Coping with water insecurity in urban Ghana: patterns, determinants and consequences

Water insecurity, defined as the lack of access to safe, reliable and affordable water at sufficient levels (Jepson et al. 2017), undermines public health and economic development in many contexts. Although water insecurity is currently a prevalent issue globally, Sub-Saharan Africa (SSA) accounts for the highest number of people without access to safe water (United Nations Children's Fund and World Health Organization (UNICEF & WHO 2019)). Currently, about 60% of the 790 million people without access to safe drinking water services reside in SSA (UNICEF & WHO 2019). Unsafe drinking water, an important dimension of water insecurity, is a well-known route to the transmission of diarrheal diseases, contributing vastly to the global burden of diseases, particularly in resource-constrained areas like SSA (Prüss-Ustün et al. 2017; Adams et al. 2020).

After decades of efforts by governments and international development agencies, particularly under the Millennium Development Goal (MDG) 7c, which aims to ‘halve the proportion of people without sustainable access to safe drinking water by 2015’ (United Nations World Water Assessment Programme 2015) and its successor, the Sustainable Development Goal (SDG) 6.1, which aims ‘to achieve universal and equitable access to safe and affordable drinking water for all’ (United Nations Children Fund & World Health Organisation 2017; UNICEF & WHO 2019), urban water coverage of piped water access in many SSA countries remains a challenge. For instance, the proportion of people with piped water access in SSA declined by 50% between 2000 and 2017 (UNICEF/WHO 2019). This is mostly due to population growth and inadequate infrastructural investments by governments to meet the water needs of the growing population.

Apart from the global and regional disparities in water access, there are access gaps between rich and poor households, particularly in SSA. For instance, Armah et al. (2018) report that wealthy urban households are 329% more likely to access improved water services compared with poor urban households in Ghana. The burden of these inequalities is disproportionately felt by women and girls who, in addition to other social reproductive roles, are primarily responsible for collecting water at the household level (Bisung & Elliott 2014). Reports indicate that over a quarter of households in SSA spend at least half an hour on a single trip to collect water (WHO/UNICEF 2015), and women and girls bear water hauling as a primary responsibility in seven out of every ten households without piped water on their premises (UNICEF/WHO 2015).

Unlike most SSA countries, Ghana achieved its MDG 7c 5 years before the 2015 deadline (Alagidede & Alagidede 2016). However, some scholars argue that given the challenges associated with measuring access to water service, there is a high likelihood that this achievement is overstated (Martinez-Santos 2017). This is likely the case given that a staggering proportion (over 40%) of the urban population remains without access to clean and quality drinking water (GSS 2019). Even for houses with access, the water supply is sporadic in many urban neighborhoods, thereby compelling households to resort to a patchwork of mostly unsafe sources for their daily water supply.

In the face of water insecurity, households resort to alternative water sources for daily survival (Majuru et al. 2016). Two reviews recently synthesized evidence on water coping strategies – i.e., responses to water insecurities – worldwide. The first review by Venkataramanan et al. (2020) classified individual and household-level water insecurity coping strategies into four domains including access, use, quality and/or reliability of water. Like Majuru et al. (2016), the review found that most households engage in several activities to cope with water insecurity. These activities include sharing water, purchasing water, changing household routines and relocating to other areas. They further identified the benefits of coping strategies in some instances to include strengthened social bonds and reduced risk of water-borne diseases. Water sharing can also result in households’ reliance on unsafe water sources (Stoler et al. 2019). The second review by Achore et al. (2020) examined how households cope with water insecurity and the socioeconomic consequences of water insecurity coping strategies, focusing on only qualitative studies. The second review classified water insecurity coping strategies into soft and hard coping strategies (Achore et al. 2020). These two review studies identified two important research gaps and future research directions that form the foundation of the current study. First, the literature on water insecurity coping strategies mainly originates from rural settings in three countries: India, Kenya and Bangladesh (Venkataramanan et al. 2020). It is imperative to explore how poor urban households in other settings like Ghana respond to water insecurity, recognizing that knowledge of water insecurity coping strategies is situated and context-dependent. Second, the reviews emphasized the importance of identifying the predictors and consequences of water insecurity coping strategies to fully capture the complex linkages between water insecurity and other social characteristics such as poverty and gender (Achore et al. 2020; Venkataramanan et al. 2020). A better understanding of determinants of water insecurity-related coping strategies could help practitioners identify vulnerable groups and provide targeted interventions that seek to build or strengthen coping strategies in the short term.

