Autor(es): Sampaio, Cristina
Data: 2024
Identificador Persistente: http://hdl.handle.net/10400.5/97414
Origem: Repositório da Universidade de Lisboa
Autor(es): Sampaio, Cristina
Data: 2024
Identificador Persistente: http://hdl.handle.net/10400.5/97414
Origem: Repositório da Universidade de Lisboa
Gewily et al report on the significant modification of the well-established progressive supranuclear palsy rating scale (PSPRS), a process that was prompted and guided by feedback from the United States Food and Drug Administration (FDA). This study comprehensively compared the original (full PSPRS, 28 items) and the shortened only and shortened and rescored (PSPRS-10 and rPSPRS-10) versions of the PSPRS using item response theory (IRT). The study used data from 979 patients across four interventional clinical trials and two registries. The findings revealed that the PSPRS-10 contained 76% of the information from the full scale while demonstrating better unidimensionality and inter-item correlation. Some items in the full PSPRS showed poor correlation with disease severity, supporting the reduced scale to evaluate longitudinal disease progression. Power analyses indicated that PSPRS-10 was more effective than rPSPRS-10 in detecting treatment effects. The study concludes that focusing on the PSPRS-10 items is justified for assessing PSP severity and treatment effects in clinical trials, although rescoring may potentially weaken its sensitivity.