The efficacy of a medical test is determined by three numbers.
1) The real incidence rate - what percent of the population (after the fact) actually has the condition.
2) The false positive rate.
3) The false negative rate.
The problem with the PSA test is that while the real incidence rate is relatively high, the false positive and false negative rates are extremely high.
1) The incidence rate varies with age and ethnicity. According to the CDC (and wikipedia, for what it's worth), (http://www.cdc.gov/cancer/prostate/statistics/race.htm), the age-averaged rate is 100 per 100,000 for asians, 160 per 100,000 for white and 250 per 100,000 for black men. But they don't recommend the test for men under 45. And age really is the determining factor. (http://seer.cancer.gov/publications/prostate/inc_mort.pdf). So, let's assume an incidence rate of 10% for 55 year old men for purposes of this exercise. - for 75 year old men, it's probably closer to 90%, for 20 year old men, essentially zero.
2) According to the National Cancer Institute (http://www.cancer.gov/cancertopics/factsheet/detection/PSA), the false positive rate is 65-75%. Giving the test the best chance, I'll take the lower limit of 65%.
3) I haven't found a definitive source for the false negative rate, but wikipedia cites a paper giving a 25% false negative rate. Let's give it the benefit of the doubt and call it 20%.
^ Thompson IM, Pauler DK, Goodman PJ, Tangen CM, Lucia MS, Parnes HL, Minasian LM, Ford LG, Lippman SM, Crawford ED, Crowley JJ, Coltman CA (May 2004). "Prevalence of prostate cancer among men with a prostate-specific antigen level or 4.0 ng per milliliter". N. Engl. J. Med. 350 (22): 2239–46)
So - give a population of 1 million men a PSA test and here's what you get.
100,000 men have prostate cancer
900,000 do not have prostate cancer.
Of the 100,000 men who DO have prostate cancer...
a) 20,000 test negative (a problem, but what are you going to do? This was my father's case (see below))
b) 80,000 test positive (okay, but do you need treatment? Odds are you'll die of something else first (see below))
Of the 900,000 who do NOT have prostate cancer.
c) 585,000 test positive (the real problem)
d) 315000 test negative. (good on yer.)
The real problem is that honking huge false positive rate. If you test positive, there's still less than a 50% chance that you actually have prostate cancer, and even if you do, it's probably not going to make a damn bit of difference over the course of your life, but it's still very very scary and you get a biopsy or have radiation treatment and risk impotence and/or incontinence and possibly seriously reduce your quantity of life for the rest of your life for no good reason. It's even more complicated by the fact that the PSA level goes up naturally as you age. If your level goes from 4 to 10 over 10 years, what does it mean? Flip a coin.
That said, given that my father, both his brothers and my paternal grandfather all died from prostate cancer (between the ages of 90 and 94, I'll grant you - that's the thing. The vast majority of men will die of something else before the prostate cancer kills them), my doctor recommends continuing to take the test every 5 years.