Interpreting epidemiological research results with caution

Every so often, we come across articles that offer a “teaching opportunity” for the laymen. Such articles are often associated with what feels like over-interpretation by scientists of the results of their research.

A recent article by Shavers et al. certainly seems to fall into this category.

The authors start with two suppositions:

  • First that “Men who do not perceive themselves to be at risk of developing and dying from prostate cancer may be less likely to utilize prostate cancer screening.”
  • Second that, “This coupled with variation among organizations in recommendations about screening with prostate-specific antigen (PSA), may contribute to confusion for providers and/or patients making prostate cancer screening decisions.”

The authors examined data on 1,075 African American, Hispanic, and non-Hispanic white male respondents to the 2003 Health Information National Trends Study (HINTS) to see if there was any association among demographic characteristics, perception of the risk of developing prostate cancer, and PSA test utilization among men aged 45 or older. (And note that an “association” does not necessarily imply a cause and effect relationship.)

The authors state the following results and conclusions:

  • African American men less frequently, while Hispanic men more frequently, perceived their risk of developing prostate cancer to be higher than the average same-age man compared to non-Hispanic white men.
  • Overall, men who perceived their likelihood of getting prostate cancer as very low to moderate … or perceived the likelihood compared to the average same age man as less or about as likely …  were significantly less likely to have received a PSA test in a model adjusted for age, marital status, education, and health insurance coverage.
  • These findings suggest that all men, but particularly African American and Hispanic men, could benefit from information regarding their specific risk of developing prostate cancer before making a decision about prostate cancer screening.

However, what we actually know from this study is only that the men said certain things in response to certain specific questions in 2003. We do not know why they said them. We do not know if their responses were accurate. And we do not know whether the study data were sufficiently accurate to allow the conclusion suggested.

The “New” Prostate Cancer InfoLink is not, in fact, aware of any reason to believe that African Americans and/or Hispanics are at any greater need than whites for “information regarding their specific risk of developing prostate cancer before making a decision about prostate cancer screening.” There are, for example, recent data to suggest that African American men are more likely than white men (on average) to undergo PSA testing (although that may not have been true in 2003, when the original data were collected).

We would suggest that the data in this study are at best hypothesis-generating and at worst potentially misleading. The study could all too easily lead to newspaper headlines saying things like, “African Americans less likely to understand need for PSA testing,” but the data certainly wouldn’t appear to justify such a headline.

The full paper may resolve some of these issues — but the abstract alone is problematic. Over-interpretation of the data in a scientific study, and particularly the over-interpretation of the data in an abstract of the full paper, is always a potential problem. The lay person will almost inevitably treat such “scientific” statements as if they are some form of “truth.” But commonly they aren’t. They are simply the authors’ attempt to summarize a series of very complex data, and they need to be read and interpreted with the greatest caution, because they may be strongly influenced by the authors’ own perceptions and beliefs.

One Response

  1. Mike,

    People tend to trust people who they identify with and also the more you say something the more people believe it is true. Who do they trust and what is the message that they are hearing? How often do they hear the message? It would be nice if everybody made decisions or trusted the facts but so much has to do with who is delivering the message and how much that person is trusted. A man’s fears play a role in the whole discussion. Somehow facts seem to get in the way. Just think about the current debate about health care or go to a town hall. If someone emphasizes only one side of the facts and you try to give both sides and the first person is someone you trust, then you have a very difficult time getting a balanced message out. The problem is experienced when the people on the extremes are the ones with the loudest voices and get the most press. If those people are in a position of trust, and are part of a group that you trust, then a message may not necessarily get out in a balanced manner. For those of us who can understand the subtleties and understand that the messages are not proven but instead are based on incomplete knowledge. Unfortunately we do not have the information right now. Most people do not understand those subtleties. They want a definitive statement to be made. If the messages are conflicting, as human beings we tend to go back to our own experiences to make a decision.

    Right now there is messaging is that men are over-treated and most men will not die from prostate cancer. If that message is delivered by someone in your own ethnic or cultural group and cancer scares you, then wouldn’t it make sense for you to avoid testing? As someone who works in the community, the results of this study doesn’t surprise me. As one Princeton educated AA man said to me at a jazz festival, I won’t get tested because I have good karma. If the prostate cancer community has a problem, it is that both sides push their ideas based on their experience. The “right” thing is probably somewhere in the middle, but given the polarity that we have in our country right now, I am not sure that is possible.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: