Phenom,
Good questions.
Phenom said...
1) The difference in mortality between the treated and untreated cohorts you state as 2.6% of the total. But couldn't the data also be interpreted as showing a 50% increase in untreated mortality over treated mortality? (31 untreated died vs. 21 treated died out of approximately the same number of patients.)
When the statistical analysis shows that two numbers are "not significantly different" from one another, it means that the 2.6% could just as easily been found to be 0% if the full population and not just a sample had been used in the study. In statistical inference-speak, one has to reject the hypothesis that the numbers are different. So you can see why one can't manipulate the numbers as you propose. They aren't 50% different, they are no different as far as we know.
Phenom said...
2) Did it really show that only 17% of men presenting with PC symptoms (pre-PSA era) died from it over the next 30 years? Even for an older cohort, that seems like a very low mortality for these far advanced cases. And do we know if these men had any treatments?
Here's what it says (p.429):
"Patients were followed for 32 yr and only 3 (1%) of the 223 patients were alive at the end of the study. Distant metastasis occurred in 41 of 90 patients who experienced local progression (44.4% of those with progression, 18% of total). We observed 38
(17%) deaths due to PCa. The mean times until development of metastasis and PCa death were 9.2 and 9.5 yr, respectively. Among 99 patients who were aged ≤70 yr at diagnosis, 53 (53.5%) experienced local progression and 25 (24%) had cause of death attributed to PCa. Of 87 patients who were aged >75 yr at diagnosis, 21 (24.1%) experienced local progression and 8 (9.2%) died of PCa. The frequency of progression and PCa death was similar between patients with palpable and nonpalpable tumors. In the low-risk group, nine patients (12.9%) progressed to metastatic disease and died from PCa. Five of nine patients (55%) with poorly differentiated cancer (grade III) died from PCa and the remainder died from other causes."
They were not curatively treated, but they were palliatively treated:
"We used a population-based cohort of men diagnosed with early, initially untreated PCa in central Sweden as described in detail previously... 223 patients were diagnosed with localized disease by digital rectal examination (DRE) with absence of distant metastases (T0–T2 tumors) and were untreated... All 223 patients were followed from diagnosis until the date of death or the end of the study (June 30, 2010). No patients were lost to follow-up... Men with symptomatic local progression or metastasis were treated with castration (surgical or gonadotropin-releasing hormone analogs), exogenous estrogens, or antiandrogen. Two men were treated with palliative radiotherapy instead of castration."