The truth about the health of millions of people around the world is being lost in statistics as a result of the way health data is collected.
The report on the new framework for international development calls for a 'data revolution', and its emphasis on disaggregated data is extremely welcome. It means that progress against any new goals will be tracked in sub-groups including income level, rural/ urban divide, regions, and gender. This is excellent and will help us understand better how poor health and poverty affect the most vulnerable people around the world. Breaking down health data into groupings allows us to see what disparities there are within a country regarding health and what can be done to resolve them, but there is one area the report fails to be strong on, that could make a huge difference to the most marginalised people around the world.
Progress is being made to improve the health of women and children across the world - in just one example, Ethiopia, which has one of the highest rates of child deaths, has seen under-five mortality drop to 88 per 1,000 from 118 ten years ago. Yet the way many countries measure health outcomes masks the failure to improve the health of women from marginalised ethnic groups. To return to Ethiopia, the above statistic represents the average for the country; if we look specifically at the rate of child deaths in the eastern Somali Region, where a high proportion of the country's nomadic groups live, we find out things have not improved at all in the last decade, with 122 deaths per 1000. The use of aggregate data by many countries presents an image of progress when for some people from the most poor-off communities there has been little change.
Here in the UK, disaggregated data allows us to discover the inequalities there are in health among our own population - in London the gap in life expectancy between rich and poor is as high as 17 years. Using this data, action can be taken to target resources in the areas that need them most and, furthermore, uncover the underlying causes for such inequality.
Using disaggregated data can help improve health in developing countries in a similar way - particularly in middle-income countries which frequently have obscenely high levels of inequality.
But the recent report, and many international discussions about breaking down data, fail to consistently include one category. That is ethnicity, which is one of the key variables of a person's health and wellbeing. Disaggregating data by ethnic group would help strengthen some of the world's most marginalised communities, especially with regards to maternal health.
Some disaggregation of data is taking place, and having a positive impact. Many countries now break down data by wealth, region, sex or education background - but ethnicity is far too often missed from health surveys. Even though the tools for collecting data on ethnicity may exist, governments often don't use them, or only capture headline figures rather than delving deeper. In Laos for example, a survey was used to collect data by the language of the head of the household but failed to include religion or ethnicity. In Sierra Leone data was collected on the religion of the head of the household but neither language nor ethnicity. As a result, governments are only able to put together part of the picture of health for ethnic groups. And where ethnicity is captured, for example in Ethiopia's most recent health survey, it is often only published as background information on the population, not broken down in terms of how those ethnic groups' health outcomes differ.
There are understandable concerns regarding breaking down data by ethnicity; many governments are reluctant because they fear it will confirm negative stereotypes or fuel ethnic tension. Other barriers to progress include lack of technical expertise in analysing data in some countries, data collection being impossible due to war or disaster, and having the resources to repeat surveys regularly in order to build up sufficient data.
But when you look at the impact of not highlighting the true health situation of marginalised communities, these issues become details to work through.
The movement for better health services for pregnant women could especially benefit from the disaggregation of data by ethnicity. For many women around the world there are cultural and ethnic barriers to them exercising their right to maternal health: from discrimination and language barriers among health workers, to the nearest health centre being too far to travel to. By reporting data disaggregated by ethnic group it is possible to reveal what challenges these women face and create ways to implement real change in their health.
To make disaggregating data by ethnic group more common - to make sure some of the most marginalised women worldwide count - all of civil society has a role to play. Governments must prioritise building up their national health information systems with the full involvement of indigenous, ethnic and cultural groups. For development organisations, donors and international institutions, the goal must be to ensure all progress towards the new post-2015 goals is disaggregated by ethnicity and that national level mechanisms for collecting and analysing health data are well supported. The UK Government should champion the disaggregation of health data by ethnicity and support marginalised communities worldwide in the movement for health justice. Only by taking these steps collectively will we reach a stage where women from indigenous groups and ethnic minorities no longer simply disappear in the data.