Introduction

Reliable access to sufficient water is critical for individual and community health, well-being, and economic development [1]. The importance of water security is highlighted by the United Nations’ recognition that the right to safe drinking water is fundamental for the fulfillment of all other human rights [2] and the inclusion of water-related targets in the Sustainable Development Goals (SDGs), including SDG 6.1, “universal and equitable access to safe and affordable drinking water for all”. Although governments have committed to these targets and invested in the development of improved water treatment, storage, and distribution infrastructure, the world is not on track to achieve universal water security by 2030 [3,4]. Substantial gaps in water accessibility, affordability, and adequacy for household use persist globally; these tend to be most acutely experienced by women, individuals living with disabilities, minoritized groups, and households in low-resource settings [5].

Progress toward safe water for all is predominantly tracked using directly observable measures, sometimes referred to as “resourced-based” metrics [6]. These measures include the number of pipes installed, the number of households with access to improved drinking water services (based on water source type and time to collection), the number of service beneficiaries, the cost and quality of available water, and water system functionality [7,8]. Such measures convey useful information but do not capture whether water is accessible when needed, meets users’ preferences, or is of sufficient quantity for critical domestic activities – that is, whether individuals are experiencing water security [6,9].

Having physical access to improved water services is not equivalent to being water secure. For instance, households classified as having a “safely managed drinking water service” – the highest service level in the schema developed by UNICEF and the WHO to track SDG 6.1 – sometimes report supply interruptions, high costs, and quality concerns that influence how they use water [10]. Many households also rely on multiple water sources and store water to fulfill domestic needs and manage shocks, but such coping strategies are not routinely measured or reported [11–13]. Further, most monitoring programs only collect information about a household’s primary drinking water source, yet many individuals in low- and middle-income settings use multiple sources for drinking and other uses (e.g., washing clothes) [14,15].

There is thus an acknowledged need to measure people’s lived experiences to understand how water insecurity impacts human well-being [16]. The Household (HWISE) [17] and Individual Water Insecurity Experiences (IWISE) Scales [18] measure issues with water access and use that are not captured by resource-based indicators (e.g., impacts on daily routines, meal preparation, and water intake). They also capture psycho-emotional aspects of water insecurity (e.g., worry, anger, shame). These 12-item scales ask people to report how frequently they (in the case of the IWISE Scale) or members of their household (HWISE Scale) experienced water-related problems (e.g., skipped washing hands due to water issues, went to sleep thirsty) during a given recall period. Responses are summed to produce water insecurity scores that can be compared across contexts and time.

Momentum to measure water insecurity is growing [19,20], partly because the information provided by experiential measures is germane for improving water governance and management [21]. Whereas standard water indicators primarily focus on infrastructure availability and performance, experiential measures are more dynamic and sensitive to diverse factors that contribute to water insecurity (e.g., pricing changes that reduce accessibility, shifting consumer preferences). Further, data can be disaggregated to identify disparities and better allocate resources.

Data generated from use of the WISE Scales have offered a comprehensive understanding of who is (un)able to benefit from the right to water (21). Inclusion of the IWISE Scale in the 2020–21 Gallup World Poll provided the first nationally representative estimates of water insecurity experiences. Across 31 countries, the prevalence of moderate-to-high water insecurity among adults ranged from 3.6% in China to 63.9% in Cameroon [22]. These data motivated the WHO/UNICEF’s Joint Monitoring Programme (JMP) to recommend inclusion of the abbreviated IWISE-4 Scale [23] in global monitoring efforts to track gender disparities in water access and use [24]. Additionally, the WISE Scales have revealed that greater water insecurity is associated with a higher risk of diverse health and well-being outcomes, including food insecurity [25,26], water-fetching injury [27], intra- and inter-household conflict [28], gender-based violence [29], HIV viral loads [30], and psychological distress [31]. Finally, the WISE Scales have been used by governments and non-governmental organizations to design projects that address the causes of water insecurity. In Australia, for instance, a community-based organization implemented the HWISE Scale and demonstrated that the prevalence of moderate-to-high water insecurity among Aboriginal Australians living in one town was much higher than the national average [32]. Documentation and communication of this disparity garnered increased political attention and financial investment to improve the local water distribution network and address high salinity issues [33,34].

