OBJECTIVE: To examine the psychometric qualities and develop the clinical utility of the Sydney Psychosocial Reintegration Scale (SPRS) as a measure of participation in people with traumatic brain injury. DESIGN: Data generated from previous prospective studies. SUBJECTS/PATIENTS: Convenience samples of healthy community-based volunteers (n=105) and people with severe brain injury (n=510). METHODS: (i) The equivalence of a new 5-point version of the SPRS was determined vs the original 7-point version; (ii) construct validity was tested using Rasch analyses; (iii) normative and comparative data tables were produced, and data examined for floor/ceiling effects; (iv) a reliable change index score was generated. RESULTS: Patterns of psychometric properties for the 5- and 7-point versions were almost identical (e.g. total scores rs=0.98). Rasch analyses on Forms A and B found good fit to the model, for person (3.36 and 3.03, respectively) and item (7.78 and 7.25, respectively) separation; reliability coefficients were high (all ≥ 0.90). Mean infit statistics met standard criteria (between 0.7 and 1.3). No floor/ceiling effects were detected. The reliable change index value was calculated for the total score for Form A using logit scores, and a conversion table provided. CONCLUSION: The 5-point version of the SPRS demonstrates strong psychometric qualities as a measure of participation after traumatic brain injury.
OBJECTIVE: To examine the psychometric qualities and develop the clinical utility of the Sydney Psychosocial Reintegration Scale (SPRS) as a measure of participation in people with traumatic brain injury. DESIGN: Data generated from previous prospective studies. SUBJECTS/PATIENTS: Convenience samples of healthy community-based volunteers (n=105) and people with severe brain injury (n=510). METHODS: (i) The equivalence of a new 5-point version of the SPRS was determined vs the original 7-point version; (ii) construct validity was tested using Rasch analyses; (iii) normative and comparative data tables were produced, and data examined for floor/ceiling effects; (iv) a reliable change index score was generated. RESULTS: Patterns of psychometric properties for the 5- and 7-point versions were almost identical (e.g. total scores rs=0.98). Rasch analyses on Forms A and B found good fit to the model, for person (3.36 and 3.03, respectively) and item (7.78 and 7.25, respectively) separation; reliability coefficients were high (all ≥ 0.90). Mean infit statistics met standard criteria (between 0.7 and 1.3). No floor/ceiling effects were detected. The reliable change index value was calculated for the total score for Form A using logit scores, and a conversion table provided. CONCLUSION: The 5-point version of the SPRS demonstrates strong psychometric qualities as a measure of participation after traumatic brain injury.
Authors: David N Borg; Jennifer Fleming; Joshua J Bon; Michele M Foster; Elizabeth Kendall; Timothy Geraghty Journal: BMC Health Serv Res Date: 2022-04-05 Impact factor: 2.655