Date Presented 03/28/20
Confirmatory factor analyses were performed on models of participation depicting multiple assessments of poststroke participation as measuring the same construct. Results suggest that the Activity Card Sort, Reintegration to Normal Living Index, and Stroke Impact Scale–Participation subscale are tapping the same construct. A professional assumption is highlighted, and similar studies are encouraged to strengthen the profession’s ability to validly demonstrate its impact on participation.
Primary Author and Speaker: Matthew Crowley
Contributing Authors: Annie Fox, Marjorie Nicholas, Lisa Connor
PURPOSE: Conceptualizations of the participation construct vary, which raises the question of whether different assessments of participation are truly capturing the same underlying, unitary construct (Badley, 2008; Eyssen, Steuljens, Dekker, & Terwee, 2011). Stroke is a prevalent health condition that consistently results in a myriad of participation restrictions (de Graaf et al., 2018). The Activity Card Sort (ACS), Stroke Impact Scale – Participation subscale (SIS Participation), and Reintegration to Normal Living Index (RNL) are three common measures of post-stroke participation. These assessments have each demonstrated internal consistency, but have never been shown to jointly have construct validity within a single population. To demonstrate validity of the measurement of participation as an outcome across assessments, it is essential to rectify this shortcoming in the literature. The purpose of this study was to confirm whether multiple assessments of post-stroke participation are measuring the same underlying construct.
DESIGN: This study was quasi-experimental and cross-sectional by design. Participants (n = 105) were recruited from the Aphasia Center and Occupational Therapy Center at the MGH Institute of Health Professions, the Washington University in St. Louis Cognitive Rehabilitation Group Stroke Registry, and stroke support groups throughout Boston, MA. Inclusion criteria included: ≥ 18 years old, ≥ 6 months post-stroke, able to tolerate 6 hours of testing over 2 sessions, and able to travel to testing site. Exclusion criteria included: history of multiple strokes, traumatic brain injury, on-going and uncontrolled seizure disorder, pre-stroke disability, pre-existing neurological condition, and self-reported severe medical or psychiatric illness.
METHOD: The ACS, SIS Participation, and RNL were administered to all participants with modified presentation methods to facilitate the responses of participants with aphasia. Confirmatory factor analysis was used to determine the extent to which the collected data fit models of participation constructed on the assumption that the three measures are tapping the same, single construct. Analyses were performed on a three-factor model of participation consisting of RNL, SIS Participation, and ACS total % activities retained score, and a six-factor model of participation consisting of RNL, SIS Participation, and ACS total % activities retained in the four domain scores (Social, Low demand leisure, High demand leisure, Instrumental ADL).
RESULTS: The three-factor model of participation resulted in significant factor loadings of each variable onto the latent construct of participation (range of β = 0.72-0.81, p<.001). Participation accounted for a sizable amount of variance on each of the measures (range of R2s = .515 - .657). The six-factor model of participation yielded acceptable model fit statistics (x2 (8) = 12.38, p =.14, RMSEA = .07, CFI = .99, TLI = .98), and all factor loadings were statistically significant (range of β = 0.59-0.92, p<.001).
CONCLUSION: Preliminary evidence suggests that the ACS, SIS Participation, and RNL are measuring the same latent construct of participation. The results support the use of the RNL, SIS Participation, and ACS as valid measures of the same participation construct in stroke rehabilitation. Replication of the findings with larger sample sizes are needed to strengthen this conclusion. This study brings to light a significant professional assumption that has not been empirically confirmed, and encourages further examination of the validity with which participation is being assessed in practice. Production of such evidence will strengthen the profession’s ability to measurably demonstrate its impact on participation.
References
Badley, E. (2008). Enhancing the conceptual clarity of the activity and participation components of the international classification of functioning, disability, and health. Social Science & Medicine, (66), 2335-2345. doi: 10.1016/j.socscimed.2008.01.026.
de Graaf, J.A., van Mierlo, M.L., Post, M., & Achterberg, W.P., Kappelle, L.J., & Visser-Meily, J. (2018). Long term restrictions in participation in stroke survivors under and over 70 years of age. Disability & Rehabilitation, 40, 637-645. doi: 10.1080/09638288.2016.1271466.
Eyssen, I., Steuljens, M., Dekker, J., & Terwee, C. (2011). A systematic review of instruments assessing participation: Challenges in defining participation. Archives of Physical Medicine & Rehabilitation, 92, 983-992. doi: 10.1016/j.apmr.2011.01.006.
