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

Introduction: Currently, an estimated two thirds of the world population is water insufficient. As of 2015, one out of every five people in developing countries do not have access to clean sufficient drinking water. In an attempt to share the limited resource, water has been distributed at irregular intervals in cities in developing countries. Residents in these cities seek alternative water sources to supplement the inadequate water supplied. Some of these alternative sources of water are unsafe for human consumption, leading to an increased risk in water-borne diseases. Africa contributes to 53% of the diarrheal cases reported globally, with contaminated drinking water being the main source of transmission. Water-borne diseases like diarrhea, cholera, typhoid, amoebiasis, dysentery, gastroenteritis, cryptosporidium, cyclosporiasis, giardiasis, guinea worm and rotavirus are a major public health concern. The main objective of this scoping review is to map the available evidence to understand the sources of water among residents in cities in Africa and the relationship between clean water sufficiency and water-borne diseases in urban Africa. Methods and analysis: The search strategy will identify studies published in scientific journals and reports that are directly relevant to African cities that have a population of more than half a million residents as of 2014 AND studies on the ten emerging water-borne diseases, which are diarrhea, cholera, typhoid, amoebiasis, dysentery, gastroenteritis, cryptosporidium, cyclosporiasis, giardiasis, guinea worm and rotavirus. Ethics and dissemination: This scoping review did not require any formal ethical approval. The findings will be published in a peer-reviewed journal.


Introduction
Urbanization in sub-Saharan Africa (SSA) is growing at 4% annually, with the population living in urban areas projected to double by 2050 1 . This rapid urbanisation outpaces the development of infrastructure in these cities leading to inadequate access to basic amenities, including good housing, adequate social amenities, and continuous supply of safe drinking water to city residents 2 . Poor access to clean water driven by rapid population growth, increased water demand, infrastructural constraints, and consumer response to cope with insufficient supply of safe water through use of alternative water sources is associated with ill-health 3 . This rapid urbanisation is anticipated to accelerate demand for water 2,4 , which lies at the nexus of food security, poverty reduction, economic growth, energy production and human health 5,6 .
Currently, an estimated 40% of the global population is water insufficient 7 . The World Health Organisation (WHO) international benchmark states a minimum water requirement per person per day of between 50 and 100 litres in order to meet basic domestic needs 8 . However, by 2017, only half of the population residing in urban SSA had access to safely managed drinking water that was free from contamination. Goal 6 of the Sustainable Development Goals (SDGs) aims to attain sustainable management and availability of sufficient clean water and sanitation for all 7 . This aligns with Aspiration 1 of the African Union Agenda 2063 objectives, on sustainable development in Africa 9 .
Residents of urban areas in Africa depend on both improved and unimproved water sources including piped water, boreholes, wells, vendors and surface water 8 . However, the Africa Infrastructure Country Diagnostic report by the World Bank categorised piped water as the only major source of improved water in SSA 10 . In 2017, piped water was only accessible to 230 million people (61%) in urban areas in this region 11 . In an attempt to share and ration limited resources, piped water has been distributed in cities in developing countries at irregular intervals, known as intermittent water supplies (IWS) 12 . Residents have responded to these challenges by seeking alternative sources of water, some of which are unsafe for human consumption 13,14 . Half (53%) of the water sources in Africa are faecally contaminated, predisposing the people to the risk of diarrheal diseases 15 .
In 2016, more than half a billion deaths in SSA were attributed to diarrheal diseases with contamination of drinking water identified as one of the leading risk factors 16 . Major pathogens, such as Escherichia coli, cryptosporidium, aeromonas spp, shigella and entamoeba, often found in unsafe water, are associated with moderate-to-severe diarrhea which is especially life threatening to infants 17 . Disease surveillance and response in SSA lists cholera, diarrheal diseases and typhoid fever as some of the priority diseases associated with poor quality water, that should be regularly monitored and reported 18 . Previous reviews have explored the prevalence of intermittent water supplies in low income settings 19 , household water availability across Africa 20 , the global burden of diarrheal diseases 16,21 and implications of intermittent water supply on gastrointestinal illness 22 . We have found no review focused on urban areas in Africa and the implications of water-borne diseases as a result of intermittent piped water in this region.
This manuscript details the protocol to a review of the link between insufficient piped water supply and waterborne diseases and syndromes in urban Africa. In doing so, it seeks to address the following research question: "what is the proportion of residents with safely managed water in cities in Africa and what is the correlation with water-borne diseases and the symptoms?" This will be achieved by synthesising findings of studies on: a) water sufficiency in cities within Africa; b) consequences of rapid urbanisation on water sufficiency in African cities; and c) the linkages between water sufficiency and water-borne diseases and their symptoms in Africa.
This work should provide information to guide policies that aim to help Africa achieve one of its Agenda 2063 aspirations on provision of adequate basic necessities for urban populations in the region 9 .

