One of my New Year's resolutions was to finally get 'round to editing down this wonderful long letter from economist Jishnu Das of RISE-PAK on his trip to Pakistan in December for earthquake disaster relief. (My previous post on RISE-PAK was Asim Khwaja: “The nice thing about computers is that they don’t go into shock.") Here is Jishnu's December 13th letter, a response to my asking how his trip went:
Part of my trip involved working with Lahore University of Management Studies faculty and students on a field trip. For me, working with volunteer students from LUMS in the field was an incredible experience. They would wake up at 5:30 every morning, brew tea, cook breakfast and hike off to the villages for a full day before daybreak. On walks, they would be crossing landslides, talking to each and every person they met and returning well after dark by torchlight. These are some of the most committed and enthusiastic people I have been fortunate enough to work with and their commitment to information and transparency is amazing (were this a holiday hike, I would definitely have viewed being pulled out of a warm sleeping bag at 5:30 as a dastardly act...)
Cheers
Jishnu
DATA ISSUES
1. Creating a universal list of settlements: One big problem in compiling and understanding the data is that villages are divided into multiple settlements, and there is no universal list of settlements available. Since text (specially with translation from Urdu to English) is not standardized, it is impossible to tell, for instance, whether the relief provided to "Berbazar" is the same as that provided to "Berbush" and which village this settlement belongs to. I worked closely with the UN-HIC towards developing their gazetteer of locations. Unfortunately, things are almost as prelim there as they were 1 month ago, with everyone still stuck with settlement names issues. Piet and I will be working on this more this week, and we should have something that will be made public by the end of this week. We are also working with the Population Census Organization in Pakistan on finalizing this. By the way, we came across www.fallingrain.com, and this contains geo-locations for millions of locations around the world. If someone can send out an html crawler and capture the database that would be great (we did Pakistan).
2. Villages versus settlements: There is problem with what is a "village" and what is a settlement, but I am not sure that it is really bad. 3 villages that I covered in a recent survey are in the database as villages--Batangi, Gajoo Khokhar and Basantlok. Indeed, so are the villages that Jawad's group followed (Sund Ban, Chamata, Doba and Harama). The one problem is a village called "Muslundi" which is on the other side of a smallish stream (so batangi is at the start of this side-valley; basantlok is further down on the same valley. On the other side are Ratanser, which is in the census list). While this is NOT in the Noura Seri Patwar Circle list of village, IT IS in the Seri Dara list of villages--Seri Dara is the neighboring PC. So, my impression is that someone who is aware of the mauza-settlement issue and has a list of mauzas can sort this out pretty easily, but this is based on a very very limited sample. (One problem with going the settlement route is that most villages will have a Dhana, which literally means "top" and a kayer, which means "ridge").
3. Google Earth: Unfortunately, (a) no-one is aware of the VBR's (I told everyone I met, and sent them the link), (b) they work too slow on the broadband in Pakistan. I took the UN-HIC compound guys through it fairly carefully, and hopefully they are using it now.
RISEPAK AND WHERE IT STANDS
1. RISEPAK was set up as a self-coordinating enabling environment, where all relief actors and those affected by the earthquake could come on a common platform by posting information about damage and relief. Constantly updated, these postings would provide regular information that could help target future relief to those who need it most.
2. By a number of accounts, RISEPAK has achieved a lot of what it set out to do. Within 2 months of its launch (its now 7 weeks), there are 1800 messages that have been posted, and updated information on 950 villages out of around 2500 that were thought to have been affected (close to 40%). In addition, the RISEPAK site has also proved useful in a number of other ways. Organizations have used our pre-prepared forms to organize their own information systems; most organizations have worked closely with our maps, which were the most detailed available at the time and bulletin board posts have allowed sellers and buyers to get in touch with each other. Some anecdotes:
a. One organization that we went to had not heard about RISEPAK. They insisted that they were very organized in collecting their data at the village level, and were using standardized forms to record this information. It turned out that the forms were the RISEPAK damage and relief forms about villages!
b. In a recent pilot (more on this below), Shandana (a faculty member at LUMS) was speaking to the army major in charge of a particular area. The major was adamant that they were doing a great job and were making their information transparent and accountable through their own website. When asked about the website, he said that they were using RISEPAK---something that he had developed a full sense of ownership over.
3. At the same time, a lot more can be done. What is very clear is that smaller organizations in the relief effort have used and posted to RISEPAK on a very regular basis. For them, RISEPAK has turned out to be a boon---it has developed the trust of most players by acting in a non-partisan manner, and organizations who are regularly posting to the site are able to point out the work that they are doing to the entire world. What they are doing is transparent, accountable and verifiable; at the same time, it allows for massive benefits in coordination among the various relief actors.