Despite broad uptake of the WISE Scales, there is limited evidence about how water insecurity experiences vary in response to water service interventions (e.g., before and after installation of a water system). Information about a measure’s sensitivity is valuable for program evaluators, practitioners, and policymakers as they determine which tools to use to assess project impact [35]. Additionally, the added value of including the WISE Scales alongside standard resource-based indicators has not been well characterized. Existing indicators used in the water sector provide detailed insight into the presence and functionality of water infrastructure, but these data do not capture peoples’ experiences and may miss the extent to which intended beneficiaries’ lives are improved through infrastructure upgrades.

We therefore analyzed data from two interventions that sought to improve water access and use. They were led by different organizations (Helvetas and Oxfam GB) and implemented in socially and hydrologically diverse regions (rural Nepal and urban Sierra Leone); both used the HWISE Scale to assess intervention effectiveness. Our first aim was to evaluate whether the HWISE Scale could detect changes in water insecurity experiences following intensive, multilevel interventions that improved local water services. Our second aim was to examine whether project impact varied by sociodemographic characteristics to understand which, if any, groups were being “left behind”.

Methods

Key variables

In both settings, the accessibility and safety of available drinking water sources was summarized using a five-level drinking water service ladder, a standard resource-based indicator developed by the JMP to track progress toward SDG 6.1. Respondents provided information about their primary drinking water source and the roundtrip travel time required to collect water from that source, which was used to classify each household’s water service as “surface water”, “unimproved”, “limited”, “basic”, or “safely managed” using criteria established by the JMP [44]. Necessary water quality data for distinguishing between a basic and safely managed service were not collected, such that the highest service level that water sources could be classified as was “at-least basic”.

Household water insecurity, the main outcome of interest, was measured using the 12-item HWISE Scale, which has been validated for use in many low- and middle-income countries [17]. Respondents were asked to report how frequently in the prior four weeks they or anyone in their household experienced water-related challenges, such as changing foods prepared or being unable to wash their hands due to problems with water (Table 1). Response options were “never” (scored as 0), “rarely” (1), “sometimes” (2), “often” or “always” (3), or “don’t know”. Responses were summed to create an aggregate score (range: 0–36), with higher scores indicating greater water insecurity. If respondents skipped a question or answered with "don't know", aggregate water insecurity scores were not able to be calculated. Households were classified as having no-to-marginal (scores of 0–2), low (3–11), moderate (12–23), or high water insecurity (24–36) [45].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. The 12 household water insecurity experience items in the HWISE Scale [17].

https://doi.org/10.1371/journal.pwat.0000395.t001

We explored whether the impact of the intervention on water insecurity experiences varied by sociodemographic characteristics. Variables included self-reported respondent gender (man/woman), age (categorical), and whether households had a single mother or any members with a disability (yes/no, in Sierra Leone only). In Nepal, we also looked at variations in impact by district given regional differences in climate and topography.

Results

Participant characteristics

Approximately 60% of participants in Nepal were women, and most respondents across sampling waves were younger than 50 years of age (Table 2). At baseline, respondents resided in Surkhet (31.3%), Kalikot (26.5%), Jajarkot (25.3%), and Dailekh Districts (16.9%). In Sierra Leone, there were slightly more men than women in the comparison group (54.1% vs. 45.9%), but approximately equal percentages in the intervention group (48.9% vs. 51.1%). Respondent ages were similar across groups, with most aged 20–39 years.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Demographic characteristics, household water sources, and water insecurity experiences of program participants, by site and intervention status.

https://doi.org/10.1371/journal.pwat.0000395.t002

Water infrastructure and access

At baseline in Nepal, all households used drinking water sources that were improved per JMP criteria, but 39.2% were classified as having a limited water service level because their roundtrip water collection time was greater than 30 minutes. Further, although technically improved, the piped water used by 68.7% of sampled households at baseline came from an unauthorized connection to an unprotected source, often transported using inadequately installed piping. After the intervention, all sampled households used a managed, on-premises piped water system that was sourced from a protected spring and distributed through gravity-fed or mechanized-lift systems (i.e., had an at-least-basic drinking water service) (Table 2). The mean roundtrip time to households’ primary drinking water sources decreased from 28.2 minutes (sd: 16.9) to 2.4 minutes (sd: 1.7) after the intervention.

