Skip to main content
Loading Explore guides…
Engineering & Global Poverty · Acedra
Home
Explore
Engineering
Engineering & Global Poverty
Sign in to save
Chat
If YouTube asks you to sign in, open the video on YouTube.
Watch on YouTube
Reload player
Transcript
76 segments
Copy
Hide video
0:01
I'm going to share today a few different things: one is the scope that the Mortensen Center is attempting to help make a contribution to. So the scope of global poverty in the world today and the role that engineers can play and trying to alleviate these challenges of poverty in low and middle income countries and then we're specifically some of the work we do in the Mortensen Center and more specifically than that the particular work my research group
0:29
does and has done and we'll be doing going forward. So here we are, looks pretty familiar. What's cool about this is it's actually it's a series of pictures, it's not a graphic, it's a every place on earth on a cloud free day but looks pretty familiar. So here's another way of looking at the planet, just humor me, what do you see here? What does that tell us? People, rich people, right, we see us, we see Europe, we see India, Britain, China and merging, that's the Nile but
1:04
you see most of sub-saharan Africa essentially in the dark. So here's another way of looking at the planet. What are these? Fires right, some of these are forest fires but most of these are cook stoves. Three million people in the world still use biomass so wood and charcoal on open campfires for their daily energy needs and respiratory disease associated with indoor and outdoor air pollution of which biomass burning is a major contributor to the
1:33
global burden of disease. Five million children die every year from respiratory disease. We still have two billion people in the world without safe sanitation there's still pretty much a billion people they don't have access to clean water. The official numbers is lower than that, it's like 700 million but the real number is closer to a billion when you look at actual service delivery and quality of delivered water and 99 percent of the children who die under
2:01
the age of five every year were born in a low-income country. So these are issues of air quality, water quality, sanitation and safe energy access in developing countries that still exist today. There's another way of looking at the plan. These are aerosols so these are sandstorms and these are biomass aerosols. There's a snap shot of one one day on the planet and while the trend in overall in the world is positive people are
2:37
getting healthier and wealthier, there's more stability, there's fewer number both proportionally an absolute number of people without access to clean water safe sanitation and energy, in some of the areas of the world things are actually getting worse attributed with the climate change. In East Africa there's 30 percent less water annually now than historic norms, so droughts which used to be every 10 or 15 or 20 years is now nearly every year in the
3:04
Horn of Africa attributable to climate change. Malaria is a parasite that lives on mosquitos, it's killed more people in human history than any other cause, respiratory disease, diarrhea, war, anything else, malaria hands down has killed more human beings and the one saving grace of malaria is that it was too cold for the parasite to survive in mosquitoes above about five or six thousand feet, so there are population centers like Nairobi, Kenya, they came
3:33
into being because people observed that you were not dying of malaria there anymore and now just within the past five years there's persistent and chronic malaria in Nairobi. So we actually have disease patterns changing attributable to climate change. This is another way of looking at the planet: it's the world population proportional to landmass, looks pretty familiar, so it's not a lot of discussion there, although you've course see Indian, China with a billion people each jumping
4:03
out and then you have economic growth over the last quarter of the 20th century and you see most of Africa almost entirely disappears, most LDC, least developed countries, are in sub-saharan Africa still today and you see Europe and Asia and the United States and North America growing. Similar to economic growth are proportional carbon emissions. It's almost the same map with the exception of South Africa, it's nearly the same map, so there's been
4:36
a direct coupling through the Industrial Revolution basically until now, between emissions, energy use and economic growth. It's only in the past two or three years that there are signs of decoupling between increased energy consumption and increased economic growth. In fact Anthropocene, a magazine published here out at SEEC had an article in their last issue basically suggesting and proposing that the United States is finally decoupling. There were
5:04
reducing our per capita energy use while our per capita GDP is increasing, so just in the past few years.
