After leaving Eshisiru Secondary School in Kakamega, where he was a volunteer teacher in 1985 and 1986, Michael Kremer made a fateful return journey that has come to redefine his life.
“I did my PhD (Harvard, 1992) in Economics, and then got a job where I was now earning enough money, and I could afford to come back to Kenya. So I just came on holiday,” he explains.
He reconnected with some old friends. One was a former headteacher now working with a small NGO, which supported seven schools around Kakamega with textbooks, deworming pills and hiring of teachers.
When the friend mentioned this, Kremer suggested that he chooses twice as many schools and then begin working in half of them and later compare to measure the impact of what they were doing.
This approach is called a randomised controlled trial. It is similar to the trials that pharmaceutical companies use to test new medicines.
The first results released in 2004 were disappointing. Kremer had expected schools that received textbooks to perform better. They did not.
Children who received deworming pills had better health and education outcomes, including a 25 per cent decrease in school absenteeism compared to those who did not.
The field experiment helped Kremer, then a lecturer at Harvard University, share the 2019 Nobel Prize in Economics with two economists from the Massachusetts Institute of Technology, for pioneering the use of field experiments to study which policies best improve the lives of the poor.
These results inspired Kenya and India’s ministries of health and education, and private donors, to support the now highly successful school-based deworming programmes.
According to some reviews, this was the largest public health programme ever attempted.
Kremer did not stop there. He was able to track 84 per cent of the Kenyan pupils after 20 years to understand how they were faring in life.
“Individuals who received two to three additional years of childhood deworming experience an increase of 14 per cent in consumption expenditure,” he said in results published last year.
There is also an increase of “13 per cent in hourly earnings and nine per cent in non-agricultural work hours, and they are nine per cent more likely to live in urban areas”.
“Given deworming’s low cost, a conservative annualised social internal rate of return estimate is 37 per cent,” he said.
The Ministry of Health has since committed to continuing to support deworming activities in schools.
Prof Kremer is now a professor at the University of Chicago. He spoke with the Star's JOHN MUCHANGI about the new results.
Water treatment, just like deworming medicines, is very cheap. If you buy a bottle of Waterguard, it's very inexpensive, that lasts for a month for a household
What was the government's response to the surprising 2004 results of the study in Western Kenya?
When I presented the results to the Permanent Secretary of Education, he was very excited about them and wanted to move things forward.
The World Bank was very supportive. And the Kenyan government eventually launched what became the national school-based deworming programme.
And this is a combination of the Ministry of Education and, at that time, the Ministry of Public Health. Because we had had such strong evidence of impact, I also helped share that evidence with donors, such as the World Bank. The evidence meant there was a very strong case, and donors have provided some of the costs that would not otherwise be covered.
So my understanding is that the Kenyan government is currently reaching about six million children every year.
The study was done in Western Kenya. Does it have relevance for other settings?
There are many places in Kenya where worms are an issue. Not everywhere in Kenya, the environment has to be right for the worms. But that includes Central, Coast. Some of the initial work was done in Busia but it applies all along the lake because the Schistosomiasis and other things make it an area where there are lots of worms.
Dr Charles Mwandawiro from Kenya was very involved in setting up the Kenya programme. We spoke to people at the World Bank about this, in Kenya, and World Bank programmes in other parts of the world. And they communicated with state governments in India. They also have a worm problem. And they were interested in the Kenyan model. And so Dr Mwandawiro went from Kenya to India to brief them on how these programmes could be established and run through government systems.
So Kenya played a direct role in the establishment of this in the state governments in India. And now [India PM] Narendra Modi has made this a national programme.
And you followed some of the children into adulthood. What did you find?
The initial results [released in 2021] were on staying in school. And we saw that there were about one-quarter of the absence from school was reduced.
Later, we followed them into young adulthood. It's now 20 years later, so it's been been a while. We're finding that they’re both earning and consuming about 13 per cent or 14 per cent more. The cost of this [deworming] programme is about 50 US cents (Sh57) per child per year. Maybe the price has even come down.
