Data Distractions: Medicine and Metricide

Often it's not how we measure a subject but what we measure that really matters.

Debates about measurement efficacy and data quality tend to dominate discussions . While getting that right is important, the battle in many sectors is much more fundamental: overcoming opposition to any sort of measurement and then fighting tenacious rear-guard actions against objective data gathering and iterative improvement processes that might make someone or some organization look bad.

In medicine, as the following article explains, death - or rather, its absence - is often used as the baseline metric for demonstrating that things aren't going too badly. And while it may sound callous to say that death is beside the point, preventing it requires investment in data gathering and interpretation much further up the value chain. JL

Jeremy Smith reports in the New York Times:

Nonfatal conditions are responsible for a majority of health spending — and of human suffering. Every time we see a doctor and don’t die afterward, we demonstrate the inadequacy of counting deaths to track health.

What’s worse: Ebola or AIDS? Measles or malnutrition? Lung cancer or low back pain? As individuals, as a nation, and as a global community, where should we focus our time and money to improve lives the most?

The way we usually answer these questions is to count the number of deaths: The more people killed, the more important the problem. Counting deaths is so familiar that few have thought to question it. But death toll alone says nothing about how long people live, and good health is much more than not being dead.

Every year, for example, more than six million people worldwide die of stroke. Only about 300,000 people die worldwide of meningitis. So is stroke 20 times worse for humanity than meningitis? Not necessarily — because most people who die of stroke are age 75 or older, while those most likely to die of meningitis are infants. Death is an inevitability, but a death in very early childhood is a tragedy. All else being equal, saving the lives of infants should still be one of our global health systems’ top priorities.

Now consider everything that doesn’t kill people. If you are blind or deaf, anxious or depressed, disabled, disfigured or simply sick, your pain does not show up in death records or life expectancy statistics. Yet nonfatal conditions are responsible for a majority of health spending — and of human suffering. Every time we see a doctor and don’t die afterward, we demonstrate the inadequacy of counting deaths to track health.

Faced with these issues, health economists have in recent years developed new summary measures of personal, public and global health, perhaps chief among them a unit they call disability-adjusted life years, or DALYs. DALYs (rhymes with tallies) are akin to one of the advanced statistics — like Wins Above Replacement — that have revolutionized professional sports. Except, in this case, the wins and losses are years of healthy life.

DALYs are calculated first by measuring how many potential years of life are lost when a person dies. DALYs then incorporate the total years lived with disability — a measure based on international estimates of how much each nonfatal condition detracts from perfect health. Being paralyzed, for example, is considered close to half as healthy as perfect health, so every year you live with paralysis, you have lost the equivalent of half a year of healthy life.

Focusing on disability-adjusted life years may sound convoluted compared with simple lives lost. But it more closely aligns with most people’s intuitive sense of how health really works. Say, for example, that I’m in a paralyzing car crash at age 35, but survive until age 65, when I die (alas) from a heart attack. Then, assuming an ideal life span of 85 years, I lost 20 years of potential life to the heart attack. But, because I lived 30 years with paralysis after the car crash, the crash cost me the equivalent of 15 years of healthy life. My total health loss is 35 DALYs, a figure that accounts both for the health loss from the fatal heart attack and the health loss from the nonfatal car crash.

Just as baseball and other sports have been transformed by our understanding of new numbers, public and global health can be, too.

In late 2012, the British medical journal The Lancet published a new ranking of the world’s leading health problems comparing deaths and DALYs by cause. In terms of death, scientists reported, lung cancer kills about 200,000 more people than road injuries annually. But measured by DALYs, road injuries are almost two and a half times worse for humanity. That’s because most fatal victims of lung cancer are in their 60s, 70s and 80s, while those most likely to die of road injury are in their 20s and 30s — and road injuries cause almost 40 times more disabilities. If you are an international policy maker or aid agency choosing how much to invest in road safety relative to antismoking campaigns, that’s vital information.

The same kind of analysis works even better at a national level, where most health spending takes place, because current health care gaps usually correlate much more with the leading causes of DALYs than the leading causes of deaths. As Mexico moved to a universal health care system in the last decade, it used this type of analysis to prioritize those treatments — like medications for childhood cancers, and emergency care after a car accident — that reduce DALYs the most.

Australia has used its own DALY calculations to direct close to $900 million in public health program spending since 2009, focusing successfully on curbing tobacco use, childhood obesity and diabetes.

Now people everywhere can bring “Moneyball” to medicine. A few months after releasing their global numbers in The Lancet, the same scientists supplied the underlying figures for 187 nations. These statistics will be updated again later this year. At last report, in the United States, measured by DALYs, the third-largest health problem was low back pain. Fifth is major depressive disorders. Eleventh is neck pain. Thirteenth is anxiety disorders. None of these maladies kill anyone directly, so they don’t even show up on a list of leading killers. But they still cause huge amounts of pain and suffering, and cost our economy billions of dollars in lost productivity.

When will low back pain get the research funds and attention given to lung cancer, just below it in a DALY ranking? The toll from major depressive disorder, No. 5, is estimated to be 20 percent worse than that from stroke. Why don’t we promote early detection in the same way, on public billboards and ad campaigns? Health loss from anxiety disorders is estimated to be 80 percent higher than that from breast cancer. Do advocates for anxiety treatment even have their own colored ribbon?

These are provocative questions, and new statistics in any field are inherently threatening to the status quo. But just because we have always looked at things one way is no reason not to consider alternatives, especially when it comes to something as fundamental as improving how we all live and die. We don’t have to take new measures like the DALY as the final word, but they highlight areas for health gain we might otherwise miss. Advanced stats are too important to leave to professional sports teams. We can all try to choose the prevention, detection and treatment strategies that best add years to life and life to years.