Rethinking household surveys reveals true picture of health inequality

Previous blogs in this series have highlighted the lack of data on health and other development issues in informal settlements – despite the fact that, as in Nairobi, they house more than half the population (and workforce) in many cities.

One exception to this is APHRC’s research programme in Nairobi which has greatly increased our knowledge of the most serious health issues that women, men and children face there – and importantly their causes.

It has done this by developing new methodologies to fill data gaps. APHRC’s household survey (PDF) of all Nairobi’s informal settlements in 2000 revealed how bad conditions were – reflected in very high infant, child, under-five and maternal mortality rates. A follow-up survey in 2012 showed where conditions and health outcomes had or had not improved.

Why the data deficit?

It’s frustrating for those of us who work in cities to have so little relevant official data for planning, managing and governing. This inhibits city governments and civil society in their work at community, city and city region level to meet UN Sustainable Development Goals – as well as meeting goals for climate change adaptation and mitigation.

Data is also vital for addressing the current pandemic while building capacity to better cope with any future ones. The net result is ever-increasing responsibilities for local actors without the data they need to do their jobs.

The lack of data is even worse for those living or working in informal settlements. A previous blog described how relevant data is not collected or not available or not used. Household surveys have sample sizes too small to provide meaningful information for cities or city districts. When this issue is raised, the usual response is that it would be too expensive.

There is often no system in place to access data from hospital and healthcare records. It’s too difficult to get vital registration systems functioning in informal settlements so a key source for cause of death is lost. There is also the difficulty faced by any agency trying to collect data in informal settlements when there are no maps, no street names or addresses – and suspicion from residents.

Without relevant data, the very poor living conditions and health outcomes in informal settlements get hidden in aggregate statistics for urban and rural populations.

City-wide averages will usually appear better than rural areas as this is where wealthier households are concentrated. This does not mean that poorer groups in cities derive any health benefit from their urban location.

A truly groundbreaking study

The first APHRC study (PDF) managed to overcome these difficulties. This survey of households living in informal settlements in Nairobi had a large enough sample size to provide accurate population, development and health statistics for settlements in different districts.

Its primary objective was "to document population and health problems among the residents of Nairobi’s informal settlements and to compare these with indicators from national surveys for other sub-groups of the Kenyan population" (follow-up survey, page xvii).

The survey chose to have the same questions as Kenya’s Demographic and Health Surveys (DHS) so its findings could be compared to other population groups. The results showed how misleading the DHS programme was in not highlighting the very large differentials in health risks and outcomes within cities.

Aggregate statistics for Nairobi show it scores better than other areas, but a very different picture emerges from its informal settlements. If this study had not used the DHS structure, it could have included other neglected topics such as child accidents – but it would have lost the impact the comparisons brought.

The results highlighted just how bad conditions were. For instance, “the excess mortality and disease burden among the urban poor compared to any other subgroup in the country; their limited access to health care and family planning services; and the debilitating environment that characterises their physical living conditions, including inadequate access to water and sanitation, poor housing conditions, poor livelihood opportunities and the near-absence of public sector services” (second report, page 2).

These in part explain the very high under-five mortality rates in all the settlements, with more than one in ten dying before their fifth birthday – for Embakasi it was one in four. These are 20 to 40 times the rate in high-income nations.

For diarrhoea with blood for children under three years old (which signifies serious systemic infection) in the two weeks prior to the survey, all the informal settlements had much higher percentages than both national and rural averages. Many other indicators showed comparable disparities.

Had things improved 12 years on?

Twelve years after the first survey, Nairobi’s informal settlements were revisited to take stock of the changes that had taken place since 2000. It was useful both for what it shows for 2012 and for what had changed since the previous survey.

The second report showed “marked improvements in environmental, health and educational indicators among informal settlement dwellers. However, these improvements were not uniform, with subgroups of younger women and women without formal education being consistently disadvantaged. Additionally, slum residents remain generally disadvantaged in comparison to the rest of Nairobi and Kenya” (second report, page xix).

The table below shows under-five mortality rates declining in all but one settlement – and quite substantially in several. But most still had unacceptably high rates. Prevalence of diarrhoea with blood for children under three years old was much higher in the informal settlements in 2012 (8%) than in Nairobi (0.6%), and the country as a whole (3.3%) in 2008-09.