So in the paper this week we have coverage of the dismal state of global vital registration data, that is, the data that governments collect on cause of death among their citizens and report to the WHO.
I am more sympathetic than most to concerns about the state of national data on health, however it’s possible that my experience of working with these data has left me in a sort of utilitarian existential crisis. We all say over and over that we need these data in order to decide where to expend our efforts in global health, and yet I have a creeping sense that we’re all squinting terribly hard down the wrong end of the telescope.
There are intractable problems with attributing deaths to single causes. It so happens that both of my major diseases of interest lose out in the present system – people with HIV who die of tuberculosis are counted as HIV deaths, while people who die of hepatitis-C associated liver cancer are counted as cancer deaths. What we chose to count as a primary cause of death is not a perfect science, but the way we count has far reaching impacts in a system where everything comes down to numbers; dollars spent, lives saved.
Lopez and colleagues have coined the term “garbage codes” for cause of death reporting that is tautologically stupid, for example when the cause of death is coded as “heart failure”, which is something akin to saying a car accident was caused by a collision with a tree. You have to take a few steps further back in the chain before you get to anything meaningful. And yet there’s a sense in which saying somebody died of tuberculosis is just as much a nonsense as saying they died because their heart stopped beating.
Nobody but the supremely unlucky dies of tuberculosis in a country with a functional economy and an adequate health system, any more than a perfectly healthy person’s heart just stops beating one day. So when I hear my colleagues say “We need to know how many people die of TB every year before we can do anything”, more and more I think to myself, “Really, do we?”. Does it actually matter if it’s 500,000 or 1.5 million? Does it matter whether or not TB outpaces maternal mortality in young women? However you count them, young people are dying needlessly in poor countries. They are dying for want of basic necessities, simple preventative measures, and access to adequate healthcare when catastrophe strikes. Whether they die in childbirth, or of TB, or from diabetes, the underlying cause is the same – poverty and weak health systems.
Certain problems demand targeted responses, of course. Deaths in childbirth require different interventions to deaths from TB, but the basic infrastructure required to deliver care is the same, and in many poor countries in 2015, it’s not there. When we say “We need to know where to direct our resources”, the choice should never really be between maternal mortality, TB, and diabetes, because it’s rare that one particular disease is neglected while the rest of a health system thrives. The lack of funding for TB services in high burden countries isn’t an oversight that can be corrected with better data – it’s a sign that the health system itself is failing to cope.
It’s easy to argue, as Lopez and Demaio do, that this data is particularly shortchanging those dying of non-communicable diseases. Perhaps that’s true. But once you start using data to make moral arguments you can make it say anything you want. You can argue for the disease that takes the most lives, or the most healthy life years. You can lump all your cancers together or split them up into ever smaller categories. You can tell people that such and such is now “the leading cause of death in Nation X”, but so long as we are mortal we must all die of something. It would be a triumph if Alzheimer’s became the leading cause of death in Australia – it would mean that the average person had reached old age after avoiding cholera, malaria, tuberculosis, premature cardiac trouble, suicide, and accidental injury.
We could argue for the rest of our own lives about whether we should spend our aid money on tuberculosis or ebola, on maternal deaths or diabetes. Such arguments are compelling, and sit comfortably within the vertical approach to health that we all know to be disastrous. We can continue to have these debates, but much like a doctor scribbling nonsense words about heart failure and respiratory arrest, we will have failed to apprehend the underlying disease.