.
In this type of situation, at least three PACs can be defined. Indeed, let's go back to the lottery facing the individual: In this lottery the results are represented by a two-dimensional utility where: -T 0 is the survival time in the event of illness and D represents the financial loss resulting from the illness; -T 1 is the survival time in case of good health.
Faced with this risky situation, we can define the CAP: 1) to reduce p0; 2) to increase T 0 and 3) to reduce D. It is important to note that the most widespread concept of CAP is that corresponding to 1). It was "popularized" by Drèze (1962) and Jones-Lee (1974) as part of their work on the social utility of human life (Eeckhoudt & Hammit, 2001). However, people could also be interviewed about 2) and 3).

Estimate of Willingness to Pay
The participation of households in the financing of health care poses the problem of the supply and pricing of public health goods and services. Two main theoretical approaches are available to clearly estimate households' willingness to pay for health care: -the indirect approach: it consists of using data on the use of goods or services to assess consumers' responses; -the direct approach: it is used to ask individuals how much they are willing to The closed questionnaire method consists of starting with a starting amount and asking the respondent if he is ready to pay this amount or not (bidding game). For example "would you be willing to pay X CFA francs for a pre-financing system for your health care?" As for the open questionnaire, the respondent is asked to express their maximum willingness to pay for health care.
For example, "What is the maximum amount that you will be willing to pay to participate in a pre-financing health care system?" To elucidate the willingness to pay of rural households for access to health care we use for the needs of our study, the method of contingent analysis at the level of which we only use the dichotomous technique which responds better to the bargaining strategies practiced on the African markets.

Modeling of Factors Determining Willingness to Pay
The information resulting from the technique of dichotomous choices ("bidding-game") is a set of negative or positive responses resulting from a specific question on discrete values, it is true that the values offered by households within the framework of a community health care pre-financing system are continuous variables. Thus, the dependent value obtained by the dichotomous choice procedure is not the maximum value that the household would be willing to pay for its health care but, rather, the range in which the true value of the willingness to pay would lie. Insofar as there is a risk of violation of the assumptions relating to the error term, the use of the ordinary least squares (OLS) method would not be appropriate for the explanation of the determinants of the willingness to pay contrary to the case where the dependent variable is quantitative. The ordered probit model is therefore the one we will use and whose estimation will be made by the maximum likelihood method.
From the above discussions, consider V m as the maximum amount that household I agree to pay to participate in a community health care pre-financing system. Based on the consumer demand theory we assume that V m is a linearly dependent variable of the explanatory variables. If X m represents the socioeconomic factors of rural households determining V m , we have: where α and β are parameters to be estimated from the model and e m the error term.
Equation (1)  However, based on the responses from the interviews, the area in which V m defines itself is known. Let us also consider R 1 , R 2 , R t , the t values which share the area of definition of willingness to pay in t + 1 categories and Y m a categorical variable such as: , it follows, from Equation (1), that: If we assume that e m follows a standard normal distribution, we have: (4) is the ordered probit model that will be used to explain the variation in the value of willingness to pay. Empirically, Equation (1) above can be written in the form: with, V i the cumulative function of the reduced normal law and X i the socio-economic and health characteristics of the rural household i.
Previous work has shown that the variables representing socio-economic (Kusi et al., 2015;Sanoussi & Ametoglo, 2019) and health (Cisse & Sauvain-Dugerdil, 2018) characteristics are likely to determine consent to prepay for health care rural households. These variables are described in Table 1.
The sign of their coefficient makes it possible to assess their effect on the consent to prepay for health care. The sign (+) indicates that the expected effect of the variable concerned on the willingness to pay of households is positive and the sign (−) indicates that the expected effect of the variable concerned on the willingness to pay of households is negative.

Study Area and Database
This study was conducted in northern Benin, in the Borgou department and more specifically in the commune of N'Dali. The latter is limited to the north by the municipalities of Bembèrèkè and Sinendé, to the south by the municipality of Parakou, to the east by the municipality of Nikki and to the west by the municipality of Djougou. With a density of around 12.1 inhabitants per square kilometer and a population of 113,604 inhabitants (INSAE, 2013), the study area covers an area of 3748 km 2 , which represents 3.33% of the national territory. The data used for the analysis were collected using a structured questionnaire and group interview from one hundred and twenty (120) rural households via household heads. The latter were chosen at random in three (03)

Socio-Economic and Demographic Characteristics of Households
Men make up more than half (60.83%) of the sample size. Women (39.17%) are much less represented than men. The proportion of heads of household whose age is between 20 and 50 not included is 72.50% compared to 27.50% for those   Socio-economic and health realities can also explain the amount agreed to pay by the households of this commune. It is therefore important to be interested in the determinants of the CAP for a good understanding.

