Data-driven analyses of behavioral strategies to eliminate cysticercosis in sub-Saharan Africa

Abstract

Background: The multi-host taeniosis/cysticercosis disease system is associated with significant neurological morbidity, as well as economic burden, globally. We investigated whether lower cost behavioral interventions are sufficient for local elimination of human cysticercosis in Boulkiemdé, Sanguié, and Nayala provinces of Burkina Faso.

Methodology/principal findings: Province-specific data on human behaviors (i.e., latrine use and pork consumption) and serological prevalence of human and pig disease were used to inform a deterministic, compartmental model of the taeniosis/cysticercosis disease system. Parameters estimated via Bayesian melding provided posterior distributions for comparing transmission rates associated with human ingestion of Taenia solium cysticerci due to undercooking and human exposure to T. solium eggs in the environment. Reductions in transmission via these pathways were modeled to determine required effectiveness of a market-focused cooking behavior intervention and a community-led sanitation and hygiene program, independently and in combination, for eliminating human cysticercosis as a public health problem (<1 case per 1000 population). Transmission of cysticerci due to consumption of undercooked pork was found to vary significantly across transmission settings. In Sanguié, the rate of transmission due to undercooking was 6% higher than that in Boulkiemdé (95% CI: 1.03, 1.09; p-value < 0.001) and 35% lower than that in Nayala (95% CI: 0.64, 0.66; p-value < 0.001). We found that 67% and 62% reductions in undercooking of pork consumed in markets were associated with elimination of cysticercosis in Nayala and Sanguié, respectively. Elimination of active cysticercosis in Boulkiemdé required a 73% reduction. Less aggressive reductions of 25% to 30% in human exposure to Taenia solium eggs through sanitation and hygiene programs were associated with elimination in the provinces.

Conclusions/significance: Despite heterogeneity in effectiveness due to local transmission dynamics and behaviors, education on the importance of proper cooking, in combination with community-led sanitation and hygiene efforts, has implications for reducing morbidity due to cysticercosis and neurocysticercosis.

Fig 1. Compartmental model structure for multi-host taeniosis and cysticercosis disease system.

A proportion of people with cysticercosis (I_H2) are assumed to be coinfected with taeniosis. Both autoinfection among those with taeniosis as well as consumption of infected pork by those with cysticercosis were recognized as mechanisms of coinfection.

Fig 2. Model fit to data.

The white circles and grey error bars are the point estimates and 95% CIs from the data. The violin plots (with embedded boxplots) are the model results. The point estimates, which were used for the likelihood, are close to the medians and means of the model output. The distribution of model output for cysticercosis in pigs and taeniosis in humans generally falls within the data CIs for those variables. The model is producing a larger range of values for active human cysticercosis than is observed in the data.

Fig 3. Results of Bayesian melding model fitting procedure for the three study provinces in Burkina Faso: Sanguié, Boulkiemdé and Nayala.

Four, province-specific quantities (i.e., monthly rate of decay of egg viability in the environment (ρ), the product of the probability of transmission of T. solium cysticerci from pork meat to humans and the rate of undercooking in markets (β_Mc_k), the product of the probability of transmission given ingestion of each T. solium egg and the rate of consumption of fecally contaminated soil by pigs (β_Eχ_P), and the ratio of the rate among humans of ingestion of food or water contaminated by eggs in the environment and contact with eggs directly on another human carrier versus the rate of pigs’ consumption of fecally contaminated soil (χ_F/χ_P)) were fit using data on three outcomes (i.e., active cysticercosis in humans, active cysticercosis in pigs, and current taeniosis in humans).

Fig 4. Impact of behavioral change interventions on active cysticercosis infections.

(A) Changes in the prevalence of active cysticercosis due to reductions in consumption of T. solium cysticerci in undercooked pork meat. (B) Changes in the prevalence of active cysticercosis due to reduced environmental exposure to T. solium eggs such as through use of latrines. Bars represent the average prevalence across simulations with 1,000 parameter sets. Error bars represent one standard deviation above the mean. Horizontal, gray bars represent average, pre-intervention prevalence of active cysticercosis by province (See Table 1).

Fig 5. Combined impact of pork cooking and latrine use interventions on active cysticercosis.

(A) 5% reduction in exposure to T. solium cysticerci in undercooked pork meat, in combination with 5–50% reductions in exposure to T. solium eggs in the environment. (B) 10% reduction in exposure to T. solium cysticerci in undercooked pork meat, in combination with 5–50% reductions in exposure to T. solium eggs in the environment. (C) 25% reduction in exposure to T. solium cysticerci in undercooked pork meat, in combination with 5–50% reductions in exposure to T. solium eggs in the environment. (D) 50% reduction in exposure to T. solium cysticerci in undercooked pork meat, in combination with 5–50% reductions in exposure to T. solium eggs in the environment. Bars represent the average prevalence across simulations with 1,000 parameter sets. Error bars represent one standard deviation above the mean.