4. Key to the success of RISEPAK has been the central role of the Lahore University of Management Studies (LUMS) faculty and students. Early on, we realized that the RISEPAK effort was a bit rushed. If the system had been set up before hand and key organizations had been trained in its use, information would have been posted regularly without much prompting. As it is, we were working on the fly. This meant that relief actors had to be taken through the site, trained on using it, trained on the importance of data at the village-level and data had to be constantly obtained from these groups.
5. The LUMS faculty and students took this challenge heads-on. Instead of celebrating Eid with family and friends (Eid is somewhat like Christmas, only larger, since it comes after a month of fasting) faculty and students headed out to Islamabad and the affected districts to get this data collection exercise moving. In Islamabad they developed close networks with relief organizations, helped them systematize their data and start sending it into RISEPAK. They set up a team of volunteers who took in this data---from fax transmissions, e-mails and the website itself---and parsed, collated, and updated it on the website. Their field-teams also visited the affected district headquarters and started working with the district governments, the UN and relief organizations in the field. The activity led to a huge increase in postings to the website---RISEPAK had updated information on 200 villages before the LUMS team went out; within a week of their return, the numbers went up to 500, and now stands at just above 950.
6. I was fortunate to be a part of the next such team that went up; again, the students and faculty taking off from their hectic schedules in their quarter-break (7 days) to head up to quake affected districts and villages. Key to the medium term reconstruction and relief in the region was an assessment of how well government compensation programs have worked so far, and I went to Pakistan to work with the government and the UN and to visit the quake affected areas to arrive at some assessment.
7. In Islamabad, I met up with the team from LUMS (close to 35 students and 10 faculty members); we then headed out into different directions---one team went to Bagh district, another to Mansehra and a third (including myself) to Muzaffarabad. The team I was in consisted of 15 people including myself; some of these would work in the district headquarters, others would head out to villages (both those that are more and less accessible) to assess the state of compensation and data.
The Post-Earthquake Household Survey In the five days of the field-trip, much was accomplished.
1. The 12 teams that went out to the villages surveyed close to 3000 households in 18 villages---easily the largest independent survey of households in the post-quake scenario by an independent group. We branched out into Bagh, Muzaffarabad and Mansehra/Balakot and then chose villages according to a stratification based on close to road/far from road and large relief activity/low relief activity. I was in a group that went to "low relief activity and both close and far from road". In addition, I was also part of a team that visited a relief-camp in Muzaffarabad. These data are currently being collated, and will be made public for all agencies to use fairly shortly on the RISEPAK site.
DATA-COLLECTION STRATEGIES IN VILLAGES: ARE SURVEYS POSSIBLE (SURPRISINGLY, THEY TURN OUT TO BE FAIRLY EASY)
1. People are very, very used to making lists in the villages we went to---indeed, it turned out to be harder to do a focus group than to make lists. The moment we sat down to do focus group, people would start gathering with ID cards to get their names entered. NOT entering names is, basically, NOT an option--we would have people running down to ensure that their names are on the list.
2. At the same time, there is very little movement across settlements within the same village. People are able to, fairly accurately, give names and rough family composition (total members and children) for families in the settlement, but not across settlements in the same village. Batangi, for instance is 3 settlements---lower Batangi (an abbasi settlement), main Batangi (chaudhuris) and dhana (mostly abbasis). The first day we went to batangi, drew up a list of settlements and then went to dhana. We made the list of families with the school teacher on the advise of locals and then went to every sub-settlement. The school-teacher missed out some families (abbasis) right on top---nevertheless, it would have been hard to miss them out all together, since the moment we sat at a central location, this was pointed out to us.
3. Another village, we went to is similar, though spread out over a larger distance. The village contains two settlements at a 1/2 hour walk from each other, and it is impossible to get names of households in one settlement sitting in another. To get a sense of how fast a village could be covered with just basic household-level information and damage, we split up and went to each settlement. We then asked people to gather at these settlements and completed close to 120 households in 4 hours or so. We then verified that households had not been left out, though am not a 100% sure.
Relief Camps Relief Camps are also reasonably easy to survey in---again, people are used to surveys and the kind of information we are asking about. There are problems with split families---don't know what we can do about them in terms of verification. People should e-mail the women who went to the relief camp (Nadia and Erum in the team I was with--erum is copied on this note) and ask them about their experience. Stories were very different depending on who was telling them---the relief-camp organizers, men, or women. There is absolutely no sanitation or toilets in these camps, and women are having a horrendous time. This is something that I know a lot less about...
