When you look at the population of the U.S., the latest Gallup polls, belief in God, membership importance, and attendance, this is done by age; you can see the different categories. When you look at belief in God, it's straight across: about 95 percent of the population.
Membership changes, but among the over-65 population, between 75 and 80 percent are church members. With regard to the importance of religion, among the over-65 age group, about 75 percent indicate that religion is very important to them. Of course, as people become sick and ill and go into the hospital, it becomes even more important to them. It's amazing to me that in the over-65 population, we are looking at rates of 55 to 60 percent who are attending church weekly or more often.
Many people, especially those over 65, are religious and turn to religion for comfort, support, and hope when they become sick. The medical profession has largely ignored this.
With regard to mental health, prior to the year 2000, there are a number of studies looking at well-being, hope and optimism, purpose and meaning in life in the 20th century, and these are the studies that show a positive relationship between religion and these various things. (See Chart 1) You can see depression, anxiety and fear, marital satisfaction, social support: 19 of 20 studies on substance abuse. The strongest effects are found in stressed populations. It's important to remember that.
Since the year 2000, there's been a large, growing interest in this area. Entire issues of various secular journals have been devoted to this topic as well as a growing amount of research and discussions. Between 1980 and 1982, there were 101 articles in the Psyc Lit data base; by 2000 to 2002, there were over 1,100 articles. It had gone up by almost tenfold. These are not all research studies, but they involve discussions and at least are a reflection of the interest in the area.
There are reasons why religion can influence coping. These are logical, rational: It provides a positive, optimistic world view; provides meaning and purpose to life; helps people to psychologically integrate negative things; gives people hope; enhances their motivation; personally empowers them and gives them a sense of control.
By praying to God, they feel they can influence their outcome, so they are not as helpless. Religion also provides role models for suffering--Job, for example--as well as guidance for decision-making, which helps to reduce stress; answers to ultimate questions that science cannot answer; and social support, both human and divine. Most important, it is not lost with physical illness or disability.
Better mental health in turn is related to better physical health. In the last six months, there have been major studies in JAMA, Proceedings of the National Academy of Sciences, and the Lancet showing the connections between better mental health and better physical health--depression in particular, affecting health-related quality of life in coronary artery disease (CAD), affecting Interleukin-6 levels (an indicator of immune functioning) two to three years after the death of the patient. Depressed patients have nearly double the mortality in CAD, and there is experimental evidence that negative affect (or negative mood) influences immune function.
Therefore, we have a logical reason why religion might influence physical health through mental health, through enhancing social support, through influencing health behaviors, all affecting physical health outcomes.
Now let's look at how religion is related to physical health and medical outcomes. There are many studies out there: different populations, different samples, different investigators, different time periods, and different disorders. Many of these studies have methodological weaknesses, but not all of them. Almost all are epidemiological; there are very few clinical trials, except for in meditation.
This gives you a sense of the research that is out there. In three of three studies you find a connection between religious involvement and immune and endocrine function; in five of seven studies, the religious experience lower mortality from cancer; in 14 of 23, they have significantly lower blood pressure; in 11 of 14, they have lower mortality; and in 12 of 13, clergy mortality is lower. In addition, numerous new studies are now in review that are currently being evaluated by journals.
Let's look at the strength of this effect. Odds ratios are hard to understand, but binomial effect size helps to explain the magnitude of the impact in lay terms. When 50 percent of a population has died, the number of additional people alive per 100, or the number of people dead per 100, because of the activity equals the binomial effect size. The binomial effect size can be determined from odds ratios.
Here's an example. Exercise rehab following coronary artery disease--these effects are all the results of meta-analyses. The odds ratio is 1.35, which means a 35% greater chance of being alive in coronary artery disease patients who undergo exercise rehab. This also means 3.7 people are alive per hundred as a result of that behavior when 50 percent of the mixed population has died.
Now, considering that there are almost 13 million people with CAD, you divide that by 100, and multiply that times the binomial effect size of 3.7, and this results in almost 500,000 people with CAD who are alive because of exercise rehab. For psychosocial treatments following CAD, the binomial effect size is 6.6 people per hundred, with slightly more than 850,000 people with CAD alive as a result of psychosocial treatments in CAD.
For cholesterol-lowering drugs and CAD, again affecting almost 13 million, with an odds ratio of 1.35, this means that almost 500,000 people with CAD are alive because of drugs like Lipitor. For hazardous alcohol use, 1.24 is the odds ratio, translating into 2.6 extra deaths per hundred, resulting in--given the high prevalence of hazardous alcohol use--about 750,000 fewer people alive.
Let's look now at weekly religious attendance. Here is a single religious variable, looking at a single outcome, mortality. The McCullough meta-analysis published in 2000 has the best odds ratio for the effect of religious attendance on mortality. It was 1.37, meaning a binomial effect size of 3.9. Given that there are 122,650,000 people attending religious services weekly or more often in the United States, this results in 4,783,380 fewer deaths as a result of religious attendance (if this relationship is causal).
The NIH Consensus Conference, whose results were published in 2003, with confounders only in the model (the best estimate of the true effect), resulted in an odds ratio of 1.43, which translates into a binominal effect size of 4.5, with even a greater potential number of people affected (5,519,284). Confounders mean age, sex, race, health status. The odds ratio for the full model (i.e., with explanatory variables such as social support, health behaviors, mental health, etc. in the model) is 1.33, with 4,415,428 more people alive. This means that even when you control for factors by which we think religion exerts its effects on health (social support, etc.), you still end up with an impact involving nearly 4.5 million people that cannot be explained.
