How Studies Are Conducted
There are several factors to think about when looking at various studies and their results. Often, it is not enough just to know the result; we see this commonly in media headlines—“Soda pop found to cause deaths in people with diabetes” and when we dig deeper, we find that someone with diabetes would need to drink more than 48 bottles of nondiet soda a day to cause any effect. (That example is for explanation purposes only and is not real.) The point is that you have to know more than the headline whether it is results of a radiation study or any other study.
We are going to explain a few of the issues that complicate how to interpret the results that make it difficult to say whether radiation is good, bad, or neither.
1. In Vitro versus In Vivo
In vitro experiments are those performed outside of a living system (a human or an animal) usually on cells or tissues. In vivo experiments are those performed on a living system, usually a human or an animal. The questions here are, Can we take results from radiation exposure to cells sitting outside a human body or from radiation exposure to an animal and determine the impact on a human? Can we say that the effect on a cell outside the body would be the same if that cell were inside the body? When experiments such as these are conducted, scientists try to create a setting for the cells that is as true to what it would be in the human body as possible, or they pick an animal that most closely resembles a human for the organ system they are studying. Scientists work hard to let the reader of the study know the assumptions they had to make to come to any conclusion about the effects of low radiation doses on humans. Even with this, the questions still linger.
2. Population Studies
Based on Item 1, it would seem that looking at populations of people who have been exposed to radiation might give us better answers about low radiation dose effects than studies in cells or animals. True. But these are not without their complications, too. Each person has different activities and different family traits: some might smoke, some might exercise, some might live in a poor air-quality city, some might eat a lot of fruits and vegetables, some might have a family history of cancer, and on and on.
If a group of people is exposed to radiation, how do we find a good “control” group of people (a group not exposed to the extra radiation) with which to compare them? Do they all eat the same thing? Were they all raised in the same area? Do they all have the same job? This is one of the complications that make it difficult to see differences in disease rates when we are dealing with low radiation doses. Most studies try to match the control group with the exposed group by having people about the same age, about the same gender distribution, and usually from the same region of the world, but many other factors cannot easily be accounted for.
Probably the largest complicating factor in a population study looking at harmful effects of low radiation doses is that normal incidence rates for diseases are very high. That means we need a lot of people to study to get good statistical results. This is equivalent to rolling the dice more than six times to decide if they are loaded or if the results are statistical variation. Results from a low-dose radiation study with 10 people in it probably won’t be seen by scientists as being any good. Results from a low-dose radiation study with 1,000 people or more in it would be good. Results from a low-dose radiation study with a million radiation-exposed people in it and a million matched controls would be the best . . . but we’ll never have that because it is economically impossible to identify, characterize (smoker or nonsmoker, etc.), and follow that many people. So what we do is look at several of the smaller studies to see if they’re getting similar results and, if so, draw some conclusions based on that.
Studies known as ecological population studies involve large populations. An example is studying the cancer rates in the entire population of Maine compared with the population of Colorado, which has natural radiation levels more than two times higher than Maine’s natural radiation levels. This type of study is interesting, but it cannot answer the question of how low doses of radiation affect disease in an individual because it doesn’t include any information about the individuals in the populations of those states. For example, the cancer rate in Colorado is 11 percent lower than in Maine, but the radiation dose to the population is about 100 percent higher. However, this does not provide scientific evidence that higher radiation exposure reduces cancer rates. It just means that other lifestyle differences between the populations of Colorado and Maine are more important in determining cancer rates than low-dose radiation effects.