Humanitarian Water at Risk

It’s funny how you can get used to anything. As difficult as it was living in the camp, when I left there was a sense of loss.

Well, now it’s really been a long time since my last blog post. Almost exactly a year ago to the day. I look back on a post I wrote when I was in South Sudan still, the entry titled 'Pause'. A whole lot had happened back then, in those three months between entries. But now it’s been a year and a whole lot more has happened.

I finished up the research in South Sudan and left the field at the end of April last year. It’s funny how you can get used to anything. As difficult as it was living in the camp, when I left there was a sense of loss. I was tired, super burnt out I think, but I knew I was going to miss Jamam. When I look back on it now, the nostalgia is almost painful. The people and the place are burned into me.

I don’t want to dwell on it too much, but coming home was rough. I’ve travelled a lot, have worked for many years overseas. But coming back from South Sudan was a different kind of beast. It’s funny…I spent eight months in the refugee camp and hardly once got sick (thankfully!). But the moment I got home, BAM! Hit by the worst fever I’ve had in my life. I guess it was finally coming off the adrenaline after eight months and all of the stress, exhaustion, everything that I had pushed down, finally catching up with me, all at once. I was totally bedridden, unable to move for two weeks. And I don’t think it was just the physical that knocked me out, the fever coalesced with an emotional malaise, and made it hard to move.

Eventually I knocked myself out of it and got back to living. I began to work on the research, and damn that was a helluva bigger job than I thought it would be. I worked on figuring out how to analyze the chlorine decay data from South Sudan and writing up the report all summer, all the while applying for new jobs, trying to figure out what would be my next step. I really wanted to continue somehow doing research in emergencies, research that could help protect the lives and well-being of the most vulnerable. I was looking into academic positions that would enable me to do this kind of work. Eventually, I came across a position for a postdoctoral fellowship at the University of California, Berkeley. It was a new fellowship out of the Development Impact Lab and the Blum Center for Developing Economies at UC Berkeley which sought to bridge the physical and social sciences on international development projects. They also wanted to support research related to disaster relief. As I read the call for applications, with each sentence, it became clear to me that this position was made for me, and I was made for it. I put in the application, did an interview, and by end of the summer got an invitation to head to California.

I arrived at Berkeley in November and moved into a new spot in Oakland. This is one helluva beautiful and messed up place, but that’s another topic. For the fellowship, I proposed to expand the research that I had started in South Sudan to other camps around the world. I’ll come back to that in a minute, let me explain what the research from South Sudan showed first.

So through the summer I was struggling to figure out how to analyze data. One day, sitting in the car with my younger brother Saad who’s a math whiz, I was explaining the problem I was struggling with. He just sat for a moment and said, alright, that’s doable, we’ll just write some code in MATLAB (a mathematical programming language). We got home, he got on his computer and knocked it out. Working with Saad, we developed a novel analytical approach for this new kind of problem (there isn’t any other published literature on this problem that I can find!). I wrote a report and with MSF we iterated it several times to get to the final product.

So if you look back at the entry entitled “Science Camp” you can get an idea of the question the research was trying to answer. Analyzing the chlorine decay data from South Sudan, we found that the current humanitarian guidelines for FRC at tapstands offer insufficient residual protection post-distribution. We found that the chlorine decay data showed a high-level of agreement with a second-order rate reaction, corroborating what much of the literature says about chlorine decay. Of the various parameters we analyzed in the field, it emerged that the initial FRC concentration, electrical conductivity, and ambient air temperature were significantly associated with FRC decay. This makes sense with what we would expect: decay rate should be higher when the initial concentration is greater, as FRC decay is nonlinear; electrical conductivity is a proxy measure for metals content of water which would accelerate FRC decay; and higher temperatures would be associated with greater reaction rates. The research also looked at associations between FRC decay and water handling practices: of these we found statistically significant evidence that covering water containers in the home reduced FRC decay, as we would expect to find.