Ghana is an ideal case to examine water insecurity coping strategies among the urban poor for many reasons. First, while Ghana achieved the MDG water target ahead of schedule, significant urban inequalities remain. For example, 64.1% of the urban poor population use water packaged in plastic sachets as their drinking water source (Wardrop et al. 2017). Second, previous research showed that the aggregate data on Ghana's MDG 7c achievement masks regional, district and community differences (Ainuson 2010). Third, Ghana is experiencing rapid economic growth (Owusu-Manu et al. 2019) and unprecedented rural–urban migration (Yeboah 2021). An increase in rural–urban migration is reportedly associated with increased inequalities in municipal services that disproportionately affect informal settlements and the urban poor (Awumbila & Owusu 2014).

A state-owned limited liability company, Ghana Water Company Limited (GWCL), is responsible for water supply to urban areas across Ghana. According to a study by the PURC (2015), most poor people do not receive service directly from the GWCL, but rather from informal service providers and sources made available by Non-Governmental Organizations (NGOs). As such, in 2007, the GWCL in accordance with the recommendations of the Millennium Development Goal (MDG) developed a national water policy, which is designed to provide a framework for the sustainable development of Ghana's water resources (National Water Policy 2008). The overall goal of this project is to improve the health and standard of living of all individuals, including improvement in access to water and sanitation. Although this policy has been implemented, more than 4 million urban households in Ghana lack access to safe drinking water, and the poor in rural and urban areas are disproportionately affected.

The cross-sectional data were collected in February–March 2020. To estimate the sample size, the study used a 5% error margin and a 95% confidence interval. This yielded a sample size of 385 and 362 households for Accra and Tamale, respectively. We expanded the sample size to 1,200 people, evenly distributed throughout the two research areas. A total of 1,192 adult active water collectors within the household (483 men [260 in Accra and 223 in Tamale] and 709 women [367 in Accra and 342 in Tamale]) took part in the survey, representing a response rate of 99%. We adopted a systematic sampling technique and used the EPI random path method to randomly select the first house and direction to survey (Milligan et al. 2004). To ensure that all households in the selected neighborhoods had a fair chance of being selected, enumerators were required to survey the first household in every fourth and third house in Accra and Tamale, respectively. The sample size in each city was determined using a probability proportional to population size (PPS) sampling (Skinner & Skinner 2016).

Households in this context are defined as a person or group of related and unrelated persons who live together in the same dwelling unit(s), who acknowledge one adult member as head of the household and share the same housekeeping arrangements (GSS 2019). The survey was mainly administered to an adult household member regarded as a water collector and a primary maker of decisions around water within the household. Women were overrepresented in the sample because they are primarily responsible for water collection in Ghana and disproportionately bear the social and economic burden of water insecurity (UNICEF/WHO 2015). The data collection process took place over 4 weeks. A team of six enumerators from the University of Ghana in Accra and the University of Development Studies in Tamale, fluent in the local languages (Twi, Ga, Dagbani and Hausa), were recruited and trained in data collection, including skills in culturally appropriate ways of asking questions. The survey was administered face-to-face with all participants.

Questionnaires were first pre-tested with 20 participants in each study site to ensure the questions were understandable, relevant and captured the various constructs of interest. The questions captured general patterns of water insecurity, coping resources, total number and sex of household members, household assets and essential community resources. The survey was administered to a random sample of households from selected neighborhoods with water security challenges. These neighborhoods included Chokor, Jamestown and KorleGono in Accra, and Lamashegu, Kukuo and Vitting in Tamale. After the day's work, the data enumerators met with the team lead for debriefing. This created a space for them to highlight some of the challenges they faced in the field; solutions and guidance were offered, smoothing the data collection process. Ethics review and clearance were obtained from the Queen's University General Research Ethics Board (GREB) (GREB Ref #: GSKHS-340-20; TRAQ # 6028559). Verbal informed consent was requested from all study participants before the start of each interview.

