If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Animal and Human Health Program, International Livestock Research Institute, Nairobi, KenyaInstitute of Infection, Veterinary & Ecological Sciences, The University of Liverpool, Leahurst Campus, Neston, UK
Department of Infectious Disease Epidemiology and London Centre for Neglected Tropical Disease Research (LCNTDR), Faculty of Medicine, School of Public Health, Imperial College London, UKSCI Foundation, Edinburgh House, London, UKMRC Centre for Global Infectious Disease Analysis, Department of Infectious Disease Epidemiology, Faculty of Medicine, School of Public Health, Imperial College London, London, UK
This mapping study used spatial statistics and a systematic literature review.
•
The spatiotemporal variation in Taenia solium risk and prevalence is evident in Uganda.
•
The high prevalence and risk areas show the need for urgent focal T. solium control.
•
A mapping protocol is presented to support country planning for T. solium control.
Abstract
Objectives
The lack of subnational mapping of the zoonotic cestode Taenia solium in endemic countries presents a major challenge to achieving intensified T. solium control milestones, as outlined in the “World Health Organization neglected tropical disease roadmap by 2030”. We conducted a mapping study in Uganda, considered to be endemic, to identify subnational high-risk areas.
Methods
T. solium prevalence data, adjusted for diagnostic sensitivity and specificity in a Bayesian framework, were identified through a systematic review. Spatial autocorrelation and interpolation techniques were used to transform demographic and health survey cluster-level sanitation and poverty indicators, overlaid onto a pig density map for Uganda into modelled porcine cysticercosis (PCC) risk maps.
Results
A total of 16 articles (n = 11 PCC and n = 5 human cysticercosis (HCC) and/or human taeniasis) were included in the final analysis. The observed HCC prevalence ranged from 0.01% to 6.0% (confidence interval range: 0.004-11.4%), whereas the adjusted PCC ranged from 0.3 to 93.9% (uncertainty interval range: 0-99.8%). There was substantial variation in the modelled PCC risk factors and prevalence across Uganda and over time.
Conclusion
The high PCC prevalence and moderate HCC exposure estimates indicate the need for urgent implementation of T. solium control efforts in Uganda.
The endemic transmission of the zoonotic cestode Taenia solium continues to result in a substantial global human health burden and economic losses, particularly to communities and poor smallholder pig farmers across sub-Saharan Africa, South and Central America, and across Asia [
Humans are the definitive hosts harboring the adult worm (taeniasis) after consuming undercooked or raw pork infected with the larval stage of T. solium. Scavenging pigs are exposed to eggs dispersed in the environment through open defecation by taeniasis carriers. This results in the larval-stage cysts developing in the porcine intermediate host (porcine cysticercosis; PCC). Environmental contamination with T. solium eggs in areas with poor sanitation and hygiene provides a pathway for humans, as aberrant hosts, to accidentally ingest eggs, with the larval stages encysting in striated/heart muscles and subcutaneous tissues of the human host (human cysticercosis; HCC). There is also a potential for encystment in the central nervous system, including the brain, causing human neurocysticercosis (NCC).
NCC is a highly pleiomorphic disease in humans, largely due to variations in the location and number of cysts. However, the major burden is experienced due to seizures and other neurological sequelae due to the degeneration of viable cysts. In endemic settings, T. solium has been estimated to cause approximately 30% of all epilepsy cases [
], and in 2010, T. solium accounted for the greatest burden of any food-borne parasite, with around 28,000 deaths per year and 2.8 million disability-adjusted life-years [
], no existing national T. solium control programs are in operation in endemic countries. Under the 2012 World Health Organization (WHO) roadmap on neglected tropical diseases (NTDs), T. solium control was identified for scale-up by 2020 [
World Health Organization Accelerating work to overcome the global impact of neglected tropical diseases: a roadmap for implementation executive summary.