And if you can invest 50 cents (Sh57), and then wind up earning 13 or 14 per cent more, that's a wonderful investment and a great rate of return. And it's even a greater rate of return for Kenya as a whole.
If people are earning more, they're paying more VAT, and the government's making the money back, so it's even good for the government budget.
So the government could have borrowed the money for the programme and it would still be ahead.
You have received a lot of praise for the use of these field experiments to fight poverty. Have you encountered criticism of the same?
Yeah, some people point out that you can't answer every question like this.
Some very important questions can't be analysed with these techniques, but I say this is one technique among many techniques. Like anything else, you know, there's the initial version, and then it improves over time.
So as I mentioned initially, there were seven schools in treatment. That's a very small sample size. There are many technical refinements over time. We should be working constantly to improve things.
You have done a similar experiment on the benefits of water treatment
As a background, lack of access to safe water is a big issue in Kenya and also in many other parts of the world.
So there are more than 2 billion people who still drink water that's contaminated with faecal material. It's estimated that one and a half million people are dying from diarrhoea every year, mostly children.
Water treatment, just like deworming medicines, is very cheap. If you buy a bottle of Waterguard, it's very inexpensive, that lasts for a month for a household.
It's like deworming, it's a problem that sometimes falls between sectors. So the Ministry of Education and Ministry of Health need to cooperate as they did in Western Kenya.
Most studies are not designed to measure child survival, they’re designed to measure diarrhoea. So we searched for every study that was a randomised trial on diarrhoea. And then we wrote to the authors to say, do you have any data on child survival? Some of them did.
Now, any one study with, you know, one case of child death or two cases, that's not enough to put on their article. But if we combine all of this, we had enough evidence to say that there was a statistically significant reduction in mortality after water treatment.
And it was much much bigger than I would have expected and I think many people would have expected. There were about 15 studies and across these 15 studies, on average, we found about one-quarter reduction in child death. So this is all-cause mortality.
That means the combination of very inexpensive treatment plus such a large effect means it's very cost-effective. If the government or donors are looking for how can they save the most lives with a limited amount of money, this is right up there with other highly cost-effective things like vaccination and mosquito net distribution to families with young children.
We guess that about 500,000 lives could be saved, at a cost of about $1.5 billion, or Sh150 billion. These are big amounts of money, but these are amounts of money that the world spends. Obviously, that's a target for the future, but it just suggests the potential impact of this. I just presented some of these preliminary results at a Centre for Global Development event.
We guess that about 500,000 lives could be saved. Obviously, that's a target for the future, but it just suggests the potential impact of this
What is the proposed way to achieve this?
Kenya has really been a leader in innovative solutions to water treatment. So there's this one approach, which is dispensers. This is very common in Western Kenya, where there are many naturally occurring springs near Lake Victoria.
Just set up the dispenser full of Waterguard, but instead of having to pay, whoever's collecting water, they just turn a tap and it releases the right amount for 20 litres jerrican. And then when they're walking home, it mixes into the water, and then the water is safe. This gets about half more people using it.
Currently, about 2.2 million people have access to the dispensers in Kenya, and another two million in Uganda, and Malawi. That's done by the NGO Evidence Action.
But there's another approach that I think is very promising, which is the use of coupons or vouchers. So the idea is that a household gets vouchers for Waterguard. And then they can use the voucher to redeem the Waterguard for free at clinics, possibly even at shops.
In our discussions, the Ministry of Health suggests we use the ministry’s channels. An advantage of this approach is there's already a lot of emphasis on maternal and child survival. Women are already getting some things like antenatal care, vaccines and mosquito nets for free. So they could also pick up those vouchers at the clinic.
That has the advantage also of targeting the young children, who are most at risk of diarrhoea.
Not everybody is going to use the Waterguard. Some households will choose not to, maybe they don't like the taste.
With the coupon system, if the households go to the trouble to redeem the coupon, they will actually use the Waterguard, which is, obviously, one brand.
I think, one natural next step is to try to work within government systems and see if this can work at scale the same way. And if it does, then I would hope the government would consider this for further action, but obviously, that's a government decision.