- Learning About New Ways to Prevent Cancer
- Look at where the information came from
- Look at the science behind the prevention method
- Types of human studies on cancer risk
- Studies that observe humans
- Human testing: Clinical trials
- A closer look at the evidence
- Other questions about studies on new ways to prevent cancer
- What does this mean to you?
- To learn more
- Appendix A
Studies that observe humans
A study that simply looks at people is not really testing a method or compound to prevent cancer. The researchers are only looking at whether they can find differences in cancer risk that might be linked to something that the people already do.
They might observe that some of the people who did one thing (exercise and eat fruits and vegetables, for instance) were less likely to get cancer than those who didn’t. Observation studies can also find out how strongly a factor is linked to a certain disease. But a study like this can’t prove that a certain thing a person did caused them to get cancer, or that something else prevented cancer.
What’s a population study?
Population studies look at large groups of people. Researchers may study how often a group gets certain types of cancer and compare their cancer risk to certain lifestyle factors. There’s more than one way to do this: cohort studies follow the same group over time, and cross-sectional studies look at a group at a single point in time. Since cancer takes so long to grow, looking at the same group over time is often the best way to learn about cancer prevention methods.
Cohort studies: These studies take a group of people and watch them over time, testing them or asking them questions. (A cohort study may also be called a follow-up or longitudinal study.) A cohort study usually observes groups and compares people with different possible risk factors to see how these factors affect their outcomes.
Cohort studies can be prospective, meaning that the researchers select a group and follow them through time. They can also be retrospective, which means that the researchers find people (or often medical records) and look back at the group over time. For instance, a retrospective cohort study may look back at people who were exposed to radiation in Hiroshima to find out how many of them got cancer compared to similar people who were not exposed to radiation. A prospective cohort study looking at cancer risk may keep up with a healthy group of people over time and ask them about how much they exercise, what they eat, whether they take vitamins, or even take their blood to look at chemistries. They keep this information (and often collect more) while they wait and watch to see if there’s a difference in their cancer rates that might be linked back to any of these factors.
Cross sectional studies: These studies look at people at just one point in time. They look at how certain factors might relate to each other, but there are some drawbacks. There’s usually no way to find out what happened in the past without counting on a person’s memory. And there’s no way to find out what happens to the people after the study.
Cross sectional studies may observe people and look for links between their actions and cancer. They often give researchers ideas about what might be causing more cancers in some people, but these kinds of studies cannot show exactly what caused the cancers. Figuring out the cause requires further research, unless the link has already been proven.
For example, a cross sectional study that looks at the level of a certain vitamin and cancer may find that people with lower vitamin levels are more likely to have cancer than those people with higher vitamin levels. Can we assume just from this study that the vitamin protects against cancer? No, because we can’t tell from this study which came first, the cancer or the low vitamin levels. We also don’t know if the group had a high vitamin level because they ate healthy diets with lots of fruits and vegetables (which would contain many other things besides the vitamin). There are many other things that this kind of study can’t tell us. But it does give researchers ideas about what to look at next, so this kind of information can be a good place to start.
Case control studies: These studies look at people who already have a disease or condition, like cancer, and compare them to an otherwise similar group of people who don’t have the disease. Then, the researchers look at eating habits, exercise, drugs, or other factors to see if the groups are different in any ways that might explain why one group got cancer and the other didn’t. Most case-control studies are retrospective (meaning that they look back at the group over time).
A common problem with these types of studies is that people often remember events or habits from years ago in different ways based on what has happened more recently. If a person has cancer, for instance, he or she may recall having had worse eating habits than those who are well – even if their habits were the same. A person who’s still healthy may report better eating habits in the past than what actually happened, because there’s no reason to worry about them or try to remember the details. This is known as recall bias, which in this case is a type of misclassification error – it ends up putting people into the wrong groups.
In some studies, the poor recall of those being studied may be more random in its error. For instance, suppose you ask people whether they were exposed to high doses of a certain mineral, and they have trouble remembering. In this case, there may be nothing that pushes a number of people more toward one error than the other. This could mean that a number of people end up classified as having the exposure when they didn’t, and others are classified as not having the exposure when they actually did. This can dilute the groups enough that there’s no difference found between the groups, even though a difference would have been found if everyone was in the right group. This is another type of misclassification error. In this case, it’s another way recall bias can cause false results.
When they are done well, case control studies can be helpful in producing ideas about cancer causes and risk reducers. But conclusions about cancer prevention methods, even when based on a number of case control studies, are not as strong as those based on clinical trials.
How are observation studies misunderstood?
A study that only observes people cannot prove what factor caused an illness, but that doesn’t stop people from trying to guess at the cause and even writing about it as if the guess were fact. For instance, there were studies years ago that linked gum disease with heart attacks. News reports talked about this link, with many theories about how gum disease might cause heart attacks. The problem was that these were observational studies that could only show links, not find causes. The missing piece in many of the early study reports was that smokers are much more prone to gum disease. A lot of people with severe gum disease smoke; smoking causes both heart disease and gum disease. So the real culprit in some cases was smoking, not gum disease. Another explanation may be that people who don’t care for their teeth are less likely to eat well or get good preventive health care. That’s not to say that there’s no other link between gum and heart disease, but studies that simply observe people have trouble controlling for all the differences between people with and without the disease being studied.
There are other kinds of statistical observation studies that may get a lot of attention and lead to confusion. For example, people from a country where fish is eaten 3 to 5 times a week may have a lower risk of certain diseases than people from another country. A person reading about such a study might believe that the oils found in fish are responsible, and take fish oil supplements. Or they may sell the supplements, and cite the statistics as evidence. But it may turn out that the reason the people in the study had lower disease risk is that they ate less red meat, that they walked more each day, that they weighed less, or some other factor that wasn’t even discussed in the study. A closer look may even reveal that the people with the smallest cancer risk were not the ones who ate more fish.
As you can see, there are often many possible explanations for these types of findings in observation studies that the reader may not know about. It’s no wonder that people – and even reporters – can be confused by the news reports on observation studies.
Last Medical Review: 10/09/2014
Last Revised: 05/21/2015