Water insecurity: Water insecurity was measured using the Household Water Insecurity Experiences (HWISE) Scale (Young et al. 2019). The cross-culturally validated HWISE scale consists of 12 items meant to elicit water insecurity experiences (Young et al. 2019). Using a Likert scale ranging from 1 = Never to 5 = Always, participants responded to a series of questions that focus on being worried or anxious about the lack of water, safety and reliability of their water source, and forgoing or economizing water use because of inadequacy (refer to Supplementary material A).

Wealth: The wealth index was measured using a modified version of the DHS wealth measure (Rutstein & Johnson 2004). The wealth items were modified to account for assets in the Ghanaian context. Participants were asked to indicate if they owned any of 20 predetermined assets (e.g., cattle, houses, lands, motorized vehicles, bicycles, fridge, TV, animals, etc.). A weighted score for participating households was calculated and weighted by the total number of people in a household. We further grouped these scores into the 1st, 2nd, 3rd and 4th quartiles.

Primary water source: Participants were asked to indicate their primary source of drinking water (e.g., piped water in the building, shared piped water close to the house, public boreholes, protected dug well, protected spring, water from water vendors, water tankers owned by a private individual, rainwater, bottled water, rivers, streams or lakes, unprotected dug well, unprotected spring, spring). We then categorized the indicated sources into two categories: improved (e.g., piped water in the building, shared piped water close to the house, public boreholes, protected dug well, protected spring) and unimproved (water from water vendors, water tankers owned by a private individual, rainwater, bottled water, rivers, streams or lakes, unprotected dug well, unprotected spring) source based on the JMP water ladder, a ladder used to benchmark and compare water services across countries (WHO/UNICEF 2019).

Seasonality: We asked participants to indicate their primary water source during the dry and wet seasons. Ghana has two seasons, the wet and dry seasons, with the dry season spanning from December to February in the South and November to March in the Northern part of the country. We decided to add seasonality because we believe the type of season heavily influences water insecurity coping strategies.

Other independent variables: In addition to the above variables, we collected data on other independent variables that we presumed might impact water insecurity coping strategies. These include the number of people in a household, occupation, marital status, gender and education level. These variables have been shown to predict water insecurity in many contexts (Adeniji-Oloukoi et al. 2013; Majuru et al. 2016; Abubakar 2018; Kangmennaang et al. 2020). The number of men and women 18 years and above in a household was included because we assumed households with many adults would be able to use a diverse set of resources to cope.

Coping strategies: Coping strategies were identified by synthesizing qualitative studies (see Achore et al. 2020) and other reviews (Majuru et al. 2016; Venkataramanan et al. 2020). Before data collection, a list of coping strategies was shared with local women's group leaders for feedback and context validation. The women were asked to identify the most common and relevant coping strategies in their neighborhoods, which were then synthesized and used to revise our household questionnaire. To capture coping strategies, survey participants were asked the following question: if water from your primary source becomes consistently unavailable in a month, how often would you take extra measures to (insert coping strategy)? The idea was to capture what people use if their regular water source becomes unavailable. Response options included the nine coping strategies and their frequency of use (never, rarely, sometimes, often and always). As described in Section 2.4, we grouped coping strategies into three categories using principal component analysis (PCA).

Statistical Package for Social Sciences (SPSS) software (version 24.0) was used in analyzing the data. We created water coping strategies index using a PCA (Elhaik 2022). PCA is a statistical technique used to transform large sets of variables into categories based on some underlying relationship among the various items (Wold et al. 1987; Abdi & Williams 2010). We employed a PCA after running various tests to confirm its suitability for analyzing the data. We performed a Kaiser–Meyer–Olkin test (KMO) measure of the sampling adequacy and obtained a KMO value of 0.701, exceeding Kaiser's (1974) recommended value (0.6). We also found a significant Bartlett's test of sphericity (p < 0.05), meaning that our correlation matrix is factorable. We then created a coping strategy score for each household using principal component extraction with ProMax rotation. We retained coping strategy factors (components) using two criteria: components with an eigenvalue greater than 1.0 and a scree plot. The PCA extracted two components with an eigenvalue greater than 1.0. The scree plot also showed that the line started to break after the first two components; details of these terms (eigenvalue and scree plot) have been described elsewhere (Wold et al. 1987; Abdi & Williams 2010). The two components accounted for 67.7% of the data variance, with component 1 accounting for 57% of the variance. Lastly, we used a minimum component loading of 0.4 to determine if a variable contributes meaningfully to a component.

