The state of diabetes drugs has gotten to the point where it is good news that a drug does not produce worse outcomes than a placebo.

I am not kidding.

The New England Journal of Medicine published two trials, each testing whether a diabetes drug increased the risk of heart attacks and death from heart disease. The trials are a result of FDA guidance for companies to assess the cardiovascular safety of new anti-diabetic therapies.

The SAVOR study found that saxagliptin, an anti-diabetes drug approved in 2009 in the U.S. and co-marketed by Bristol-Myers Squibb and Astra-Zeneca, did not increase or decrease the rate of heart attacks, death from heart disease or stroke compared with placebo. The drug did slightly lower the glycolated hemoglobin levels, a reflection of glucose levels. The drug also increased the risk of hospitalization for heart failure by 27%.

The EXAMINE study found that alogiptin, another antidiabetes drug of the same class that was approved this year and sold by Takeda, did not increase or decrease heart attacks, death from heart disease or stroke compared with placebo. This drug also had slightly lower glycolated hemoglobin levels.

Now, don’t get me wrong, I am glad that the companies are testing whether these new drugs increase the risk of heart attacks. We went through enough debate about Avandia (and in some quarters the debate continues) to know that this is a potential problem.  But the message that these studies did not show that the drugs increased the risk of heart attacks, death from heart disease or stroke should not obscure the point that we hope that these drugs actually improve outcomes.

The point is that we still do not know if these drugs confer any benefit that would be felt by a patient. Improving blood glucose levels is not enough to tell you that the drugs will reduce the risk of health problems that patients would like to avoid ­ and do so without causing other health problems.

Besides that, saxagliptin caused a 27% increase in the risk of hospitalizations for heart failure. The alogiptin study was much smaller than the saxagliptin study and so may have been unable to detect this problem. In any case, a substantial increase in the risk of heart failure is a problem and has been seen in other drugs such as rosiglitazone (GlaxoSmithKline) and pioglitazone (Takeda).

We need comparative effectiveness studies of different regimens designed to determine how they fare on patient outcomes. Diabetes affects about 7% of our population and pharmaceutical sales are in the billions of dollars.

Rather than mandating narrow studies of cardiovascular studies, we should be encouraging studies that let us know which regimen is best for which patients, based on what they actually do for patients.

I am glad that these drugs do not cause more heart attacks and strokes than placebo. But patients and doctors need to know more in order to make wise choices about which drugs to use.  We need to know if they really produce benefits that are worth any risks.

Not inferior to placebo is good to know, but does not begin to meet what we need to know.

Harlan M. Krumholz is a professor of cardiology, epidemiology and public health, Yale University School of Medicine. He blogs at Forbes, where this article originally appeared.