Steven sits down to breakfast: a large plate of pancakes with blueberries and maple syrup. Delicious! He has eaten them every day for the last five years - ever since he read an article by a nutritionist saying that it was the perfect meal to wake the body up.

But then, he opens his newspaper and reads a headline which shocks him: "Health risk from pancakes with blueberries and maple syrup!"

Steven is not alone in feeling confused. Every day, experts seem to come out with contradictory new guidelines for how we should live our lives: what we should or should not eat or drink and how much exercise we need to take.

Last week it emerged that an NHS "healthy weight" calculator had been giving the wrong advice for five years. It was designed to tell people how many calories they should eat per day to lose weight gradually.

But the calculations included a basic mistake about how much exercise the person using it was taking. As a result, anyone following its advice could have put on almost two and a half stone per year.

Now two researchers from Harvard Medical School have written an article in the New York Times arguing that much of the advice we are given cannot be trusted because it is based on erroneousMistaken. research.  

Anupam Jena and Christopher Worsham give moderate drinking as an example. For decades experts argued that this had health benefits, with statistics showing that those who drank a reasonable amount of alcohol lived longer than those who did not drink at all. Red wine was recommended because it had a high concentration of an antioxidantSubstances that prevent or delay some types of cell damage. which protects the body.

But a new review of 100 studies which reached this conclusion claims that they were carried out wrongly. It points out that people who drink moderately are likely to live longer for many other reasons.

They tend to be well off, to eat healthily and to take plenty of exercise - whereas people who do not drink at all tend to be poor and badly nourished. Some "non-drinkers" were former drinkers who had been forced to give up alcohol because it severely damaged their health.

The latest studies show that it is best not to drink at all. Ireland has just passed a law saying that all alcohol must carry a health warning.

Jena and Worsham are particularly interested in whether artificialMade or produced by human beings rather than occurring naturally, especially as a copy of something natural. sweeteners damage your health. Many studies have relied on "randomisation" - choosing people at random and asking them to take sweeteners instead of sugar with their meals.

The trouble is that the trials often just involve a small number of people for a short time - "which makes it hard to reach reliable conclusions about their long-term effects."

There are also "observational" trials, which follow people for a longer time to see how they use the sweeteners and whether they develop health problems like diabetes. The problem is that the diabetes might be the result not from the sweeteners but from something completely different.

Research, Jena and Worsham say, needs "a credibilityThe quality of being believable or dependable. revolution". Health is something that you cannot buy - but it is still our most valuable resource.

Yes: The people who give it are constantly arguing amongst themselves, and do not even have any faith in their own conclusions. The research they use to make their recommendations is not carried out in a truly scientific way.

No: Experts may sometimes get things wrong, but on balance it is best to trust them. Everyone accepts that smoking or staying in the sun for too long without protection can lead to cancer.

Or... We should follow the motto of the ancient Greeks: "Nothing in excess". As a general rule, if you partake in anything to a sensible degree you do not have much to be afraid of.