Lyn, This turned out longer than expected.
I’m sure there is good ground for what you say but there is also good ground for another viewpoint on psa thresholds:
- It’s long been said that you should go to the same assay for your blood tests recognising that results can be different at different places.
- There are theories about the value of tracking psa levels to determine whether it’s fast growing or if it comes on quickly after the operation. That gives an indication about what the problem might be and an indication of prognosis, if you want it. An example is below from John Hopkins and there is similar on Harvard Prostate Knowledge:
.............................................................................................................................................................
‘If you have a Gleason score of 5-7, Your PSA increased more than two years after surgery AND your PSA doubling time was greater than 10 months:
Your chance of not developing metastasis(having a bone positive scan) in: Three years is 95 percent, Five years is 86 percent, Seven years is 82 percent’
In support of the higher threshold it also says on another page:
‘Sometimes, there is such a thing as too much information. According to Daniel W. Chan, Ph.D., the only thing that really matters is at which PSA level the concentration indicates that the patient has had a recurrence of cancer. The key threshold is 0.2 ng/ml, which indicates biochemical recurrence. However, it might take months or even years before there is any clinical physical evidence of symptoms.
In the laboratory, Dr. Chan trusts the sensitivity of levels down to 0.1 ng/ml or less. However, the results could vary from day to day. In routine practice, if the PSA is less than 0.1 ng/ml, it is considered the same as nondetectable or zero.’
(Note that the above is based on a report over 10 years old and nearly 20 years in parts which could mean some data is 30yrs old)
.............................................................................................................................................................
Also data is often scattered yet statisticians still determine trends. It must surely be better to find a trend earlier, and there is at least one person on here who has benefited from that. To have no data beneath 0.1 could be losing valuable insight, even if for most it’s of negligible implication as this can be a debate about the threshold of probability.
Also with regard to the point in your note about researchers not being concerned about the cost impact of their research. That could be the case but often the researchers don’t know where their research is heading. Give it to an accountant and they’ll spot an opportunity not in the researchers intention. Give it to a doctor and they might say something else.
Having read previous discussions on this it seems, to me, 0.03 is a compromise where inaccuracies are less significant and enable a trend to be established well before it gets critical. That being said I’m quite happy with 0.05 at the hospital I attend and you seem to be happy with the 0.1 threshold.
Hope that reads amicably enough. Regards
'I'm just a patient and only know feelings'