The last four largest studies that controlled for all these variables got an average odds ratio of 1.37, again agreeing with the McCullough meta-analysis.
The Strawbridge study, looking at women, found in the full model this odds ratio (1.52), resulting in a binomial effect size of 5.2 per 100. Given that 69 million women attend religious services weekly, this means that over 3,582,000 additional women might possibly be alive as a result of weekly attendance. Compare this to the number of lives (2,252,900) that cigarette smoking takes among women who smoke.
In comparison to the number of lives potentially impacted by religious attendance (i.e., 5,519,284), the population of Washington, D.C., is 572,000, and the circulation of Newsweek magazine is almost 3.2 million.
Is the effect that religion has on health causal? There is limited evidence from clinical trials that it is. Religious interventions in religious patients with depression, anxiety, bereavement, and pain caused depressive symptoms, anxiety, and bereavement to become better more quickly. This is not only Christian interventions, but also Buddhist as well as Islamic interventions.
The are also clinical trials looking at meditation's effects on lowering blood pressure, reducing cortisol, cholesterol levels, and cardiac arrhythmias. These studies are not always perfect in terms of the methodology. I'm sure we'll find out later more about their weaknesses. But just because a study is weak doesn't mean it doesn't provide any useful information. In all, the information we have from clinical trials provides some evidence to support the huge amount of evidence from epidemiological research that this relationship may be causal. Epidemiological research by itself, however, can also contribute to causality.
In epidemiology, Hill's criteria for causation provide guidelines on determining whether a relationship is causal. What is the strength of the association? For religion and health, the strength is moderate. What is the consistency of the relationship? The relationship between religion and health is moderately consistent. What about specificity? Religious attendance particularly affects cardiovascular disease and stress-related diseases, as you would expect, and therefore is specific.
What about the temporality? In prospective studies, it appears that religious attendance predicts mortality in the future, providing evidence for temporality. What about a biological gradient? In both the Hummer study and the Musick study, as frequency of church attendance increased, the effect size on mortality also increased, providing evidence for a biological gradient.
What about plausibility? It is strong--highly plausible that religion influences physical health. We have a model of how religion might do this, acting through mental health, social support, and health behaviors. What about coherence? Yes, it is also coherent. The effects of religion are strongest in stress-related illness.
What about experiment? This is the only one of the Hill criteria in which the evidence is limited at present, given the relatively few clinical trials that have been done in this area. What about analogy? Yes, other psychosocial constructs, such as depression and stress, influence disease course, as we saw earlier.
What should physicians do about this? We can no longer justify that religion is usually irrelevant to health, neurotic, or health-damaging. But, while this is not sufficient to justify a physician's prescribing religious advice or recommendations, there are other reasons to justify limited physician involvement.
Religious beliefs impact medical decisions. This is an important reason for clinicians to address religious issues as part of routine clinical care. Studies show that 66 percent of medical patients indicate that religious beliefs would influence their medical decisions should they become seriously ill. Here, in making a decision about whether patients with end-stage lung cancer should receive chemotherapy, family and patients ranked "faith in God" as second in importance, even ahead of whether or not the chemotherapy would effectively treat the cancer. When 300 oncologists were asked this question, they ranked "faith in God" dead last among seven or eight other important influential factors.
So there's a difference here between what patients are saying affects their decisions on whether or not to receive chemotherapy and what physicians think affects patients' decisions in this regard. Physicians underestimated the importance of religion in influencing patients' medical decisions with regard to chemotherapy.
End-of-life decisions relate to religious beliefs and can cause serious conflict. You see here a study conducted in North Carolina, a random sample of women over age 40. If they discovered a breast lump, what would they do? Forty-four percent would trust more in God to cure their cancer than medical treatment, and 13 percent believed that only a religious miracle could cure cancer, not medical treatment.
With religious beliefs having such a profound influence on medical decisions, how can doctors practice good medicine without communicating about these issues with their patients?
So what do I recommend? Take a spiritual history. Because religion influences coping with illness and medical decision-making, doctors ought to take a spiritual history; respect, value, and support the beliefs and practices of the patient; and orchestrate the meeting of spiritual needs. Praying with patients is more controversial, although in certain circumstances, I feel it is appropriate.
In taking a spiritual history, what do you ask? First of all, you need to introduce the subject to the patient. Why is the doctor asking these questions? This needs to be explained so the patient won't be surprised or wonder why the doctor is asking questions about religion. The kind of information you want is as follows: Do religious beliefs or practices provide comfort, or do they cause stress? Don't imply that religion is either good or bad, only that it can provide comfort or can cause stress.
How might beliefs influence medical decision-making? Doctors need to know that. Are there beliefs that might interfere with or conflict with medical care? Is a person a member of a religious or spiritual community, and is it supportive? Are there any other spiritual needs that someone ought to address?
Not recommended: Do not prescribe religion to non-religious patients; force a spiritual history if the patient is not religious; coerce patients in any way to believe or practice; spiritually counsel patients; engage in any activity that is not patient-centered; or argue with patients over religious matters, even when they conflict with medical care or treatment. Even so, many complex situations can arise.
In summary, a religion-medical connection is not new or unnatural. Many patients are religious and use it to cope with illness. Religion is related to mental health, social support, and health behaviors. Better mental health, in turn, and better social support are related to better physical health.
Thus, religion should be related to physical health. And when you examine it, it is. The relationship is only moderate in strength, but it has a huge impact given the number of people who are religious. There is growing evidence that the relationship may be causal. Religion affects coping with illness and medical decisions. Thus, physicians should communicate with patients about these issues, but there are important boundaries and limitations.
by Harold G. Koenig
Heritage Lecture #816
December 22, 2003