Q&A: Dr. Death on data for health

Professor Alan Lopez is an Australian epidemiologist. Photo: supplied

Professor Alan Lopez is sometimes called Dr Death. But the moniker is one Lopez embraces. For him the name demonstrates his four decades as an epidemiologist are making a difference. Today, his work focusing on cause of death statistics is at the core of the Data for Health Initiative, which is helping developing countries collect the data that improved health systems desperately need.

For 25 years, the Australian epidemiologist worked at the World Health Organization as director of the Epidemiology and Health Statistics section and chair of the board of Health Metrics Network — though he initially intended to be there for just three months.

Analyzing global health data, Lopez quickly identified the limitation of information on causes of death and thought there must be something more that can be done to improve it.

His ideas led to a grant from the Bill & Melinda Gates Foundation to strengthen the quality of cause of death data. This has had a cascading effect.

Michael Bloomberg, former mayor of New York City, became aware of Lopez’s work and asked him about improving the quality of death data for a focused number of countries. This led to the design of the entire Data for Health Initiative, in collaboration with Bloomberg and his colleagues. And Lopez introduced Australia’s Foreign Minister Julie Bishop to the program to expand funding and collaboration.

See more stories on Going for Goals:

How MIGA believes blended finance can help achieve the SDGs

Opinion: More money alone won't meet SDG 3

Innovative financing for health: Insights from the Grand Challenges Award Repository

Opinion: High returns await those investing in population health

How venture capital can help finance the SDGs

“It is an enormously satisfying feeling, personally, to see an area of scientific and methods development that I’ve been involved in all of my life is now finding an application to assist people in developing countries to agitate government for better data so their health policies can be better targeted and more effective,” he told Devex.

Now operating out of the University of Melbourne at the Melbourne School of Population and Global Health, Lopez calls himself the “technical mastermind” behind the Data for Health Initiative. While the implementation is conducted by Vital Strategies based in New York, Lopez and his team are a unique global capability providing knowledge, strategy, training and support to improve global death data to save lives.

Speaking with Devex, Lopez discussed his work on the Data for Health Initiative, the challenges it poses but the potential it brings. Here is the interview, edited for length and clarity.

What is the role of the University of Melbourne and your team in this program?

There are three main pillars of what is driven out of Melbourne.

One is that people who die in hospital, we take for granted that their doctors know what they died from. It’s important to know what they died from because those single statistics on the death of a person are aggregated and then used by governments to identify national health issues, including outbreaks of diseases.

So they have an important health function. Yet most doctors are not well trained to understand that public health function and correctly certify the cause of death.

We teach doctors medical certification of causes of death — how to correctly certify the underlying causes of death. We use standard procedures established by WHO that have not been promoted well in countries. We are the agencies of promotion for these good medical statistic practices in countries to get the right information for policy.

We are using country-tailored tools to work with medical associations and ministries of health within countries, and we are beginning to see large numbers of doctors being trained in these tools.

Secondly — what about the deaths that occur outside of hospitals without doctors? These account for the vast majority of deaths in developing countries. Understanding these deaths is really what is important for government and public policy.

We’ve developed a tool called verbal autopsy where we go to the family of the deceased with a tablet and ask a simple questionnaire for about 20 minutes about signs and symptoms and then using statistical methods and algorithms and IT advances, we take that interview and predict the most likely cause of death — there is a very good relationship between symptoms and causes of death.

This works. We are finding in repeated countries where we are rolling out verbal autopsy that it is quick and it’s highly accurate in getting the right cause of death. You’d be surprised that even the remote bush of Papua New Guinea has really good IT services and electricity. And it’s cheap. So the time is right to use these techniques to get information about causes of death that is useful for policy among populations where there is no doctor.

Thirdly, we are trying to build capacity in these countries to analyze data. These people have incomplete and incorrect data sets in front of them and we are helping them to extract the truth about who is dying of what in their populations.

You are operating in a diverse range of countries. What was the selection process in choosing countries and what are the different challenges they bring to this initiative?

The Department of Foreign Affairs and Trade had a significant say in the countries we work in.

Solomon Islands, Papua New Guinea, Indonesia, Philippines, Myanmar, Bangladesh and Sri Lanka are some priority countries for DFAT. And because Bloomberg was the former mayor of New York City, he wanted us to have cities. So we are also working in the city of Mumbai and Shanghai.

That is the Asia-Pacific component. But there are also countries in Africa and in Latin America.

Part of the dialogue in choosing countries was that I wanted the project to be able to demonstrate improvement across a range of initial development statistical systems.

So we have PNG and Solomon Islands, which are very rudimentary. And then we go across to Shanghai which has a mortality and cause of death data system not significantly different to Australia. But the problem in Shanghai is that they have not used the data for public policy. The Shanghai Center for Disease Control, our partner there, wanted a structured, formal audit of the quality of their data. So the University of Melbourne has done this and applied this expertise

From Shanghai, we are not seeing the same kind interventions that are needed in PNG or Sri Lanka, but rather a system that will benefit from a formal, rigorous statistical audit that will point to directions where they need to improve.

Of all the countries we are working in, the only one we are having trouble in getting traction because of the political environment is Indonesia.

It’s been 18 months, so I have had a long time to mull over and reflect on the reasons for this.

I think it is a combination of factors. The government is not really certain of the value of the project, despite being signed up to it. As the project is primarily technical assistance — focused, continuous technical assistance — Indonesia has been slower than other countries to appreciate. And it is just a very bureaucratic government system in Indonesia in terms of intersectoral responsibility. This project is very much intersectoral. We have many players in countries, and in Indonesia it has been difficult to get them around the table.

Countries around Indonesia are really making progress and we hope this will be influential.

They are hearing from other countries in the Asia-Pacific about progress under Bloomberg and I am beginning to see more momentum — it still hasn’t happened — but we are getting more frequent reports out of Jakarta that things are happening.

Are there plans to expand beyond the countries you are currently working with?

No! We are operating in 20 countries which is taking up all my waking hours.

In the past, the Health Metrics Network I was chair of at WHO tried to do too many countries too quickly. Here, at the initial discussions with Bloomberg I suggested about 50 or 60 countries because I am ambitious as I am nearing the end of my use-by date. But Bloomberg suggested four or five countries, saying it needs to be focused.

We went mid-way and settled on 20. But even that is a very big ask for this project.

I want to get it done well, get traction over a period of four years and show that we can make an impact and change data quality in 20 countries.

That will be a huge achievement.

At the end of the four years of the Data for Health Initiatives, what are your expectations that the interventions will be permanently embedded and sustainable within the systems of the countries you are working in?

We are running like mad, as fast as we can, and we’re in countries as often as we can. Our presence there, being honest brokers and seen by the countries as purveyors of good advice and good reasons and selling that to the countries as often as we can is not difficult.

But the hard part now is whether we will be leaving behind, in now 21 months, a permanent change. I think it will be a mixed scorecard.

I have no doubt Myanmar, Sri Lanka, Philippines and Solomon Islands are countries where we will have achieved our goals. In other countries, we will be close, but I would have liked more time — PNG will be one like that and Bangladesh another. And in other countries I will feel that the task was just too big that we have really not made the progress I would like to make. Maybe I was just hoping for more — Ghana and Tanzania are examples.

A mixed scorecard, but I have a team of 12 people here that are fantastic and committed, and I have great confidence that they will do what they can in the time available.

Read more international development news online, and subscribe to The Development Newswire to receive the latest from the world’s leading donors and decision-makers — emailed to you free every business day.