From a Scientific American article:

The three-year Study of the Therapeutic Effects of Intercessory Prayer (STEP), published in the April 4 American Heart Journal, was the largest-ever attempt to apply scientific methods to measure the influence of prayer on the well-being of another. It examined 1,800 patients undergoing heart-bypass surgery. On the eve of the operations, church groups began two weeks of praying for one set of patients. Each recipient had a praying contingent of about 70, none of whom knew the patient personally. The study found no differences in survival or complication rates compared with those who did not receive prayers.

An interesting quote:

Dean Marek, chief chaplain at the Mayo Clinic, saw the problem as a possible flaw in the study design: “The sense of community was not there. You could call it impersonal prayer rather than intercessory prayer.”

Stopping short of suggesting that the healing power of prayers by friends and family might reside in the personal connections rather than in the prayers, the authors stated that they have no plans for a follow-up study. This one, sponsored largely by the John Templeton Foundation, cost $2.4 million.