Is religion for more than just the soul?

Research into the role religion plays in our health behaviors and outcomes is lacking, but are there opportunities to learn?

Photo by Ben White on Unsplash

In polling, some 36% of American adults say they attended a religious service “at least once a week.” But as we all know, what people say they do doesn’t always line up with what they actually do. When people respond to surveys about their habits, they don’t always answer accurately, either because they can’t remember or because they, consciously or subconsciously, want to give answers that make them look good.

University of Chicago economist Devin Pope was interested in understanding what people actually do when it comes to participating in organized religion. One reliable way to see where people are actually going? Follow their cell phones. Mobile phone location data is generated whenever your phone gives its location to a service, such as an app that tells you the local weather or provides driving directions. Unlike surveys, mobile phone data doesn’t forget, and it’s not going to judge you for not showing up for worship as often as you’re supposed to.

Pope recently published some of his initial analyses on Twitter, where he estimated, based on mobile phone location data looking at places of worship on their typical days of worship, about 30 million Americans attended a place of worship on a weekly basis. This estimate corresponds to less than 10% of the population—a far cry from the 36% in surveys (though Pope is working on improving these estimates)—but still a sizable group.

Looking at Protestant Christians, who make up the biggest chunk of the 30 million, there were little dips in attendance on long holiday weekends, and big spikes on Christmas and Easter (similar spikes were seen for all Christian denominations studied except for LDS, who had similar weekly worship throughout the year except for weeks of online conferences).

Religiousness and health

While Pope’s analysis suggested that a smaller number of people actually attend weekly worship than say they do, there are still millions who go—and many millions more on Christmas and Easter. With participation in organized religion being a major aspect of life for so many, it’s only natural to wonder what impact it may have on our health.

There are lot’s of ways that religion may impact health. One is through the organizations that it permeates. For example, one study showed that when hospitals fall under Catholic ownership, rates of tubal ligations—a form of birth control opposed by the church—fall by about 31%.

Religion may also impact our health at the individual level, through the lessons it teaches, requirements it imposes, community it brings, or the faith and comfort it provides. But studying the effect of religiousness on health for individuals, however, is extremely difficult, and the current limited evidence base on this subject reflects this.

Comparing health of people who are more or less religious presents problems, since these people will be different in ways other than their religiousness that are hard to measure and account for. If we find better health behaviors or outcomes in people who are religious, we can’t say that religion was the cause. Researchers can’t circumvent this issue by randomizing patients to be religious, either. So we’re left looking for situations where patients are more or less religious due to accidental random factors: random acts of religion if we’re being cheeky.

Here’s something we tried. Looking at Pope’s results, it appears that there are about 10 million extra people going to a Protestant church on Easter and Christmas than normally go every week. If Easter is timed randomly with respect to a patient’s health, we see that these people get an as-good-as-randomly timed “extra dose” of religion around the holiday.

We thought this little bit of extra religion might impact health behavior surrounding contraception. Would young women be less likely to fill a prescription for a birth control pill after going to an Easter service at a church that may frown upon the use of birth control pills and/or premarital sex? In a quick analysis of insurance claims data, we found that rates of birth control pill prescription fills (initial prescriptions and refills) were constant before, during, and after the week of Easter, suggesting against an Easter effect on oral contraception behavior. (We did see increases many weeks after easter for 17 and 18 year old girls, which we think might be related to leaving home.)

Some untested ideas

Looking for naturally occurring randomness with religion is challenging. Here’s a few other ideas that came to us, that maybe one of you will run with.

The first idea that came to our mind as hospital-based physicians was the impact of religion on decision making at the end of life.

One study showed an association between involvement of hospital chaplains (who oftentimes are able to provide guidance to patients of any faith) and lower rates of in-hospital deaths, with corresponding increases in enrollment in hospice care, suggesting chaplains might be helping hospitalized patients navigate the limits of medical care at the end of life. But it could also be that patients who involved chaplains in their care were already more likely to seek hospice services.

One way to see if chaplains had a true effect on use of hospice services would be to find some occasion where, by chance, chaplains were more or less available. For example, we could look at hospice use in the times just before and just after an additional chaplain was added to the hospital staff—a randomly timed event as far as individual patients are concerned. Or we could look at periods of hospice use by patients when the hospital chaplain is on vacation.

Alternatively, we could look at differences before and after the addition of an electronic order for a physician or nurse to request a hospital chaplain’s services. Electronic orders are much easier to place than paging a chaplain or making a phone call, and they might lead to more chaplain involvement—in a way that, again, is randomly timed as far as patients are concerned.

The second idea relates to the Pope—not the economist Devin Pope that we mentioned above, but the actual Pope. The Papal Inauguration is an intensely religious experience for many around the world. Might we expect to see some health behaviors change in the weeks or months following the installation of a new Pope? Or what happens when the Pope comes to town? Visits to cities by the Pope may re-invigorate religious sentiments among many. Might health behaviors change?

Clearly, there are many, many ways in which religion can impact health, both on the larger population and for the individual. Looking at the data in the right way, finding natural randomizing experiments when we can, might help us better understand what religion might do for the body beyond the soul.