In an editorial in the IJE, David Leon examined life expectancy trends in Europe over the last half century, focusing on the differences between Western and Eastern European countries. Towards the end of the piece, he raises a question that I find particularly interesting: Why have Eastern and Western European life expectancy trends been (mostly) running in parallel since 1990?
The fact of the east-west mortality gap in Europe is a well-established fact. Its source can be traced to the stagnation in life expectancy gains behind the Iron Curtain in the 1970s and 1980s, while Western trends continued upward. It was particularly the working-aged population’s mortality rates and cardiovascular mortality rates that remained high in the East compared to the West. After the fall of the Berlin wall in 1989, life expectancy in Central Eastern European countries regained upward momentum. (The situation was different in the Baltic States, where the chaotic disintegration of the Soviet Union in 1991 precipitated a mortality crisis that lasted until the mid-1990s.) However, there is one curious thing about these newly upward trends in life expectancy in Central Eastern European countries: For the most part, they run in parallel to life expectancy trends in Western countries.
One of the striking features of Western European trends is how similar they are. Moreover, since the early 1990s the CEE countries have been following the same rate of improvement, with some suggestion that this is now also the case for females in the former-Soviet Baltic States. Why is this? The differences between countries in their health service organization, national wealth, culture and associated individual behaviours are many. Some of these may explain the vertical displacement of each line, but they do not seem to have a strong effect on the rate of increase in life expectancy.
Why is it the case? Leon identifies a few hypotheses that may be driving this parallelism in life expectancy.
This parallelism may be a product of the sort of diffusion of knowledge and ideas that occurs in a connected world implied by Preston, whereby improvements in public health or medical treatment and personal behaviours all make small but incremental contributions whose net effect is to reduce mortality. At any point in time the precise combination of factors that are driving life expectancy upwards may differ between countries. There are of course more distal influences that may also contribute to these parallel increases in life expectancy. These reflect longer term parallel improvements in levels of education and standard of living of most European populations.
I would organise established drivers of life expectancy trends into three groups: prevalent behaviours, available mortality-reducing technlogies (including medical and public health interventions), and broader societal factors.
The proximal drivers of mortality in Eastern Europe have been thoroughly investigated and alcohol and tobacco consumption identified as the key culprits. Other factors in this group include diet and stress, but these are less well supported. (This may be due to historical diets and mental health being more difficult to ascertain.)
The availability of advanced health care and public health interventions were key to the West’s reduction in cardiovascular mortality rates. It is therefore not surprising that healthcare quality (whatever this means, perhaps reflecting the available mortality-reducing technologies) was found to be associated with both the stagnation in life expectancy in Eastern Europe before 1990, as well as with its renewed upward trend as it improved. The way that this particular hypothesis was most commonly investigated was through the use of amenable mortality as a proxy for health system performance.
The last group is represented by the well-established association between GDP and life expectancy, the so-called Preston curve. In a nod to the previous driver, Preston also noted that at each level of wealth, life expectancy varies, and suggested that some process of diffusion of medical and public health knowledge could be driving this difference. Of course this group also includes the prevalent level of education, the level of inequality, and other societal characteristics.
All three groups interact and overlap. I would argue that it could be useful to think of them in the framework of complex causality as necessary but insufficient conditions of life expectancy growth. How would this perspective help us better understand the issue at hand?
When the barriers to diffusion of knowledge were eliminated with the fall of the Iron Curtain (and the collapse of the Soviet Union), we may have expected that Eastern European countries would rapidly introduce tobacco and alcohol policies, blood pressure medication, and coronary bypass surgery that have proven their effectiveness in reducing cardiovascular mortality in the West. This should have then lead to a precipitous drop in mortality in Eastern Europe, a rapid convergence in life expectancy between the two country groups, and the elimination of the east-west mortality gap. Why did this not happen?
The example I provided focuses on the diffusion of mortality-reducing technology, which is just one of the necessary but insufficient conditions of mortality convergence. Its diffusion from West to East is of course important, but its effect on life expectancy is mediated by how this knowledge is operationalised in a given health system and whether it is made broadly accessible to the population. This is the part that makes it an insufficient condition for mortality convergence.
We can all imagine how this works in practice. Knowing what interventions reduce mortality is different from a robust implementation of these interventions. The latter requires (at least) appropriate government policy, trained personnel, and physical infrastructure. Another missing piece is making these implemented interventions widely available to the public, as well as the public using them. This requires a well-functioning health insurance system that supports people in seeking treatment, as well as sufficient levels of health literacy and disposable income to engage in health protection activities, for example a healthy diet and regular exercise. These considerations clearly place us in the domain of health in all policies, including education, social welfare, and economic policy.
How does seeing mortality convergence as an outcome of interconnected necessary but individually insufficient conditions help us forward?
First, this perspective predicts that the growth in life expectancy in Eastern Europe in the 1990s and 2000s will be approximately as fast as it was in the West in the 1970s and 1980s. I would argue that there is no reason to expect that improving accessibility to mortality-reducing interventions should be much easier or faster in Eastern Europe in the 1990s and 2000s than it was in Western Europe in the 1970s and 1980s. That is because it is not the technology but governance (passing legislation and steering investment) and widespread uptake that are the key bottlenecks. This prediction can be tested by comparing the life expectancy trends Western countries after 1970 with those in Eastern countries after 1990.
Second, this perspective predicts rapid convergence in life expectancy between the East and the West to occur only after all the necessary conditions are in place and the uptake of the mortality-reducing interventions rapidly picks up steam in the population. This prediction can be tested if we examine cases that have orchestrated growth in life expectancy that surpassed parallelism: East Germany, Slovenia, and more recently Estonia. As subnational life expectancy data becomes more readily available, certain regions in Eastern Europe will also be categorised in this group.
Finally, this perspective predicts barriers of life expectancy growth, not usually examined in the literature: poor legislative capacity, poor investment capacity, and poor individual capacity to utilise health care, etc. These factors can be explored by examining cases where life expectancy continues to stagnate.
Perhaps the necessary but insufficient conditions perspective explains (at least to some extent) why life expectancy trends between East and West Europe are not converging. However, I do not think that it can explain why very different countries experience such similar rates of life expectancy growth. Perhaps the trends were not so very similar after all? Or perhaps there is a mechanism at play that sets an upper limit to the rate of life expectancy increase? I hope to return to these questions in a future post.