Health outcomes matter. We were reminded yet again this week that entirely-plausible, widely-accepted, even vociferously-advocated interventions still need to be supported by rigorous scientific data. Headlines in the Washington Post screamed: "Deaths Halt Part of Diabetes Study. Scientists Fear Heart Attacks, Strokes Were Tied to Treatment. "

The headlines were similar to those when the feds halted part of the ALLHAT trial of anti-hypertension drugs because, while lowering blood pressure, they weren't reducing heart attacks or strokes. The headlines mirrored those when the government embarrased itself advocating Hormone Replacement Therapy (HRT) for post-menopausal women only to find it was killing them. After all, scientists had offered untested assurances that the interventions "can't hurt anyone."

Soon, the same headlines may be written to apply to advice to lower dietary salt.

Public health leaders are so anxious to overcome serious and perplexing health problems that they lose their discipline. They set aside standards of evidence-based medicine that demand scientific support before approving an intervention. Everyone wants heroes. Certainly the researchers want success. So do the intervention sponsors, whether a pharmaceutical company or an activist lobby group like, in this case, the American Diabetes Association. So do the media; overcoming a dread health threat is important news. All of us, frankly, want these interventions to work. But we cannot ignore the discipline of the scientific process in our enthusiasm and zeal. And we need to temper our noisy presumptions, the arrogance that we know all the answers and now have only to force implementation of strategies never fully tested, with a profound humility. Pride goeth before the fall.

So this week's news was that adverse findings "stunned and disappointed experts." Rob Stein reported in the Post story:

Aggressively driving blood sugar levels as close to normal as possible in high-risk diabetes patients appears to increase the risk of dying from a heart attack or stroke ...

The findings are the second major blow to widespread assumptions about how to protect against heart disease -- the nation's leading killer. Another recent major study found that driving blood cholesterol levels as low as possible did not necessarily slow the progression of heart disease.

Many earlier studies had shown that tightly controlling blood sugar significantly reduced the risk of many complications. The new study -- known as the Action to Control Cardiovascular Risk in Diabetes, or Accord, trial -- was designed to convincingly test whether various aggressive treatment strategies reduced the risk of heart disease -- the main cause of death among diabetics.

The study involves 10,251 patients ages 40 to 82 at 77 sites in the United States and Canada at high risk for heart disease for any reason, such as high blood pressure, high cholesterol or obesity. About half were placed on a regimen combining diet and exercise with commonly used drugs designed to lower their blood sugar levels to those of the average person with diabetes, while the other half were put on a more intensive regimen designed to drive levels closer to those of someone without diabetes. The patients were further divided into those who also received blood-pressure-lowering medication or drugs to improve their cholesterol levels.

Over about four years, about half of the patients getting intensive treatment achieved blood sugar levels close to normal, and about half the patients in the standard treatment group achieved levels close to the average diabetic.

But a special 10 member panel that was monitoring the study alerted the organizers that 257 patients r eceiving intensive treatment had died, compared with 203 receiving standard treatment, a difference of 54 deaths -- or 3 per 1,000 participants per year, officials said. About half the excess deaths were from heart disease.

Stein interviewed Cleveland Clinic cardiologist Steven Nissen, who seemed chastened, admitting: "This suggests that there are things drugs do that we don't understand."

The potential for unanticipated (and adverse) consequences is why we give lip service to supporting evidence-based medicine. Then along comes a plausible answer to an important problem and we must show discpline and humility to say: sounds great, let's find some money and test it out to see if it's true.

In August 2006, international experts joined the Salt Institute in urging the U.S. Department of Health and Human Services to find funding for a controlled trial of the question: will lowering dietary sodium improve health, specifically cardiovascular risk and mortality? We were told, in effect, that this was our problem and we should fund the study despite the fact that HHS's own U.S. Preventive Services Task Force has determined there is no evidence supporting population advice to reduce dietary salt.

It's time for some discipline and humility, HHS. And some funding.