In Sierra Leone, a greater percentage of participants in the intervention group primarily used an improved drinking water source compared to those in the comparison group (65.2% vs. 54.4%) (Fig 1). Accordingly, a lower percentage of households in the intervention group had unimproved drinking water services (34.8% vs. 45.6%), and a higher percentage had at-least-basic drinking water services relative to those in the comparison group (48.2% vs. 33.0%). The mean roundtrip time to households’ primary drinking water sources was similar between groups: 18.6 minutes (sd: 17.5) in the comparison group and 16.2 minutes (sd: 27.6) in the intervention group.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Drinking water source and service level used by households in Sierra Leone, by intervention status.

Estimates calculated from multivariable models using propensity score matching. * p < 0.05, ** p < 0.01, *** p < 0.001.

https://doi.org/10.1371/journal.pwat.0000395.g001

Water insecurity experiences

In Nepal, the frequency of water insecurity experiences declined after the intervention. The prevalence of moderate water insecurity (HWISE Scale scores of 12–23) declined from 18.3% at baseline to 1.4% at endline (p < 0.001), and low water insecurity (HWISE Scale scores of 3–11) declined from 59.2% to 8.5% (p < 0.001). Those experiencing no-to-marginal water insecurity (HWISE Scale scores of 0–2) increased from 22.5% to 90.1% (p < 0.001, Fig 2A). Across all households, water insecurity scores declined by an average of 6.1 points (sd: 4.9).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Water insecurity category (A, C) and mean scores for each HWISE item (B, D) among households in Nepal (n

 = 71) and Sierra Leone (n = 981), by intervention status.

https://doi.org/10.1371/journal.pwat.0000395.g002

Although the majority (87.3%, 62/71) of participants experienced a decrease in water insecurity, there were a few exceptions (Fig 3). Scores did not change for six households, which had baseline HWISE Scale scores of zero (n = 5) or one (n = 1). Water insecurity scores increased by two, four, and eight points for three households. These households had low baseline HWISE Scale scores of zero (n = 2) and four (n = 1); the latter household was the only to shift from low to moderate water insecurity. In contrast, the mean baseline HWISE Scale score for the 62 Nepali households that experienced a decline in water insecurity was 7.2 (sd: 4.1), suggesting that these households originally experienced greater water insecurity and had greater potential to benefit from the intervention than those whose scores did not change or increased.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Change in HWISE Scale scores among 71 households in Nepal after a multilevel water intervention.

https://doi.org/10.1371/journal.pwat.0000395.g003

In Sierra Leone, the prevalence of moderate-to-high water insecurity was lower following intervention (24.8% in the intervention group vs. 66.3% in the comparison group, p < 0.001) and no-to-marginal water insecurity was higher (39.9% vs. 12.9%, p < 0.001) (Fig 2C). On average, the intervention caused an 8.0-point (sd: 1.1) decrease in water insecurity scores (p < 0.001).

Affirmation and frequency of specific water insecurity experiences

In addition to having lower aggregate scores after the intervention in Nepal, households affirmed fewer water insecurity experiences (Fig 4). For instance, 59.7% of households at baseline reported changing plans in the prior month due to problems with water compared to only 11.7% at endline. The most affirmed experiences at endline included being unable to wash clothes due to water problems (25.6%), worrying about not having enough water for all household needs (15.4%), and feeling angry about one’s household water situation (13.9%), although each of these were lower compared to baseline (65.4%, 71.8%, and 70.9%, respectively).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Reported frequency of water problems experienced in the prior month among households in Nepal, by intervention period.

https://doi.org/10.1371/journal.pwat.0000395.g004

The frequency and mean scores of experiences with water insecurity also declined across sampling periods in Nepal (Fig 4, Fig 2B). In contrast to the frequent affirmation of “often or always” at baseline, no experiences were reported as such at endline; most affirmed water insecurity experiences were identified as only occurring “rarely”. For instance, 22.8% of households at baseline reported “sometimes” or “often or always” feeling ashamed because of their water situation. At endline, no households reported water-related shame with such high frequency; only one participant reported “rarely” feeling ashamed and the rest reported “never” feeling ashamed.