]. Although this goal was missed, T. solium was included in the new WHO NTD 2021-2030 roadmap, which calls for “intensified control” in hyperendemic areas of 17 countries by 2030. Endemicity at the country level has recently been updated [
]. However, there is little understanding of the distribution of T. solium at the subnational level, which hinders the ability to efficiently target the existing toolbox of interventions in hyperendemic areas. The intervention toolbox includes human mass drug administration with the taenicides (praziquantel, niclosamide, and albendazole). In addition, further interventions in the porcine host include pig treatment with oxfendazole, pig vaccination with TSOL18, and improved pig husbandry and hygienic practices through health education [
], countries are required first to conduct a rapid assessment of endemic/high-risk areas using the WHO toolkit, which guides the collection of simple, standardized indicators for the presence of infection to create a composite risk score at the subnational level [
]. These mapping efforts can be complemented by spatial statistical approaches developed within this study.
Uganda is the leading producer and consumer of pork, with the highest per capita pork consumption rate in East Africa at 3.4 kg per year, with demand continuing to rise locally and regionally [
]. The national pig herd is predominately raised in smallholder systems, providing an important livelihood source to >1.1 million households in the country as of 2008 [
]. These systems are characterized by partial housing and outdoor rearing of pigs on pasture and poor sanitary conditions, which are known risk factors for T. solium [
] recently highlighted the large burden associated with T. solium in Uganda, estimating that in 2010, NCC resulted in more than 170,000 disability-adjusted life-years, with economic losses of more than $75 million. However, there has yet to be a systematic approach to understanding the variation in the level of risk for the presence or prevalence of T. solium across Uganda to support the targeting of interventions. Therefore, this study aimed to map all the available T. solium prevalence data and the PCC risk factors within Uganda, building a complete insight into the T. solium landscape and subnational variation of indicators.
Methods
Systematic literature review: T. solium prevalence data
The key elements of the review question are as follows: population (humans and pigs focusing on HCC and human taeniasis (HTT) in humans and PCC in pigs), interventions (baseline infection data measured or cross-sectional studies), context and study design (the review focused on any study conducted in any part of Uganda), and outcomes (prevalence and incidence infection markers). The time frame includes any relevant studies conducted in Uganda up to June 21, 2021, with no lower limit.
Search strategy: The search was conducted on June 21, 2021 and followed the Preferred Reporting Items for Systematic Reviews and Meta Analysis guidelines [
] on conducting systematic literature review (SLR), with a protocol developed before the implementation of the search. Full details on the search strategy, including information on the identification of additional studies, literature databases searched, search terms used for each database, inclusion/exclusion criteria, literature screening/reviewing, and data extraction (including variables extracted in Supplementary Table 1) are included in the Supplementary materials (pp. 12–13).
Prevalence confidence interval and adjustment method
All analyses were conducted using the R environment for statistical computing version 4.1.3 [
R Core Team. R: A language and environment for statistical computing. Version 4.0.5 [software]. 2018. [cited 2022 Jan 17]. R Foundation for Statistical Computing, Vienna, Austria. Available from: https://www.R-project.org/.
]. The binomial confidence intervals not provided by the authors were calculated using the inbuilt package ‘stats’. Informed PCC district-level prevalence was estimated in a Bayesian framework [
] from the reported observed prevalence estimates extracted from the studies. To perform the adjustment, literature-based estimates for sensitivity and specificity of each diagnostic test were identified and used to specify uniform priors (Supplementary Table 2, adapted from Braae et al. [
]. Apparent prevalence was only adjusted for each study where information on the number sampled, and the number of positive individuals was available for each diagnostic used.
Porcine cysticercosis risk factor mapping
The information on risk factors and data sources are provided in Table 1, with PCC risk factor maps for Uganda produced in 2001, 2006, 2011, and 2016.
Table 1Risk factor, data source and threshold values for modelled porcine cysticercosis risk mapping.
Year
Risk factors & data sources
Threshold values to define proportion of population between low/high
The variables for each year of the 2001-2016 Uganda Demographic and Health Surveys (DHS) were aggregated at the cluster level (approximately 20 households) to represent the proportion of households with each variable. First, the spatial autocorrelation for each clustered variable was assessed by computing a semivariogram and fitting variogram models (spherical or wave models to approximate the relationship between semivariance and distance; the parameters used for each model are found in Supplementary Table 3 and the model fits are provided in Supplementary Figure 1) with the gstat package (version 2.0.7) in R [
]. Spatial interpolation methods were then used to transform cluster-level point data into a smooth map of the proportion of households with the variable of interest across Uganda, using the ordinary kriging procedure (with the gstat package in R) and accounting for spatial autocorrelation assessed in the previous step. Pig population densities were already available as a smooth distribution map across Uganda, so the spatial analysis steps were not performed for this variable. For all three variables, the interpolated data points were categorized as high or low based on each variable's upper third (66th centile) distribution of the data (Table 1). Next, each variable (risk factor) was plotted as a binary variable, categorized as either high or low and overlayed into a single composite risk factor map for each year (2001-2016). In the composite maps, the risk factors are shown in yellow (poor sanitation coverage, risk factor A), purple (high pig density, risk factor B), and red (high levels of poverty C), with the highest risk zones indicated in brown (all three risk factors, ABC). The full reproducible code for this analysis can be found at: https://github.com/SCIFoundation/Uganda_porcine_cysticercosis_risk_mapping.