Determinants of Willingness to Pay of Rural Households
Estimating Equation (4) allowed us to identify the factors that influence the willingness of rural households to pre-finance health care for future consumption. The thirteen (13) explanatory variables were not taken into account in the estimation of the model. Indeed, after simulations, some variables were found to be non-significant. This is the gender of the respondent (GENDER, household health status (STATSAN), availability of essential drugs (DISPOMED) and marital status (SMAT). Thus, the following variables were used in the estimation of the model: the age of the respondent (AGE), the education level of the respondent (EDUC), the number of children in the household (CHILD), the level of monthly income of the respondent (INCOME), speed in welcoming patients (HOME), membership of an association or rural community action organization (ASSOCOM), ownership of the health center (CLEANLINESS), presence doctor at the local health center (DOCTOR) and the tradition of using the local health service (TRAD). After the elimination of the non-significant variables, the results of the estimation of the model by the maximum likelihood method are presented in Table 3.
Reading the table tells us that 86.3% of the variations in the consent of households to prepay for health care is explained by the variables taken into account in the model (R squared (R2) = 0.863). The student t statistic which is the test of the significance of the partial regression coefficients makes it possible, through the calculation of the probability linked to each student t statistic, to give the level of significance of the corresponding coefficients. In light of the results recorded in Table 3, all the variables are significant at 5% because all the probabilities are less than 0.05 (Prob < 0.05). From the regression results, it appears that the willingness to pay of households is influenced both by the characteristics specific to the respondents and those specific to the local health services.
-Characteristics specific to respondents The results in Table 3 show that variables such as the respondent's age (AGE), the respondent's education level (EDUC), the number of children in the household (CHILD), the household income level (INCOME), and the tradition of using the local health service (TRAD) significantly affect at the 5% threshold the willingness to pay of households to collectively pre-finance their health care. All these variables each display a positive coefficient, therefore of the same sign as its expected effect. Age is positively significant. This variable is therefore relevant in its contribution to explaining the consent to prepay of rural households. The older the respondent, the higher the value of their approval to pay. This is explained by the deterioration in health status with age.
This conclusion invalidates the results of Sigue et al. (2019) that "âgé the older the head of the household, the less likely he is to pay for technology".
The level of education being significant, confirms the results of Racodon et al. (2018) who had shown that physical activity after cardiovascular rehabilitation depends on the academic level of the patients.
The level of household income (INCOME) and the tradition of using local health services (TRAD) are significant and each displayed a positive coefficient; which is consistent with their expected effect. The higher the income, the more the farmer is ready to contribute to his health care. Indeed, "The more the income of a household that uses the local health service increases, the more important it is their consent to pre-finance their health care".
This result confirms the work of Perronnin and Louvel (2018) according to which the level of household income and the social environment are factors that predispose and facilitate the rate of community health care coverage.
The variable (CHILD) representing the number of children in the household displays a coefficient with the same sign as its expected effect; this variable is therefore relevant in explaining the consent to pre-finance rural households.
This result also confirms that of Mao: "The larger the household size, the less the landscapes are willing to pay for the medical cooperative". The variable (ASSOCOM) representing the membership of a household in a rural community organization or association displays a sign coefficient contrary to the expected effect. Since its probability (0.0007) is less than (0.05), it is therefore relevant in explaining the approval to prepay of rural households.
Characteristics specific to the local health services These are variables such as: the speed with which patients are received (HOME), the cleanliness of the health center (CLEANLINESS) and the presence of a doctor at the local health center (MEDECIN). All these variables each display a coefficient with the same sign (positive) as the expected effect. The variables (HOME) and (CLEANLINESS) are significant and very relevant in ex- and therefore very relevant for the use of health care by rural households. It appears that: "The cleaner the health center and the warmer the reception for patients, the more raised the willingness of households to prepay for health care." The variable (DOCTOR) representing the actual or periodic presence of a doctor in the health service is also significant. To involve households in the pre-financing of their health care, it is important that the doctor or nurse is regular at his post.

Conclusion
The consent of households to pre-finance their health care is not the result of chance but rather of socio-economic and demographic parameters such as the age of the respondent, the level of education of the household, the number of children in the household, membership of a household in a rural community organization or association, tradition of using local health services, speed in welcoming patients, cleanliness of the health center, the actual or periodic presence of a doctor in the health service and especially the level of household income. In short, the determinants of consent to prepay rural households in the commune of N'Dali are the socio-economic factors and the quality of services.
Specifically, older respondents are more in favor of consent to prepay than their younger counterparts; similarly, the wealthiest households are in favor of pre-financing than the poor. The value of the willingness to pay which would encourage the adhesion of rural households to the pre-financing health care system is 7000 FCFA. So, social and practical implication is the best way to resolve the problem of pre-finance. According to the research results found I notice that households should enforce their pre-finance for their health care. So, this study is significant for all households. It helps them to understand clearly how to pre-finance.

Conflicts of Interest
The author declares no conflicts of interest regarding the publication of this paper.