Perceived control predicting the recovery of individual-specific walking behaviours following stroke: Testing psychological models and constructs

Abstract

Perceived control predicts activity limitations, but there are many control belief concepts and how these are defined and measured has implications for intervention design. This study examined whether individual-specific activity limitations and recovery were predicted by theoretically derived control conceptualizations, the Theory of Planned Behaviour and an integrated model (Theory of Planned Behaviour with the World Health Organization ICF (International Classification of Functioning, Disability and Health) model). This predictive cohort study used measures of impairment, intention and perceived control (perceived behavioural control, Theory of Planned Behaviour; self-efficacy, Social Cognitive Theory; locus of control, Social Learning Theory), assessed 2 weeks after hospital discharge, to predict walking limitation (UK SIP: FLP) and recovery after 6 months. Theoretically derived items were individually tailored for patients' baseline walking limitation. Two hundred and three stroke patients (124 men and 79 women; mean age = 68.88, SD = 12.31 years) Walking limitation and walking recovery (respectively) were predicted by perceived behavioural control (r = -.36(**), .26(**)) and self-efficacy (r = -.30(**), .22(**)), but not locus of control (r = -.07, .02). Both theoretical models accounted for significant variance in walking limitation and recovery--but not beyond that explained by perceived behavioural control. Predictive power was not improved by modifying the control component or by including impairment in regression equations. Results suggest that perceived control predicts individual-specific disability and recovery and that reductions in activity limitations may be achieved by manipulating control cognitions. In addition, reducing impairments may not have maximal effect on reducing disability unless beliefs about control over performing the behaviour are also influenced.