Results
The database search generated 86 records, 69 of which were retained after removing duplicates. After screening titles and abstracts, 35 full-text manuscripts were reviewed with 16 articles retained for data extraction (n = 11 pig studies and n = 5 human studies; see Supplementary appendix: Full data extraction tool). The search and screening strategy results are shown in Figure 1 and described in detail in the Supplementary material (p. 12–13).
Figure 1Preferred Reporting Items for Systematic Reviews and Meta Analysis flow diagram detailing the number of studies identified, screened and assessed for eligibility, and included in the Uganda Taenia solium prevalence data review.
There was an increased frequency of publications from 2015 onwards in comparison to previous years, with one study per year in 1996, 1997, 2002, 2008, 2009, 2013, 2015, 2018, and 2020 and two studies per year in 2017, 2019, and 2021 (Supplementary Figure 2).
The spatial distribution of the studies (Figure 2) shows a relative concentration of earlier studies in the central, mideastern, and western areas, followed by a larger concentration of later studies (i.e., 2010 onwards) in northern regions of the country (HTT/HCC studies in Table 2 and PCC studies in Table 3). The areas in northeastern and southwestern Uganda are devoid of any T. solium studies.
Figure 2Spatial distribution of Taenia solium studies in Uganda.
HCC, human cysticercosis; HTT, human taeniasis; PCC, porcine cysticercosis.
Report on nodule prevalence, not able to calculate population level prevalence from this data. Authors report results for Taenia solium based on exposure (Abs) to the larval-stage (cysticercosis) using western blot, and the adult tapeworm (taeniasis) using rES33 antigen [53], although not specified if immunoblot or ELISA-based-with T. solium positivity based on both tests used.
Report on nodule prevalence, not able to calculate population level prevalence from this data. Authors report results for Taenia solium based on exposure (Abs) to the larval-stage (cysticercosis) using western blot, and the adult tapeworm (taeniasis) using rES33 antigen [53], although not specified if immunoblot or ELISA-based-with T. solium positivity based on both tests used.
Ag, antigen; Ab, antibody; ELISA, enzyme-linked immunosorbent assay; HCC, human cysticercosis; HTT, human taeniasis.
a Confidence intervals calculated by systematic literature review authors.
b Report on nodule prevalence, not able to calculate population level prevalence from this data. Authors report results for Taenia solium based on exposure (Abs) to the larval-stage (cysticercosis) using western blot, and the adult tapeworm (taeniasis) using rES33 antigen
The map was produced by finding the longitudinal average of the published informed occurrences. Districts with a zero prevalence for PCC included Kayunga, Kampala, Lwengo, Nakasongola, Wakiso, Mpigi, Bukedea, Kamuli, and Kumi. Owing to the higher number of PCC entries, the prevalence was divided into five class ranges. The other records represent the single-point estimates (circles with dots). The maps were produced in QGIS version 3.24 [
QGIS Development Team: QGIS Geographic Information System. Open Source Geospatial Foundation Project. Version 3.24 [software]. 2013 [cited 2022 March 22]. Available from: http://qgis.osgeo.org.
], whereas the other studies focusing on HCC sampled all the age groups. The prevalence of HCC ranged between 0.01% and 6.0% based on indirect immunofluorescence test, immunoblot, and antibody enzyme-linked immunosorbent assay (ELISA) (assuming that Ngugi et al. [
] reported HCC only and not a combined HCC/HTT estimate, which was unclear from the study). Furthermore, in an onchocerciasis nodule prevalence survey conducted in Moyo and Kanungu, Katabarwa et al. [
] found that a proportion of 33.3% and 66.7% of the nodules, respectively, were T. solium cysts, indicating the presence of T. solium in this area, though without being able to determine prevalence.
