In 1948, in the aftermath of the partition of India and Pakistan, a journal article on the health situation of refugees and internally displaced persons stated that “this report is based entirely on impressions”. Another of the earliest articles about a humanitarian emergency, the East Bengal cyclone in 1970, stated that “relief supplies and volunteers poured in, but no one knew the magnitude or geographic distribution of losses and needs.” How did these volunteers know what skills were needed? What supplies and commodities to distribute? How did they know where to go or who to help? Impressions, best intentions, and customary practices were the rule at the time, and health interventions were rarely supported by epidemiological or clinical studies that provided evidence of effectiveness. Disaster relief was guided by charity: people affected by emergencies have lost everything, so anything given as aid could only help.
By the 1980s, this governing perspective began to shift to a competence-based model: thresholds for what could be achieved in terms of improved crude and age-specific mortality rates were established, as were quantifiable indicators that were considered to be measures of an improved quality of life for those affected by natural disasters, human-made emergencies, or both. Trends could be monitored to determine if a relief effort was effective in meeting people’s basic needs for food and nutrition, water and sanitation, shelter, and health care. This transition was most evident in the publication of the Sphere Handbook, most recently published in 2011 and currently being revised, which defines humanitarian action as a human-rights-based discipline and suggests minimally acceptable standards of performance.
During the late 1980s and the 1990s humanitarian relief, in some sectors more than in others, became more adept at determining and meeting the needs of populations affected by emergencies. Repeated epidemiological assessments led to the identification of fairly constant patterns of morbidity and mortality and, consequently, to the general recognition that some interventions should be accorded priority over others, at least in the initial phases of a relief effort. As has been pointed out about humanitarianism, “the transition from [relief] based on anecdotal evidence to methods that prioritize hard data has led to astonishing improvements in performance”. Still, even today, as was too often the case in the past, anecdote all too often remains the basis for health interventions. For this reason, the new Lancet Series on health in humanitarian crises is timely and most welcomed.
As Karl Blanchet and colleagues remind us in their Series paper, although scientific publications on health in humanitarian crises have increased with time, the number of studies done in humanitarian settings, their quality, and measures of effectiveness of the interventions are insufficient. Their review of the literature shows that only a small proportion of the indicators in the Sphere Handbookis based on strong scientific evidence—most have come into being on the basis of humanitarian workers considering them to be “best practice”, perhaps consistent with what more rigorous studies would show, but perhaps not.
Some people complain that the growing accumulation of studies, statistics, and standards risks burying the “humanity” of humanitarian assistance beneath a mountain of public health and clinical data. They may have a point. So, this Comment is not intended to be yet another call for more scientific evidence and for the use of that evidence in humanitarian decision-making processes. Those calls are already plentiful and have yielded modest results. The most recent of these is the eloquent and comprehensive Series paper by Francesco Checchi and colleagues that is essential reading for all relief workers.
This Comment is a call to humanitarians working in the medical and public health arenas to recognise that the objective of this work is not to produce evidence, but rather to provide the most effective interventions for the most marginalised and most vulnerable in our societies while paying full respect to their human rights and to their ability to live their lives with dignity. For this reason, not all kinds of research are appropriate in humanitarian settings—ethical standards must remain high and uncompromised when contemplating undertaking research in compromised populations, as Dónal O’Mathúna and Chesmal Siriwardhana highlight in an accompanying Series Comment. Two fundamental questions should underlie all evidence-generating activities: which interventions of established effectiveness in non-emergency settings remain effective in emergencies and, if the established means of delivering those interventions need to be modified, what is the best way to ensure that the population in need derives maximum benefit from them?
Two of the Series papers make it clear that the current system is not working optimally. Sandro Colombo and Enrico Pavignani9 offer a sweeping, historical analysis of repeated failures, ascribing them to factors at play within and extrinsic to the existing structure of the humanitarian health sector. Paul Spiegel points out that the world has changed considerably since the current system of humanitarian relief came into being. He offers four concrete recommendations, not all of which are health-specific, for improving the humanitarian system in the future, recognising that the interests of both public and private sector relief agencies are sometimes placed before those of the populations they purport to serve.
As distasteful as it might be, political interests, available funding, and other non-scientific factors will always be part of the humanitarian decision-making process. Yet, our work can only be accomplished if the evidence-based principles that are so fundamental to medicine and public health are given the greatest possible consideration in determining what gets done. In the practice of humanitarianism, however it might be defined, science is clearly not sufficient but is, more than ever, necessary.