Page 439 - An Evidence Review of Active Surveillance in Men With Localized Prostate Cancer

Appendix Table C3.2. KQ3 multivariable analyses (continued)
Author Factors Data source Duration Analyzed Population WW/AS Methods Results as described in paper
yr sample characteristics definitions
PMI
Wolters 171 Clinical Post hoc 1993- 8010 low, Not Polytomous 1.  Age OR=1.19 (CI 1.17, 1.21)
2010 analysis of 2006 (completed intermediate explicitly logistic 2.  PSA OR=0.30 (0.23, 0.39)
19739124 ERSPC data set) and high risk provided regression 3. T2 vs. T1 =OR 0.33 (CI 0.28,
cancer predicts AS 0.39)
compared to 4. PSA 50+ vs. ≤4.0 OR=1.73 (CI
RP 1.02, 2.94)
5. Gleason ≥8 vs. ≤6 OR=0.20
(CI 0.13, 0.32)

NS: study arm; lymph node
involvement
van den Patient Analysis of 2007- 129 Men who Closely Multivariable Patients who perceived that
Bergh 181 2009 preference men in PRIAS 2008 decided on AS monitoring regression to physician played the most
19637245 (PSA ≤10 for disease predict important role in shared decision-
ng/mL, PSA progression decisional making process also had more
density <0.2 to decide conflict doubts (high decisional conflict)
ng/mL/mL, on initiating regarding the choice for AS.
localized or curative
nonpalpable therapy Involvement of physicians in the
disease, ≤2 decision-making process was
biopsy cores; assessed by a non-validated
≤3+3 Gleason) instrument.
Sommers 178 Patient Survey of 428 2004- 167 had T1, T2N0M0, Not Logistic 1. Desire to avoid side effects
2008 preference eligible men 2007 eligible + not yet treated explicitly regression main predictor of choice of WW
18704993 with newly analyzable provided predicts (logistic regression coefficients not
dx’d localized data choice of provided, P<0.05)
cancer from 2 WW vs. other
RT and 2 treatments or 2. “Current bowel problem” was
urology clinics undecided also a predictor of choice of WW
in Boston (logistic regression coefficients not
provided, P<0.05)

C-137