Page 129 - An Evidence Review of Active Surveillance in Men With Localized Prostate Cancer

• The type of insurance (e.g., Medicare vs. private insurance) may be a determinant in the
choice of observational management strategies versus other treatments.
• Residing in an urban area (versus a rural area) may affect the probability of men
receiving observational management strategies versus active treatment.
• The use of decision aids may be informative and could encourage more active patient
involvement in the treatment decisionmaking process.

Key Question 4. What are the comparative short- and long-term outcomes
of active surveillance versus immediate treatment with curative intent for
localized prostate cancer?
In conformity with the Key Question, we systematically reviewed only comparative studies
that directly compared AS (or other observational management strategies) to immediate
treatment with curative intent. In order to understand the effectiveness of AS relative to active
treatment options, studies of AS need a control group for comparison. As such, we did not
include single-arm AS cohort studies, which cannot address comparative effectiveness questions.
Thus, this review does not address evaluations of long-term clinical outcomes reported in
noncomparative cohort studies, which are beyond the scope of the Key Question.

Clinical Outcomes
We did not identify any studies reporting clinical outcomes specifically of AS management
strategies compared with immediate definitive treatment. No study evaluated AS where the
intervention was employed in a predefined group of patients using predefined monitoring
methods to identify patients who would potentially be eligible for treatment with curative intent.
Faced with a lack of studies comparing AS to immediate treatment, we elected to evaluate
studies that compared other observational management strategies (largely resembling WW) with
immediate treatment. One can argue that efficacy results from studies of WW may represent the
low bound of the potential efficacy of AS.
In addition to previously published systematic reviews and evidence reports, our searches
identified updated results from a multicenter RCT (2 publications, one on clinical outcomes and
one on costs), and 12 cohort studies (2 prospective and 10 retrospective). Notably, the majority
of evidence for this key question came from retrospective analyses of observational studies. Due
to the differences in patient characteristics and risk profile between patients treated with
observational strategies and those who received active treatments, confounding by indication is
likely in these studies. Temporal changes in patient characteristics at diagnosis may affect the
magnitude of the reported treatment effect sizes in observational studies. Of the studies we
reviewed only two performed statistical analyses that accounted for year of diagnosis. 207,208
Another study included only patients diagnosed in a single year. 159 One study used pathology
information based on a contemporary reading of available pathology material in its statistical
209
adjustments.
Findings from Previous Systematic Reviews
We examined two recent systematic reviews of treatments of men with clinically localized
8,9
prostate cancer. Of note, these systematic reviews included some studies of men receiving

75