5:14
Proportional population and poverty: it's almost the inverse map so we'll go back further here for a second. So for proportional carbon emissions, proportion of population and poverty. And these numbers are a little bit old but it's still the status quo. But things are getting better. You've seen I'm sure, I've stolen this piece from Hans Rosling, I'm sure you've seen his much more dynamic TED talk the stats of the Sweedish statistician, but this is data of income and life
5:46
expectancy so life expectancies linear and income is on a log scale over 200 years and effectively the trend is positive for every region and every country in the world, although you can see some incredible impact so there's those the Second World War, you saw the first world war a couple seconds before that. But every country in every region of the world has been getting healthier and wealthier, although you still see a large cluster of African countries behind. And
6:12
I highlighted Rwanda in particular because our team has done a lot of work in Rwanda with colleagues in this room, I'll come back to it. And this is the state of the world today. These substitutions made the case in a book recently called Fact Fulness and the Gates Foundation has recently adopted this new paradigm of dropping this idea of developing and developing developed countries. There's been lots of different euphemisms used
6:42
over the years, you know third-world countries or at least developed countries or global north and global South. It's a little bit more boring, it doesn't roll off the tongue as easily but it's more accurate to look at income levels. There are clear stratifications of income levels and you do of course see clustering in the low level one, so it's a little bit more challenging to try to succinctly describe the conditions you're trying to address but
7:09
we're trying to adopt within the Mortenson Center either this framework or the UN adopted framework of low and middle-income countries, for a variety of reasons, the main one being there's no such thing is that the economy between developed and developing. This is the global burden of disease in the world between high socio-demographic income countries so to rich countries and low SDI countries, poor countries and this is the mortality, this is mortality among children under five.
7:41
So congenital defects or neonatal issues will kill you if you die under the age of five in the United States or in Europe. If you're born in a developing country you're in a much higher risk overall of dying under the age of five and if you do chances are its respiratory disease or diarrhea or malaria, which are environmental hazards.
8:07
So that's the condition of the world and the Mortensen Center and our research team and a lot of you are already engaged in trying to see how engineers can contribute to addressing these poverty challenges. But the status quo for most development efforts and there's a lot there's a lot billions upon billions tens of billions of dollars are spent every year both formally and informally in development programs but it looks like this: money comes in
8:34
from a donor from the World Bank, from UNICEF or USAID or all the way down to a Rotary Club or Engineers Without Borders but there's some donor putting money in to programs that often have a technology component: water filters, cook stoves, bio gas generators, latrines. Not all development efforts are technology-based but many have a technology component to them and these interventions are supposed to positively impact people's health and livelihood.
9:00
The challenge is that long term measured impact from many of these interventions is often hard to find, it's very short lived. Roughly half of water sanitation infrastructure that's installed every year in developing countries with foreign aid money is broken within the first 18 months. About half of in destruction is broken at any given time and it's partly because of this status quo. Their owners will pay for infrastructure, there's an expectation
9:27
the communities and nations pay for ongoing but the reality is massive service gaps. How can we change this, how can we influence this system so that the impact in developing countries with partners is aligned to the payments and then we're able to sustain services. What we want to create is this more virtuous cycle where you have a service like a water pump. You continuously monitor that service, you use that monitoring to prioritize and enable maintenance and
10:00
other services operation and maintenance activities and you pay. You have money from communities, from governments, from multilateral agencies, tied to services. People are willing to pay for services but they're much more willing to pay for reliable water services so you can in fact have greater cost recovery when you professionalize and deliver better services but also there's probably long-term role for aid. Now there are a lot of people that work in the sector
10:29
that will say our goal is to work ourselves out of a job, we're only going to do the heavy lift of paying for the dam or the water pump of the borehole and countries have to pay for it. In reality forty billion dollars a year is spent on WASH by foreign aid. We can help that money be much more cost-effective. So our team has had a few different mechanisms for trying to influence this system. One is by creating commodities things like carbon credits, health
10:54
credits, development impact bonds, other performance-based contracting that is ties payments to services that are delivered and then the other is through improved monitoring. So our particular niche has been satellite and cellular and connected sensors that remotely monitor this infrastructure so we put sensors in water filters, cook stoves, latrines, on water pumps. We've used them on a sampling basis within impact evaluations and randomized control
11:22
trials and then we've also used them operationally to try to influence day-to-day operations of service and sometimes we're doing both of those at the same time. So here's a few different examples of where we've done this. This is getting a little messed up.