The Promax rotation indicated a high loading on five variables on component 1 and three variables on component 2 (Table 1). A variable (borrowing water from neighbors) was loaded on both components 1 and 2 and was removed to improve the reliability and rigour of the coping indices (Boateng et al. 2018; Young et al. 2019). Dropping this variable resulted in an overall increase in Cronbach's alpha from 0.69 to 0.87. Component 1 indicated a high loading on the following variables: buying water from private vendors, conserving water, storing water, purifying unsafe water and harvesting rainwater. Component 2 loaded high on the following variables: making an illegal connection to a public water pipeline, digging a borehole or well and acquiring water from distant sources. We named components 1 and 2 as behavioral and physical coping strategies, respectively. While physical coping strategies require physical efforts and hardware usage, behavioral coping strategies require behavioral adaptations. Adding borrowing water to the behavioral coping strategies resulted in a reduction in the Cronbach's alpha from 0.86 to 0.62. Adding it to the physical coping strategies reduced the Cronbach's alpha from 0.60 to 0.47, indicating that the removal of this variable improved each scale's reliability.

Component scores for behavioral and physical coping strategies were generated for each household, with positive (+) scores indicating the use of a coping strategy and negative (−) scores denoting no coping strategy. Based on this, the outcome variable was categorized into the following groups: no coping mechanism (when a household scores negative on both behavioral and physical coping strategies), only behavioral coping strategies (when a household scores positive on behavioral coping strategies and negative on physical coping strategies), only physical coping strategies (when a household scores positive on physical coping strategy and negative on behavioral coping strategy) and both behavioral and physical coping mechanisms (when household scores positive on both behavioral and physical coping strategies).

Since the outcome variable had more than two categories, a multinomial logistic regression was used in performing our analysis with ‘no coping strategies’ as the reference outcome. We performed both adjusted and unadjusted multinomial regression to identify predictors of coping strategies (see Table 3 and Supplementary Table 4). The selection of the predictor variables was influenced by relevant literature on water insecurity coping mechanisms. We used an odds ratio (OR) with 95% CI to interpret results from the multinomial regression analysis, which is expressed as the odds of a particular outcome given the presence of one or more predictors/determinants.

This study examined predictors of coping strategies employed by households in the face of water insecurity in urban Ghana. Respondents were asked to indicate coping strategies they employed when water from their primary source becomes unavailable. The study shows that 43.7% of households use sachet water as their primary drinking water source (see Figure 2). However, when water from the primary source becomes unavailable, participants resort to various activities and strategies to cope with the situation (see Figure 2).

Our findings indicate that most urban households in Ghana face water insecurity challenges. As a result, they often employ behavioral, physical and a mix of behavioral and physical coping strategies to mitigate the situation. This finding is in line with those of previous studies (Adeniji-Oloukoi et al. 2013; Abubakar 2018), including two scoping reviews (Majuru et al. 2016; Venkataramanan et al. 2020). Like our findings, these studies found that in the face of water insecurity, households resort to either a single or diverse portfolio of coping strategies, with some households, according to Venkataramanan et al. (2020), taking drastic actions such as completely relocating their homes. Residents would ideally not employ many of these coping strategies (e.g., illegal connections) if safe and reliable drinking water is readily available. Although few (see Figure 2), it is surprising to see people report illegal connections as a coping strategy knowing that it is an illegal activity. Reporting this is probably a way of protesting the inequalities in the system that forces them to use illegal connections in the first place. Conversely, the number of illegal connections might be underreported because of fear or social desirability bias. Thus, it is vital to engage with the government and other stakeholders (e.g., Community leaders, Chiefs and NGOs) to tease out the broader socio-political and economic factors and water governance systems that systemically exclude poor areas and informal settlements from service delivery.