In Sierra Leone, mean scores for each HWISE Scale item were lower among households in the intervention compared to comparison group (p < 0.001) (Fig 2D). For instance, the mean score for “going to sleep thirsty” was 0.9 among households in the comparison group compared to 0.4 for those in the intervention group. The greatest difference (0.8 points) was observed for worrying about water insufficiency: the mean score was 1.6 among households in the comparison group compared to 0.8 among those in the intervention group. Experiences with the highest mean scores in the intervention group included water supply interruptions (0.8), worrying about water insufficiency (0.8), being unable to wash clothes due to water problems (0.7), and feeling angry about one’s water situation (0.7).

Discussion

Experiential water insecurity measures have been used to understand the scope and impacts of water problems globally, but their added utility for program monitoring and evaluation in relation to directly observable water access and safety indicators has only begun to be documented. Based on data from two program impact evaluations conducted by Helvetas in Nepal and Oxfam GB in Sierra Leone, we found that experiential water insecurity metrics provided actionable information that complemented traditional resource-based water indicators.

Data from both case studies indicated that HWISE Scale scores and related indicators (e.g., percentage of households experiencing no-to-marginal water insecurity) are sensitive to interventions. In Nepal and Sierra Leone, mean water insecurity scores declined, the reported frequency of individual water insecurity experiences decreased, and the percentage of households experiencing moderate-to-high water insecurity declined in response to the respective interventions. For instance, in Nepal, the percentage of households experiencing moderate water insecurity declined from 18.3% at baseline to 1.4% at endline. Project impact did not vary by respondent gender or age (i.e., individual-level characteristics), which was anticipated given that the interventions were implemented at the community and household levels.

We know of only three other studies in the peer-reviewed literature that have reported how HWISE Scale scores vary in response to different water interventions; none found meaningful impacts on water insecurity scores or prevalence [48–50]. The null findings aligned with authors’ expectations, however, given cited shortcomings in intervention implementation and the limited dimensions of water insecurity that were addressed. For instance, water insecurity did not decrease in response to a water kiosk project that was implemented by the United States Millennium Challenge Corporation in Freetown, Sierra Leone; the authors posited that this was likely due to low kiosk functionality (many beneficiaries reported insufficient or irregular water availability) and limited adoption among studied households [48]. As such, findings from prior studies and those presented here indicate that water insecurity scores are sensitive to changes in the local water environment and appropriately reflect the effectiveness of interventions, satisfying key indicator criteria for monitoring WASH [35].

Observable, resource-based water indicators also improved in response to the multilevel WASH interventions implemented by Helvetas and Oxfam GB. Two years after program rollout in Nepal, all surveyed households were classified as using at-least-basic drinking water services (i.e., water from an improved source that has a roundtrip collection time of 30 or fewer minutes). Yet, 25.6% of households were limited in their ability to wash clothes due to water problems, 15.4% worried about water insufficiency, and 10.4% experienced water supply interruptions despite infrastructure improvements. These issues may have persisted because water meters to calculate payments for water service were installed in the months prior to the endline evaluation, leading families to limit domestic water consumption to reduce water bills. Importantly, reductions in water insecurity were heterogenous across districts, with greater impact observed in Jajarkot and Kalikot compared to Dailekh. The differing impacts by region suggest that the project may need to be tailored to better meet local conditions in some settings, with further follow-up needed to understand the reasons for these observed differences. The information generated by the HWISE Scale prompted Helvetas to partner with the University of Cambridge to develop a harmonized framework for visualizing HWISE Scale data as well as a process for using experiential data to inform practical recommendations for advancing WASH programming. Further, these data led local government officials to create municipal WASH service units.

In Sierra Leone, HWISE Scale data gave more comprehensive insights into how the project impacted people’s lives compared to traditional supply-side indicators. The intervention substantially increased the percentage of households that primarily used a basic drinking water service by 31.5% (from 33.0% to 48.2%). Meanwhile, the percentage of households experiencing moderate-to-high water insecurity decreased by 62.6% (from 66.3% to 24.8%). Similar declines in water insecurity experiences were observed among households with a single mother and those with a member with a disability, suggesting that program impact was equitable. Estimated project impact was likely higher using the HWISE Scale because it captures more nuanced information about the overall accessibility, sufficiency, and reliability of available water services for domestic uses, going beyond the physical characteristics and proximity of household water supplies. Furthermore, specific experiences measured in the HWISE Scale revealed important, but less tangible, ways by which the project improved people’s lives, including worrying less about water insufficiency and feeling less angry about one’s water situation. Such information is useful for determining which projects have the greatest impact on human well-being and could be adapted to other settings.