11:42
Okay, broken hand pump, an engineer's instinct is I can build a better hand pump. That pumps not broken, the proximate cause is because those bolt sheared off but really the ultimate cause if you take a failure analysis framework is that the system surrounding that pump failed. We are trying to influence that through improved monitoring. We put hundreds of censors all across hand pumps in Western Rwanda a few years ago. Chantal Iribagiza who's here and is
12:13
a PhD student in this department, at the time was the program manager for Living Water International, so she was the professional staff whose job was to make sure that these pumps stayed operating. So in fact Chantal's team was the subject of our study. We put sensors on several hundred hand pumps in one group the pumps were treated to basically the ad hoc business as usual maintenance, which meant the implementer in the county of the district government
12:39
would get around to maintenance when they could. The next group was a circuit rider model, which is considered the best practice in the sector, where there were dedicated technicians with vehicles and supply chain whose jobs were to go around and fix hand pumps. In the third group, and we had sensors in all these groups but only in one one of the groups was the data actually made available to the implementer, to Chantal's team, who had a dashboard that they
13:04
could see when a pump is broken. So we were passively monitoring two groups and then actively providing feedback on one. It had a pretty dramatic impact on water system functionality. In the baseline when a pump broke it was broken for more than 200 days. so more than half the year. Which is essentially the status quo in this sector as well, that's about typical. The circuit rider model got that down to about sixty days but only with the actual active feedback
13:33
were we able to get that, or Chantal was able to get that down to twenty days. So almost order of magnitude improvement in water in response time and moving the mean functionality for about 50% to 90%. The other thing is that it's cheaper to do this, even with the added costs of things like sensors and supply chain and dedicated staff. If you're metric is cost per person or cost per liter of water it's cheaper to maintain than it is to go off and build new pumps
14:02
all the time, half of which are broken in a few years. But the problem is well however self-evident that may be, it's not the status quo in this sector. We've been applying some machine learning tools to this as well, I'll be brief on this, but basically we want to drive functionality to 100%. This is one, each dot here's one day of a functional pump delivering water. Using some reasonably sophisticated machine learning tools, we can actually see when a pump is in
14:32
distress, when a water pump, the tire pressure warning light for a hand pump, and deploy technicians during this period of time. That's almost a week of a pump working, delivering water, before it actually breaks. Using this technique we've been able to demonstrate that 99 percent uptime is feasible without an increase in cost, because the maintenance activity was going to be required anyway, pump was gonna break anyway. You're just lowering that gap, shortening that window.
15:01
Okay next, here's what we're doing most actively now within our research team and this work also has overlays with the very large sustainable WASH systems grant that Karl and Amy have, that a number of you are involved in. Which is trying to improve functionality of services at scale in East Africa. We're currently monitoring about a million, 1.3 million people's water supply on a daily basis in Ethiopia and Kenya. So we've covered every single water point
15:30
in a far Ethiopia. We're expanding the Somali region now for all the emergency drought infrastructure and we've covered all of the emergency drought infrastructure in Northern Kenya. By the middle of next year, we'll be monitoring about five million people's water supply daily. These are satellite connected sensors that everyday are providing feedback on water point functionality, repair intervals, water yields. We also compare
15:57
it against other data sources like rainfall because there's you know, the the demand for the ground water in this region is inversely correlated to available surface water and rainfall. We have a whole bunch of different users of this data so we have the National Drought Management Authority in Kenya that pays for water trucking and they prefer that just keep bore holes running. We have UNICEF that spends a huge amount of money on drought relief and
16:21
would rather keep the bore holes running. We the Afar regional water bureau, whose job it is to maintain infrastructure functionality but these places are so remote that they don't know if that system is broken or not. So they're our customer, using this data - we reduced the maintenance intervals. This is sort of what it looks like, this is in Kenya, this is a five year 30 million dollar program that we have a piece of, operating all across northern Kenya. That's an example the
16:48
pump house. That's an example of actually a new system that failed within about a week and our sensor detected that, this particular failure. Then this is Ethiopia, a similar sort of environment, incredibly arid, people are nomadic pastoralists predominantly, historically they've relied on surface water. People keep their wealth in their camel herds and even within the you know within short few windows, here over the past couple years, you can see people
17:16
switching from camels to goat because they're more resilient to climate change. So this is a place where climate change is being felt in very real terms and there's a few other examples of our operating context in Kenya.