Definitions
For the purpose of this protocol and the planned review, key terms are defined and classified as follows: • Waterborne disease: includes cholera, typhoid, amoebiasis, cyclosporiasis and giardiasis diseases [16][17][18] • Symptoms of waterborne diseases: focus on diarrhea, dysentery and gastroenteritis 16,17 • Etiological agents of diarrheal diseases: include cryptosporidium and rotavirus 16,17 • Water insufficiency: is classified as having less than 50 litres per person per day 8

Amendments from Version 1
The current version includes a more detailed definitions section which standardises the key terms used in this protocol. In addition, the title, abstract and the inclusion and exclusion criteria have been amended to ensure the aim and the scope of the study is clear and precise. Details on the variables to be extracted from the articles for full-text screening have been clearly stated to capture indicators on water sources, sufficiency and safety with occurrence of waterborne diseases in cities in Africa. Finally, the presentation of results has been improved to ensure the study locations of the studies reviewed will be linked to the coordinates from peer reviewed and publicly available water scarcity map from the Water Footprint Network to allow us to make observations on the trend between cities with prevalence of waterborne diseases and water scarcity. The resultant maps will enable researchers to identify areas that have gaps in knowledge and research needs.
Any further responses from the reviewers can be found at the end of the article

REVISED
• City: an urban area with a population of more than half a million residents 23

Methods
The scoping review will use the Joanna Briggs Institute methodology guidance 24 and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines for conducting systematic reviews and meta-analysis (http:// www.prisma-statement.org). These methodologies have been used for published scoping reviews 25,26 .

Inclusion criteria
The review will include the following criteria: 1) Studies undertaken in African Union member states.
2) Studies describing the water situation in cities, classified as urban areas with greater than half a million residents.
Since the classification of a city and urban environments is not standardised 2 , we use areas with a population greater than 0.5 million people to be consistent with the UN report that estimates one in every three people will reside in cities with at least half a million inhabitants by 2030 23 (Figure 1). The list of the cities that meet this criterion have been selected from the United Nations World Urbanisation Prospects of 2014 27 .
3) Studies focusing on water-borne diseases, symptoms and etiological agents:

Search strategy
Comprehensive literature searches will be undertaken in Embase, MEDLINE, Web of Science and Google Scholar databases. These four databases have been identified as the optimal combination of databases that will guarantee adequate coverage of studies for this scoping review 28 .
The search strategy will take a three-step process. The first step will involve carrying out a limited search in MEDLINE, Google Scholar (first 500 results), Embase and Web of Science databases. The text and index terms that are used to describe the articles will be assessed. The second step will include searches using the keywords and index terms. In the final step, we will go through the references to identify key articles that might have been missed in the first two steps. The search terms used in the study are seen in Table 1.

Study selection
Once the searches have been undertaken in the databases, the title and abstracts will be extracted from the articles. Duplicates will be removed, and the review team will screen the studies using two levels: initial screening and full-text screening. During the initial screening process, three reviewers will read the abstracts of the studies captured by the search terms and assess their relevance in light of the inclusion criteria. To ensure consistency, 10% of all the studies will be randomly selected and independently reviewed by one other reviewer. Any inconsistencies between the primary and secondary reviewers will be discussed and a consensus reached.
Full text articles will be obtained for the studies that pass the initial screening stage. Microsoft Excel (version 16.36) will be used to store the extracted data. Table 2 shows the characteristics that will be extracted from each study. Of the data extracted, 10% will be randomly selected and independently reviewed by one other reviewer. Any inconsistencies amongst the reviewers will be discussed and an agreement will be reached.
All irrelevant studies will be removed and the reason for their exclusion will be recorded. In this stage, another 10% of the studies will be sampled and shared with the secondary reviewer who will exclude or include the studies based on their relevance to the study objective. Consensus will be reached for any discrepancies in the studies among the reviewers.

