We’re piloting a new format for EpiAnalysis starting this week. Rather than writing long essays every month, we’re going to post short, weekly summaries of interesting new data or controversies relevant to the epidemiology community, offering direct links to new papers and concise analysis of emerging issues. We hope this format makes the blog more accessible and quicker to read.
This week, we’re highlighting recent discussions about the “human microbiome”, controversies about child mortality data, and the flurry of universal healthcare programs being introduced in middle-income countries like Mexico and India.
From microbe-fighting to microbe-farming
The launching of the Human Microbiome Project earlier this summer set off a flurry of excitement about transitioning medicine from a microbe-fighting profession to a microbe-farming activity. The idea is that microbes aren’t all dangerous pathogens; some also save us from sickness. Having a healthy garden of gut bacteria, for example, prevents opportunistic infections like Clostridium difficile (“C. diff.”), a killer pathogen often brought-on by excessive use of antibiotics. The “hygiene hypothesis” similarly suggests that kids who are not appropriately introduced to microbes in their early years may generate immune systems that promote allergies and asthma (via modifications to the polarized Th1/Th2 responses); the inflammatory abnormalities associated with low parasite exposure in clean environments also seem related to various rheumatologic and gastroenterologic diseases as well as autism. So scientists have been testing approaches that attempt to promote “good” microbes in the system, and even re-introduce microbes into people with inflammatory disorders that range from arthritis to bowel disease. Indeed, the “fecal transplant” for patients with C. diff. infections seems remarkably effective, however disgusting. And as crazy as it sounds, introducing modified parasites into autistic adults is under study as a means to reduce auto-immune nervous system inflammation.
Some of the key recent papers in this area include a study of neighborhoods of skin bacteria in last month’s Science that found bacteria produced localized protective immunity that may also explain why inflammatory skin diseases such as eczema in certain parts of the body. “Feeding” protective bacteria could help prevent drug-resistant Staph infections from invading the skin. A related study of human guts shows that different peoples’ gut neighborhoods of bacteria can be classified into broad groups that may be used for early disease diagnosis. A broad overview of the microbiome initiative is summarized in a recent article in Nature, and a “manifesto” from some of the field’s leaders was recently published in Science Translational Medicine. Common questions for the field are how genotyping large arrays of microbes may lead to targeted therapies that don’t promote excessive microbial overgrowth (i.e., how do we limit or control how much the introduced bacteria grow?), and how to capture the impact of variable living environments as a confounder to determining what sets of microbes are adaptive or healthy for a given individual (i.e., if I move to Kenya, won’t the microbes on my skin change, and how do I know which subset of the new microbial neighborhood is healthy change?).
Tracking child deaths
The journal PLoS Medicine this week introduced a series of articles on estimating child mortality. Some groups announced to great fanfare that child mortality rates had been decreasing worldwide. But the data upon which these claims are based appears to be highly controversial. The key insights into these limitations come from Dr. Kenneth Hill, who currently chairs the Technical Advisory Group of the United Nations Interagency Group on Mortality Estimation (UN IGME). Hill points out that only about 60 countries have reliable civil registration systems from which to derive accurate child mortality statistics; the other 130 countries have limited or no surveys, and so various imputation methods are used to derive their child mortality statistics.
Major papers in the PLoS collection by Hill and his colleagues reveal that when time series regression models are fit to the available data, mortality among children less than 5 years of age does seem to be falling, but not at a rate suitable to reach the Millennium Development Goal. However, this conclusion may be debated as different groups seem to interpret similar data in alternative ways based on differing assumptions about how to deal with missing data, the effects of epidemics like HIV, and how much to assume that neighboring regions have similar mortality trends. Regardless of the methodological assumptions, North Africa seems to be advancing surprisingly well in reducing child mortality rates, while sub-Saharan Africa, southern Asia, and Oceania continue to fall behind, and girls appear to be systematically discriminated against in many locales. Some of the papers in the series highlight specific methodological challenges with estimating child mortality, such as the fact that HIV introduces a selection bias because traditional methods of child mortality calculation assume no correlation between mortality risks of mothers and those of their children; that using shorter time intervals for surveys of child mortality may be useful for both reducing bias and capturing the effects of major disasters, epidemics and conflict; that there are potential inaccuracies in using traditional life tables to estimate infant mortality; and that common indirect methods of inferring child mortality from household surveys may be inaccurate when mortality rate changes are not gradual. These papers are useful for those statistically correlating the impact of global health policies and programs with child mortality rates.
An editorial in The Lancet last week celebrated Mexico’s declaration of having achieved universal health coverage for its population. Mexico’s policy fundamentally changed from viewing healthcare coverage as an employment-related benefit to a citizenship-related right.
Next week, The Lancet will introduce a full series of papers discussing universal healthcare in a number of other regions. Among the most interesting papers already exploring the issue include the review of Rwanda’s program by Ranu Dhillon (which questions the role of community-based health insurance) and an introduction to India’s universal healthcare proposal by Anna Marriott of Oxfam. Some concerns have been raised that India’s scheme may rely excessively on subsidizing care delivered by private providers; our recent systematic review of the subject found that this approach may lead to problems with efficiency, access and medical quality. Indeed, while universal coverage is a laudable goal, several concerns about quality and the avoidance of a two-tiered system or “infrastructure inequality trap” (in which public funds are diverted to private systems, as the latter have more “absorptive capacity”—creating an inferior public sector) need to be addressed by such healthcare expansion schemes.