17:33
Looking at data that we had from about a million peoples water supply over 18 months, we worked with the NASA severe program to look at satellite collected rainfall data. So that's shown in blue against our borehole data and it's most obvious here in Kenya, although the trend was identical. There's an inverse correlation between rainfall and ground water demand, it's actually, it's kind of obvious. But the scale the data we've collected is pretty phenomenal and
18:00
they independently had basically the same result, which is that if there's been rain in the previous week, there's a 20% less use of borehole water now. That's just kind of an FYI, what this really means is that when there is no rainfall and there's no surface water the bore holes become the primary water source. When those bore holes are broken it leads to these emergency conditions which can be quite severe. Okay moving out of bore hole world, ground
18:27
water world. Let's give another example where we've used some of these techniques. This is a Rwandan household. 80% of Rwandan homes are rural and 80 to 90 percent of primary fuel use in Rwanda is biomass, on open either play stoves or three stone fires. Air quality, from air quality issues associated with biomass burning, is the leading cause of illness and death in Rwanda. The second is diarrhea associated with with water sanitation. People in Rwanda, rural Rwanda,
19:03
typically get their water from these types of hand pumps again. So out of a program that actually originally started, frozen here ok, I know the program that originally started here at CU-Boulder, who's involved with Engineers Without Borders or has been involved. So back in 2001, actually sorry 2002, we started getting involved in Rwanda and it grew from a very small EWB project funded by Rotary clubs and fundraising activities, to what over a 15
19:36
year period became a nationwide program. Oops I'm totally frozen. Where we were able to distribute over 350,000 cook stoves, high efficiency cook stoves, and over a hundred thousand water filters covering 1.7 million people in total. This is in 7500 villages, 350,000 households, 1.7 million people and all entirely privately financed. So we had private funding from a company to support distributing these stoves and water filters. A business model, I'm gonna have to keep ad-libing this thing is
20:13
totally dead, the business model was to continuously monitor the intervention. Monitor the stoves, monitor the filters and earn carbon credits under the United Nations tied proportionally to ongoing use and function of the water filters and cook stoves. We were the first to come up with his idea and apply it to water treatment. We did it first in Rwanda, moved that to Kenya and then brought it back to Rwanda. The model was effective, the problem is there were two
20:47
big problems that came out of this work. One is the standard monitoring standards that are required by the carbon credit authorities are much lower than the standards that academics researchers required. So there was actually a big gap between what auditors would say was happening within these types of interventions, and not just ours, and what the level of proof that an RCT would give you. So as a result there's quite a bit of controversy in this
21:15
sector about the appropriateness of using things like carbon credits for funding water and cookstove interventions. Since that time there's been a few other responses. One, in our program in Rwanda we brought in Tom Klostin, the London School of Hygiene and Tropical Medicine. Tom's now an advisor for the Mortensen Center. He's the leading expert on water filter evaluations. He now has a 30 million dollar grant from the NIH to do the definitive cookstove evaluation. So
21:42
his team was the right team to evaluate our program. Then further there's been other new mechanisms that have emerged in the past few years to try to pay to incentivize implementers on a performance basis. So for example there's things called ADALY credits, adverted disability adjusted life years, which are credits that are proportional to an estimate of health impact, which you can earn from water from any cookstove intervention without necessarily
22:12
measuring health. You measure proxy indicators you measure use, performance, efficiency and personal exposure to particulate matter and from that you can earn health credits that you can monetize and sell. So there's been other new interventions that have come in - or innovations that have come in to try to pay for performance. Another good example of this are our colleagues at Bridges to Prosperity, who are going to be putting pedestrian foot bridges all over Rwanda.
