By creating a new and innovative way to look at massive amounts of patient data worldwide, one man hopes he can change the way public health crises are managed, as Cynthia Graber reports.
Imagine you are a foreign aid worker trying to persuade a senior politician in a developing world country to introduce a pneumococcus vaccination programme. It’s not just a case of stressing how the bacterium causes diseases including pneumonia, meningitis, and sinusitis, and kills over a million children under the age of five every year worldwide. The politician has to decide how to allocate scant resources. How does the death toll compare with malaria and AIDS? Aren't road traffic accidents a bigger problem? Has vaccination been a success in neighbouring countries?
These statistics exist, but you don't have the relevant reports and academic papers to hand. And even when you do have the information, a list of numbers may not the best way to express the strength of your case.
By creating new and innovative visual displays out of oceans of data, Christopher Murray hopes his tool can change this situation for the better. Called GBD Compare, users can rapidly determine which diseases are most harmful to children in Africa, or view how the developing and developed worlds compare in terms of heart disease, all with a few clicks of a computer mouse.
The data viz tool processes data from the Global Burden of Disease (GBD) report, which compiles statistics, charts and graphs on causes of death and disease. “The thing that’s really neat about the visualisations is they allow people to see the problem in context – in the context of all the other problems, how it’s changing over time, how it compares to other countries,” says Murray, director of the Institute for Health Metrics Evaluation (IMHE), based in Seattle.
When Murray shows this tool to people outside the academic world of public health, Murray says, they immediately get it. “That just totally changes who you can engage in a thoughtful discussion about what are the key health problems and where they’re going,” he says.
The new tool has the enthusiastic backing of no less an advocate than Bill Gates, and, just three months after its launch, it's already leading to changes in health policies.
Murray’s interest in international public health was sparked at age 10, in 1973, when his father, then a cardiologist at the University of Minnesota, decided to take a sabbatical year and volunteer in Africa. The family raised donations, flew to the UK, bought a couple of Land Rovers, and, recalls Murray, drove out across the Sahara to eastern Niger to take over a hospital that had been built but never opened.
“We all had jobs,” says Murray. “Mine was – being at the bottom of the totem pole there – in charge of the pharmacy.” While his father met with patients and wrote prescriptions, the young Murray organised and stored the pills and dished the appropriate ones out. That year overseas was “extraordinarily influential on everyone in the family,” says Murray. His parents repeatedly returned to volunteer in Africa, and both his siblings now work in health. Murray went on to study medicine and health economics, eventually working for the World Health Organization (WHO) and heading Harvard University’s Initiative for Global Health.
In 1991, the World Bank sought his help in creating a comprehensive policy document on international health problems. The result of Murray’s partnership with Alan Lopez at the WHO was the first GBD. The GBD released official health statistics on the burden of disease, beginning with 1990 data on 50 causes of death across seven international regions. A decade later, the team once again compiled health data for the WHO, along with legions of maps, charts, and graphs, based primarily on what the authors thought would be useful.
The latest version of the GBD is both bigger and better. For one thing it includes more than a billion pieces of information covering 291 causes of injury, disease, and death, for 187 countries, measured in 1990, 2000 and 2010 (published in a traditional form in a series of studies in the Lancet in December).
More radically, the team (now overseen by the IHME, which Murray founded in 2007 and still heads) has now made the information accessible to a wider audience by presenting it in visually engaging ways on an interactive website. Launched in March, their visualisation tool shows causes of disease, injury and death within a rectangular pie chart made up of blocks whose sizes are proportional to the numbers they represent.
Categories are highlighted in different coloured blocks, which are sub-divided into specific causes in different shades, according to whether they are on the increase or decrease. Red signifies communicable diseases, such as malaria or diarrhoea, as well as maternal, neo-natal and nutritional disorders. Blue is for non-communicable diseases, including cancer, heart disease, and diabetes, while injuries are represented in green.
Users can select the entire developing world, for example, and pull the lever from 1990 to 2010 to see that communicable diseases overall have decreased as a percentage of mortality. They can view metrics across the whole world, regions or individual countries, focus in on specific time frames and ages, or look at outcomes for men or women only. Other visualisations allow comparisons and rankings of causes of death, disease and injury in different countries.
