|
|
|
 |
|
| Making Sense of Scientific Studies: Questions to Ask |
|
|
Article date: 1999/01/25
|
- Where was the study originally reported? Well-known journals, such as the New England Journal of Medicine, Science and Cancer use strict peer review criteria before they publish a study. This process lends more credibility to scientific reports.
- Where did you read or hear about the study? Network news programs, national newsmagazines and city newspapers usually have science reporters who carefully translate scientific articles into reports for the general public. But, given their time constraints, they may be short on specifics.
- Where was the research done? Most cancer research requires sophisticated scientific training, facilities, and equipment, so it's no surprise that most advances in cancer research come from well-known cancer centers, hospitals, and universities. Smaller hospitals and medical centers do generate valid research, but most research that has resulted in changes in clinical care are large, long-term studies carried out at major medical institutions.
- Does the study support or contradict past research? The more evidence there is for something, the more likely it is to be valid.
- Who were the research subjects? Some news items report studies from animal research - important but not always immediately pertinent to humans. Sometimes specific age, racial or ethnic groups are examined, such as African-American men over 75 with a high risk of prostate cancer or Ashkenazi Jewish women with a tendency to develop breast and ovarian cancers. Specific findings from such studies may not always apply to other groups of people.
- What type of study was it? A randomized clinical trial, where groups of individuals randomly assigned to different treatment groups are compared, generally is considered the "gold standard" in evaluating new treatments.
- If the study is a clinical trial, what phase is it? Phase I clinical trials are intended to examine the safety of the treatment. Phase II trials examine the effectiveness of a treatment. Phase III trials are planned to accurately compare the new treatment with the standard treatment to see which is more effective.
- How long did the study last? How large was it? Generally, studies affecting medical policy or guidelines are the result of years of research examining a large number of people. It usually takes a number of such studies before the experiment or trial is considered definitive.
- Do you understand the terminology used to describe cancer risk in the study? Terms such as relative risk and absolute risk are stumbling blocks for many non-scientists. For example, if a certain behavior or exposure is associated with a 100 percent increase in the risk of developing cancer, it sounds ominous at first. However, if only one person out of 100,000 currently develops that particular type of cancer, a 100 percent increase in risk means only 2 in 100,000 people. The relative risk (100 percent) may sound impressive, but the absolute risk (2 per 100,000) may be quite small.
- What is the statistical significance? This term is a measurement to describe how often the particular result would occur simply by chance, if the study or experiment were repeated many times. The most common minimum level of significance accepted for statistical significance is 95 percent; meaning if the study were repeated many times, the probability of this result occurring due to chance alone would be 5 percent or less.
- Does the study promise a "magic bullet" for preventing or treating all cancers? Unfortunately, most beneficial changes cannot be obtained easily with pills or potions. A healthy diet, exercise, and other good health habits are the best strategy for reducing a person's risk of cancer. Current cancer treatments are selected by doctors based on careful classification of the type and extent of cancer in each patient. Claims of one treatment that cures all cases of all cancer types should be viewed skeptically.
ACS News Center stories are provided as a source of cancer-related
news and are not intended to be used as
press releases.
|
|
|
|
|
