Article type
Year
Abstract
Background: The overall therapeutic effects from randomized controlled trials (RCTs) are often biased as investigators seldom take the treatment-related harms (treatment-related morbidity and mortality) into account. To date there has not been a systematic assessment of the quality of reporting of treatment-related harms (TRH) and its impact on overall therapeutic effects of cancer treatments.
Objectives: To assess the quality of reporting of TRH and its impact on overall therapeutic effects.
Methods: We reviewed all consecutive phase III RCTs conducted by National Cancer Institute’s co-operative groups from 1960 to 2007. The quality of TRH reporting was classified as good (both numerator and denominator for TRH are reported), intermediate (qualitative reporting or either numerator or denominator are missing for TRH) or poor (no TRH reporting). Association of overall therapeutic effects with quality of TRH reporting was evaluated using chi-square tests and meta-analytic techniques.
Results: We evaluated 672 RCTs (754 comparisons, 223 084 patients). The quality of reporting of treatment-related morbidity was ‘good’ in 23% (175/754), ‘intermediate’ in 63% (475/754), and ‘poor’ in 14% (104/754) of studies. The quality of reporting of treatment-related mortality was ‘good’ in 58% (435/754), ‘intermediate’ in 8% (64/754), and ‘poor’ in 34% (255/754) of studies. There was no association between treatment superiority as concluded by trial investigators and quality of reporting of treatment-related morbidity (p = 0.95) or treatment-related mortality (p = 0.93). Meta-analysis according to quality of reporting of TRH also did not show an association in terms of favoring experimental or standard treatments for the outcomes of overall survival, event-free survival or the primary outcome (Fig. 1).
Conclusions: We found no evidence of systematic bias attributed to quality of reporting of treatment-related harms on overall therapeutic effects. Investigators judged both harms and benefits when they drew conclusions about treatment superiority in this cohort of studies.
Objectives: To assess the quality of reporting of TRH and its impact on overall therapeutic effects.
Methods: We reviewed all consecutive phase III RCTs conducted by National Cancer Institute’s co-operative groups from 1960 to 2007. The quality of TRH reporting was classified as good (both numerator and denominator for TRH are reported), intermediate (qualitative reporting or either numerator or denominator are missing for TRH) or poor (no TRH reporting). Association of overall therapeutic effects with quality of TRH reporting was evaluated using chi-square tests and meta-analytic techniques.
Results: We evaluated 672 RCTs (754 comparisons, 223 084 patients). The quality of reporting of treatment-related morbidity was ‘good’ in 23% (175/754), ‘intermediate’ in 63% (475/754), and ‘poor’ in 14% (104/754) of studies. The quality of reporting of treatment-related mortality was ‘good’ in 58% (435/754), ‘intermediate’ in 8% (64/754), and ‘poor’ in 34% (255/754) of studies. There was no association between treatment superiority as concluded by trial investigators and quality of reporting of treatment-related morbidity (p = 0.95) or treatment-related mortality (p = 0.93). Meta-analysis according to quality of reporting of TRH also did not show an association in terms of favoring experimental or standard treatments for the outcomes of overall survival, event-free survival or the primary outcome (Fig. 1).
Conclusions: We found no evidence of systematic bias attributed to quality of reporting of treatment-related harms on overall therapeutic effects. Investigators judged both harms and benefits when they drew conclusions about treatment superiority in this cohort of studies.
Images