Seasonality greatly influences access to water in Ghana. The dry season is accompanied by a decline in rainfall, surface and groundwater, resulting in a decreased water supply. Interestingly, the results show that respondents were more likely to use coping strategies in the dry season than in the wet season. This is perhaps because households do not have a variety of surface water sources, which are mostly unimproved, at their disposal during the dry season as they do in the wet season. In giving reasons for similar findings in their study, Adeniji-Oloukoi et al. (2013) argue that households always opt for reliable water sources to cope with shortages in supply and are therefore likely to use a diverse set of sources in the wet season. Our finding also corroborates those of other scholars (Kaptué et al. 2013; Pearson et al. 2016; Kelly et al. 2018; Dongzagla et al. 2021). Pearson et al. (2016), for instance, in examining the relationship between seasonality and change in primary water sources in Uganda and Tanzania, found that households mostly use boreholes in the dry season to cope with water insecurity. Similarly, in examining the relationship between seasonality and drinking water contamination in two Ghanaian municipalities, Dongzagla et al. (2021) found that water contamination was higher in the wet season.

While men generally use physical coping strategies, women opted for behavioral coping strategies. Women perhaps use behavioral coping strategies because of the financial commitments of physical coping strategies. This is likely attributable to decision-making processes in most Ghanaian households, where women make less input into water-related decisions at the household and community levels (Dickin et al. 2021). To undertake activities that require a financial commitment, women mostly consult their male partners within the household. This lack of financial independence perhaps results in women using behavioral coping strategies, which do not require much financial commitment, if any. Thus, water practitioners should continue to create empowerment programs to promote gender equality.

The results unexpectedly show that level of education is not a significant predictor of the use of a coping strategy. This is in contrast with other studies (Pattanayak et al. 2005; Katuwal & Bohara 2011; Majuru et al. 2016; Abubakar 2018) that identified the level of education as a determinant of the use of a coping strategy. This is surprising because attaining higher education is reportedly linked to salaried employment and high income resulting in better access to improved water services or the use of a mix of behavioral and physical coping strategies. On the other hand, it is less surprising when the current levels of unemployment (8.4%) in Ghana are considered (GSS 2019). The lack of employment opportunities means that individuals, albeit with high educational levels, do not have access to well-paying jobs. This hinders their ability to afford water from alternative sources that require a financial commitment when their regular source becomes unavailable. Of note, the study was conducted in resource-constrained settings in Accra and Tamale.

Residents in the relatively poor category were significantly less likely to adopt a mix of behavioral and physical coping strategies. In addition to using behavioral coping strategies, relatively wealthy individuals can easily adopt physical coping strategies that might require some form of financial commitment. This finding corroborates those of other studies, which indicate that high-income households are more likely to construct boreholes to make up for unreliable piped water supply (Adeniji-Oloukoi et al. 2013; Abubakar 2018). Adeniji-Oloukoi et al. (2013) found that although most poor households desire to adopt a strategy like digging wells and boreholes that can last several seasons, the lack of financial resources prevents them from doing so. Spending money on boreholes/well construction will reduce spending on other household essentials like food. Thus, the unwealthy shy away from these coping strategies. Conversely, Angko (2013) found a high number of dug wells among low-income households in the Boli in the Wa municipality of Ghana. The reason for this contrasting finding could be attributed to the social and study locations. The Boli area is rural, where there is likely to be a high sense of community; individuals collectively come together to help their neighbors dig wells. This is not the case in urban areas. For instance, Adeniji-Oloukoi et al. (2013) indicated that when networking for water development was suggested to their urban study participants, they refuted the idea, arguing that water is a social service. Thus, the onus lies on the government to provide them with water. Other coping strategies such as storing and fetching water from long distances also require some financial commitments. For example, storing water might require buying storage tanks, installation fees and perhaps constant electricity flow to pump the water when needed.

