Improved health-related water management could prevent one tenth of the current global disease burden and investments in improved access to safe drinking water may realize at least ten fold economic returns [1]. Yet lack of access to safe water remains the third most significant risk factor for poor health in developing countries [2]. The health impact of various interventions to improve access to safe water has been extensively reviewed, but primarily by focusing on rates of acute infectious diarrhoeal illness to evaluate outcome [2–5]. It is likely that more health impacts of sub-optimal water supply are frequently overlooked or underestimated, because effects other than acute diarrhoeal illness are not usually considered [1].

Understanding the health impacts of sub-optimal water access more broadly is crucial for appropriate and sustainable water resource development. The benefits of investing in interventions to improve safe water access may be underestimated and, therefore, such interventions not prioritised, if the broader health impacts are not comprehensively evaluated. For example, many people must still collect and physically carry water from a source distant to their home, which may have important health consequences for those who perform the task [6].

Water filled containers are often carried on the head, however, transportation with wheel barrows, animal drawn carts or by rolling filled containers has also been observed [7, 8]. These methods obviously create physical demands on the body and the potential for adverse physical stress from regularly carrying loads of water has been recognised [6, 9, 10]. Physical loading of the body within an individual's capacity for adaptive responses may lead to tissue strengthening, however, frequent loading beyond capacity for adaptation or repair may lead to injury through fatigue failure, accumulation of fatigue damage [11] or early degenerative changes in bone and soft tissues [12].

Assumptions have been made that water carrying is detrimental to health and associated with musculoskeletal disorders, such as spinal pain or other joint problems [8, 13, 14]. Such assumptions are supported by strong evidence that the physical demands of work such as handling heavy materials, bending, twisting and lifting, are risk factors for onset of simple low back pain [11, 15] and other musculoskeletal disorders [16, 17]. In particular carrying heavy loads on the head by professional porters has been documented to cause catastrophic injury, such as spinal fracture, dislocation or death [18] and has also been associated with early onset of degenerative changes in the cervical spine [12, 17, 19, 20].

Although head loading due to occupational activities has been associated with degenerative changes in the cervical spine, the relationship between symptoms such as neck pain and activities which require head loading is not clear. Despite a much higher prevalence of upper cervical osteoarthritis in porters (91.6% in male porters compared to 6.8% in the control group) Badve et al. [17] stated that an association between symptoms and radiological changes was not found. Similarly, a recent systematic review did not find evidence that cervical disc degeneration is a risk factor for neck pain and reported variable evidence for a relationship between radiographic signs of degeneration and neck pain prevalence [21]. A recent study found that degenerative changes observed in cervical plain films were poorly related to the severity of symptoms or neck dysfunction in women with chronic pain and working in sedentary occupations [22].

However, very few studies have specifically investigated water carrying as it is performed by women and children in developing countries and used appropriate methodologies to investigate its association with health generally or musculoskeletal disorders specifically [10, 23, 24]. Most studies investigating the health impact of physical loading are of male adult workers [17, 25–27] or are situated in high income countries [28] such that existing evidence may not be applicable to women and children who typically collect and carry water for domestic use [29]. Importantly, women and children have reduced injury tolerance for physical loading through the cervical spine compared to men [30–36] and in rural areas may be particularly vulnerable to physical injury due to high levels of poverty, poor health and chronic disease [37–42]. Therefore, it is not clear whether regularly carrying containers of water for domestic use leads to detrimental effects such as accelerated degenerative changes in the spine and other joints and whether or not any such effects are symptomatic and impact on health related quality of life.

Two recent reports indicate that some people may experience high rates of perceived exertion and pain sufficient to limit their capacity to carry water containers [7, 43]. Reduced capacity of women or children to collect water due to pain or fatigue may have serious implications for the health of their families. However, water carrying is a physical activity which might also lead to beneficial health effects in some individuals. In researching the health impacts of water carrying, it is important to consider health impacts broadly [7] and recognise the limitations of applying existing evidence to this special activity and population group. Researcher assumptions about risk factors and health effects may introduce bias into research methodology in terms of determining the questions asked and outcomes measured and consequently how study participants report the health impacts of water carrying. For example, the use of leading questions or outcome measures which assume an association with symptoms such as pain might influence participant responses and their description of the health effects of water carrying.

As there is a lack of empirical data specifically related to water carrying, the aims of this study were to better understand how water carrying is performed and experienced by people who perform the task, identify health risk factors potentially related to carrying water and gain insight into the possible health effects of the task. The following research questions are addressed in this report

A mixed methods approach was taken, utilising both quantitative and qualitative data to better understand domestic water carrying as it is performed and experienced by adults and children in Limpopo Province, South Africa. Ergonomic principles were used to develop the approach to quantitative data collection. An 'ergonomic' evaluation of work incorporates assessment of a broad range of potential risk factors related to the environment, organisation of work, the nature of the task or the individual [44].

