In the last few weeks I have spent time with a dear friend who was preparing for and recovering from a significant medical procedure.

Earlier last month he was told he would need either repair or replacement of a valve in his heart. I have health care experience as a nurse, administrator and health care educator, so we talked quite a bit. Everyone involved in the management, funding or making of health care policy should spend time at the bedside. It is an enlightening experience.

I have thought a fair amount about health care decisionmaking at the macro level. But watching it done in real time by someone you love is a whole different matter. Decisions are often made from a sense of urgency, sometimes backed by medical necessity, and as often by personal anxiety. In this case it was a little of each.

My friend was scheduled for two diagnostic procedures on a Monday; their results would drive the decision about the appropriate procedure and its timing. Prior to those tests, we talked about options, the relative invasiveness of the two options, as well as their success rates, recovery time and long-term effectiveness.

On that Monday, he was prepped and medicated and underwent both diagnostic procedures. Late in the day, just prior to release, he was introduced to a surgeon. Together with his spouse, my friend decided to have surgery that Friday, four days hence. Thankfully, the procedure went well, and he is on his way to recovery. In retrospect, it was a good decision.

My reflections have to do with how such critical health care decisions are made. I observed to my friend during his hospitalization that he had spent more time with the carpet salesman during a recent remodeling than he did with the surgeon whose hands were to be in his chest.

And without casting aspersions on individual providers, is anyone surprised that a surgeon would recommend surgery? If you have an expensive and finely tuned hammer, problems are more likely to look like a nail.

Much has been made of a famous study at Dartmouth that demonstrated dramatic variations in the prevalence of certain procedures — based on provider preference. This accounts for adjoining communities performing hysterectomies or C-sections at far different rates that cannot be explained any other way.

I'm not implying anything insidious. Patients are looking for the reassurance that a confident, well-trained provider brings. Providers are doing the job they are exquisitely trained to do, and heath care organizations are trying to meet the needs of the communities they serve.

As is so often the case, the problem is with unintended consequences. Costs soar and quality suffers because side effects from marginally or unneeded procedures require rehospitalization, longer recovery times, and increases in morbidity and mortality. All of which support a system based on providing sick care and an incentive to do still more. For new delivery models meant to promote health and decrease hospitalizations, these effects are counterproductive and unacceptable.

So what is to be done? The most prolific health care observers refer to the need for protocol-driven care, done at places that do it often by providers experienced with that protocol. Patients have an obligation to be informed about the options, including quality, cost and patient experience (how good are these institutions and staff at meeting the personal needs of the patient?). Providers must understand the limits of their expertise and learn to live with the burden those limits bring.

As health care organizations become larger and more integrated, they must balance the ability to do more with the social imperative to do only what is necessary.

Thomas Gilliam is an administrator in the University of Minnesota's School of Public Health.