22:41
They've typically done that with private donors giving the money to do one bridge at a time, has anybody been involved with Bridges to Prosperity? Now, they've created a development impact bond or performance bond. B2P will use private investors to go build the bridges, the Mortensen Center will run the evaluation of if those bridges are impacting health and household income and a bond buyer will pay B2P, not based on the completed bridge but based
23:11
on actually proving that they've had an economic impact in communities. So this is a huge new sector in development that's trying to move towards performance-based payments. Okay, another sufficient ad lib to get me back on track. The entire Eastern Province and the yellow areas were the control sectors for the RCT. That's the stove, high efficiency cook stove, reduces wood fuel use considerably doesn't reduce air quality enough to
23:40
have a health impact by most measures, although the water filter had a phenomenal health impact but again has to be paired with extensive behavioral messaging. So while these are technology interventions, this is a public health program and so the program was driven by public health theory and methods and public health professionals. The technology was subordinate, technology has to work but it's really not sufficient. We had about 4,000 community
24:08
health workers working nationwide with us. This is the type of materials that we've developed. So this has evolved since then to a new program that Abby is getting involved in and Laura's getting involved in, where the Gates Foundation is taking what we know about how to promote household water treatment and we're incorporating that into a standing Ministry of Health Program for community hygiene club. So we're taking these learnings and applying them
24:33
forward in Rwanda. One other thing we did was use sensors, we put sensors inside stoves and filters to try to directly measure use. It's a big messy middle in public health interventions where engineers can design stoves, filters, bed nets that work. Public Health people have a pretty decent way, epidemiologists can measure health impact. Pretty sophisticated tools, not very sophisticated tools for measuring use or compliance or adherence whatever you
25:01
want to call it. Same type of water filter stove or latrine or bed net in similar context, while every high reported adoption, people tell you they're using them, even however fancy you get with your survey tools. If you pay somebody to sit there and take notes when people use a stove or filter or latrine, really high level of compliance. Yet wildly varying health outcomes, which up till recently has tended to implicate the technology. "Oh
25:30
water filters don't work" or "household water treatment doesn't work" when it appears based on data we've collected that the bigger challenge is whether or not people are really using the interventions. So to try to address this with an art program, we hid sensors inside our water filters and in one arm people knew the sensor was there and in another arm that didn't know the sensor was there. This is the impact we saw. Epidemiologist using the best
25:58
available WHO survey tools reported that people used about a liter and a half of water per person per day, which is exactly the correct answer. When people know what sensors in their household they're only using about a liter of water a day. When people don't know there's a sensor in their home, they're only using about half a liter a day. So there's two things happening here. People are reacting to be monitored, even if passively monitored and this persists
26:26
for a month, more than a month and there's biases in the reporting. The only time that our survey data has actually matched our sensor data is during the periods of time in the day that surveyors are there. So it's not that surveyors are Christmas treeing their reports, they are inducing the behavior.
26:47
So based on that RCT we doubled the number of households with clean water, reduce indoor air pollution by 70%, reduce diarrhea by 30% but also reduced cero conversion, which is a blood indicator of exposure to Cryptosporidium by forty eight percent among children under five. So these are probably conservative numbers. We also reduced respiratory disease by thirty percent, which is surprising because most other cookstove interventions haven't seen
27:14
this type of impact. We didn't have independent arms, we didn't have cook stoves over here and filters over here and both over here. There's a combined intervention. So the best we can say about this is either by being healthier from reduced diarrhea you're more resistant to respiratory infections and or we were effective getting people to cook outdoors. So while personal exposure didn't reduce directly, the indoor air pollution did and that would
27:41
have impacts on the non cooks. We can't desegregate that. Then we're also able to estimate both morbidity mortality associated with the intervention. Another place where we've taken this forward and Chantal is involved in this study, is within this very large-scale NIH and Gates funded program called HAPIN. It's 2018 and we actually don't know how healthy you can get if you don't use biomass stoves. The HAPIN trial is
28:10
funded by NIH and Gates led by M raids in four countries, Rwanda, Peru, Ecuador, no Rwanda, Peru, Guatemala, India, that's the fourth one. In over five years and the as an efficacy trial they're paying people, they're giving away free LPG stoves, free LPG whole bunch of other treats. Lots of follow-up, visits every single week, they're taking ultrasounds from pregnant women, they're following children's from minus six
28:41
months to age five years. It's an everything in the kitchen sink effort to establish the maximum benefit of clean cooking, which will be the point, the dose-response point on the curve that no intervention ever reaches again. Our piece of this puzzle is to use PM 2.5 sensors that are actively monitoring indoor air pollution, provides as feedback to households using this pictogram of a child's lungs with also the implication of the environmental
29:09
benefits. This was a human centered design process that Chantal led, to come up with both the triggers and the message. So you get persistent feedback on current air quality and then if we detect a baseline stove use, it plays music, it plays these Rwandan drums. It's basically an incredibly inexpensive smoke alarm and our study is to see if this kind of data can encourage exclusive use of LPG stove. Okay last thing: sanitation. We've also used
29:40
instrumentation within sanitation interventions both operationally in part of studies. This is Sanergy inc. Anybody worked with Sanergy? They're one of our practicum partners. Sanergy has about 800 of these franchise toilets across Nairobi in two different informal settlements and their business model is they provide the latrine, they service every day, franchisees, buy the latrines but then charge for use. Sanergy wanted us to help them
30:09
optimize their service area and their maintenance interval or their service intervals because they're losing money on servicing every day. So we put sensors inside them that included our ID tags. The sensors would estimate fill rates so service intervals for latrines the ID tags could be used by the franchise owner to request service like an out of schedule service and the waste collectors would be required to scan their badge to market service intervals
30:37
market service events and we integrated all of this within Salesforce and we're able to optimize and demonstrate the optimization of these latrines although it wasn't cost-effective the sensors are expensive it was still cheaper to just pay people to visit latrine every day. The other places where reviews these are across a number about eight different Gates funded health trials of latrine interventions in India and Bangladesh.
