Blog Feed

We have a guest blogger today, Dr. Robert Reiman[1]. Enjoy.

 

 

We’ve read articles or heard news reports where the ending is something like “This study suggests that “A” causes “B”.”

 

What does that mean exactly? In statistics terminology, it could mean a few different things: that “A” is likely the cause of “B”, that “A” is somehow related to “B”, or that right now, all we’ve looked at about “A” is to see if it somehow influences “B”.

 

Let’s put words on the “A” and the “B”. Let’s say that “A” is cooler temperatures and “B” is the peach crop. So when you read an article or hear a report saying “This study suggests that cooler temperatures cause a decrease in the number of peaches that ripen,” it could mean any one of the three relationships listed above.

 

The point is that when the word “suggests” or a similar word is used, the link between “A” and “B” isn’t proven. When an article or news report uses the term “suggests,” we should hear or read it as “It is being looked at; other things must also be considered” or “Further study is warranted.”

 

Why? Well, according to a recent news item, a recent State of Michigan health report suggested that there was an increase in cancer rates in one county where a nuclear plant was located. The health report itself did not indicate any relationship to radiation; however, the news article was quick to link the report data to radiation from the plant. Let’s take a closer look.

 

From the report: During a ten year period, the number of cancers for the population under the age of 25 in the affected county rose at a rate more than three times faster than that for the rest of the state; it rose from 18.5 cases to 23.4 cases per 100,000 people. In the rest of Michigan, the rate rose from 20.2 cases to 21.9 cases per 100,000 people.

 

Looking closer: The report cited the normal number of cancers in this population to be "18.5 cancer cases per 100,000 people" during the ten year period. This means that, on average during this time period, 18.5 people out of 100,000 people would get cancer in any given year. Note that 18.5 cases is an average; from year to year, the average number of cases per 100,000 people under 25 could vary between 10 and 35 (in one year the rate might be 10 cancer cases and in the next year it might be 30 cancer cases, but it averages to 18.5). We see this variability in Las Vegas. If you roll two dice 1,000 times, most of the time the toss will total 7, but about 50 of the rolls would be “snake-eyes” (both ones) or “boxcars” (both sixes).

 

Since there are about 50,000 people under the age of 25 living in this particular county, the normal number of cancers in people under 25 is an average of nine; about nine people under 25 in this county would be expected to have a cancer in any given year. From year to year, this could vary between 5 and 17 cancer cases. This is the normal number of cancers with no additional radiation exposure.

 

The increase in the rate of cancer (18.5 to 24.3 cases per 100,000), as discussed in the health report, is based on three additional cases of cancer. This difference, 9 cancer cases versus 12 cancer cases, is not out of line of what we would expect based upon random variability. We would expect up to 17 in any given year.  

 

From the report: Critics of the plant and nuclear power say more study is needed.

 

Looking closer: When the number of people getting a disease is close to or outside what is expected, more study may be needed. When rates fall close to what is expected year after year, more study is a wasteful use of taxpayer dollars.

 

From the report: A professor at a local university stated that cancer among young people should be viewed as an indicator for radiation problems associated with nuclear plants.

 

Looking closer: All of the state and federal reports regarding the nuclear plant in question show that any plant releases of radiation are below what is allowed by law. Statements such as the one from the professor leave out an entire host of possible reasons why the number of cancers might be increasing: what are these young people eating and drinking, what are they exposed to in their homes, what are their living conditions, what chemical or other environmental pollutants might they be exposed to, what other radiation are they exposed to, and so forth.

 

 

While the cancer rates in this county are not outside what would be expected, the issue is that by focusing solely on the plant and the belief that radiation emitted from the plant is causing all the disease issues in the area, real reasons for increases in disease will be overlooked.

 

So when a study result suggests something, be careful not to read that the study has proven something.

 

---