The health of people who live in slums and the trouble with estimating neighbourhood effects

Slums are a large and growing feature of urban areas in low and middle income countries. But, despite the ease with which you might picture what such informal settlements look like, there is no consensus about what exactly defines a slum. UN-Habitat defines ‘slum’ at the household level as a dwelling that lacks a number of important amenities such as improved water and sanitation while UNESCO defines it in terms of an urban area with certain features. In a new series of two papers, we discuss slum health and interventions designed to improve the health and welfare of people who live in slums. We make the argument that the UN-Habitat definition of slums is inadequate as slums exhibit unique neighbourhood effects.

Neighbourhood effects are those effects caused by the shared physical and social environment in which a person lives. Identifying and measuring such effects are important for public and health policy. In the context of slums, we argue that such neighbourhood effects determine the effectiveness of an intervention. For example, the benefits of provision of water and sanitation facilities are dependent on the already existing infrastructure, density of housing, and social structure of the area. The intervention may therefore be effective in some places but not in others, or require a certain level of input to reach a ‘tipping point’. However, estimation of the causal effect of a neighbourhood on population health and well-being can be difficult.

For certain outcomes causal neighbourhood effects are fairly easy to discern. Consider John Snow’s map of the 1854 outbreak of cholera below. The plot of cholera cases enabled John Snow to identify the infamous water pump from which cases of cholera were being contracted. In this instance, the causal neighbourhood effect of the shared water pump is clear, and the effects simple to measure. It’s not as if people contracted cholera and then decided to move closer to the water pump.

snow-cholera-map-1

John Snow’s map of the 1854 cholera outbreak in London

A similar exercise can be conducted using survey data. The map below shows an estimate of the spatial distribution of cases of diarrhoea in the under 5s in Nairobi, Kenya, with notable slum areas marked by the letters A to E. There is clearly an strong correlation between slum areas and risk of diarrhoea. It would not be a strong assumption that there was a common cause of diarrhoea in slum areas rather than people who were more likely to get diarrhoea choose to move to those areas.*

imageedit_2_8243201114

Estimation of the risk of diarrhoea in the under 5s in Nairobi, Kenya. Disease risk is estimated by applying a spatial filter across a regular lattice grid and then estimating a binomial model to predict disease risk at each point. Red areas indicate higher risk and turquoise areas lower risk. Blue lines indicate areas with a >80% probability of having higher risk than the city average.

Inference about the effects of higher or lower wealth or socioeconomic status or more ephemeral characteristics of neighbourhoods on health and outcomes is more difficult. It is generally not possible to randomise people to neighbourhoods; individuals of lower socioeconomic status are more likely to move to poorer neighbourhoods. The exception is the Moving to Opportunity experiments in the US, which showed that better neighbourhoods improved adult health and improved the health and economic outcomes of their children.

J. Michael Oakes has a detailed discussion of the issues involved in the estimation of causal neighbourhood effects. He identifies four key problems. Firstly, due to social stratification between neighbourhoods, the “selection” equation that sorts individuals into neighbourhoods is likely to be nearly identical for all people in the same neighbourhood. Modelling selection therefore removes most of the variation between neighbourhoods. Secondly, even if neighbourhood effects were emergent properties of the interactions between individuals, such as the epidemiology of infection, they would still not be necessarily identifiable as the expression of those emergent properties is dependent on the neighbourhood level variables. Oakes likens it to trying to estimate the incidence by controlling for prevalence. Thirdly, neighbourhood level effects are not likely to be exchangeable, an assumption widely used in statistical inference. And fourthly, neighbourhood effects are not likely to be static. Arguably, quasi-experiemental methods such as instrumental variable or regression discontinuity designs and more sophisticated models may help solve these issues, but convincing applications can still remain elusive.

The points above contribute to the argument that the effectiveness of a community-level intervention, even when measured in a randomised trials, depends on neighbourhood effects. As we have discussed in other contexts, development of a causal theory is clearly required for the appropriate design of studies and interpretation of evidence. From a health and public policy standpoint innovations in methods of causal inference for neighbourhood effects can only be a good thing.

 

*Of course, there are other factors that may explain the correlation. For example, slum areas were more likely to be sampled in the rainy season. We also therefore examined childhood stunting, which showed the same pattern. See the paper for more detail.

Fermi problems and public health

Enrico Fermi was a physicist well known for his ability to make good approximations to difficult questions. A well-known Fermi problem is ‘how many piano tuners are there in Chicago?’ Answering these sorts of questions involves using estimates of simpler quantities that, if correct, yield the correct answer when combined. Indeed, we could consider these calculations as arising from a simple model. Many public health policies can be approached in this way and yet, despite large amounts of journal space devoted to these issues, the debate usually revolves around the quality of the evidence. So why aren’t these sorts of calculations commonplace as a start to informing the debate? Consider the following two examples.

E-cigarettes

As we’ve noted before, e-cigarettes are a contentious public health issue. But we don’t want to wade into the vagaries of this debate, rather we’d just like to consider the benefits of maintaining a regulated e-cigarette market versus prohibition of e-cigarettes. If we assume that under a prohibition, ex-smokers who use e-cigarettes would switch back to smoking, and that the costs of enforcing a prohibition are balanced by the gain in tax revenue. Then the costs of prohibition and the benefits of regulation could be equivalent, and equal to the product of:

  • The number of switchers from cigarettes to e-cigarettes. Approximately 850,000 people.
  • Relative risk reduction of major smoking related illness. Let’s say 95% reduction.
  • The risk of a premature death. About 1.1% of smokers die prematurely each year.

