Les Roberts, Ph.D., trained in physics, environmental health, environmental engineering and epidemiology and is a consultant engineer/epidemiologist, working for, amongst others, the Refugee Health Unit, National Center for Environmental Health, Centers for Disease Control and Prevention, Atlanta, USA. He also lectures at the Department of Geography and Environmental Engineering at the Johns Hopkins University, Baltimore, USA.Most of what is known about preventing illness during emergency situations comes from the study of illnesses in populations during peaceful, stable circumstances which permit careful scientific investigations. While the psychological and nutritional stress seen in many populations in crisis does make them unique, it is generally assumed that the basic immunological and physiological responses observed by medical and public health researchers in stable populations apply to those in crisis as well. Thus, establishing the best practices and standards among emergency relief workers is usually an issue of “how to do” something rather than “what to do”. When vaccinating children against measles, the only judgement involves the “how to” issues associated with organization and logistics. Once the “how to” questions are worked out, the “what to do” questions of how much vaccine to inject, what equipment to use and where on the body to vaccinate are already predetermined. Moreover, the efficacy of conferring public health protection once the vaccine is injected into a population is usually similar to that seen by researchers in stable populations.Unfortunately, when it comes to providing water, the issues of “how to do” and “what to do” are, for the most part, interwoven. Goals can be set regarding the cleanliness of the water, the volume provided, how many hours a day the service will be available and how close to the people water will be delivered. Yet, to a greater extent than with the provision of food or medicine, the service provided is dependent on environmental circumstances. For example, deep groundwater sources are conducive to providing microbiologically safe water, but perhaps not in vast quantities. These environmental constraints give the water professionals a great deal more latitude, but at the same time a greater need for judgement than most other specialists working during emergencies. In particular, the standard for water quantity provision is seen as somewhat flexible. If an NGO only manages to provide 1,400 kcal of food per person per day to a population with no other food supply, this is viewed as an unsuccessful programme. Often this occurs because of circumstances beyond the control of the relief community or because donor funding was insufficient; and people are comfortable in saying that a programme is not acceptable without needing to blame any of the response personnel. In the water sector, rather than saying that a water provision programme is inadequate, there is a tendency to say that the standard is flexible and that these specific circumstances permit lower quantities of water to be provided.UNHCR’s  guideline regarding water quantity in the 1960s was adopted from that of the WHO , 30 litres per person per day (l/p/d). This standard was not adapted from empirical data contrasting healthy populations with unhealthy populations, but was established by “experts” estimating the amounts of water needed for cooking, washing, drinking, etc., and then adding up the total daily requirements. In the 1970s the WHO standard was reduced to 20 l/p/d and the UNHCR target followed suit. In 1982 UNHCR declared the desirable goal to be 15 to 20 litres  and by 1992 it was calling for a minimum allocation of 15 l/p/d. But “when hydrogeological or logistic constraints are difficult to address, a per capita allocation of 7 litres per person per day should be regarded as the minimum ‘survival’ allocation. This quantity will be raised to 15 litres per day as soon as possible” . Again, the new target was based on what was actually occurring given the relief communities’ funding and workload, and was in no way based on human health data. At the time, the average water provision to an African refugee was perhaps only 12 l/p/d, in spite of the UNHCR standard . While the new target is a “survival” ration, millions of refugees throughout the world currently receive between 7 and 15 l/p/d. The new “survival” target enables project managers to say that, while we are not providing an ideal quantity of water, provisions are well above the survival level. Thus, the diminishing quantity standard is fundamentally to provide political protection rather than public health protection.There are many who will point out that water