« Neat reads | Main | The paradox of screening 2 »


Feed You can follow this conversation by subscribing to the comment feed for this post.

Pierre-Hugues Carmichael

As I understand it, positive screening tests are usually followed by more rigorous diagnostic tests, that is to say tests that are more sensitive than screening tests, right ? Am I also right in assuming that screening tests are designed to have as high a specificity as possible ?

I agree that screening tests should only be administered to "sufficiently at risk" individuals, that is to say that there is probably no need to screen young men for prostate cancer. However, I am left wondering if the problem of false positives is purely economical (that is to say these false positives cost more to society because they have to undergo the more expensive diagnostic tests). Is past information taken into account when analysing the results of a screening test ? Say your first three screenings are negative and your fourth is positive, does that increase the chances that this last result is a false positive ? I would love to read your thoughts on the question.


PHC: Love the thoughtful comment.

I tried to find some numbers for you. It appears that the PSA test has sensitivity of 80%+ and specificity of 33%. That means, someone who is healthy has a 2/3 chance of testing positive. (This comes from a Hoffman, Gilliland, et. al. Albuquerque study, cited by Prof. Jeff Douglas at Illinois.) This is worse than randomly flipping a coin.

For the PSA test, a negative means something but a positive means nothing so if you had three negatives in a row, I wouldn't be too worried. But... I'm assuming these are tests taken in quick succession because if these tests are taken years apart, there is no reason to believe the health status of the patient has not changed.

Economics is an important consideration since money is a finite resource. But for prostate cancer, we don't need an economic argument. As Dr. Ablin pointed out, clinical trials showed that those receiving PSA screening had the same death rates as those who didn't (both very low).

So, to summarize, this test costs $3 billion a year, tells 2/3 of healthy people they have cancer, and does not help the identified sick people prolong their lives. Pretty damning, huh?

Jon Peltier

Positive screening tests (especially mammograms and PSA tests) are followed-up in one of two ways:

1. Wait-and-see: have a repeat test in 3 or 6 months.

2. Biopsy.

The comments to this entry are closed.


Link to Principal Analytics Prep

See our curriculum, instructors. Apply.
Kaiser Fung. Business analytics and data visualization expert. Author and Speaker.
Visit my website. Follow my Twitter. See my articles at Daily Beast, 538, HBR.

See my Youtube and Flickr.

Next Events

Jun: 5 NYPL Public Lecture on Analytics Careers, New York, NY

Past Events

Apr: 2 Data Visualization Seminar, Pasadena, CA

Mar: 30 ASA DataFest, New York, NY

See more here


Numbersense: Statistical Reasoning in Practice, Principal Analytics Prep

Applied Analytics Frameworks & Methods, Columbia

The Art of Data Visualization, NYU

Junk Charts Blog

Link to junkcharts

Graphics design by Amanda Lee


  • only in Big Data