Developing a Model for Suicide Ideation after Severe Traumatic Brain Injury Using Structural Equation Modelling

Abstract

Objectives: The risk of suicide ideation (SI) after severe traumatic brain injury (TBI) is double the level within the general population. Little is known about the mechanisms underlying this elevated risk. This study aimed to develop an explanatory model of SI after severe TBI, based on the interaction among distal (predisposing) and proximal (precipitating) risk factors, protective factors and SI. Methods: Participants (n = 90) with very severe TBI (Post Traumatic Amnesia > 7 days) were recruited from the Liverpool Brain Injury Rehabilitation Unit in Sydney Australia. Participants completed a battery of neuropsychological tests targeting executive functions as well as self-report measures of SI, positive and negative mood states, stress, self-awareness, self-esteem and problem-solving. Treating clinicians rated participants on measures of aggression, frontal systems dysfunction and participation. Data were analysed using structural equation modelling. Results: The final model demonstrated good fit indices (NFI = 0.924, IFI = 0.994, TLI = 0.986, CFI = 0.993, RMSEA = 0.03). The model accounted for 34% of the variance in SI. One distal risk factor, poor problem-solving, had a direct relationship to SI. Other distal risk factors including post-injury aggression and disinhibition had direct and indirect links to depression (a proximal risk factor), which in turn mediated their relationship to hopelessness and SI. The protective factors (social support, hope, community participation, self-esteem) mediated the relationship between distal/proximal risk factors, hopelessness and SI. The protective effect of social support and community participation was mediated through their association with increased levels of hope and self-esteem. Conclusions: The current model demonstrates the direct and indirect effects of risk factors and protective factors associated with SI after severe TBI. The model provides several suggested targets for clinical intervention to build positive mental health and reduce suicide risk. The model can now be tested prospectively to evaluate its predictive validity.