disparities1Those working in global health and development are used to hearing about inequalities and disparities in health outcomes. We know that poor or disadvantaged populations around the world are more likely to have poorer health outcomes across almost all measures of health. It is promising to see large donors such as the Ford Foundation recognizing this and shifting their funding to focus on addressing global inequality.
But many questions remain about the best way to achieve the health outcomes we seek. For example, are some interventions having positive outcomes while exacerbating existing disparities within a population? Is this necessarily a bad thing? What if specifically focusing on disparities can sometimes slow overall progress to achieve a public health goal?
Let’s look at tobacco control
There have been many successes in tobacco control over the past few decades. Many of the strongest tobacco control measures have been enacted in high-income countries, and as such, now nearly 80% of the more than one billion smokers worldwide live in low- and middle-income countries, where the burden of tobacco-related illness and death is heaviest.
Recognizing this, large donors such as Bloomberg have entered the global tobacco control space with the Bloomberg Initiative to Reduce Tobacco Use. The initiative has been effective in reducing tobacco use worldwide. Since its launch in 2007, 32 countries have passed 100% smoke-free laws, 20 countries have passed laws requiring graphic warning labels on tobacco packages, and 16 countries have passed bans on tobacco advertising and sponsorship. Additionally, 2.3 billion people are now protected by at least one of six tobacco control policies advanced by the WHO, which are known as MPOWER. Such overall successes should certainly be celebrated, but breaking down the numbers further can tell a different story.
Should it matter if these programs are helping the rich more than the poor?
The success of many international tobacco control initiatives is commonly measured by declines in smoking rates – which tell a compelling story. For example, Brazil saw a drop in smoking prevalence from 31% in 2000 to 17.2% in 2008. At the same time, we see that people making less than a quarter of the minimum wage smoked at a rate of 23.1% compared with those 13.3% of those making two minimum wages or more.
We see this trend across a concerning number of countries, where lower socioeconomic classes continue to smoke at much higher rates while the prevalence among the highest classes is falling. In the Philippines, the rates of smoking in the lowest socioeconomic quintile fell from 42% to 38% between 2003 and 2009, while rates in the highest quintile fell from 28% to 17%. That is a nearly 300% faster decrease in smoking rate for those in the highest quintile compared to the lowest. In Bangladesh, the lowest socioeconomic class smokes at a rate of 59%, compared with 28% in the highest. Even more concerning is that while the rate in the highest quintile has fallen since 2003, rates of smoking in the lowest quintile have actually increased from 54% in 2003.
Even in Australia, which has led the way with some of the strictest tobacco control laws in the world, we see similarly shocking statistics. Though overall smoking rates have been declining, data from 2013 shows that people living in areas with the lowest socioeconomic status were three times more likely to smoke daily than people in the highest (19.9% compared with 6.7%). Similarly, between 2010 and 2013 there were no changes in smoking rates among people who were unemployed, though significant declines in daily smoking rates were seen nationally during that period. Indigenous Australians were also two and a half times as likely as non-Indigenous Australians to smoke daily (32% compared to 12.4%).
What this means is that when programs improve population averages, these overall successes may not necessarily be benefiting vulnerable populations. In fact, studies looking at tobacco control measures in Europe have found that untargeted smoking cessation interventions may have contributed to reducing adult smoking but are likely to have increased inequalities in smoking. The use of public education to reduce smoking and tobacco use harms is a common approach that is likely to exacerbate disparities as the more educated are more likely to respond to public education about smoking cessation, while more targeted education efforts with specially targeted content will be most effective for sub-populations with the highest smoking rates.
As noted by the UN’s panel report on the post-2015 development agenda, excluded populations and vulnerable groups must be specifically targeted within global health interventions – and their success should be measured using data broken down by income, education and other specific metrics.
What is the good? Asking the right questions when addressing disparities
The problem is that we often lack the data needed to properly assess the impact of these interventions to work out what exactly is contributing to the apparent disparities. For example, much of our global data comes from the WHO Global Adult Tobacco Survey, which doesn’t provide data on smoking by income level.
As well as the need for more and better data, a number of important questions remain as we shift more and more attention to addressing disparities in the global health context. For example, in relation to the tobacco examples above:

  • Can we tell what other factors are at play that contribute to the higher prevalence of smoking among vulnerable populations? For example, is the tobacco industry specifically targeting vulnerable populations through marketing or other practices? Were rates in those populations higher to begin with?
  • Are vulnerable populations still seeing significant benefits from the program, even if they are being overshadowed by the success of the intervention among the higher classes? Is this a problem in and of itself?
  • Are some interventions that increase disparities still preferable? What if smoking is eliminated entirely from one part of a population and is lowered in another?
  • Is it better to reduce smoking by 20% among the poor but increase disparities, or reduce smoking among the poor by only 10% but with smaller disparities and inequalities?

In the case of the last question, many would agree that the former is better for the poor. Yet many discussions of disparities rarely consider these questions or their implications, which can lead to an excessive focus on reducing disparities when the goal should be to reduce overall smoking and tobacco use harms as quickly and efficiently as possible.
Programs or interventions that focus just on addressing disparities may not reduce smoking and tobacco use harms as effectively as other more broad-based interventions such as increasing tobacco taxes, lowering nicotine levels in cigarettes or strengthening marketing and sale restrictions. Such measures are probably more effective in reducing disparities than more specific programs focused on measures particular to the sub-populations alone.
We need to do a better job of identifying the goals, and quantifying and measuring the outcomes we seek within programs focused on disparities, as well as within programs that aren’t specifically focused on inequality but rather on a broader health issue such as a tobacco control, obesity or alcohol. Progress needs to be monitored not only through population averages, but with a specific focus on vulnerable populations to make sure disparities aren’t being further exacerbated – unless that program helps the vulnerable populations more than interventions seeking to reduce disparities.