This gives around 9,000 premature deaths avoided each year. And similarly at a cost of about £350 per person per year (£3b annual NHS costs divided by 8.8m smokers), a saving of approximately £14.5m per year. As a further cost to a regulated market we could add uptake of e-cigarette use by never smokers – about 56,000 people. Taking this into account, we get 8,600 premature deaths avoided and £13.5m saved. But, the benefits are likely to rise as public information increases and more people aim to quit smoking, which already seems to be the case (figure, left). Smoking consumption among e-cigarette using smokers (1.3m in the UK) may also be declining (figure, right).

smokeprevalenceconsumption

Figure. Left – smoking prevalence by year in the UK. Right – daily cigarette consumption among smokers in the UK. Source data.

Seven day NHS

Seven day NHS services are a key aspect of current government healthcare policy. The aim is to reduce the so-called weekend effect: the increased risk of mortality associated with weekend admission versus weekday admission. Rachel Meacock described her paper on this blog that attempted to value those ‘excess’ deaths to see if the seven day policy met typical cost-effectiveness standards (they did not). But, as they and others have noted, not all of these deaths are preventable. We can attempt to come up with an estimate of how many of those deaths are preventable based upon what we know. So, we multiply:

  • Number of emergency weekend admissions: 1.23 million
  • Risk of a preventable adverse event: 4.5%
  • Relative risk increase of an adverse event associated with weekend admission: 20%
  • Probability an adverse event lead to death: 7.4%

Which gives 820 preventable deaths. This is 15% of the total number of deaths associated with weekend admission, which would suggest that the benefits of preventing these deaths are 15% of the size estimated by Meacock et al.

Informing the debate

Certainly, these are not thorough statistical analyses and do not capture the uncertainties involved. They are Fermi Problems and they are a simple attempt to inform the debate. The quality of the evidence is of course key to ensuring the estimates informing the calculations are correct. For example, the 95% reduction in risk associated with e-cigarettes is disputed by some. But, that figure would have to be close to zero, or there would have to be some very large unaccounted-for cost to regulation in order to favour prohibition.

There are, of course, moral, political, and legal dimensions to policy formation. Policy analysis simply provides the figures to inform that debate. In some cases no further evidence might be forthcoming, leaving such calculations as the only option. For example, policy aimed at mitigating pandemic influenza requires studies from a pandemic, and it is too late when one occurs. Even a back of the envelope calculation like this is preferable to basing policy on intuition and whim alone.

Rugby, Rugby, Rugby

This morning saw headlines reporting the publication of an open letter to MPs calling for a ban on contact rugby in schools. The letter argues that the high impact collisions in rugby lead to a number of serious injuries including concussion and fractures. For a child that plays a full season of rugby there is a 28% risk of experiencing and injury and an 11% risk of concussion.

These high numbers may seem to warrant a ban on full contact rugby as the letter argues. Indeed, the expected harms that befall a rugby playing teenager may be much higher than those associated with other prohibited actions, such as the use of cannabis. Rugby may therefore be over a ‘threshold’ required to enforce a prohibition on contact rugby in schools. Such a threshold might exist where the marginal costs of enforcing a prohibition were less than the marginal benefits. The costs may be quite low as schools would presumably be compliant with the ban, while the benefits are high in terms of harms avoided. However, this is obviously not how prohibitions are worked out; the marginal costs, for example, of a prohibition on recreational drug use are arguably very high relative to the potential benefits since the ban is particularly ineffective as a harm reduction strategy.

The choice then is political. The median voter enjoys rugby and sees it as a part of their cultural heritage. Opponents of contact rugby may be more likely to lie at one end of the political spectrum and have little effect on sports policy in a representative democracy. For many people the perceived loss of a cultural institution in schools, the loss of liberty for their child, or the overreach of the state is a great cost. (Just view some of the comments on today’s news reports). Even in the face of strong evidence of harms, many are likely to be recalcitrant in their views, and even be annoyed that evidence was presented at all. Prof. David Nutt discovered this the hard way when he was fired from the Advisory Council for the Misuse of Drugs for comparing the risks associated with ecstasy alongside those of horse riding.

Nutt’s strategy is the correct one though. To make informed choices about the policies that are chosen from amongst the set of possible policies in any given context one must be able to compare them. People are well aware that full contact rugby carries risks, but for most people they have never had personal experience of any life changing risks from the sport. Saying that there is a risk of 1 adverse event every ~4 exposures is potentially meaningless. But add in the information that horse riding carries an equivalent risk of 1 in ~350 exposures and ecstasy 1 in ~10,000 exposures and individuals are better prepared to understand the risks.

When considering investment in healthcare technologies, a benefit of 10 QALYs, or cost-effectiveness of £10,000/QALY is meaningless in isolation. The choice to invest takes place against the background of a myriad other choices. Trying to implement harm reduction strategies in sport or elsewhere is the same. When placed alongside other potentially harmful activities rugby appears to be one of the highest risk and the costs of prohibiting full contact would be small, thus the burden of proof would seem to fall on those who oppose any prohibition. However, I would suspect that the letter as currently written would have little impact on many people’s views and therefore little impact on public consensus or policy.