Search

Additional file 2: Text S1. A simple analysis of the impact of treatment and heterogeneity on worm aggregation among hosts.

Additional file 1. Figure S1. Distribution of prevalence from sampled schools. Table S1. Monte Carlo maximum likelihood estimates for geostatistical model parameters for S. haematobium (a) and S. mansoni (b). Figure S2. Simulated country and distributions of districts, subdistricts and schools. Table S2. Consumable and equipment costs included for sampling 50 children using urine filtration (A) and Kato-Katz stool analysis (B); life expectancy and quantity needed per team is included for capital items. Table S3. Total survey mapping costs per school sampling 50 children. Table S4. Mean survey costs per school sampling 50 children by country, year and diagnostic method. (DOCX 950 kb)

Additional file 1: Figure S1.A. lumbricoides prevalence over time (a–c) and aggregation (parameter k) over time (d–f), in this plot we show 500 out of 1000 simulated villages. a and d: Random non-compliance; b and e: Semi-systematic non-compliance; c and f: Fully systematic non-compliance. Figure S2. Age-prevalence over time for a selected village in which elimination of A. lumbricoides infection was not achieved before mass drug administration was implemented (a) and post-MDA and one in which elimination was not achieved (b). Age groups (years): 1, 0–5; 2, 6–10; 3, 11–15; 4, 16–30; 5, 31–50; 6, 51+. Figure S3. The proportion of individuals who still harbour worms, either hookworm (a) or A. lumbricoides (b) after each round of mass drug administration (T1–5) for random compliance, semi-systematic compliance and systematic compliance. Figure S4. The proportion of individuals who harbour both male and female worms (A. lumbricoides) and produce fertile eggs (grey shaded color) and the proportion of individuals who harbour either only male or only females worms and produce no fertile eggs (orange shaded color) for random compliance (a), semi-systematic compliance (b) and systematic compliance (c). Figure S5. Results from analytical approach for predicting change in aggregation after treatment under semi-systematic non-compliance. a Comparison of mean worm burden observed from the stochastic model immediately post-treatment and that expected from the simple probability model. b Comparison of k observed from the stochastic model and that expected from the simple probability model immediately post-treatment. Note: For a, the gray line represents the equation x = y (perfect agreement).

Additional file 1.

Background\ud Buruli ulcer can cause disfigurement and long-term loss of function. It is underdiagnosed and under-reported, and its current distribution is unclear. We aimed to synthesise and evaluate data on Buruli ulcer prevalence and distribution.\ud \ud Methods\ud We did a systematic review of Buruli ulcer prevalence and used an evidence consensus framework to describe and evaluate evidence for Buruli ulcer distribution worldwide. We searched PubMed and Web of Science databases from inception to Aug 6, 2018, for records of Buruli ulcer and Mycobacterium ulcerans detection, with no limits on study type, publication date, participant population, or location. English, French, and Spanish language publications were included. We included population-based surveys presenting Buruli ulcer prevalence estimates, or data that allowed prevalence to be estimated, in the systematic review. We extracted geographical data on the occurrence of Buruli ulcer cases and M ulcerans detection from studies of any type for the evidence consensus framework; articles that did not report original data were excluded. For the main analysis, we extracted prevalence estimates from included surveys and calculated 95% CIs using Byar's method. We included occurrence records, reports to WHO and the Global Infectious Diseases and Epidemiology Network, and surveillance data from Buruli ulcer control programmes in the evidence consensus framework to grade the strength of evidence for Buruli ulcer endemicity. This study is registered with PROSPERO, number CRD42018116260.\ud \ud Findings\ud 2763 titles met the search criteria. We extracted prevalence estimates from ten studies and occurrence data from 208 studies and five unpublished surveillance datasets. Prevalence estimates within study areas ranged from 3·2 (95% CI 3·1–3·3) cases per 10 000 population in Côte d'Ivoire to 26·9 (23·5–30·7) cases per 10 000 population in Benin. There was evidence of Buruli ulcer in 32 countries and consensus on presence in 12.\ud \ud Interpretation\ud The global distribution of Buruli ulcer is uncertain and potentially wider than currently recognised. Our findings represent the strongest available evidence on Buruli ulcer distribution so far and have many potential applications, from directing surveillance activities to informing burden estimates.

Additional file 1: Figure S1. Trends in mean monthly overall patient attendance per year, stratified by site. Figure S2. Trends in mean monthly attendance of patients not suspected of malaria per year, by site. Figure S3. Trends in mean monthly suspected malaria patients per year, stratified by site. Table S1. Changes in attendance of patients suspected versus not suspected of malaria over-time, comparing mean monthly attendance between first and last calendar years of study duration. Figure S4. Age distribution of test confirmed malaria cases, by sex and site. Figure S5. Age distribution of patients not suspected of malaria by sex and site. Figure S6. Age distribution of patients that tested negative for malaria, by sex and site. Figure S7. Adjusted marginal probability of test confirmed malaria, by sex, intervention period, age, and site. Table S2. Multivariable association between age (in three categories) and covariates of interest among malaria confirmed cases, accounting for effect modification of intervention periods on sex. Table S3. Association between age (in three categories) and covariates of interest among malaria confirmed cases, fitting an interaction between diagnostic test used (B/S vs. RDT) and intervention duration. Figure S8. Trends in the annual proportion of RDT use among tested participants, stratified by site. Figure S9. Scatter plot of age with test positivity for LLINs only sites, stratified by intervention period. Figure S10. Scatter plot of age with test positivity for (LLIN plus IRS) sites, by intervention period.

Additional file 1: Figure S1. Trends in mean monthly overall patient attendance per year, stratified by site. Figure S2. Trends in mean monthly attendance of patients not suspected of malaria per year, by site. Figure S3. Trends in mean monthly suspected malaria patients per year, stratified by site. Table S1. Changes in attendance of patients suspected versus not suspected of malaria over-time, comparing mean monthly attendance between first and last calendar years of study duration. Figure S4. Age distribution of test confirmed malaria cases, by sex and site. Figure S5. Age distribution of patients not suspected of malaria by sex and site. Figure S6. Age distribution of patients that tested negative for malaria, by sex and site. Figure S7. Adjusted marginal probability of test confirmed malaria, by sex, intervention period, age, and site. Table S2. Multivariable association between age (in three categories) and covariates of interest among malaria confirmed cases, accounting for effect modification of intervention periods on sex. Table S3. Association between age (in three categories) and covariates of interest among malaria confirmed cases, fitting an interaction between diagnostic test used (B/S vs. RDT) and intervention duration. Figure S8. Trends in the annual proportion of RDT use among tested participants, stratified by site. Figure S9. Scatter plot of age with test positivity for LLINs only sites, stratified by intervention period. Figure S10. Scatter plot of age with test positivity for (LLIN plus IRS) sites, by intervention period.