Porcine cysticercosis
A total of 11 studies on PCC were conducted between 2002 and 2019 (Table 3). All studies reported the prevalence of PCC using different diagnostic techniques, including partial and full carcass dissection (used on n = 4 occasions), tongue/lingual palpation (n = 4), postmortem meat inspection conducted by meat inspectors/obtained from abattoir records (n = 1), antigen (Ag)-ELISA (n = 4), or a combination of these (n = 1). Only one study included a T. solium control intervention [
], with baseline prevalence estimates reported in Table 3.
Informed porcine cysticercosis prevalence variation across Uganda
Informed prevalence of PCC ranged from 0.3% to 93.9% (Table 4; grouped by district), with the districts with the lowest/highest informed prevalence (Figure 3) closely following the spatial distribution of apparent prevalence across districts (Figure 2). Large uncertainty ranges were observed around most estimates.
Table 4A table showing the informed PCC prevalence ranges by district. Ordered alphabetically by district.
District
Year range
PCC Informed prevalence range (95% uncertainty intervals)
Note: The average (median) PCC value, where a range exists for a district (in Table 4), is calculated and then plotted in Figure 3, based on the appropriate median PCC prevalence band in the legend of Figure 3.
The average minimum and maximum prevalence values were calculated and plotted for districts with informed prevalence ranges. The maps were produced in QGIS version 3.24 [
QGIS Development Team: QGIS Geographic Information System. Open Source Geospatial Foundation Project. Version 3.24 [software]. 2013 [cited 2022 March 22]. Available from: http://qgis.osgeo.org.
Spatiotemporal variation in porcine cysticercosis risk factors between 2001 and 2016
The composite PCC risk maps for 2001, 2006, 2011, and 2016 show the varied risk factors geographically and over time (Figure 4; Supplementary Figures 3-6 show the individual risk factor maps for each year). In 2001, high levels of poor sanitation (A; yellow) and poor sanitation with high pig densities (AB; green) predominate in the northeast and northcentral regions, respectively, with pockets of high pig densities combined with high poverty levels (BC; pink) in the northwest, central, south, and southwest regions. Large parts of the northeast transitioned from high poor sanitation (yellow) to combined with high poverty (AC; orange), whereas the northcentral/east region transitioned to all three risk factors (ABC; brown). There were additional pockets in 2006 of high poor sanitation and poverty (orange), and all three (brown) emerged from the northwest. High poverty disappeared from combined high poverty pig densities (pink) in those pockets found from 2001. However, new pockets of high poor sanitation and high pig densities (green) emerged in central/southern parts of Uganda. In 2011, areas with all three risk factors (brown) spread into the northcentral regions. The areas of high poor sanitation with high poverty (orange) disappeared from the northeast tip and northwest/western regions. These re-emerged to some degree again in 2016 in these areas of Uganda, whereas areas with all three risk factors retreated largely from northcentral areas in 2016.
Figure 4A map showing the distribution of porcine cysticercosis risk factors over time in Uganda.
The distribution and confluence of risk factors (A = high poor sanitation, B = high pig density, C = high poverty) are indicated. The water bodies (lakes/rivers) are included in blue. Districts relevant for 2020.
Discussion
To the best of our knowledge, this is the first study to collate and assess national T. solium prevalence data in Uganda and explore the modelled spatiotemporal variation in PCC risk factors. This review has shown increasing interest in research on T. solium in Uganda in recent years. However, most studies focused on PCC and were cross-sectional in nature, aiming at establishing the prevalence and risk factors. Cross-sectional studies are often preferred in the study of disease prevalence because they are easier to conduct and require fewer resources and time than other the types of studies [
]. This is unlike the neighboring Tanzania, where many studies have been conducted for various reasons, including establishing prevalence, incidence, risk factors, and co-morbidity with other diseases, such as HIV and epilepsy, as summarized by Ngowi et al. [
]. The greater volume of research work conducted in Tanzania could be attributed to the existence of initiatives supporting research on T. solium, including the Cystinet Africa research consortium (https://www.cystinet-africa.net/), which has funded several research projects over the recent years.