Projects on this scale have only recently become possible thanks to the enormous increase in data, greater computing power, and new ways of doing statistical analyses, according to Peter Speyer, data director at IMHE. "The technology was not advanced enough even five years ago to have an easy way to visualise all this data," he says.
The initial spark of inspiration came in the autumn of 2010 when Kyle Foreman, then working as a research fellow at the IMHE as part of his Masters degree in public health at the University of Washington, got fed up with boring charts. Foreman taught himself D3, the program the team still uses to create the visualisation, and made a prototype for internal use.
“It didn’t take long before we realised just how powerful these tools are,” says Murray. “That’s when we realised that we needed to invest in doing this in a way that everyone could use.”
The effort to collect and interpret the data was immense. Speyer says that in addition to the nearly 500 official authors from 50 countries, the IMHE worked with thousands of organisations around the world to access the data.
Despite that, it’s not perfect. The results are only as good as the data collected on the ground, and Murray has heard occasional criticisms. “There are always people saying, ‘No, you haven’t done a fair job for blindness, or you haven’t got cholesterol right in Indonesia because you missed my favorite study’,” he says.
That’s an easy fix, he says. If and when new data becomes available, his team can integrate it. There are also plans for more general updates at least once a year. Some problems, however, can’t be solved by larger data sets. For instance, some scientists claim that malaria is over-diagnosed in hospital deaths in Africa. “That’s a tough one to deal with," adds Murray. "You have all the data on one side and expert opinions on the other. The bottom line is, we won’t really know until more studies are done.”
The tool is already leading to policy changes. Researchers in the UK teamed up with Murray to produce a report on how it compared with 19 other wealthy countries, in both 1990 and 2010. Published in March in The Lancet, the study demonstrates the UK has successfully reduced cancer deaths, but lags behind in addressing cardiovascular disease and tackling the increasing rates of alcohol- and drug-related illnesses.
“From a policy perspective, that’s extremely important,” says Sir Michael Richards, director for reducing premature mortality at National Health Service England. “Instead of saying how fantastic it is that we’ve improved, this says that we’ve still got a lot to do if we want to be among the best in the world.”
Chinese scientists have also worked with the IMHE to compare health statistics in their country today with those from 20 years ago, and to other major economies in the G20. In a study published in the Lancet earlier this month, the team found that while China has made rapid improvements in reducing infant mortality and improving life expectancy, it is facing a growth in diseases related to poor diets, high blood pressure, tobacco use, and environmental and household air pollution.
The tool has proven so useful that other countries are now working with the IMHE to collect and analyse their own health outcomes on the regional level within each country. Murray says Indonesia’s minister of health and office of the President are “very excited about the visualisations” and want to drill down into local results with a larger team of Indonesian researchers. China, Australia, Brazil and the UK are also interested in more detailed, within-country versions, and the government of Saudi Arabia has begun a new collaboration with the IMHE to track the health of its citizens and inform future policy decisions.
For now, the website is designed with policy makers in mind – government officials and scientists. Some curious non-experts may feel a little overwhelmed by the volume of information available. For these people, the team is considering adding a simpler layer that will be even more accessible for the general public, “taking away all the complicated controls, the great degree of detail, and just hammering home key points,” says Speyer. The goal, he continues, is to “make it more intuitive and more fun,” to encourage even more people to engage with the data.
Bill Gates, whose foundation funds both the IMHE and the GBD, raved about the site during a speech to mark its launch at his foundation's headquarters in Seattle, calling it “one of the best efforts that has been done” in data visualisation.
He went on to stress how important it is in the work of the Bill & Melinda Gates Foundation, which seeks to improve healthcare and reduce extreme poverty internationally, by, for example trying to convince the governments of countries including India to adopt the pneumococcus vaccine.
"When you have a tool like this, you can even drill down and see the various studies that have been built up to support this information," said Gates. "This is going to help us tell that story and get better health policies more rapidly than we’ve been able to do in the past.”