Some highlights on the situation in villages
i) In Islamabad, a lot of people felt that villages had emptied out. This is far from reality. Despite the large number of casualties and injuries, the percentages are not as large as one might believe a priori. For instance, the 80,000 deaths means that 1.5% of the areas population died. I was working in one of the hardest hit areas, where all houses had been completely destroyed, yet on average, 2% of individuals in the villages died, another 2% were injured and another 1% were in relief camps. This left 95% of the original population intact in the villages----more critically, not one person said that they were planning to leave for the winter. This requires that the means to construct emergency shelters are made available immediately to the large population with destroyed houses, who cannot spend the winter in tents (since you cannot light fires in them). Ensuring the arrival of corrugated iron sheets (something that people have been pointing out for a while) will definitely save many, many lives.
ii). At least in the area I was in, the government and army have done an incredible and very fair job of giving compensation exactly according to government guidelines. Every person whose house has been destroyed received Rs.25,000 (roughly $400) and every household with a death received Rs.100,000 ($1,600). I believe that other field-teams are finding the same thing.
iii) While people feel that livestock are a major source of income in the area, they actually are not----close to 70% of the households interviewed did not own a single animal (buffalos, cows or goats).
iv) One key advantage of working with the LUMS team is that there were female students as well, who could talk to widows and other women, usually left out of the survey process (it is culturally difficult for a man to talk to a woman alone). Women's concerns were usually different from men's---while men were very concerned of the need for shelter, women, who have to deal with the everyday process of living were also anxious about the lack of warm clothing and blankets for children, and the lack of cooking utensils---something that was causing them a lot of unneeded hardship. Here is what one student has to say:
"Women’s responses tend to be more detailed. Men leave out what they feel is unnecessary, I personally found women more willing to take the time to communicate the smallest of details. However nuances such as the issue of dependence on other family members for a widow, or feelings of marginalization and perceptions of being harassed or mistreated require some probing before they are brought out. Again, these will seldom be expressed in the presence of male members of the community. Finally, women tend to have a better idea of other vulnerable women in the community, such as single mothers, and were helpful in identifying them. As with the other sex, one-on-one interactions tend to be more honest and informative. As the group size increases and people struggle to get a voice, responses tend towards the more "rehearsed" type. It is always better, I found, to initiate spontaneous conversation with individuals rather than wait for the more vocal members of the community to gather and lead the discussion" (erum haider)
Some thoughts on winter from Sadia Qadir 1. Clearly, this is going to be the large war. Here are some impressions from a student (sadia qadir)
SHELTER
1. Normal, un-insulated tents are not useful any more and the idea of moving to lower altitude areas is almost next to impossible. When asked them what are you planning to do one snowfall begins, they offer the plan that women and children are going to stay indoor while the men will take care of the outdoor chores. For fuel, they are depending upon (wrongfully so perhaps) the logs they have stockpile from the rubble of their houses. They plan to use it all through the winter season. Even though they anticipate they might run out earlier on, than expected.
2. The next best thing is shelter made from CGI-sheets. I had the opportunity to see one. This particular one was made in the triangular shape as that of a tent. It was however much more spacious. It was made with 20, 12-14 square feet sheets. I was told this is the minimum number of sheets required to build a shed that size, and sheets any fewer than 20 are useless. Quality (thickness) of these sheets is also an issue.
3. A small fire-place that was set in a corner and was also being used as a kitchen. Similar hearths have resulted in horrifying hazards in the usual fabric-tents. This is apparently their best chance to salvage from the extreme cold once snow falls. A family of 14 was staying here and I was told all of them fit in nicely at night.
4. Another, important issue is that of Kora (or Kori), which is a thin layer of suspended frozen moisture that layers the ground in this season. According to the people, sleeping arrangements comprising of floor-beddings, causes this frostiness to seep through the layers of bed linens and blankets and does not go away. As Shandana, suggested, perhaps providing char-pais (beds) will help combat this problem.
5. One of the limitations of initiatives like distributing CJI sheets is that people are selling them. Probably it's the fuel available to them, which leads over-estimating their ability to endure the winter. What ever the reason, the trend is been observed and confirmed by locals, the NGOs as well as other authoritative and operational bodies working in the region. This is perhaps, one of the major reasons many of the organizations are selectively distributing (and therefore accused of bias) the insulating tents and / or CJI-sheets.
