The nexus between improved water supply and water-borne diseases in urban areas in Africa: a scoping review

Background: The sub-Saharan Africa has the fastest rate of urbanisation in the world. However, infrastructure growth in the region is slower than urbanisation rates, leading to inadequate provision and access to basic services such as piped safe drinking water. Lack of sufficient access to safe water has the potential to increase the burden of waterborne diseases among these urbanising populations. This scoping review assesses how the relationship between waterborne diseases and water sufficiency in Africa has been studied. Methods: In April 2020, we searched the Web of Science, PubMed, Embase and Google Scholar databases for studies of African cities that examined the effect of insufficient piped water supply on selected waterborne disease and syndromes (cholera, typhoid, diarrhea, amoebiasis, dysentery, gastroneteritis, cryptosporidium, cyclosporiasis, giardiasis, rotavirus). Only studies conducted in cities that had more than half a million residents in 2014 were included. Results: A total of 32 studies in 24 cities from 17 countries were included in the study. Most studies used case-control, cross-sectional individual or ecological level study designs. Proportion of the study population with access to piped water was the common water availability metrics measured while amounts consumed per capita or water interruptions were seldom used in assessing sufficient water supply. Diarrhea, cholera and typhoid were the major diseases or syndromes used to understand the association between health and water sufficiency in urban areas. There was weak correlation between the study designs used and the association with health outcomes and water sufficiency metrics. Very few studies looked at change in health outcomes and water sufficiency over time. Conclusion: Surveillance of health outcomes and the trends in piped water quantity and mode of access should be prioritised in urban areas in Africa in order to implement interventions towards reducing the burden associated with waterborne diseases and syndromes.


Introduction
The sub-Saharan Africa (SSA) has experienced the highest annual urban population growth rate (more than 3.5%) in the world 1 . However, the growth of urban infrastructure has been slower, leading to populations without access to adequate resources including water services, health facilities, and housing 2,3 .
Globally, it is estimated that one in every two people will be living in water stressed areas by 2025 increasing the challenge of water supply 4 . As of 2017, only half of the population residing in urban areas in SSA had access to improved water sources which included piped, boreholes, protected wells or springs, rainwater or packaged water 5 . However, going by The World Bank categorisation of piped water as the only major source of improved water in urban areas in SSA 6 , only 56% (230 million people) residing in urban areas in this region have access to clean water 7 .
More than half a million deaths in SSA have been attributed to diarrheal diseases, with water contamination being one of the key risk factors 8 . The global enteric multicenter study identified Escherichia coli, Cryptosporidium, Aeromonas spp, Shigella spp and Entamoeba histolytica to be associated with increased risk of death among children younger than 24 months with moderate-to-severe diarrhea 9 . Due to their high burden, several waterborne diseases including cholera, bloody diarrhea and typhoid are included in the Integrated Disease Surveillance Strategy used in most African countries to improve countries speed of detection and response to public health threats 10 .
The United Nations Sustainable Development Goals (SDGs) 3, 6 and 11 that focus on good health and wellbeing of populations; clean water and sanitation; and sustainable cities and communities directly or indirectly address this problem associated with rapid urbanisation in SSA 11 . The African Union Agenda 2063 aspires to have an African continent that is based on inclusive growth and sustainable development 12 . To reduce the burden of waterborne diseases in the context of an urbanising population, a good understanding of the relationship between water and these health outcomes is required.
Previous reviews have focused on water quality 13,14 , water availability 15,16 and the reallocation of water from rural to urban regions in Africa 17 . Other reviews have also focused on the environmental determinants of waterborne disease outbreaks in Africa 18 , the link between waterborne diseases and water resource development in Africa 19 and climate change globally 20 . To ensure a medium level of health concern, an access of at least 50 litres per person per day is required 21 . However, there is a gap on insufficient access to piped water (less than 50 litres per person per day) in urban areas in Africa and the association with waterborne diseases and syndromes in the African continent.
Here, we conduct a scoping review to assess the link between sufficient access to piped water supply and waterborne diseases and syndromes in African cities. Specifically, we answer the following questions: i) How has the relationship between waterborne diseases and piped water sufficiency been studied in Africa? ii) Are there under-utilised study designs, under-studied metrics of water sufficiency or under-studied syndromes or waterborne diseases?

Literature search methods
This scoping review was conducted following the Joanna Briggs Institute methodology guidance for scoping reviews 22 and the preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) extension guidelines for conducting scoping reviews 23,24 . Briefly, this approach involves: i) conducting a systematic literature search to identify articles that meet the inclusion criteria, ii) assessing the relevance of the articles to the study question(s), iii) assessment of the full text articles iv) data extraction and synthesis. The scoping review protocol for this study is published and available 25 .