31:04
Billions of dollars have been spent by the Indian of Bangladesh government, by the Gates Foundation, by AID on trying to reduce open defecation in India and Bangladesh and in areas where it's a cultural norm so it's not shameful, totally okay, but we also know it's a big disease vector so it's a big contributor to the burden of disease. There's an article in the New York Times a couple years ago out of one of these trials that said latrines in India do not
31:32
improve health and it was missing a couple words because what we've seen across all eight of these studies and thousands of thousands of latrines we've monitored is that what people say they do is not really what they do. The only time again that the actual behavior matches what people say is when enumerators are there watching. And the other thing is it's just a huge error bar, some people are over reporting, some people are under reporting. There is no
32:01
decent survey tool for estimating compliance with latrine use. Okay that's it. I'm exhausted. That's a cross-section of our project. Some old ones and most of those are current ongoing activities. Any questions? yeah thanks! Churps was developed by, there's a program funded by NASA and USAID called severe and they've got five regional hubs around the world. One of them based in Nairobi, they are essentially GIS and remote sensing experts, trying to take remote sensing
32:46
data which is produced by NASA and the European Space Agency predominantly and apply it or develop tools for countries that don't have that kind of data resource. One of them is chirps which is a station corrected rainfall estimate that's been it's also used by a bunch of other things including the famine early warning network to look at crop yields and project project yields and forecast famine. Chirps is a rainfall estimate, it's one of their main
33:16
indicators. It's four kilometers. Yeah so in our data, we pick the nearest pixel to where our hole is but as you saw in that map we actually have huge geographic coverage of our sensors at this point. We're expanding this work now with Fusenet because even though the premise of Fusenet and severe is that many developing countries don't have local data sources like weather stations so this is a way to augment that but all
33:46
remote sensing is much stronger windows grunters. So our data is a ground truth source for these services now or we're working on that we're starting to work with them on that. There are questions.
34:11
Yeah so the idea is there's only basically one implementation of this now, it's for cook stoves and I came out of Berkeley where there was one there was one really mega cookstove study that was done in Peru called Inspire over multiple years and they developed the dose-response so personal exposure to PM 2.5 what's how's that what's your response and health impact and it's a really steep curve. So you really have to reduce
34:37
emissions considerably before you start really accruing health benefits. Berkeley took that model that they developed and created a tool where anybody can go in and input their study, their cook stove, their location, their compliant and high adherence rate and personal exposure which is an active measure and it'll kick out a daily estimates. So 1/8 daily one disability adjusted life year is one person not being sick for a year or 12 people not being sick for a month or 24
35:08
people not being sick for a week, right? Two weeks. And then the gold standard which is a voluntary training mechanism that got started using carbon credits, created a methodology that says we will then issue you implementer a daily credits so you get these 180 daily's because you prove this health impact, then you're going to go sell them to somebody. Which can be the world bank who has been a buyer revealing credits, there are foundations that are buyers,
35:38
it's not that sophisticated a market yet but the idea is towards a commodity that pays based on measured performance. Some of that has been superseded by taking conventional contracting performance-based contracting like bonds and applying into this context.
Here to help you learn
Sign in to keep a private conversation with your course tutor.
Ask me anything about the material…
Sign in to chat