The most prevalent drinking water source among the participants was sachet water. Like most SSA countries, the demand for water services in urban Ghana surpasses the supply. The lack of access to public water services compels the underserved and unserved urban residents to use alternative water sources, including packaged (sachet) water (Williams et al. 2015; Smiley & Stoler 2020), to meet their daily water needs. This finding mirrors those of Dzodzomenyo et al. (n.d.) and Stoler et al. (2012b, 2015). Stoler et al. (2012a), for instance, in their study found that the uptake in water packaged in sachets is mostly a reflection of households’ response to the government's inability to supply them with quality drinking water. The use of sachet water as a primary water source has two implications. First, sachet water is marred with several controversies and debates regarding its quality due to the unhygienic conditions under which the sachet water is produced and distributed (Dzodzomenyo et al. n.d.; Stoler et al. 2012a). Second, apart from quality, the use of sachet water means households, particularly poor households, expend their limited financial resources on buying water. This can widen the existing inequalities between the rich and poor and exert financial stress on poor households.

In addition, vended water is managed and supplied by private enterprises with their power asymmetries. They determine their products’ prices and thus randomly change their prices, draining the income of the poor urban households who depend on their services to meet their water needs (Stoler et al. 2012a; Amankwaa et al. 2014). In addition to widening the inequalities gap, some scholars have argued that the constant change in water prices and water access negotiation is a significant source of distress for urban poor households (Tutu & Stoler 2016; Adams 2018). Given the widespread use of sachet water, the government and policymakers should look for alternative ways of protecting consumers’ health and well-being while considering the critical role private water vendors play in meeting the needs of households that lack access to safe and reliable drinking water. One way of doing this is to educate water vendors on properly disinfecting their storage tanks and producing water under hygienic conditions.

This study has some limitations worth highlighting. First, the study employed a cross-sectional design, which restricts our ability to establish causation between the lack of water and coping strategies, preventing us from making causal inferences. Second, in addition to being a cross-sectional study, we surveyed neighborhoods in Accra and Tamale that faced water insecurity. Our study sample's socioeconomic and demographic makeup is thus not representative of other areas of Ghana and even Accra and Tamale. Therefore, interpretation of the results should be limited to water-insecure neighborhoods or other neighborhoods with similar demographic and socioeconomic profiles as our study areas. Future studies can conduct a comparative analysis by collecting data in selected water-secure and water-insecure neighborhoods in Ghana. Future studies can also regress the predictor variables on the individual outcome variables to evaluate their impact. Third, the data used for this analysis is self-reported and thus subject to recall bias. Also, we could not explore the predictive capacity of income or poverty on coping strategies due to the absence of the income variable. That said, we used wealth as a proxy for income/poverty in our analysis (Arias & De Vos 1996; Poirier et al. 2020). Wealth was calculated using households’ durable goods. However, this methodology is critiqued in the development literature (McKenzie 2005; Poirier et al. 2020). Also, given that seasonality plays a critical role in water insecurity coping strategies, the season within which data were collected should be considered when looking at the results. Although they might be using varieties of coping strategies, households likely reported only coping strategies used during the particular season (dry season) in which the data were collected. We primarily targeted adult household members who were active water collectors within the household. Thus, future studies should target children as they are active water collectors for most households in the study context. Finally, we acknowledge that the two categories of coping strategies are not mutually exclusive.

Given that most urban households in Ghana use coping strategies to mitigate the unreliable water supply, it is imperative to understand the range of socio-ecological and economic factors that influence the use of these coping strategies so that interventions can be better targeted. The projected increase in the urban population in Ghana and other SSA countries is likely to pose additional challenges to the municipal water systems, consequently exacerbating water insecurity issues. Thus, governments should implement a long-term, reliable water solution that takes gender, households’ socioeconomic status, and seasonality into account to meet Ghanaians’ immediate and future water needs. Our evidence also points out that most households in less wealthy neighborhoods are disproportionately affected in terms of water insecurity and lack the financial resources to implement short-term coping strategies. This is both a social and environmental justice issue and demonstrates that we cannot achieve SDG 6 if we do not tackle social justice and environmental equity in the water sector. Thus, it is imperative that governments and other stakeholders such as the WHO and UNICEF take these recommendations into account in their program planning and policy interventions.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.