Article #32
Newsletter list 


Previous Article 


Information Presentation in Prostate Cancer Screening

Robert Hamm
Oklahoma City, OK

I am preparing to run a study of the effect of a clear presentation of the tradeoffs inherent in prostate cancer screening on men's understanding of prostate cancer screening and on their screening behavior in the next six months.

Men will be randomized to (a) read a standard patient handout, (b) read an "extended balance sheet" that we are producing, or (c) go through the extended balance sheet step by step with the research assistant playing the role of the infinitely attentive, nonrushed patient educator.

The balance sheet will differ from the typical presentation by giving the probabilities of events and outcomes using a single reference population (1,000 men who decide to get screened). You will recognize this to be an absolute frequency type display which Gigerenzer and Hoffrage have advocated.

We will actually put up a display with a thousand circles, similar to that used by Annette O'Connor (e.g., Medical Decision Making, 1998, 18, 295-303). This display will be used to illustrate the prevalence of localized prostate cancer for men of a particular age group, screening accuracy (true and false positives and negatives), treatment effects (those helped, those for whom treatment was unnecessary because they wouldn't have gotten advanced prostate cancer anyway, and those for whom treatment was useless because they got advanced prostate cancer despite the treatment), and side effects of treatment. The summary will display the eight men (out of the 1,000 screened) who avoided getting prostate cancer, versus the 20 or so who became permanently incontinent or impotent.

The effects we will measure include whether they get screened in the next six months. (We only enroll men who have not been screened in the last year so that if they decide screening is for them, they would be "due.") Also, we have measures from the health beliefs model (concepts of perceived benefit from screening, etc., which can be mapped onto a subjective expected utility model), and we will look for changes in these measures.

Our outcome with the biggest theoretical baggage is that we will measure the men's utilities and plug them into an individualized decision analysis (Markov model). We can then see whether their action is consistent with their utilities and determine whether our extended balance sheet helps them take actions more consistent with their utilities.

As I look over this and ask whether this is Brunswikian rather than simply "judgment and decision making" or "medical decision making," I think it qualifies: People on the extended research team have come from a broad sample of academic environments, including psychology (linguistics, animal behavior, and judgment and decision making), medicine (family medicine and urology), sociology, health promotion sciences, bilingual education, human ecology, and decision analysis.

Contact Robert Hamm

Previous Article

Newsletter list