use habits vary between populations. In fact, many populations exist for extended periods with much less than 15 l/p/d [6,7]. Moreover, the water uses incorporated in the WHO minimum need estimate are not of equal importance. Drinking water and water for cooking are certainly more essential than water to wash clothes, at least over a period of a few days. Yet others have suggested that the minimum provision endorsed by international aid organizations should be 50 l/p/d . Water consumption in many Western nations makes a 50 l/p/d target seem absurdly stingy. For example, US residents consume about 410 l/p/d of municipal water and almost 6,000 l/p/d when industrial and agricultural uses are included . But people’s habits and judgement have little to do with the fundamental issue: how is water availability in an emergency associated with morbidity and mortality?Little controlled research has been done in the area of water and sanitation among refugees. The only known study to evaluate disease as a function of water quantity contrasted Mozambican refugee households who consumed < 15 l/p/d with those who consumed 16-20; 21-30; or >30 l/p/d . There was a steady association between consuming more water and experiencing less diarrhoea among children and, indeed, among all age groups combined. Households that consumed 10 to 15 l/p/d experienced 2.5 times more diarrhoea than those that consumed more than 30 litres, in a camp where faecal-oral diseases were the main cause of death. To evaluate this question further, field reports at UNHCR’s Geneva office were reviewed in October 1996; however, field reports rarely included both health and water availability data, and were often not standardized sufficiently to compare health experiences during different crises.Water can be acquired on the spot or transported to anywhere in the world. As with medicine or food, there is not so much a variation in water availability from place to place as a variation in cost. The standard dose of vaccine is not meant to be “the minimum sufficient” dose for a specific child’s weight and immunological history, but rather to be sufficient most of the time. Yet sanitary considerations are not like medicine, in that the needs are generally perceived to be most acute in the early phase of an emergency. This is because the vast majority of deaths often occur during the early weeks of a crisis [11,12] and because immunologically-naÃ¯ve people arriving in unsanitary settings are often the most susceptible to disease . Thus, from a public health perspective, our guidelines should say “15 to 20 l/p/d are required during the acute phase of a crisis, but less may be sufficient in the later phases”, rather than vice versa. There is a profound need for research to quantify the association between water availability and human suffering during crises. Until the data for proper cost-effectiveness evaluations exist, the water sector professionals should be willing to state what their minimum target is and why it is not being met. Unfortunately, the answer will most often be because someone, somewhere, with a flush toilet and hot shower, does not think that the extra investment to provide sufficient water is really worth it.Â Notes1. UNHCR: United Nations High Commissioner for Refugees2. WHO: World Health Organization3. Handbook for emergencies, part one: Field operations, UNHCR, Geneva, 1982.4. Water manual for refugee situations, Programme & Technical Support Section, UNHCR, Geneva, 1992.5. Daniel Mora-Castro, Programme for Technical Support Services, UNHCR, Geneva, personal communication.6. R.H. Gilman, et al., “Water cost and availability: Key determinants of family hygiene in a Peruvian shantytown”, A.J.P.H. 83, 1993, pp. 1554-58.7. S. Cairncross, J. Kinnear, “Elasticity of demand for water in Khartoum, Sudan”, Soc. Sci. Med. 34;2, 1992, pp. 183-89.8. P.H. Gleick, “Basic water requirements for human activities: Meeting basic needs”, Water International 21:83-92.9. Environmental Almanac, compiled by the World Resources Institute, Houghton Mifflin Co., Boston, MA, 1992.10. L. Roberts et al., “Keeping clean water clean in a Malawi refugee camp: A randomized intervention trial”, Bull. WHO, submitted.11. “Famine-affected, refugee, and displaced populations: Recommendations for public health issues”, MMWR, Vol. 21/RR-13, 24 July 1992.12. L. Roberts, M. Toole, “Cholera deaths in Goma”, Lancet, 1995, 346:1431.13. D.L. Swerdlow et al., “Epidemic of antimicrobial resistant Vibrio Cholerae 01 infection in a refugee camp, Malawi”, in Program and Abstracts of the 31st Interscience Conference on Antimicrobial Agents and Chemotherapy, Chicago, IL, 9 September-2 October 1991, abstract 529.