Sydney Psychosocial Reintegration Scale (SPRS‐2): Meeting the Challenge of Measuring Participation in Neurological Conditions

Abstract

This three‐part article presents: (1) a review of the construct of participation; (2) an overview of the 12‐item Sydney Psychosocial Reintegration Scale (SPRS) as a measure of participation and as a description of new developments resulting in its revision (SPRS‐2); and (3) as an application in different neurological groups. Psychometrically, the SPRS performs very well. There are no significant floor or ceiling effects, and both Form A (change since injury) and Form B (current status) show excellent inter‐rater and test–retest reliability. There is also substantial evidence for its validity (concurrent, discriminant, and convergent/divergent), and it shows good fit to a Rasch model, providing evidence for its construct validity. Rasch logit scores are used to provide reliable change index values to determine whether an individual patient's scores have improved or deteriorated. In Part 3, three samples were compared: traumatic brain injury (TBI; n = 130), primary brain tumour (PBT; n = 54), and spinal cord injury (SCI; n = 50) on the SPRS‐2 Form B. The TBI sample performed more poorly than the PBT and SCI samples on the interpersonal relationships and occupational activity domains. On the living skills domain, the TBI and SCI groups both performed more poorly than the PBT sample. These data demonstrate the differential levels of participation observed in different neurological groups.

Key words:

• assessment
• brain injuries
• community integration
• participation
• psychometric properties
• Sydney Psychosocial Reintegration Scale

Acknowledgements

Studies from which data were drawn for the present article were funded by the Motor Accidents Authority of New South Wales and the New South Wales Cancer Institute.

Notes

1. For example, both McColl, Davies, Carlson, Johnston, and Minnes (2001) and Reistetter, Spencer, Trujillo, and Abreu (2005) found a correlation coefficient of r = 0.34 between the Community Integration Measure and the Community Integration Questionnaire.

2. For example, Willer, Rosenthal, Kreutzer, Gordon, and Rempel (1993) reported a correlation coefficient of r = 0.25 between social domains of the Community Integration Questionnaire and the Craig Handicap Assessment and Reporting Technique; Kuipers et al. (2004) found a correlation coefficient of r = 0.41 between the respective productive/occupational domains of the Community Integration Questionnaire and SPRS.

3. Factual background data on the three SPRS domains is recorded using an additional 15 items, but this information is not scored for the purposes of the scale.

4. Tate et al. (2011) also suggest using the minimum difference of 8.23 logit scores for Form B, even though we do not have the test–retest data on the Form B sample (n = 150) to calculate the actual minimum difference. We consider that using the logit value of 8.23 will provide a close approximation for a reliable change index for the following reasons: (1) the logit scores for Form A and Form B show very close correspondence; and (2) the test–retest correlation coefficient for the original psychometric studies on Forms A and B were high and identical (ICC = 0.90; Tate et al., 1999, 2004). Clearly, however, this suggestion requires verification.