Presentation of results
If a sufficient number of studies report on the effect of water insufficiency on health outcomes, we will calculate heterogeneity  (I 2 ) for this subset. The index of heterogeneity (I 2 statistic) will be calculated from the sum of the squared deviations of the estimate of each study, from the overall estimate, and weighted by the influence of the study on the calculation of the overall estimate. We will examine the risk bias in the study level and characterize whether the metrics of water insufficiency and health are representative of the whole urban population or only a sub-group. We will use the R statistical software (version 3.6.1) to conduct this analysis 29 .
Cluster analysis will be performed to collate similar studies using Ward's agglomerative hierarchical clustering method, which is used in other scoping reviews 30 . The optimal number of clusters will be chosen to ensure the inner homogeneousness and external heterogeneousness of a cluster is balanced. For studies that focus on diarrheal disease, we will differentiate the self-reported studies from those with etiological characterisation of pathogens and input these studies into the planned cluster analysis.
The study locations will be geo-coded and the data will be presented using digital maps that will depict the water sufficiency in these different cities. The results will be linked to the coordinates from the peer reviewed and publicly available water scarcity map layer from the Water Footprint Network 31 which has been used in previous systematic reviews 32 . This will allow us to make observations on the trend between cities with prevalence of waterborne diseases and water scarcity. The resultant maps will enable researchers to identify areas that have gaps in knowledge and research needs.

Ethics and dissemination
The study does not involve any interviews or interactions with humans or animals and does not require ethical approval. The findings will be published in a scientific peer-reviewed journal.

Study status
Currently, we are undertaking the literature searches in the MEDLINE, Embase, Web of Science and Google Scholar databases and extracting the titles and abstracts from the articles which will be used in the initial screening process.

Data availability
No data are associated with this article.

expertise to confirm that it is of an acceptable scientific standard.
Use of statistics in the background of an abstract is discouraged due to the inability to cite accordingly.

Background:
Whereas the background has focused on water sufficiency (quantity), it is important for the authors to consider describing other key water indicators such as coverage, quality, cost and continuity that also contribute to the occurrence of water borne diseases.

○
The background may be strengthened by providing information on the various sources of water used in urban settings in Africa, both improved and unimproved.

○
Water statistics for 2015 are used yet more recent literature is available.

○
Whereas the study aim is on water sufficiency, other parameters as noted above have a direct contribution to water borne diseases. It is therefore not clear how these parameters are to be considered in the scoping review.

Protocol:
In the inclusion criteria, other water indicators noted above should be considered.

○
The exclusion criteria in principle should not be the opposite of the inclusion criteria but rather any predefined conditions that will be used omit any studies that would have met the inclusion criteria.

○
The exposure search terms as well as variables (Table 2) may also include the various water indicators beyond quantity.

○
The use of the term 'emerging diseases' may need to be justified.

○
The choice of selection of the 10 'diseases' also needs to be justified.

○
The outcome and independent variables to be considered in the review need to be described explicitly in the protocol.

Are sufficient details of the methods provided to allow replication by others? Partly
Are the datasets clearly presented in a useable and accessible format? Not applicable Competing Interests: No competing interests were disclosed.

Reviewer Expertise: Environmental and Public Health including water, sanitation and hygiene.
I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.
Author Response 03 Dec 2020 Mutono Nyamai, University of Nairobi, Nairobi, Kenya The background in the abstract is very wordy and needs to be more succint We have substantially shortened the 'background' section of the abstract in order to make it more succinct.

Use statistics in the background of an abstract is discouraged
We have removed statistics from the background section of the abstract.

It is important for the authors to consider describing other key water indicators such as coverage, quality, cost and continuity that also contribute to the occurrence of water borne diseases.
In order to limit the scope of this review we have specifically focused on insufficiency, as well as large African cities with population of more than 0.5 million AND selected waterborne diseases/symptoms/etiological agents AND water situation. In order to address other water indicators like cost, coverage, etc. and how they may be possible confounders, we will extract this information from the articles during the full text screening process as highlighted in Table 2, point 10.