Only one study reported the prevalence of taeniasis in school-aged children from 98 primary schools using microscopy [
]; however, no attempt was made to identify the parasite at the species level, and the infection could have been due to Taenia saginata (beef tapeworm). The study also reported co-infection with other intestinal parasites, including Trichuris trichiura, Ascaris lumbricoides, and hookworm. Unlike taeniasis, these intestinal parasites have received more attention from researchers in Uganda, especially in school-aged children [
Impact of a national deworming campaign on the prevalence of soil-transmitted helminthiasis in Uganda (2004–2016): implications for national control programs.
]. This reaffirms that T. solium research has been neglected due to limited stakeholder awareness and possibly competing public health priorities in many endemic countries, particularly in relation to taeniasis. Three studies reported HCC, two in northern Uganda, and one in the mideastern region. The possibility of NTD co-infections in some regions in Uganda was highlighted in an onchocerciasis survey [
] also identified a low level of exposure to T. solium antibodies (0.01% seroprevalence) in the Iganga and Mayuge districts. However, whether these estimates referred to HCC or HTT antibodies was not evident.
Most studies in Uganda were clustered in the mideastern and central regions, where there is a confluence of risk factors, including high pig densities, poverty, and poor sanitation [
]. Study site selection is generally guided by several factors, including the probable presence of infection risk factors and the contextual factors, to support the research in a specific study area [
]. From 2001 to 2016, a highly dynamic picture emerged of changing T. solium PCC risk as captured through the risk mapping analysis. The most significant changes are observed in the northern regions, particularly in northcentral and northeastern Uganda. In the northeast, from 2001 to 2006, there was a transition from poor sanitation in isolation to poor sanitation combined with high poverty. In addition, in 2006, all three risk factors in combination emerged in large parts of northcentral Uganda. The appearance of poverty across the northern regions in 2006 may be related to the large population movement of internally displaced persons, attributed to the civil conflict experienced in this region [
]. Lack of research in northern Uganda until recently is likely driven by previous insecurity in the region and by assumed low pig densities due to the traditional predominance of pastoral systems, with pig populations in the northern region estimated at 340,460 compared with 1,307,460 and 699,680 in central and eastern regions respectively [
]. In addition, geographic targeting of earlier studies in regions close to major universities in Kampala may have driven research priorities particularly before 2010.
Moreover, the smaller dispersed pockets of high PCC risk in northern areas also require further work to determine whether these pockets are of sufficient geographic scale to maintain the T. solium transmission cycle. For example, the local migration patterns and extensive pig trade networks are essential for maintaining endemic equilibria in other settings [
]. Within Uganda, pig trading occurs at varying geographic scales. Pigs are moved locally during trading through brokers located in the villages and then often moved from farm to farm or from village to village, before being shipped to urban centers for slaughter [
]. Although, a high prevalence of PCC in a particular village may not strongly correlate with high HCC prevalence in that village, the correlation may be high within surrounding villages and districts. Infected pigs transported to slaughter in larger cities, such as Lira in the north and Kampala, may be detected because meat inspection is relatively better enforced in these slaughter places than the rural districts. In addition, PCC cases have been shown to cluster around a tapeworm carrier [
]. This scenario may also present a risk for HCC infections within the villages due to the potential contamination of the environment with eggs from the taeniasis carrier.
One major limitation associated with the risk mapping analysis was the lack of pig density distribution maps for each of the four DHS years, with pig density mapping underpinned by data from 2007 in the Gridded Livestock of the World database. The temporal changes in pig densities may have therefore been missed, such as changing husbandry systems in northern areas, reflecting the settlement of internally displaced populations or refugees from South Sudan or fluctuating pig densities in southern areas, depending on the changes in localized pig production systems. The risk mapping, however, highlights potential gaps in epidemiological knowledge where there are varying yet high levels of risk factors for PCC, such as those identified in northern areas (e.g., Abim and Agago districts). There are additional interesting spatiotemporal patterns, such as the transition from the overlap of the three risk factors from 2006 to 2011 in the western region and northeastern tip, leaving only high pig densities or areas of no risk only for these zones. The multiple risks re-emerged again in 2016. These patterns may be a true reflection of changing risk or underpinned by changes in the coding of the DHS, such as the presence of the latrine type (e.g., uncovered with slab or without a slab) between 2011 and 2016.
