Article type
Year
Abstract
Background: TSA has been proposed as a method to assess whether the results of meta-analyses (MAs) are conclusive. However, the current TSA method can only be applied on count data and cannot incorporate time-to-event outcomes, which are prevalent in cancer studies.
Objectives: Present a novel method for the application of TSA for time-to-event outcomes. We use data from systematicreviews in multiple myeloma as an example to demonstrate the usefulness of the method. Pooled outcomes of overall survival (OS) and progression-free survival (PFS) were assessed for conclusiveness.
Methods: Extensive literature search identified 11 systematic reviews (23 MAs) of treatments for multiple myeloma. Based on the relative risk reduction (RRR) of 15% obtained from literature, we calculated the a priori anticipated information size (APIS) and the low-bias heterogeneity-adjusted information size (LBHIS) under the type I error of 5% and power of 80%. Both APIS and LBHIS incorporate the average rate of survival between the intervention and control groups, and potential patient loss to follow-up.
Results: Twenty three separate MAs (13 included OS and 10 PFS as outcome) were extracted (Table 1). The heterogeneity index I-sq varied between 0% and 85%and observed power from 9% to 100%. For OS, at a baseline RRR of 15% and using APIS, 54% of results (7/13) were false negative and 8% (1/13) were false positive. For the same RRR, LBHIS identified 77% (10/13) of results as false negative and 15% (2/13) as false positive results. For PFS, APIS identified 10% (1/10) as false negative and false positive results (1/10), and LBHIS also identified 10% (1/10) as false negative and zero false positive results.
Conclusions: For TSA, the new methods demonstrate the possibility of incorporating time-to-event data, which are important patient oriented outcomes. Time-to-event TSA can be extended to other fields where time-to-event outcomes are crucial.
Objectives: Present a novel method for the application of TSA for time-to-event outcomes. We use data from systematicreviews in multiple myeloma as an example to demonstrate the usefulness of the method. Pooled outcomes of overall survival (OS) and progression-free survival (PFS) were assessed for conclusiveness.
Methods: Extensive literature search identified 11 systematic reviews (23 MAs) of treatments for multiple myeloma. Based on the relative risk reduction (RRR) of 15% obtained from literature, we calculated the a priori anticipated information size (APIS) and the low-bias heterogeneity-adjusted information size (LBHIS) under the type I error of 5% and power of 80%. Both APIS and LBHIS incorporate the average rate of survival between the intervention and control groups, and potential patient loss to follow-up.
Results: Twenty three separate MAs (13 included OS and 10 PFS as outcome) were extracted (Table 1). The heterogeneity index I-sq varied between 0% and 85%and observed power from 9% to 100%. For OS, at a baseline RRR of 15% and using APIS, 54% of results (7/13) were false negative and 8% (1/13) were false positive. For the same RRR, LBHIS identified 77% (10/13) of results as false negative and 15% (2/13) as false positive results. For PFS, APIS identified 10% (1/10) as false negative and false positive results (1/10), and LBHIS also identified 10% (1/10) as false negative and zero false positive results.
Conclusions: For TSA, the new methods demonstrate the possibility of incorporating time-to-event data, which are important patient oriented outcomes. Time-to-event TSA can be extended to other fields where time-to-event outcomes are crucial.
Images