The background may be strengthened by providing information on the various sources of water used in urban settings in Africa, both improved and unimproved.
We agree with the reviewer and have therefore edited line 60 so that it now includes the main sources of water used in urban areas in the African content. Whereas the study aim is on water sufficiency, other parameters as noted above have a direct contribution to water borne diseases. It is therefore not clear how these parameters are to be considered in the scoping review. These parameters have been included in Table 2 which highlights the variables that will be extracted from the articles during the full-text characterisation process. We will report how they contribute to waterborne diseases.

○
In the inclusion criteria, other water indicators noted above should be considered. Point 10, Table 2 lists the other indicators which will be extracted from the articles during the full-text screening process.

○
The exclusion criteria in principle should not be the opposite of the inclusion criteria but rather any predefined conditions that will be used omit any studies that would have met the inclusion criteria.
We agree with the reviewer and have therefore modified the exclusion criteria and only listed conditions that will omit studies that meet the inclusion criteria.

○
The exposure search terms as well as variables (Table 2) may also include the various water indicators beyond quantity. Point 10 of Table 2 has been updated to include other indicators of water supply (frequency of water supply, water rationing, cost, coverage, quality).

○
The use of the term 'emerging diseases' may need to be justified. Based on the advice from the first reviewer, this term has been omitted from the protocol.

○
The choice of selection of the 10 'diseases' also needs to be justified We have modified this to exclude '10 diseases'. The 'Definitions' section now classifies waterborne diseases, symptoms and etiological agents of diarrheal diseases, and includes justifications where appropriate.

○
The outcome and independent variables to be considered in the review need to be described explicitly in the protocol.
As this is a scoping review rather than a systematic review, we will focus on the methods of the included studies, as opposed to their findings or examination of effect sizes derived for outcomes variables. This will form the inputs to the cluster analysis, so we do not have an outcome and an independent variable given the nature of the type of review we will be conducting.
definition of terms, and assessment criteria that make the rationale and objectives of the review somewhat unclear. These are outlined below.
The study design is not entirely appropriate for the research question.
The protocol requires further development and elaboration, primarily in the outcome variables of interest i.e. the waterborne diseases. The central idea in the protocol seems to be linking water insufficiency to the 'ten emerging waterborne diseases'. There are several concerns here.
The terminology needs to be clarified and standardized: -Diarrhoea is a symptom of several diarrhoeal diseases, including cholera and typhoid.
Classifying it as a disease in itself is inaccurate.
-Gastroenteritis is a set of symptoms (including diarrhoea, vomiting, nausea) arising from intestinal infection, so again, classifying it as a disease in itself is inaccurate.
-Cryptosporidium and rotavirus are microorganisms / aetiological agents that cause diarrhoeal diseases, but are not diseases in themselves.
-Guinea worm is a parasite, not a disease in itself. Diarrhoea is not a typical symptom of dracunculiasis (guinea worm disease).

1.
What criteria were used to classify the 'emerging' diseases? This should be clearly described.
-The question whether the diseases / symptoms / infectious agents listed in the protocol are 'emerging' is highly debatable. The paper cited on some of the said emerging diseases is from 16 years ago -it is fair to say that the landscape has changed a lot since then.
-The GEMS 2013 study on aetiology of moderate-severe diarrhoea in low-income countries highlighted rotavirus, ETEC, Cryptosporidium and Shigella as the main pathogens of concern. There are likely more recent studies on this, which the authors are advised to look up.
-Guinea worm disease / dracunculiasis has been eradicated in most countries, with about 30 cases per year (sometimes less) now reported from 3 or 4 countries in Africa, so it is unclear how it could be termed an emerging disease.

2.
If the hypothesis is that water insufficiency leads to waterborne disease, how is the insufficiency determined? E.g. if City X has a published water rationing schedule in which water is supplied 12 hours per day for 7 days a week, vs City Y that intermittently supplies water 2 days a week -are the households in both cities water insufficient? I have no bright ideas on this, but it may be food for thought.

3.
Most diarrhoeal cases in literature are self-reported, thus aetiological info will likely be limited.
-Is there some weighting that would be assigned to studies based on the depth of information on the waterborne disease(s) provided? Studies reporting on the aetiology of the diarrhoeal disease may arguably carry more weight than those only reporting diarrhoea / gastroenteritis, which may not necessarily be waterborne.