Qualitative enquiry in this study was influenced by the principles of phenomenology as described by Creswell [45] and used to explore the lived experience of water carrying. Individuals with direct experience of water carrying will have unique understanding of the task and can provide insight into how it might impact upon their own health and functioning. As the health effects of water carrying are unknown and might be experienced and interpreted variably by different individuals, such insights can indicate the domains of health which are relevant to people who perform water carrying and, therefore, important to evaluate for a potential association with the activity.

This report will focus on the analysis of the quantitative data, combined with some specific findings from content analysis of the qualitative data generated during individual semi-structured interviews. This approach was used to evaluate the relationship between pain, which was a specific health outcome revealed to be of concern to many of the study participants, and potential risk factors observed to occur during water carrying. More extensive and detailed analysis of the qualitative data will be reported separately [7] and will incorporate the findings of additional data generated from 'natural informal group interviews' which were conducted according to the methods described by Green and Thorogood [46].

The prevalence of back pain among this mixed group of children and adults, at 38% was higher than that reported in two South African studies included in a recent review [27], and which reported point prevalence for low back pain of 14% for children and 25% for adults. Importantly, we may have underestimated the prevalence of pain in the study sample due to our data collection methods. In keeping with a phenomenological approach, open questions about the health effects of water carrying were asked during semi-structured interviews to capture the potentially varied impacts which people who carry water might perceive the task to have. Participants complaining of pain were identified from their responses to the open interview questions and therefore volunteered pain as a health effect without direct prompting or suggestion that it would be linked to water carrying. In most studies investigating pain, structured outcome measures which directly ask about pain intensity or quality are used. Such direct questions may encourage pain reporting which might not be recalled or mentioned in response to more open interview questions.

A recent Danish study found that women are more likely to report spinal pain than men [51], therefore it is possible that the high proportion of women in this study, due to their role as water carriers, may explain the high prevalence of self reported pain. However, reasons for a potential association between pain reporting and gender may be different in this study population and are as yet unknown. It may be relevant that women in sub-Saharan Africa are disproportionately affected by HIV disease. HIV is associated with rheumatological conditions such as reactive arthritis [37], osteoporosis, fragility fractures and impaired fracture healing [42] and a high prevalence of pain, linked with significant psychological and functional morbidity [39]. How and why pain is reported will vary in different cultural and social contexts [15] and the relationships between physical, psychological and social influences on pain reporting amongst Venda women have not been determined. Future research should investigate the association between bio-psychosocial factors, co-morbidity and pain reporting amongst women who carry heavy water loads as well as pain impact, through participant ratings of pain intensity, duration, frequency and pain-related disability.

This study supports Cleaver's [52] claim that males more commonly use methods of water carrying which utilise equipment. However, in this study, two boys who used a steep and rocky pathway, which made use of any transportation equipment such as a wheelbarrow impossible, were also observed to carry water containers on their heads. Therefore, environmental factors such as path quality and incline gradient may also determine which carrying methods are used. Generally, this study suggests that women and children carry water and women are more likely to carry water in a way (head loading) which will focus and transmit forces through the cervical spine.

Others have reported load-weight as a percentage of body weight and tested for its association with outcomes such as self reported pain [53]. In the United States Moore et al. [53] concluded that backpack weights for children should remain below 10% of body weight and a recent review reported recommendations from several authors that back pack weights for children should be limited to 10-15% of body weight or less [54]. However conclusions drawn from studies set in high income countries may not be generalisable to poorer rural communities, where factors such as childhood health, development and general levels and types of physical activity are likely to differ in significant ways. Nevertheless, in comparison, the high container weights in proportion to body weight carried by women and children in this study seem a potential risk factor for self reported pain. A recent South African study found that a large majority of children who collected water and reported that their health had worsened complained of neck or back pain [23].

Compression forces generated purely by the weight of water carried through head loading in this study may be unlikely to exceed tissue tolerances described in cadaveric studies [30, 36, 55, 56], if applied briefly during a single loading occasion. Older children tended to carry heavier loads than the younger children in the study and their tolerance limits may be closer to those of adults. However, injury tolerance limits based on cadaver studies can only provide estimates of living tissue strength [31] which may be reduced by factors such as malnutrition or chronic illness [57], both of which are highly prevalent in poor rural areas [58] such as can be found in Limpopo Province. In particular, individuals living with HIV disease may suffer from osteopenia and are known to be more at risk of fragility fractures and delayed fracture healing [42] and may therefore be vulnerable to injury from regular compressive loading through the cervical spine.