A further limitation surrounds the lack of quantitative thresholds linking a level (proportion) of a risk factor that equates to a high risk for T. solium exposure. Therefore, we assumed values equal to or above the third quartile of values from distributions of DHS cluster-level values for each variable that represented a high risk for T. solium transmission. This cut-off was therefore used to select high-risk values from the modelled risk factor maps. Determining a quantitative relationship between pig density levels/proportion of high poverty, poor sanitation, and sustained T. solium transmission constitutes an important area for future research. Equally, poor sanitation, poverty, and high pig densities are currently equally weighted. Therefore, more information is required to elucidate each risk factor's relative contributions and therefore weighting toward T. solium transmission. The validation of the spatial mapping method, which remains a challenge due to the paucity of prevalence data, also forms a critical future step. For example, our group will focus on validating the method against the current epidemiological situation in Uganda. This will involve updating the risk mapping for the next available DHS and identifying or conducting prevalence surveys in different modelled risk zones to validate these different risk areas.
As with many other regions of the world, most studies (11/16) focused on PCC, perhaps because it is easier to measure the study outcomes in pigs due to their short life cycle and less stringent ethical considerations. The 11 studies reported varying prevalence across different regions of Uganda using different diagnostic techniques with different sensitivities and specificities. Only one study [
], so the prevalence was adjusted within a Bayesian framework. Notably, there was substantial spatial heterogeneity in the informed prevalence, with a number of districts recording relatively high informed prevalence, reaffirming the high endemicity in the country as illustrated by WHO [
] identified moderate PCC prevalence (10-20%) in Uganda. However, these were only based on four studies and reported the apparent prevalence estimates. Second, there were large uncertainty intervals associated with the informed prevalence estimates, for example, in Zirintunda and Ekou [
], where the uncertainty ranges reflected wide confidence intervals for the sensitivity/specificity values and small district-level sample sizes in 10 districts, where less than 10 pigs were sampled [
] used two Ag-ELISA tests in parallel; however, a breakdown in the numbers positive and sampled for each test was only available for the value chain type and rural/urban settings and not by district. In those studies where adjustment was not possible, findings of low (sero)prevalence may have instead reflected the presence of false positives arising from low specificity of the diagnostic test. More generally, the need to include all the studies reporting any T. solium infections in Uganda may have led to the inclusion of studies with potential risk for bias and studies of low quality, which is acknowledged as a limitation of the systematic literature reviews (SLR) component of the analysis.
Although, there was no extensive search of gray literature, which could be a limitation, the review highlights the dearth of data regarding the effectiveness of T. solium interventions in Uganda. Among all the papers, only one study by Nsadha et al. [
] explored an intervention focused on control of the parasite in the porcine host. This control trial study demonstrated a statistically significant PCC prevalence decrease in vaccinated pigs with TSOL18 and treated with oxfendazole compared with controls. Although the tools for control are available, no nationally supported control program has been implemented in Uganda thus far, with high maintained T. solium prevalence likely due to the absence of control efforts to date. The effective control of this parasite requires understanding the context in which the control interventions are carried out and exploring effective combinations of these control tools [
]. Therefore, this finding calls for a focus of efforts toward T. solium control while engaging and involving the national government. In addition, to fully benefit from the approach proposed in this study, there is a need to bring together the coordinators of the national NTD programs from both the national and regional government, veterinary departments, and academics or researchers to tap into the added value of the One Health approach, as described by Rüegg et al. [
]. One previous study conducted for Central America and the Caribbean basin region mapped first-level administrative subdivision PCC data (obtained from a systematic literature review and assessment of gray literature sources) with pig density distributions [
]. Our study uses multiple risk factors to identify PCC risk zones, in combination with the identification and mapping of T. solium prevalence data, to provide a comprehensive approach to developing situational awareness for the landscape of T. solium in Uganda. This comprehensive risk mapping approach can help to target (i) further or new research efforts to determine the presence/prevalence of T. solium, (ii) the implementation of control activities, and (iii) the potential to characterize at-risk human populations for serious adverse events associated with praziquantel treatment in schistosomiasis mass drug administrations, where latent NCC may also be present [
The WHO mapping tool provides an important step in supporting countries to rapidly assess the epidemiological context by collating in-country knowledge and expertise to conduct desk reviews and risk assessments, with the aim of identifying high-risk geographic units. Our approach is complementary to this by providing a protocol to conduct a more systematic and technical approach using spatial statistics. This approach can build on the rapid assessment collated through the WHO mapping toolkit, providing a more structured yet timely approach to mapping the T. solium epidemiological landscape in a country.
