Ok, let's set some things straight from a medical point of view.
, individual anecdotes (of anything) have no probative value in medicine. Individual case reports and even case series are only useful in medicine when we lack an actual study. You can relate that you've been cured by faith healing, or that vitamin C cured your cold, or that asparagus cured your cancer. But that offers no causal proof even in your own individual case -- and it has no generalizability at all.
So if you want to discuss whether faith healing "works" in a general sense -- or for that matter the therapeutic efficacy of anything -- you need to set the anecdotes aside because they just don't help.
, if we want to talk about whether faith healing works, we need to first think about how that can be proved. This first involves defining the problem better. We need to think about how we will define our target population
, our intervention
, and our outcome measures
For a target population
, we need to choose a reasonably uniform group of patients with a reasonably uniform spectrum of disease. This allows us to limit confounding variables that have to do with their illness. So whatever, take patients with metastatic non small cell lung cancer, or patients with NYHA Class IV heart failure, or patients with dialysis-dependent renal failure, or whatever. Or an acute condition like community-acquired pneumonia, or meningococcal meningitis, or supraventricular tachycardia.
For outcome measures
, pick something appropriate to the disease condition. So pick length of hospitalization, or rehospitalization rate, or cost of hospitalization, or 5-year survival rate, or recurrence rate, or pick a marker of disease progression. And define exactly when and how these outcome measures will be collected.
, you MUST find a way to isolate faith healing
in a way that makes any effect on outcome clearly an effect of faith healing and not some other confounder. Thus, you CANNOT compare people who pray to people who don't pray -- because there may be innumerable other aspects of their psychology and lifestyle that affect outcome. You CANNOT compare people who KNOW people are praying for them with people who do not have such knowledge -- because it might be only the sympathy and not the prayer that has an effect.
In fact the best way is to take patients who never pray themselves
, then either have true faith healers attend to half of them and false (placebo) faith healers attend to the other half. OR you need to have "secret" faith healing for half of them but not the other half and the patients never learn about it.
Finally, a little about biostatistics. In order to have statistically meaningful results, you first need to do a power analysis. What this means is that you need to pick a level of statistical confidence in your results and determine what sample size you need to recruit. Most clinical trials will require hundreds if not thousands of patients.
So let me ask, is there such a trial about faith healing that defines its parameters so clearly, that eliminates confounders, and that has a meaningful sample size? No, there is not. Which means that we really don't know if faith healing per se is helpful, or if the evidence in favor of it is confused by all kinds of other uncontrolled variables.
For instance, religious people might at a population level have better outcomes than non-religious people because of community support or because of optimistic outlook or because of avoidance of destructive health behaviors. So if you compare religious people with non-religious people, then it will be impossible to tell whether it's faith healing or some other factor that accounts for a difference in outcomes.