Frequent loading beyond capacity for adaptation or repair may also lead to early degenerative changes in bone and soft tissues [12]. A threshold of 250 Newtons of sustained cyclic loading (15% of failure stress, approximately 6MPa) applied to articular cartilage in vitro has been reported as a threshold above which cell death occurs and increases in proportion to the applied load [59]. Cell death in mature cartilage can lead to degradation of the tissue and is associated with onset of osteoarthritis [59]. Although the actual forces sustained by the cervical spine during water carrying have not been directly quantified, this study indicates that they are likely to exceed 250 Newtons for many individuals, when the weight of the head and effects of muscle contraction are added to the weight of water carried. Whilst pain, stiffness and functional impairment are clinical features of osteoarthritis, the correlation between symptoms such as pain and radio-graphically observed degenerative changes is not clearly established. Therefore future research should investigate the relationships between loading intensity, frequency and duration, history of physical loading exposure and symptoms such as neck or back pain and functional disability, rather than radiographic examination findings alone.

Guidance on good manual handling techniques for safety when pushing loads in high income countries suggest that worn wheelbarrow tyres and lack of grip padding on various types of equipment can increase the work of pushing and affect grip force and comfort [60, 61]. This can be particularly plausible on sandy pathways such as those along which water was frequently carried by the participants in this study. Water is also an inherently unwieldy load, which moves within the containers during handling. Although the participants had clearly developed skills to lift and balance containers, maintenance of a secure grip would be difficult during sudden or unexpected posture changes, as might occur when walking along routes shared with vehicles and domestic animals.

Sudden or unexpected posture changes may lead to injury through generation of high peak compressive forces. These can occur due to muscle action on the spine [11] which in the cervical region is required to support the weight of the head and loads applied to it to prevent spinal buckling [62, 63]. Rapidly or awkwardly lifting objects or accidents during manual handling can generate peak compressive forces higher than injury threshold, but may also create torsional, shear or bending moments which injure the spine if it is inadequately stabilised [11]. Hazards for slips, trips and falls include wet and uneven surfaces, obstacles, exposure to traffic, poor equipment and unwieldy loads [64, 65], all of which were typical environmental and task related factors of domestic water carrying.

Distance walked between water source points and the home may be a useful indicator of exposure time to sustained compressive loading. As the distance walked by those who reported spinal pain was significantly less than those who did not, our results might indicate pain-related disability. People with spinal pain may experience difficulty carrying water over distance and be more likely to enlist the help of other family members and children, or continue to carry water only if it is accessible close to home. Such functional disability may have further implications for families, for example, by leading to a reduction in the usual volumes of water collected for household use to support health and adequate hygiene.

Our study suggests that the volume of water carried and environmental factors, particularly the incline or gradient of the path along which water is carried are likely to influence the physical work of water carrying as indicated by RPE. It also suggests that people reporting neck or head pain may be those who carry heavier containers and also perceive the task to be more difficult, as for head-loaders reporting head or neck pain, the differences in weight of water carried and RPE were almost statistically significant. Distance walked whilst carrying water, volume of water carried and path gradient are therefore important quantifiable factors which might be useful to indicate the physical work load of water carrying.

This study has highlighted the potential association between spinal pain and water carrying in South Africa. This association is complex with water carrying probably contributing to the aetiology of spinal pain and spinal pain interfering with people's ability to carry water with potential impact on household water availability. Typical methods of carrying water containers as observed in this study impose physical loading with potential to produce symptoms typical of musculoskeletal disorders and related disability. Risk of musculoskeletal injury or pain may be high as it is usually a task for women and children, including those who may be affected by chronic poor health, and is often performed with inadequate equipment in potentially hazardous environments. Water carrying is not the only manual work performed by women and children in lower income countries and future research should also investigate the additional burden from other physical tasks.

Carrying distance could be used together with total volume or weight of water carried and path gradient to indicate the level of physical work imposed by water carrying. These factors together with the modified Borg scale and water carrying method should be investigated in future research, to better understand the type and strength of association between water carrying and health, particularly symptoms typical of musculoskeletal disorders such as pain and related disability. Identifying risk factors for musculoskeletal disorders and pain related to water carrying may also highlight appropriate interventions to reduce risk exposure.

Despite the small study size and associated lack of power, our preliminary findings still highlight the potential impact that carrying water may have on health, in particular through the effects of symptoms typical of musculoskeletal disorders, such as neck or back pain, and related functional disability. This is an important but neglected public health issue. There is a need for more research on the impact of water carrying on neck and back pain and how such pain impacts on the water carriers lives. There is also a need for research into how water can be carried in a way that reduces the potential for adverse im.pacts on musculoskeletal health.