NEWS: Can Internet Help Peg Healthy Lifestyles?

See allHide authors and affiliations

Science  08 May 1998:
Vol. 280, Issue 5365, pp. 795
DOI: 10.1126/science.280.5365.795b

Some health researchers are scrapping their No. 2 pencils and stacks of forms and turning to the Web to conduct surveys. Last month, epidemiologists launched the National Health Survey (, which hopes to get 20 million Americans to share their exercise habits, diets, and cholesterol levels so researchers can tie them to outcomes like life-span and heart disease. But some experts question how useful the data will be.

According to survey director Paul Williams of Lawrence Berkeley National Lab, computers offer benefits to participants that aren't possible with “paper-and-pencil” surveys. If you indicate you're a runner, for instance, his site can work out how your running times compare to those of other people your age. Or, if you wish, it will send your dietary analysis to your doctor. One huge advantage for researchers is that the survey will cost pennies per person, versus at least $10 a head for postage, data entry, and other costs of paper surveys, Williams says.

But Web surveys have drawbacks. Harvard's I.-Min Lee notes that the survey could lose its cost advantage when researchers have to track down participants who have dropped off the Internet because they're ill. In addition, the survey won't reflect health trends for the average American but only for a wealthier subset: people with Internet access and enough interest to fill out the forms. Williams predicts the surveys will reach a broader population over time: “Ten or 15 years from now, computers are going to be as ubiquitous as telephones.”

Yet another problem may be getting people to follow up, says Lawrence Kushi of the University of Minnesota, Minneapolis, who started a similar survey last June ( Like Williams, Kushi is reminding participants (1279 so far) by e-mail to update their data, but the response rate has been about 30%—less than he'd hoped. Still, he says, although Internet surveys are “definitely experimental,” he thinks they could reach far more people than conventional approaches.

Navigate This Article