Once we had used the FRC data to model the decay (Figure 1), we could “reverse-engineer” what the initial FRC at the tapstand should be in order to have a desired level of protection (i.e., FRC concentration) at a desired time after distribution from the tapstand. On this basis, we made a tentative recommendation that the general guideline for FRC at tapstands be revised from the current range of 0.2-0.5 mg/l to 1.0 mg/L, in all situations. According to the modeling, this should provide FRC protection of at least 0.2 mg/L for up to 10 hours post-distribution.

nth order decay model

 nth-order decay model for model 1 (all camps, all [FRC]i). The top graph presents the observed data and model predictions, and the lower presents the plot of residuals. On the top graph, the vertical axis represents FRC concentration in mg/l and the horizontal axis is time elapsed since the water leaving the tapstand. Blue x’s represent observed values and red circles represent the predicted values. In the legend, the reported k values represents the modeled decay rate; n represents the rate order of the decay (close to second-order); and R2 represents the goodness-of-fit of the model.

Once I got home and was able to review the literature, I found out just where the current humanitarian guidelines used by MSF, the Sphere Project, UNHCR, and others come from. Apparently, they all go back to the World Health Organization’s Guidelines for Drinking-Water Quality. But the problem is the WHO guidelines are based on experience with municipal piped-water systems—what water quality should be in Manila, Oakland, wherever—cities where you drink the water right from the tap! The refugee camp and the emergency context are radically different! For one, people don’t drink water right from the tap—they collect it, carry it home, and then store it for some time before consuming it; and all this in a relatively unsanitary environment. All of this means that there’s a lot more decay of chlorine than there would be in a city. And this is why we found that the current guidelines were generally failing to ensure safe water supply in the camp setting.

As I examined the literature, it became clear that this was the first study to ever look at the on-the-ground effectiveness of emergency safe water guidelines in a systematic way. The recommended FRC target it puts forward then is the first guidelines to be based on actual field evidence. It may better protect drinking water from pathogenic recontamination and thereby limit the spread of waterborne diseases in refugee camps.

But remember I said that this was a tentative recommendation? We still have to “ground-truth” this recommendation—demonstrate that it actually does improve household water quality outcomes, from which we would then be able to infer public health benefits. Moreover, these findings come from three refugee camps in South Sudan. What we don’t know right now is whether these findings are generalisable to other camps around the world.

This is exactly what I’m trying to do in my postdoctoral research now. I’m expanding the research started in South Sudan to other refugee/IDP camps around the world so that I can build a more general evidence base upon which to ground guidelines for safe water in emergencies. I’m going to places with different climactic, terrain, and water supply conditions, so that we get the most breadth of settings possible. I’ve partnered with UNHCR and MSF continues to support the work. This summer I’m planning to launch the study in Zaatari or the new Azraq refugee camp in Jordan. Early next year, I plan to get out to two more refugee camps to continue building the evidence base. My aim is to develop safe water guidelines that are evidence-based (finally!) and articulated for specific local conditions. By influencing the UNHCR, Sphere Project, and MSF guidelines for emergency safe water supply, the research stands to improve how safe water is done in refugee/IDP camps the world over.

But before heading to the field, I’m headed to London to present the work at the MSF Scientific Day Conference!

Sci Day is a conference to present scientific research carried out in MSF programmes around the world. Because MSF is a leader in humanitarian medical response, this conference often presents innovations that are taken up across the humanitarian sector. All of sections of the MSF movement are represented, as are many international NGOs and UN agencies. Many of the MSF field offices also watch this event. In general, it’s a remarkable exercise of MSF internally sharing and debating research, and inviting other people to join in on the process. I’ve heard that last year, over 1,700 people watched live from 92 countries. It’s a real honour to have this work selected for Sci Day. I’ll be presenting my work as part of the panel on research in settings of emergencies, violence and conflict, chaired by Dr. Paul Spiegel from UNHCR. Watch online on Friday May 23rd if you want to hear more about the research!