Development of a measure of sense of self following traumatic brain injury
Davidson, Emily Jane
Background: Traumatic brain injury (TBI) affects 450-700/100,000 people annually. Qualitative research has highlighted that change in self-identity might be important to survivors and potentially linked with adverse rehabilitation outcomes, yet there are no outcome measures addressing these issues. This thesis utilised contemporary principles of measurement construction in psychological sciences to established and test a measure of problems with self-identity. A key feature in this approach is the development of a clear conceptual definition of the construct for measurement before proceeding with instrument development and validation. Aim To develop and validate a tool for measuring self-reported change in sense of self for survivors of TBI. Methods: Concept analysis methodology by critical review of the literature was used to provide a definition of self-identity change following TBI. This was then amalgamated with the conceptual framework and draft Personal Identity Questionnaire that had been developed from prior qualitative work (Levack, Boland, et al., 2014). A second draft questionnaire was tested via two rounds of cognitive interviews to produce a final draft version of a Brain Injury Sense of Self (BISOSS) questionnaire for validity testing. Face-to-face interviews were used for data collection. Participants in New Zealand and the UK were asked to complete the draft BISOSS questionnaire, a demographic sheet, the Glasgow Outcome Scale-extended (GOS-E) and Sense of Coherence Scale (SOC-13). Responses were analysed with factor analysis, using SPSS software and Rasch analysis, using RUMM 2030 software. Results: The concept analysis provided a definition of self-identity change after TBI, drawing a clear distinction between the changes reported and an evaluation of those changes. Changes were identified in three broad categories: egocentric self, sociocentric self and self-as-shared. The items in the Personal Identity Questionnaire were amended and added to in light of the concept analysis findings to form a comprehensive item bank for further testing. One hundred and thirty-six community-dwelling participants were recruited. The age range at injury was 17-75 years, 68.4% male; median time since TBI was 84 months. Severity of injury was mild in 31 and moderate/severe in 105. Forty per cent of participants reported experiencing problems with their self-identity. Initial Rasch analysis of all 45 items confirmed multidimensionality in the data. Factor analysis suggested a valid three factor solution with significant overlap with the three theoretical domains. These three factors underwent Rasch analysis separately to develop three separate, valid, unidimensional subscales. These subscales address egocentric aspects of sense of self- (BISOSS-E), sociocentric aspects (BISOSS-S), and relational aspects (BISOSS-R). Discussion and conclusion: Three interlinked but conceptually separate components of self-identity affected by TBI were identified and scales to measure them developed. The Egocentric and Sociocentric subscales both have a good spread of items, forming useful measurement tools. The Relational subscale had only 6 items and needs development of additional items to improve precision. These new scales evaluate strength of self-identity and could be used to evaluate, for example: impact of TBI on sense of self longitudinally, effectiveness of interventions on reconstructing self-identity and the relationship between self-identity and other outcomes such as depression.
Advisor: Levack, William M. M.; Taylor, William J.
Degree Name: Doctor of Philosophy
Degree Discipline: Rehabilitation Teaching and Research Unit, Department of Medicine (Wellington)
Publisher: University of Otago
Keywords: sense of self; traumatic brain injury; self-identity; Rasch analysis
Research Type: Thesis