The educative impact of health care treatment on malarial prevention behavior for the poor in Guinea, West Africa

Malaria now stands as one of the most steadfast obstacles to the socioeconomic development of the entire African continent, with an impact that cannot be overstated (1). It is estimated that presently there are >300 million acute cases each year and >1 million deaths (2). Ninety percent of these deaths are concentrated in Africa alone (3). The economic costs are similarly severe, with nearly $12 billion of gross domestic product being lost annually in Africa from the disease (4).

There has been an increased awareness of the benefits of a coordinated health policy approach that balances drug treatment with broader prevention efforts (5). We show here that the choice of malarial health care treatment can, moreover, have a disproportionate positive effect on the preventative behavior of the most at-risk populations.

The data stem from a malaria module in the 1999 Demographic and Health Survey for Guinea. It was administered to all women surveyed between the ages of 15 and 49 years. The fieldwork was implemented by the Direction Nationale de la Statistique et de l'Information over the duration of 1 month from May to June 1999. We have selected here for a rural subset of this population. The “prevent” variable is a dichotomous response to the question “Are you doing something to avoid malaria?” The “treatment choice” variable represents where the respondent sought health care treatment upon their last episode of malarial infection. The options are health center, hospital, traditional medicine, pharmacy, and other. Living standard is defined in terms of access to a toilet facility. The three choices that are ordered by way of increasing quality are no facility, basic pit, and pit toilet.

Fig. 1 shows that a formal-education background promotes increased preventive behavior irrespective of the living standard. The gradient for the educated subset has a much higher baseline level of response and increases more steeply with improving living standard. At each facility level, there is a minimal difference of ≈20% between the educated and uneducated groups.

Fig. 1 displays a strong difference in prevention behavior between formally educated individuals and the uneducated. We next consider the effect of prior treatment choice on the present-day preventative behavior of the uneducated.

Fig. 2 shows the treatment-choice-dependent odds ratio of enacting prevention for the uneducated. Traditional medicine treatment is used as a reference. There is a noticeable reduction in prevention as one makes the transition in living standard from no facility to a basic pit setting: 34% decrease for health center, 17% for hospital, and 29% for pharmacy.

At the poorest no-facility level, it appears that individuals who have selected a formal health care option for malarial treatment are disproportionately likely to practice preventative behavior. There is an amplified effect for this particular subgroup that dissipates sharply among those individuals residing at the basic-pit environment.

The strength of the association between treatment choice and prevention is next considered across living standards by χ² analysis (Tables 1, 2, 3.) The level of significance in the association rapidly drops off with an improving facility environment. For a critical value of 9.4877 at α = 0.05 and four degrees of freedom, only the no-facility living standard demonstrates a significant association between treatment choice and prevent variables.

Table 4 provides a summary of the logit analysis performed to determine whether we can generate a representative model for dependent prevent variable. Besides including living standard and educational level as predictors, we also incorporate an expanded version of the treatment-choice variable that includes individuals who have either reported no prior malarial infection (i.e., no malaria) or no treatment sought in the case of infection (i.e., no treatment).

Table 4 shows that all of our predictor variables are significant at a level of α = 0.05 and that the model demonstrates a good fit in predicting the log odds of the dependent prevent variable.

Table 5 shows that all the treatment choices are statistically significant at α = 0.05 and carry robust parameter estimates. This finding corroborates that the dampening effect we observed, across living standards, in Fig. 2 is at least in part modulated by the treatment choice itself.

An unexpected result is the high negative parameter estimates for the no-treatment and no-malaria subgroups.

Further analyzing this, Fig. 3 shows that the percent levels of no-malaria and no-treatment respondents greatly decrease upon shifting to higher standards of living. Improved living standard appears to coincide with an enhanced ability to identify symptomatic malarias. The no-facility population resides in a sanitary environment that is expected to be most susceptible to mosquito breeding and infection. The high percentage of individuals stating no prior malarial history seemingly points to a substantive disease misdiagnosis. In a region of high malarial endemicity, the prevalence of malarial infection thus appears to be undervalued.

This research has two major implications for the no-facility subpopulation in particular: (i) prior selection of a formal health care treatment for malaria is associated with heightened preventative practices in the present, and (ii) those who do not seek any form of health care treatment are quite likely misdiagnosing malarial infection in many scenarios.

Formal health care treatment may play a more dynamic role in exposing individuals to a targeted health educational message. For individuals in the most severe poverty (i.e., the no-facility subgroup), such information seems to have an elevated instructive effect, compensating in part for the level of disease mis-recognition and treatment abstention also prevalent in that setting. Patient consultation may not only allow for more accurate disease diagnosis, but may also act as a conduit for a stronger understanding about malaria and its associated clinical aspects.

Treatment choice does not have as great an effect on the individuals at the basic-pit or pit-toilet settings. This may be due to the presence of a level of educational awareness in this population that coincides with a more sanitary household surrounding. Such heightened disease awareness is corroborated by the decrease in the percent of no-malaria and no-treatment subgroups across improving living standards.

Further work may help elucidate this educative mechanism in more detail. Thereafter, it would be valuable to examine how to incorporate the didactic benefits of formal treatment alongside the traditional medicine regimens that are observed by a plurality of this group.

It would also be productive to extend such analysis to HIV/AIDS in Africa to determine whether there exists a similar educative health care treatment effect. Although malaria and HIV/AIDS have a strong difference in treatment regimens and disease pathogeneses, it does not entirely preclude similar behavioral responses on the part of the very poor toward these infectious diseases. Regional health policy that more accurately captures the health dynamics of this population would benefit as a result.