John JuergensWho ya gonna believe?

A confidence shaking report came out a month or so ago about the long term effects of hormone replacement therapy (HRT). HRT has been the life line of millions of women for many years in managing the effects of menopause, but now a very distinguished panel of experts say it increases the risk of breast cancer.

And a recent article in the American Journal of Clinical Nutrition (2002;Vol. 76: pp 466-472) grabbed some headlines a few weeks ago with the results of a study the researchers interpreted as meaning that the positive health effects seen among some wine drinkers might be related more to generally healthier lifestyles than to the wine.

Sounds like the classic chicken and egg debate to me, so what do you do in situations such as this when you are confronted with conflicting information? First of all, pull out a nice bottle of your favorite comfort wine and pour yourself a big glass.

When it comes to serious health issues such as cardiovascular disease, HRT and breast cancer risk, and so forth, you always need to discuss this with your physician. However, when trying to evaluate the implications of reports such as those above, it is also important to think through the issues. It seems like every time we get some good health news, we don't have to wait too long for the flip side to come rolling off the wire to spoil the fun. Unfortunately, most news reports only present the controversial or emotional side of an issue and don't put it in context of what has been reported previously.

Some of the most important information we have about chronic health issues has come out of several very long term studies that have tracked hundreds of thousands of people over their life times. The Framingham Study up in Massachusetts is the preeminent such program. It was data from this study that gave us our first really solid evidence that alcohol and wine can have some positive health effects when used in moderation. Of course, this sent the anti-alcohol forces into a hissy fit and they immediately tried to poo-poo the credibility of the data ("poo-poo" actually is a statistical analytical technique frequently used by anti-alcohol groups).

However, this crack in the dike resulted in the torrent of similar information from other long term studies, which eventually lead to even the most conservative of federal funding agencies admitting that the data could not be ignored. This allowed scientists to research openly and prospectively the mechanisms of the observed effects rather than relying on surreptitious retrospective mining of data.

When trying to evaluate the relevance of a new study, whether it supports or contravenes the prevailing notion, you have to remember that it is just one study, and that in and of itself it doesn't completely prove or negate all that has come before it. You have to keep a score card and look at the new study in the context of what is already known.

For example, the nutrition journal study I mentioned above. The data were collected way back in 1994 on a group of about 4,400 middle-aged alumni from the University of North Carolina and their spouses. As a group, participants were highly educated and had similar incomes. Some of their findings included that, compared with other drinkers, regular wine drinkers generally ate more fiber and less saturated fat and cholesterol, and drank less alcohol. Both male and female wine drinkers were less likely to smoke than those who favored other types of alcohol.

Based on these findings, the researchers sort of jump to the conclusion that there are myriad lifestyle factors that could help explain the link between moderate drinking and better health, as well as the extra benefits that have been attributed to wine.

On the surface, this might seem like a valid conclusion, but it really doesn't carry much weight as far as I'm concerned. The study is riddled with defects and limitations that easily could have influenced the results.

When you look at this study in the context of all the other research that has been done on this issue, which now covers something in the neighborhood of half a million persons over more than forty years, the evidence goes against the strong prevailing trend, which is overwhelmingly supportive of the value of wine. The authors also largely ignore many of the studies that generated these insights. Contrary to the statements of the researchers in North Carolina, other studies have in fact control for the "myriad" lifestyle factors such as diet, smoking, and exercise. Furthermore, the other studies took into consideration all sorts of other potentially predisposing genetic and lifestyle factors over a panorama of the subjects' lives. From what I can tell, the North Carolina study was only a snapshot of a relatively small group of a people who probably do not represent the rest of the population, and it is impossible to tell from the study if the observed health effects are part of the egg or the chicken.

Just about anyone who has managed to survive the endurance test called graduate school and who has completed a research project knows that you can torture data until they confess what you want to hear. That's why it is so important not to have a knee-jerk reaction to any health related study that has enough controversial and emotional appeal that it gets reported on the national news. You can be fairly certain that you are not getting the whole story, and without context it is like being a Democrat in Texas or like alcohol-free wine: It doesn't matter.

The other thing to keep in mind about any study is that the results are statistical interpretations and summaries of information gathered on a group of people, and the results might not actually apply to any one individual in that group or beyond that group. For example, it's like measuring the height of a bunch of people in a room and coming up with an average of, say, 5 feet 8 inches. But when you look at the individual measurements, there might not be a single person in the room who is exactly 5 feet 8 inches tall. It is a general description of the group, but it might not apply to everyone or anyone specifically.

One of my favorite and most annoying examples of this is the infamous General Assessment Guide for rating a person's impairment level based on their blood alcohol concentration (BAC). This guide was developed back in the 1930s and has been cited as a Gold Standard reference ever since.

You probably have seen a version of this guide. In one column it lists an increasing BAC and in the next column is gives the kinds of physical characteristics or behaviors that a person will exhibit at those blood levels. Some versions will also include a relative probability that the person will be involved in an accident if he or she drives at that BAC.

I have seen sworn testimony of some physicians and police officers that these relationships between BAC and observed behavior will occur without fail 100% of the time. For example, regardless of the age, weight, gender, psychological make up, or prior experience, if you have a BAC of 0.09% (0.08% is the new limit for DUI in most states), the chart says you will be euphoric and excited, and you will appear to be drunk. The problem is, this just is not true. In fact, there are lots of clinically documented cases of people having BACs of 0.2% and higher who show absolutely no signs of intoxication or impairment.

This chart was developed through observation of a handful of people as they were given alcohol. It did not take into consideration the "myriad" factors that can affect how any one individual will behave under any given circumstance. Mind you, I'm not saying people don't get impaired or experience an altered state of awareness at some level of intoxication, nor am I advocating drinking and driving, but we need to apply some common sense here. I also don't want members of MADD, SADD, the CSPI and other such organizations to come after me with big sticks.

As we come into another fall, in college towns all across the nation there will be an influx of young folks who will be away from home for the first time. I suspect there will be some underage and first time drinkers among those new arrivals who will want to do some in-depth exploring of their tolerance for alcohol. I would wager that even a 21 year old first time drinker with a BAC of 0.07% probably is more far more dangerous behind the wheel of an SUV than a forty year old experienced drinker with a BAC of 0.09%.

So, who ya gonna believe about the wine and health stuff? I say trust your own judgment and experience, routinely apply the "straight-face" test for believability, and ask more than one "expert." Unlike man-made laws that in an instant can turn a person into a criminal or a hero, the laws of nature are pretty well set. As far as I am concerned, to date, the data overwhelmingly indicate that wine and other alcohol containing adult beverages clearly can have some health benefits for those for whom it is appropriate to drink. That's my story and I'm sticking to it.


Sept. 6, 2002

To contact John Juergens, write him at

Back to Oxford Town Wines