Some definition of terms and scope is required for the review to be replicated by others.
It is unclear how 'water sufficiency' is defined within the scope of the review. E.g. for households living in slums, water insufficiency may arise from the slum not being connected to the network.
Are they within the scope of the review? Water deficiency, water scarcity and water insufficiency are used throughout the protocol. Do these terms all have the same meaning within the context of the review?
Other comments: What is the reference for the 53% of diarrhoeal cases being from Africa? The paper by Bain et al. cited as ref #13 actually reports 53% of water sources in Africa being faecally contaminated, but does not refer to this 53% as the diarrhoeal disease burden from the African region.
1. We added a key definitions section after the introduction which aims to clarify any ambiguity and standardise the terminology.

○
We also removed guinea worm from our search strategy in There are likely more recent studies on this, which the authors are advised to look up.
-Guinea worm disease / dracunculiasis has been eradicated in most countries, with about 30 cases per year (sometimes less) now reported from 3 or 4 countries in Africa, so it is unclear how it could be termed an emerging disease. We have removed the terminology "emerging diseases" as the reference was outdated and its use as a term was ambiguous, as highlighted by the reviewer. In turn, we have changed the emphasis of the paper to mainly focus on the diseases/symptoms/ etiological agents. As mentioned, we specifically define these in the 'Definitions' section in lines 89-99.

○
We looked at the GEMS 2013 and GEMS 2019 study which highlighted cryptosporidium, enteropathogenic Escherichia, shigella and rotavirus as the main pathogens of concern. We decided to focus on rotavirus and cryptosporidium in our scoping review. Globally, rotavirus was the leading etiology for diarrhea mortality among all ages [1] while cryptosporidium had the highest number of deaths with the pathogen present in children between 12-23 months [2]. ○ We agree with the reviewer that dracunculiasis is not an emerging disease as it has ○ been eradicated in most countries hence it has been removed from our inclusion criteria. If the hypothesis is that water insufficiency leads to waterborne disease, how is the insufficiency determined? E.g. if City X has a published water rationing schedule in which water is supplied 12 hours per day for 7 days a week, vs City Y that intermittently supplies water 2 days a week -are the households in both cities water insufficient? I have no bright ideas on this, but it may be food for thought.
In this scoping review, the term 'water insufficiency' relates to water quantity, where we used the World Health Organisation's (WHO) international benchmark of less than 50 litres of piped water per person per day [3], as highlighted in lines 49-51. We also included the definition of water insufficiency as used in this paper on line 102.

○
In order to address instances such as that which the reviewer gives as an example, we will record how water insufficiency was assessed in the papers and if there are enough articles that use methods for characterising water insufficiency, a categorical variable depicting the methods used will form an input within the cluster analysis . We highlight this in point 13 of In order to address the concern raised by the reviewer on majority of the studies on diarrheal cases being self reported, we will differentiate studies with self-reports from those with etiological characterisation of pathogens. The differences will then be inputted into the planned cluster analysis of study methods (see lines 173-178).

○
It is unclear how 'water sufficiency' is defined within the scope of the review. E.g. for households living in slums, water insufficiency may arise from the slum not being connected to the network. Are they within the scope of the review? Water deficiency, water scarcity and water insufficiency are used throughout the protocol. Do these terms all have the same meaning within the context of the review?
We have now specifically outlined how we define water insufficiency in the new 'Definitions' section (line 93). We will use the WHO definition of less than 50 litres per person per day as water insufficiency (line 102). One of the variables extracted from the included studies is the use of the WHO water insufficiency definition, as highlighted in Table 2 point 14. ○ Informal settlements are within the scope of this review. Water is often distributed in informal settlements through water kiosks or public taps [4]. Study setting (informal versus formal) would be included as a study characteristic within cluster analysis, to help examine if different methods and study designs are adopted in such settings.

○
We agree with the reviewer that water scarcity, water deficiency and water insufficiency were used throughout the protocol and to remove any ambiguity, we will focus on water insufficiency but include water scarcity and water deficiency in our search terms to ensure we do not miss out any articles on water insufficiency.