Information sources and search strategy
In April 2020, literature searches were undertaken in the following four electronic databases: Embase, MEDLINE, Web of Science and Google Scholar (first 500 papers). These have been identified as the optimal combination of databases that would guarantee adequate and efficient coverage of studies for literature searches 26 . The exact dates when searches were conducted can be found in Table A1.
The search strategy consisted of a two-step process. The first step involved carrying out a limited search in MEDLINE, Embase and Web of Science databases to analyse the text words and index terms that are used to describe the articles. The second step included a keyword search in all four databases; index terms were also used. The search terms that were used in the study can be seen in Table 1. The search terms include a combination of names of all African cities that have a population of at least half a million residents as of 2014, as outlined in the protocol 25 , and terms representing the exposure (insufficient piped water supply) and outcome (waterborne diseases and syndromes). The study focused on publications that were written in the English or French language.

Data screening
Once searches were complete, the title and abstracts were extracted from the articles. Duplicates were removed and three Table A1. Exact dates when the searches were run in the databases. reviewers (NM, JM, MM) independently screened the study titles and abstracts using the following criteria: 1) Studies that described the water sufficiency or water situation in cities with populations more than 500,000 in 2014 2) Studies that focused on cholera, typhoid, amoebiasis, cyclosporiasis or giardiasis as diseases, dysentery, diarrhea or gastroenteritis as symptoms or cryptosporidium or rotavirus as etiological agents for diarrheal diseases 27,28 ; 3) Studies published in international scientific indexing (ISI) listed journals Any inconsistencies between the three reviewers were discussed and a consensus was reached on whether to include or remove articles from the study.

Study selection
Where available, the full text articles were obtained for all studies that met the inclusion criteria. Two reviewers (NM and MM) assessed and characterised the studies by analysing if they primarily targeted urban residents and had evaluated the relationship between a health outcome and a water sufficiency metric. The data extracted from this screening process were stored in an Excel spreadsheet.

Data extraction, synthesis and presentation
Variables on author(s), study period, source of funding, geographical scope, study design, population inclusion criteria, sample size and statistical methodology used, and whether or not the study investigated a disease outbreak were extracted from the studies.
To understand piped water access and quality reported by the studies, we extracted information on the nature of the piped water supply, mode of accessing this piped water, measurement of the unit cost of water, the per capita daily water consumption, proportion of the population without access to piped water and water quality indicators from water samples collected for testing. The reported coping mechanisms employed to supplement water needs were also extracted. Information on the health outcomes studied and how diagnoses was made (selfreported, clinically diagnosed or culture confirmed) was also extracted from the articles. Table 2 provides a list of the variables extracted from the articles during the screening process.

Assessment of the study quality
We used the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) checklist to analyse the quality of the studies included in the scoping review 29 . We assessed the studies based on whether the study objective was comprehensively stated, the study design, description of study location and dates of data collection were provided, provision of participant eligibility criteria and rationale given for sample size, explanation of how missing data was handled and how they controlled for confounders. No study was excluded based on it being poor quality. Connectedness of the study designs of the associations between water sufficiency and health outcomes To understand the connectedness of the different study designs with the health outcomes and water sufficiency metrics and water quality, we used the principal component techniques 30 . The main categories of the study designs employed in the selected publications were evaluated together with the health outcomes (self reported, clinically or culture confirmed) and a binary coding of assessment of water quality. The water sufficiency metrics were coded into either water access (mode of access, proportion with access, time/distance to water points) or water quantity (scheduled/ unscheduled interruptions, litres per person per day) categories. We carried out multiple factor analysis by grouping the study designs, health outcomes, water sufficiency metrics and whether water quality was assessed. We looked at the contributions of the first two axes and assessed the combinations of the variables that were connected, understudied and the outliers. The analysis was carried out using the FactomineR package in the statistical software R 31,32 .

Study selection
The initial database search revealed 3,099 articles. After removing duplicates, and assessing the abstracts for eligibility, 93 articles remained for full text review, with 32 of those studies meeting the inclusion criteria ( Figure 1).
Quality of the studies From our checklist, there were some strengths and weaknesses of the studies. All the studies had a clearly stated objective, study design and study location with date of data collection.
The eligibility criteria of the study participants were also clearly stated by majority of the studies (n=31, 97%).
Three quarters of the studies reported on the statistical methods employed (n=24, 75%). Less than a third of the studies explained how the study size was calculated (n=8, 25%), the criteria used in choosing the quantitative variables (n=2, 6%) and how the studies controlled for confounders (n=9, 28%). None of the studies explained how they addressed missing data ( Table 3).

