Article type
Year
Abstract
Background: Permanent interstitial low-dose rate brachytherapy (LDR-BT) is a treatment alternative for patients with localized prostate cancer. As evidence from research on effectiveness or safety may not be decisive, patient-reported outcomes (PRO), such as function, bother, and generic health-related quality of life (HRQL), could have an impact on medical decision making.
Methods: Bibliographic databases (MEDLINE, EMBASE, and The Cochrane Library) were searched until June 2010 for randomized as well as non-randomized parallel group studies. Validated questionnaires, 70% or higher response rate, and 80% or more data with relevant diagnosis and therapy were required for inclusion of studies. For non-randomized controlled studies, baseline data were required to be comparable between treatment groups or results were required to be adjusted for major confounders.
Results: We identified only 1 randomized study among 61 studies potentially relevant for inclusion. PRO were reported in 32 of 61 and 7 of these 32 studies were finally included. Thus, 25 studies were excluded because baseline data were not comparable and not adjusted or stratified (16studies), response was below 70% (4studies), or data were not reported separately for the accordant groups or outcomes (5studies). Of the 7studies, 7/5/5 studies evaluated function/bother/generic HRQL scores using 7/2/3 different instruments, respectively.
Conclusions: We found that the majority of identified non-randomized controlled PRO studies showed selection bias too high for inclusion. Included PRO studies lacked statistics about differences across treatment groups and various scales complicated the synthesis of results.
Methods: Bibliographic databases (MEDLINE, EMBASE, and The Cochrane Library) were searched until June 2010 for randomized as well as non-randomized parallel group studies. Validated questionnaires, 70% or higher response rate, and 80% or more data with relevant diagnosis and therapy were required for inclusion of studies. For non-randomized controlled studies, baseline data were required to be comparable between treatment groups or results were required to be adjusted for major confounders.
Results: We identified only 1 randomized study among 61 studies potentially relevant for inclusion. PRO were reported in 32 of 61 and 7 of these 32 studies were finally included. Thus, 25 studies were excluded because baseline data were not comparable and not adjusted or stratified (16studies), response was below 70% (4studies), or data were not reported separately for the accordant groups or outcomes (5studies). Of the 7studies, 7/5/5 studies evaluated function/bother/generic HRQL scores using 7/2/3 different instruments, respectively.
Conclusions: We found that the majority of identified non-randomized controlled PRO studies showed selection bias too high for inclusion. Included PRO studies lacked statistics about differences across treatment groups and various scales complicated the synthesis of results.