The Drivers of Health meeting in Detroit on September 11 focused on the research and research gaps in the connection between health and housing (presented by Roshanak Mehdipanah of the University of Michigan) and education (presented by Adriana Lleras-Muney of UCLA). The meeting wrapped up with a panel discussion including those two scholars and Robert Gordon, Director of the Michigan Department of Health and Human Services. (You’ll find a video recording and slides posted here.) The presenters covered a lot of ground and some parallel themes emerged.

One theme that cuts across housing, education, and many other social needs that affect health is the distinction between expansion of access and improvement in quality. You can provide greater access to affordable housing and, separately, you can make existing housing better (e.g., by reducing levels of lead, reducing mold, enhancing ventilation, etc.).

You can increase access to education or make education more effective. Likewise, you can expand access to health care and/or improve its quality.

Very often, you should do both. It can be detrimental to expand access to something that isn’t effective and could be harmful.

This point came up several times. For example, Ashish Jha, who moderated several sessions, recounted the fact that in some developing countries, when women are encouraged to give birth in health care facilities rather than home, infant mortality can go up, not down. Why? Because the quality of care in those facilities may be worse than what women would otherwise get at home.

Likewise Dr. Lleras-Muney discussed findings from some studies that increased education does not lead to mortality gains, contradicting other studies, including her own work. Why? Though we don’t know for sure, one answer may be that the increased education provided in some contexts may not be very good.

That brings me to a final point that came up at the meeting, emphasized by Dr. Mehdipanah, though raised by others: context matters. Whether and how much housing or education (or, for that matter, health care) improves health depends not just on its quality, but the context in which it is provided. A house can affect health, but so can the neighborhood in which it is situated along with the policies in place in the state in which it exists. What’s the nature of transportation and access to food? What’s other social support does the state provide?

For this reason, we should not expect that the results of a program implemented in one context will be the same when it is implemented elsewhere.

As Dr. Lleras-Muney said, there isn’t really an answer to the question, “What is the effect of education on health?” There are many effects, not just depending on context but on different kinds of people with different backgrounds.

It’s not a satisfying answer, but the evidence suggests it’s the truthful one.