Health knowledge made personal
Join this community!
› Share page:
Search posts:

Changing physicians’ behaviors (about PSA testing)

Posted Nov 16 2010 12:00am

It is common for patients to suggest that primary care physicians should be more knowledgeable about the appropriate use of PSA testing and about how to interpret the results of such tests. However, changing physician behavior is not an easy or a quick process, as the following article will demonstrate.

Kerfoot et al. thought that they could use a targeted, e-mail-based intervention process, which they refer to as “spaced education,” to reduce the inappropriate use of PSA testing for risk of prostate cancer in men under 40 years of age and over 76 years of age. (To be clear, there are obviously some appropriate young and old candidates for such testing but, in general, the widespread use of PSA “screening” of men under 40 or over 76 is not recommended by any set of clinical guidelines.)

To test this hypothesis, Kerfoot et al. carried out a randomized controlled trial among 95 primary care physicians practicing at  eight Veterans Affairs medical centers between January 2007 and February 2009. The physicians were divided into two groups:

  • Group A comprised the “spaced education” group, who received nine educational e-mails in four cycles over a period of 36 weeks.
  • Group B comprised a control group who received no specific educational intervention.

For the physicians in Group A, each e-mail gave a clinical case scenario and then asked whether it was appropriate to obtain a PSA test. The participants were given immediate feedback after submitting their answers.

The primary outcome of the study was the number and percentage of PSA tests actually being ordered by the participants in their day-to-day clinical practice for real patients who met the predetermined criteria for inappropriateness (as defined above). Patients who were getting PSA tests for other reasons than just testing  for prostate cancer risk (e.g., because the patient had problems urinating) were excluded using a structured and validated protocol.

Here are the results of the study:

  • During the actual intervention period (weeks 1-36), clinicians in Group A ordered significantly fewer inappropriate PSA tests than clinicians in Group B  (10.5 vs 14.2 percent, respectively).
  • During the 72-week period after the intervention (weeks 37-108), clinicians in Group A continued to order significantly fewer inappropriate PSA tests than clinicians in Group B (7.8 vs 13.1 percent, respectively).
  • The reduction in inappropriate PSA testing during the 72-week follow-up period represents a 40 percent relative reduction in inappropriate use of PSA testing to screen for prostate cancer risk.

Kerfoot and his colleagues conclude, correctly, that, “Spaced education durably improves the prostate cancer screening behaviors of clinicians and represents a promising new methodology to improve patient care across healthcare systems.”

However, it is also worth noting that among the primary care doctors in Group B, who received no targeted education at all, over 85 percent were using PSA testing appropriately to test for risk of prostate cancer (within the limits of the study protocol), and that the absolute impact of the educational initiative was only to reduce use of “inappropriate” PSA testing by about 5 percent. If we can only accomplish a 5 percent change in physician behavior using such an intensive educational method, this gives readers some idea of the problem of changing physician behavior.

There are a thousand reasons why changing the behaviors of highly educated adults can be difficult. Actually this is true in almost every profession. And it is just as true of patients as it is of their doctors.  Millions of patients know that they should eat less (or just better), exercise more, stop smoking, not drink and drive, you name it but we go on doing what we have always done anyway. It is easy to say that the primary care docs should know more about PSA testing and carry it out more appropriately. Getting them to actually do that … and to interpret the resulting data with greater sophistication (which Kerfoot et al. didn’t even try to address) is a more challenging opportunity!

Post a comment
Write a comment:

Related Searches