Sadia (who is a doctor) also writes about health-issues that are bound to arise HEALTH PROBLEMS
In days following the earthquake, the bulk of presentations were that of extensive trauma -- mainly to the head, spine, pelvis and/or limbs, requiring surgical intervention. I was also told by doctors at Ayub Medical Complex, that a large number of amputations were carried out in the remote areas because enough resources were not available for timely intervention to salvage limbs. As I observed in Abbottabad most medical NGOs/camps came into the affected areas equipped mostly to deal with trauma cases. This requires specialized arrangements - such as x-ray facility, a small surgical theatre, relevant medicines, surgeons and OR nursing staff.
From what I gathered after speaking with the health care professionals working in the affected areas and the people, there is a high risk for Respiratory tract diseases (especially pneumonia in children), Gastrointestinal diseases particularly in areas where water is contaminated (there was an outbreak in Balakot over the period around Eid Holidays), infections particularly fungal skin infections as a result of damp cold weather (a large number of children affected in Bagh) and complicated wound healing.
Even now, most of the medical camps currently operating in and around villages, for example, near Sewar Kalu and Kafl Garh, are equipped to deal with trauma only. They have declined patients with complicated wounds and fractures, Obstetric & gynecological problems (including uncomplicated pregnancy) and skin infections. The irony is that residents of these villages have medical camps and professionals around but unable to help them. They have to travel to district hospital Bagh, which is 6km away and is not fully functioning as its building is also affected by the earthquake. To make the situation worse, these conditions usually are slow to develop and pursue an indolent course making it very likely for patients to get 'used to' their ailments until they reach an irreversible stage.
A dangerous alternative is consulting the traditional medicine-men in their areas. I personally witness such a case ~75yrs old lady resident of Kafl Garh, with a forearm fracture complicated in an attempt to fix through malish (massage). The entire arm was massively swollen as the bandage was tight enough to cut off most of the blood supply. This lady was declined medical assistance at the closely located camp and was advised to go to district hospital Bagh. She was brought back home.
An important element to consider in dealing with these people is - they are constantly prioritizing and reprioritizing their survival strategies and coping mechanisms. Prevalence of post-traumatic stress syndrome (PTSS) and acute stress syndrome (ASD) in these populations is still largely unknown but can be expected to be quite high. In the places I visited (Kafl Garh, Sewar Kalu and Kot Baroli) there are no mental health interventions so far. Point being, in there given state they are very likely to miss out or even deliberately put off for later significant health and nutritional concerns that they otherwise would take more seriously.
Data Flows
1. This is a large issue, and looks very much like the game we used to play called "chinese whispers" (sometimes called the "arab telephone" for no apparent reason). A couple of examples:
2. In the UN compound at MZD, we were told that the greatest benefit in terms of information would come from visiting the lower Neelum valley, of which little was known. This team then went to the union-council of Noura Seri, about 45 minutes from the city, and where roads had opened recently. The major in charge of the union council was very systematic in his record-keeping and had detailed records of the tents and food distributed to the villages under his command. He assured us that 90% of the households under his charge had been given tents, which we verified through door-to-door surveys. Further, compensation had been allocated exactly according to the policies laid down by the government. The key question, and one that needs to be addressed urgently, is what happened to the major's data by the time it got to Muzaffarabad--a 40 minutes drive away? If the district government of Muzaffarabad takes over the relief effort at this time, either a large data-gathering effort will again have to be undertaken, or they will have to fight blind in the face of the coming winter.
3. A second team went to Balakot/Mansehra in the North-Western Frontier Province. One mandate of the team visits was to work with district officials towards systematizing their own records and setting up data-entry systems at the individual level for compensation and relief. The picture in Balakot revealed the usual problems that face nascent data systems, leading to large problems later on. Some examples: In the list of individuals who had received compensation, names had been recorded, but ID card numbers had not (as an aside, 95% of households have at least one member with an id card, which serves as a unique identifier). Worse, there was no standard for translating Urdu to English names. With close to 10 Mohammed Afzal's in every village and with non-unique spellings for the same name, trying to relate this data to future relief will be a nightmare. Records of who had received disability payments were now virtually unmatchable to people, since they had been recorded on a separate form that omitted the village-name field.
4. The problem is not that governments are apathetic or uninformed about the need for systematic data exercises. In fact, everywhere we have gone, district governments have been delighted to work with us on strengthening their information systems. The real issue is one of capacity, the ability to work in remote locations and familiarity with data and data-issues that typically come after years of learning the hard way.
5. The LUMS team has brought this expertise to the relief effort, and it is one that they plan to maintain over the longer term. Guys: they need ALL the support they can get.
Jishnu Das is one of the founders of RISE-PAK and lives in Washington, DC where he works at the World Bank. He holds a Ph.D. from Harvard University in Economics.