Characteristics of the publications
A total of 32 articles that assessed the association of water sufficiency in urban areas and waterborne diseases and syndromes in SSA were published between 1998 and 2019. These studies focused on 24 cities in 17 countries across Western, Eastern and Southern Africa, with 22% (n=7) of the studies based in urban Nigeria ( Figure 2). Seven of the articles (22%) were conducted in informal settlements [33][34][35][36][37][38][39][40] . Nearly half the studies did not report the source of funding, with government and philanthropies supporting most of the studies that provided that information (Table 4).
To understand the association between water and waterborne diseases and syndromes, the studies mainly used bivariate and multivariate methods of analysis. The common bivariate analysis methods used included the chi-square tests, Fisher tests, Wald tests and the correlation coefficient methods while the multivariate analysis methods included regression models (linear, logistic, random effects) and ANOVA models. The multivariate analysis models controlled for confounders/ effect modifiers in the analysis using independent variables which included source of water, type of water storage container, presence of water treatment, household hygiene and sanitation conditions, household characteristics which included size, income, employment, and presence of children (Table 4). A study done by Machdar et al employed cost-effective analysis methods to assess the cost-effectiveness of interventions for reducing the disease burden from consumption of poor drinking water 38 . Variable justification Explanation of the criteria used to choose the quantitative variables is included.
The objective assessment of water safety was assessed by the studies via testing water samples (n=19, 59%). The water samples were collected from the dominant water points of the study population (n=9, 47%), water stored in the households (n=7, 37%), both dominant water points and stored water in the households (n=3, 16%) or hand rinse samples (n=1, 3%). Several studies assessed water contamination by testing for coliforms (n=17, 89%), effectiveness of measures of protecting water from contamination through testing for free residual chlorine (n=7, 37%), organoleptic characteristics of water by assessing turbidity and pH (32%, n=6) and presence of pathogens which included klebsiella pneumoniae, staphylococcus aureus, pseudomonas aeruginosa, among others (26%, n=5).

Connectedness of the study designs used
We assessed the connectedness in the study design methods used by the articles to understand the nexus between water sufficiency and health outcomes, as shown in Figure 3. The axes in the biplot represented the first two principal components of the input data which explained 27% of the total variability, showing weak correlation among the study designs.
The black triangle markers in Figure 3 represent the mean centres for the health outcomes and the characteristics of piped water supply that were studied by the articles. The correlation circle is portrayed by the uncolored hollow black circle. The colored confidence ellipses, which are plotted around the group mean points, represent the study design methods employed by the studies and the size of the ellipses are based on the variance of each group. The numbers represent each publication included in our study.
From this analysis, we observed that cross-sectional individuallevel, cross-sectional ecological level and case control studies had a high variance and were the three commonly used study designs. Cross-sectional individual study designs were generally used in self-reported health outcomes while cross-sectional  ecological and case control study designs were used in assessing clinically confirmed and culture confirmed health outcomes respectively. Water quantity and quality were mainly assessed using cross-sectional individual and ecological level study designs, whereas water access was mainly assessed using crosssectional individual-level study designs. An unusual combination of self-reported typhoid and water quantity was observed as an outlier (Figure 3). Use of cohort study designs in assessing the association between waterborne diseases and syndromes and water sufficiency was under-utilised.