The approach presented in this paper uses a spatial statistics approach within a flexible framework to complement and extend existing rapid mapping efforts, such as those based on the WHO mapping tool, or scoping studies, such as that carried out by Ngowi et al. [
]. This will not only help countries prioritize where these initial pilot control efforts are concentrated but also identify other areas for expanding control efforts or undertaking further research to understand the endemicity landscape better.
Conclusion
An accurate mapping of T. solium is becoming increasingly urgent, given the WHO goal to see intensified control in hyperendemic areas by 2030. Our analysis reveals marked variation in T. solium prevalence across Uganda and, in combination with the PCC risk maps, highlights geographic areas of potential risk, especially in northern areas, where limited research has been undertaken to date. In addition, informed PCC prevalence estimates suggest high levels of PCC infection, particularly in central areas; although, these estimates are subject to substantial uncertainty due to limited study sample sizes and suboptimal diagnostic performance. Taken together, there is clear evidence for need of further research and the urgent implementation of T. solium control efforts in Uganda.
Funding
NN, DS, MK & LFT are thankful for the support by The German Federal Ministry for Economic Cooperation and Development through the One Health Research, Education and Outreach Centre in Africa (OHRECA). We thank all donors and organizations who globally supported its work through their contributions to the CGIAR Trust Fund. https://www.cgiar.org/funders/, https://www.ilri.org/research/facilities/one-health-centre. MAD is funded by the SCI Foundation (SCIF) - https://unlimithealth.org/, with support from Bayer AG and Merck.
Ethical approval
Approval not required.
CRediT authorship contribution statement
Nicholas Ngwili: Conceptualization, Methodology, Formal analysis, Investigation, Formal analysis, Writing – original draft, Writing – review & editing, Visualization, Supervision, Project administration. Derrick N. Sentamu: Methodology, Formal analysis, Investigation, Data curation, Writing – original draft, Writing – review & editing, Visualization. Max Korir: Methodology, Formal analysis, Investigation, Data curation, Writing – review & editing, Visualization. Moses Adriko: Writing – review & editing. Prudence Beinamaryo: Writing – review & editing. Michel M. Dione: Methodology, Writing – review & editing. Joyce Moriku Kaducu: Writing – review & editing. Alfred Mubangizi: Writing – review & editing. Pauline Ngina Mwinzi: Writing – review & editing. Lian F. Thomas: Conceptualization, Methodology, Writing – review & editing, Supervision, Funding acquisition. Matthew A. Dixon: Conceptualization, Methodology, Software, Formal analysis, Data curation, Writing – original draft, Writing – review & editing, Visualization, Project administration, Funding acquisition.
Declaration of Competing Interest
The authors have no competing interests to declare.
Acknowledgments
The authors would like to thank Dr. Catherine Pfeifer for providing the original code for the PCC risk mapping, which has been subsequently adapted and extended for this project. The authors also thank Geoffrey Njenga, International Livestock Research Institute (ILRI) for the support in proofreading the manuscript.
R Core Team. R: A language and environment for statistical computing. Version 4.0.5 [software]. 2018. [cited 2022 Jan 17]. R Foundation for Statistical Computing, Vienna, Austria. Available from: https://www.R-project.org/.
QGIS Development Team: QGIS Geographic Information System. Open Source Geospatial Foundation Project. Version 3.24 [software]. 2013 [cited 2022 March 22]. Available from: http://qgis.osgeo.org.
Impact of a national deworming campaign on the prevalence of soil-transmitted helminthiasis in Uganda (2004–2016): implications for national control programs.
00048-6&id=gr1.jpg){kind=link}
00048-6&id=gr2.jpg){kind=link}
00048-6&id=gr3.jpg){kind=link}
00048-6&id=gr4.jpg){kind=link}