Saving Millions for Just a Few Dollars

Sri Lanka achieved this, Levine argues, by making medical care free; training midwives so that today 97 percent of births are attended by a professional; building clinics in rural areas; setting up a system for sending difficult cases to city hospitals; and improving transportation networks.

What's especially interesting is that Sri Lanka in the 1990s was spending less each year on maternal health than it did in the 1950s -- 0.16 percent of its gross domestic product, compared with 0.28 percent earlier. But by the end of the 20th century, people had a better idea of what worked, and the nation was putting its money on proven winners.

"Others can take inspiration from the country's record," Levine wrote.

Measuring the Economics

The core of the project is a set of estimates of the cost-effectiveness of 319 different interventions. Most are for individuals -- drugs, vaccinations, surgical procedures, advice. Some are provided to the entire population, but only a minority of individuals will have their lives saved or improved. These include providing clean water and controlling mosquitoes, promoting abstinence and condom use, and banning cigarette advertisements.

The researchers in the Disease Control Priorities Project measure the effect of each intervention effect in units called "disability-adjusted life years," or DALYs. Each unit can represent a year of life saved when a fatal disease is prevented or cured. It can also be time without a disability when an unpleasant but nonfatal outcome (blindness, loss of limb, mental incapacity) is prevented or relieved -- although a year without a disability is weighted less than an extra year of life.

One of the chief insights of the original "Global Burden of Disease" report, published in 1996, was that poor countries had much higher rates of disability than rich ones. The updated report notes that 85 percent of "non-fatal disease burden" occurs in low- and middle-income countries.

"People living in developing countries not only face shorter life expectancies . . . but also live a higher proportion of their lives in poor health," its authors write.

By definition, an intervention is very cost-effective if it averts one DALY for a small amount of money. That's the case for speed bumps.

A study in Ghana showed that fatal crashes declined by more than 50 percent after the obstructions were installed at high-risk intersections. This is not a small effect; traffic fatalities are the 10th leading cause of death in developing countries, right below malaria.

Speed bumps are "probably the best thing and the first thing policymakers should do to lower traffic accidents," said Ramanan Laxminarayan, a health economist at the Resources for the Future think tank and one of the book's authors.

But neither he nor any of other experts assert that governments should simply pick the most cost-effective interventions and forget doing anything else. Issues of fairness, practicality and culture also have to be considered.

The value of interventions also differs widely, depending on how common a problem may be.

For example, tetanus shots are an incredible bargain in sub-Saharan Africa -- $14 per DALY averted -- while in Europe they're not much of a bargain -- $15,000 per DALY averted. That's because tetanus is real hazard for rural African newborns, whose umbilical cords are often cut with knives contaminated with dirt. Europeans simply do not have that exposure anymore.

That doesn't mean that tetanus shots aren't worth the investment in Europe. It simply means that a lot of them have to be given before a single case of the infection is actually prevented.

The report expresses the value of an intervention in DALYs averted down to the last dollar -- the average of many studies and many developing countries. But it is false precision.

The leaders of the Disease Control Priorities Project believe it is most useful to think in terms of rounded orders of magnitude -- what $10 will buy, compared with $100 or $1,000 or $10,000.