In Alexandre de Figueiredo and colleagues take a step in
In , Alexandre de Figueiredo and colleagues take a step in this direction with their time-series analysis of trends in vaccine coverage and a suite of socioeconomic and demographic factors across 190 countries over 30 years. The main aim was to gauge where and when acetanilide coverage might fall below levels that are safe for prevention of epidemic transmission, and to correlate such decreases with underlying socioeconomic and demographic factors. The investigators used WHO–UNICEF coverage estimates of three doses of diphtheria, tetanus, and pertussis (DTP3) vaccination and obtained . By use of a statistical framework based on Gaussian process regression and a newly developed vaccine performance index, which forecasts that vaccination coverage will be at a safe level (90%) in the near future, the analyses yield some interesting results next to the basic fact that worldwide coverage has increased. For instance, gross domestic product (GDP) and government health spending correlate most strongly with vaccination coverage in Eastern Mediterranean countries between 1980 and 2010, whereas primary school completion correlates most strongly with vaccination coverage in Africa (more so than does GDP). The analyses also provide a list of countries with high to low vaccine performance indices, showing that many of the countries at the low end of the list are in sub-Saharan Africa, the Indian subcontinent, and southeast Asia. From a global public health perspective, the list provides an objective measure that can be used to prioritise countries or regions where efforts to increase vaccination coverage are expected to be most efficient. Notably, although vaccination coverage correlates well with GDP and schooling in many regions of the world, this is not the case any more in Europe and, to a lesser extent, North America. Here, no socioeconomic factors correlated with high coverage, and one argument is that once the basic necessities of life are available, other factors such as social attitudes to vaccination might become more important. Because of the focus on socioeconomic factors, de Figueiredo and colleagues\' findings cannot add much more than speculation to this argument, and it will be interesting to see the outcomes when the set of variables is extended to encompass social indicators that might shape vaccine hesitancy. With a focus on global immunisation patterns and the relation with socioeconomic factors, the investigators have painted a picture with broad brushes, one that cannot hope to unravel patterns that are important in specific regions or countries and for particular diseases. Examples are the difficulties encountered in the push towards eradication of polio in Afghanistan and Pakistan driven by war and extreme ideologies, the struggle to achieve elimination of measles in Europe given vaccine refusal in clustered religious and anthroposophical groups, and the perceived lack of safety of the human papillomavirus vaccine fuelled by adverse events after vaccination. These examples show that a full understanding of local coverage patterns requires data and analyses at the local level. Technically, the vaccine performance index might have to be developed further. The index provides an aggregate measure that takes both vaccination coverage and changes in coverage into account. One could argue that in its current form the vaccine performance index punishes countries with systematically high but volatile vaccination coverage (eg, Norway) quite strongly. In fact, low vaccine performance indices in these countries might be due to reporting bias or small sample sizes (in cases when a national registry is not available). Indeed, in developing countries, precise figures for vaccination coverage are often not available, and estimation of vaccination coverage is not always straightforward. Future developments will probably have to incorporate the uncertainty in vaccination coverage estimates to prevent artificial increases in the precision of the correlations.