The Template for Intervention Description and Replication as a Measure of Intervention Reporting Quality: Rasch Analysis

• Main Authors: Marcel P. Dijkers, PhD, FACRM, Scott R. Millis, PhD, ABPP, CStat
• Format: Article
• Language: English
• Published: Elsevier 2020-09-01
• Series: Archives of Rehabilitation Research and Clinical Translation
• Subjects: Guideline adherence; Rehabilitation; Reproducibility of results; Research report
• Online Access: http://www.sciencedirect.com/science/article/pii/S2590109520300203

Objective: To determine whether the 12 items of the Template for Intervention Description and Replication (TIDieR) can be combined into a single summary score reflecting intervention reporting completeness and quality. Design: Systematic review and reanalysis of published data. After a systematic search of the published literature, 16 review articles were retrieved with 489 sets of 12 TIDieR ratings of experimental intervention, comparator, or the 2 combined as reported in primary studies. These 489 sets were recoded into a common format and analyzed using Rasch analysis for binary items. Setting: Not applicable. Participants: Not applicable. Interventions: Not applicable. Main Outcome Measures: Psychometric qualities of a Rasch Analysis-based TIDieR summary score. Results: The data fit the Rasch model. Infit and outfit values were generally acceptable (range, 0.70-1.45). TIDieR was reasonably unidimensional in its structure. However, the person (here: study) separation ratio was 1.25 with a corresponding reliability of 0.61. In addition, the confidence interval around each estimate of reporting completeness was wide (model standard error of 0.78) Conclusion: Several Rasch indicators suggested that TIDieR is not a strong instrument for assessing the quality of a researcher’s reporting on an intervention. It is recommended that it be used with caution. Improvements in TIDieR itself may make it more helpful as a reporting tool.