Taken together, these findings provide empirical evidence that including experiential water insecurity measures in WASH assessments can enhance program monitoring and evaluation. The demonstrated sensitivity of the HWISE Scale and actionable data it provided support the decision made by dozens of organizations to include the scale in their monitoring and evaluation plans. For example, USAID has recommended use of the abbreviated HWISE-4 Scale [51] in Feed the Future midline population-level assessments [52] and the World Bank is using the HWISE Scale to evaluate a $363 million infrastructure investment in Karnataka, India [21]. Further, these findings led two of the implementing organizations described in these case studies to include experiential measures as core project evaluation tools. Helvetas has made the HWISE Scale a "key performance indicator" in their 2025–2028 strategic framework, and charity: water has identified it as a required measure in its monitoring and evaluation framework, which is currently used by 58 WASH implementing organizations in 22 countries.

There remain unexplored opportunities for the use of experiential data in program evaluation. For example, they can be used to examine pathways by which projects improve human health (e.g., through increased ability to wash hands) and understand why an intervention did (not) have its intended impact on a distal outcome (e.g., infrastructure improved but households could not benefit from it due to supply interruptions). For this purpose, the IWISE Scale may be preferred over the HWISE Scale if water insecurity is heterogenous within the household, which has previously been observed [53,54]. For both the HWISE and IWISE Scales, we recommend examining aggregate water insecurity scores as well as the reported frequencies of specific experiences to understand which are most impacted by the intervention and contribute to changes in overall water insecurity. Further, future research should collect water quality data to determine how experiential and observational water data are related when the highest level of the JMP drinking water ladder, “safely managed services,” can be estimated.

The addition of HWISE or IWISE Scales into existing surveys should not be resource intensive as they are estimated to take only three minutes to implement, have been translated into many languages, and are available as modules in common data collection softwares (e.g., RedCap). If survey time is constrained or monitoring and evaluation budgets are limited, there are four-item versions of each tool that could be used, taking only one minute to implement [23,51]. Importantly, though, these abbreviated tools can only be used to measure the prevalence of moderate-to-high water insecurity because they do not provide sufficient variation to estimate the magnitude of water insecurity problems.

Strengths of this study include its comparative analysis of two distinct applications of the HWISE Scale in different geographic regions, demonstrating its applicability for use in diverse contexts, and the inclusion of resource-based indicators alongside experiential measures. Limitations include the small sample size in Nepal, which reduced the precision of model estimates. Further, the absence of a comparative group in Nepal means that observed changes in water insecurity may not be fully attributable to the intervention. Finally, in Sierra Leone, baseline data were not collected, although we sought to resolve this by asking participants to recall baseline conditions and using this information for propensity score matching.

Conclusions

These findings demonstrate that experiential water insecurity data provide valuable complementary insights to traditional observable indicators. By capturing people’s lived experiences with water access and use, the WISE Scales can identify challenges and disparities that resource-based indicators may overlook. Integrating experiential measures into program evaluations enables more tailored and effective interventions, ensuring that infrastructure improvements translate into meaningful benefits for communities. Additionally, these measures can help demonstrate the value of improving water security for human well-being and thereby garner greater project support. Beyond programmatic applications, these findings have broader implications for policy and global monitoring. Incorporating experiential data into water management strategies can enhance decision-making, optimize resource allocation, and promote more equitable and sustainable WASH interventions.

Supporting information

Acknowledgments

We appreciate the study staff who assisted with the project in Nepal, including Netra Bikram Thapa and colleagues from SAC Nepal, Lok Bahadur Bohara and colleagues from HRDC, Ganesh Bista and colleagues from Hurendec, and Mahendra Adhikari and colleagues from SOSEC. Additionally, we are grateful to the Oxfam staff in Sierra Leone who contributed to the evaluation process, especially Muhammad Naveed, Innocent Mutabaruka, and the enumerator team, as well as the community members who shared their experiences.