How effective is school-based deworming for the community-wide control of soil-transmitted helminths?

Abstract

Background: The London Declaration on neglected tropical diseases was based in part on a new World Health Organization roadmap to "sustain, expand and extend drug access programmes to ensure the necessary supply of drugs and other interventions to help control by 2020". Large drug donations from the pharmaceutical industry form the backbone to this aim, especially for soil-transmitted helminths (STHs) raising the question of how best to use these resources. Deworming for STHs is often targeted at school children because they are at greatest risk of morbidity and because it is remarkably cost-effective. However, the impact of school-based deworming on transmission in the wider community remains unclear.

Methods: We first estimate the proportion of parasites targeted by school-based deworming using demography, school enrolment, and data from a small number of example settings where age-specific intensity of infection (either worms or eggs) has been measured for all ages. We also use transmission models to investigate the potential impact of this coverage on transmission for different mixing scenarios.

Principal findings: In the example settings <30% of the population are 5 to <15 years old. Combining this demography with the infection age-intensity profile we estimate that in one setting school children output as little as 15% of hookworm eggs, whereas in another setting they harbour up to 50% of Ascaris lumbricoides worms (the highest proportion of parasites for our examples). In addition, it is estimated that from 40-70% of these children are enrolled at school.

Conclusions: These estimates suggest that, whilst school-based programmes have many important benefits, the proportion of infective stages targeted by school-based deworming may be limited, particularly where hookworm predominates. We discuss the consequences for transmission for a range of scenarios, including when infective stages deposited by children are more likely to contribute to transmission than those from adults.

Figure 1. The relationship between mean intensity and prevalence.

A The relationship between the mean intensity of infection, , the prevalence of infection, , and the negative binomial aggregation parameter, as described by the relationship in equation 1. B Relationship between the prevalence and intensity of infection as observed in a study of A. lumbricoides . The solid line is the predicted relationship between mean prevalence of infection and worm burden described in equation 1 and plotted in A fitted to estimate the aggregation parameter,  = 0.194.

Figure 2. School attendance for a selection of countries.

This figure was generated by data published by UNICEF for 2005–2010 .. For each country there is net attendance rate at primary, in urban (open circles) and rural areas (closed circles) and a net attendance rate for secondary schools (filled squares).

Figure 3. The proportion of A. lumbricoides worms in children aged 5–14, calculated from equation 3 .

The demography of the population, A, results in a proportion of 18.4% of the population aged 5–14 years old . Combining this distribution with B the distribution of worms per person by age gives 49.5% of worms in the school-aged group.

Figure 4. The proportion of hookworm eggs deposited by children aged 5–14, calculated from equation 5 .

The demography of the population, A, gives a proportion of 31.2% of the population aged 5–14 years old . Combining this distribution with B the distribution of egg output by age gives 15.7% of worms in the school-aged group.

Figure 5. Critical fraction of the population to be treated.

The predicted relationship between the critical fraction of the human population to be treated, , per annum with efficacy, , 0.9, and the basic reproductive number, , and parasite life expectancy, in years (from equation 11 in the main text).

Figure 6. Impact of fraction treated on worm burden, prevalence and effective reproduction number.

The impact of the fraction of the population treated, , on A the mean worm burden , B the prevalence of infection, and C the effective reproductive number , as described in equation 10. Parameter values are set for A. lumbricoides as follows:  = 0.81,  = 3,  = 1 yr,  = 0.967 and  = 0.95.

Figure 7. Effect of regular treatment on mean A. lumbricoides worm burden for different models.

A homogeneous population (left column), B heterogeneous population with uniform transmission dynamics (central column) and C heterogeneous population with greater contribution from children (right column) as in the text. The two rows represent annual and half-yearly treatment respectively. For all runs, basic reproduction number is 3 and worm lifespan is 1 year. Other parameters (as defined for equations 6 and 7): μ₂ = 5/yr, k = 0.7, z = 0.93.

Figure 8. Effect of regular treatment on mean worm burden of hookworm for different models.

As in Figure 7, the columns are A homogeneous model, B heterogeneous population with uniform transmission dynamics and C heterogeneous population with greater contribution from children, as in the text and different treatment intervals (rows). Simulations for basic reproduction number,  = 3, and worm lifespan is 2.5 years. Other parameters as in Figure 7. The two rows represent two-yearly and yearly treatment respectively.

Figure 9. Schematic illustration of the impact of school-based deworming on the transmission of parasites.

The number of parasites affected by a school-based deworming programme is never 100%, it is reduced by the efficacy of the drug, the proportion of the population of school age and their intensity of infection, as well as school enrolment and attendance on a deworming day. The impact of such a treatment programme on transmission is less easily quantified and depends on the details of transmission between different age-groups in the population. For further details, see main text.