Discussion
Our study presents the results of a scoping review on associations between water supply and waterborne diseases and syndromes in large cities across Africa. We find that majority of the studies have been published since 2005. The relationship between piped water sufficiency and waterborne diseases/ syndromes has mainly been studied using cross-sectional individual level study designs employing bivariate statistical methods. The main measures of water sufficiency used are access levels to piped water and water quality assessments while the health indicators mainly used are self-reported or clinically confirmed health outcomes. Cohort study design methods, measure of availability of piped water using quantifiable measures that include either per capita daily water consumption or water interruptions, cryptosporidium, cyclosporiasis, amoebiasis, rotavirus water borne diseases and culture confirmed assessment of health outcomes have been underutilised. Similarly, multivariate methods which are important in assessing the confounders or alternative transmission pathways have been seldomly used.
Piped water has been listed as the primary source of improved water in this region 66 , however results from this review contest to this with no evidence of sufficient piped water supply in the urban areas. Daily per capita water consumption and mode of access have been reported to be inversely proportional to the level of health concern, in outbreak and non-outbreak conditions 21 . However, these two variables were under-studied and only assessed by two studies, neither of which investigated an outbreak 51,53 .
The use of alternative or secondary water sources, that are often unimproved (as classified by the Joint Monitoring Programme (JMP) of the World Health Organisation (WHO) and United Nation's International Children's Emergency Fund (UNICEF)), have been listed as one of the prevalent transmission pathways for water-related pathogens, due to high exposure to faecal contamination 13,67 . Adequate water treatment has the potential to reduce contamination of these water supplies by half 43 . The studies included in this review reported use of alternative water sources as a key coping mechanism for poor or intermittent water supply while only a small proportion reported use of water treatment. Water contamination tests were a common assessment of water quality, contributing to the increased evidence of contamination in the predominant coping mechanisms employed by residents in urban areas.
Water storage, which was the second major coping mechanisms employed by the residents in urban areas, was observed as having the potential to increase the burden associated with waterborne diseases and syndromes. Low income earners, who account for 61% of the population in Africa, regularly practice poor water storage 68,69 . On the other hand, residents with a high income mainly invest in large storage tanks to ensure they enjoy safe storage and adequate water consumption even during periods of irregular water supply 70 . The in-depth qualitative assessment of poor water storage practices and their association with waterborne diseases was under-studied. None of the studies focused on user reported organoleptic characteristics of stored water in their households.
Diarrhea and cholera were the majorly self-reported and clinically confirmed health outcomes respectively while cryptosporidium, cyclosporiasis, amoebiasis, rotavirus water borne diseases were under-studied. These four waterborne diseases are among the major etiological agents associated with moderate to severe diarrhea in children below five years 9,71 . Additionally, clinically and culture confirmed health outcomes are the two main approaches used in case definition of diseases of public health concern, with cases confirmed through objective assessment of samples at the laboratory 72 . However, culturally confirmed health outcomes were seldomly employed in these studies, making it difficult to assess the public health burden associated with waterborne diseases.
Cross-sectional ecological and individual-level studies and case control studies were the main study designs used to understand the association between water sufficiency and health.
Cohort study designs and multivariate statistical methods were under-utilised, limiting the detection of hotspots.
One of the limitations of our study was a lack of studies in Luanda, Kinshasa, Cairo, Johannesburg, Khartoum cities that had a population of more than 5 million people as at 2014 and are expected to be mega-cities by 2030 73 . Furthermore, there were no studies on cyclopsoriasis which was one of the waterborne diseases under our study criteria. Another limitation of our study was potential bias introduced through the choice of databases to conduct the search. Furthermore, we did not omit any studies based on the quality appraisal conducted on the included publications. These limitations have also been reported in other scoping reviews 74 . The use of a nonconventional analysis method in our review may have also been a limitation assessing the connectedness of the study designs, health outcomes, water sufficiency and assessment of water quality. Similarly, our analysis methods deviated from the published protocol found here 25 where we had proposed to conduct cluster analysis to differentiate self-reported diarrheal diseases with etiological agents. This was not possible due to the diverging water sufficiency characteristics reported by the studies. We also did not present digital maps which overlayed the study locations and the water scarcity peer reviewed maps, as stated in the scoping review protocol. This is because the main outcome of our study was depicting under utilised study designs, health outcomes and water sufficiency metrics.

Conclusion
Monitoring of health outcomes and the trends in availability and mode of access of piped water should be prioritised in urban areas in Africa in order to implement interventions towards reducing the burden associated with waterborne diseases and syndromes. This will contribute towards understanding the exposure pathways. Similarly, this is an area that can be used to assess the strategies of Africa being closer to achieving the United Nations SDGs regarding sustainable cities, adequate water, good health and wellbeing of its citizens and the Africa Union aspiration of having an African continent that is based on growth and sustainable development while coping with water insufficiency.

Data availability
All data underlying the results are available as part of the article and no additional source data are required. scoping review and it is unclear how many of the final 32 studies were of poor quality. It would have been good for the authors to discuss the quality of the paper and link it to the strength of evidence these studies provided in terms of any associations between improved drinking water and water-borne diseases.

Reporting guidelines
Are the rationale for, and objectives of, the Systematic Review clearly stated? Yes Are sufficient details of the methods and analysis provided to allow replication by others? Yes

Is the statistical analysis and its interpretation appropriate? Yes
Are the conclusions drawn adequately supported by the results presented in the review? Partly