Validation of the cognitive reserve theory in patients with traumatic brain injury

Abstract

Introduction: Predicting the outcome of traumatic brain injury (TBI) is an essential task for researchers and rehabilitation professionals alike, primarily due to the importance of giving prognosis to the patients and their families. The general issue with previous research is that initial measures of injury severity did not display a high correlation with measures of TBI outcomes. Multivariate studies show a significant effect of cognitive variables on TBI outcomes, specifically some pre-injury characteristics of participants, which can be subsumed under the term cognitive reserve (Ponsford, 2013). According to the cognitive reserve theory (Stern, 2002), cognitive reserve has been proposed to account for the discrepancy between one's degree of brain damage and the perceived cognitive and/or functional deficit that is expected as a result of such a pathological process. The main objective of this study is to verify whether cognitive reserve variables contribute to improved prediction of short-term functional outcomes as well as participation outcomes one year after the TBI, and to check the interrelationship between predictor variables. Problems and hypothesis: The specific questions to be answered are: 1. Is there a correlation between cognitive functioning five months after TBI, injury severity and cognitive reserve measures? H1: Considering previous research, we can assume that people with minor injuries and higher cognitive reserve will have better results on all measures of short-term cognitive outcome, i.e. on measures of current cognitive functioning five months after TBI. 2. To examine the possibility of predicting short-term functional recovery based on initial injury severity, initial functional recovery, cognitive reserve and measures of current cognitive functioning. H2: Based on the findings so far, it is expected that all predictors will significantly contribute to the explanation of criteria variance, but will not predict the share of individual predictors. 3. To examine the possibility of predicting long-term participation outcomes based on initial injury severity, cognitive reserve, measures of current cognitive functioning and measures of functional recovery at the end of rehabilitation, as well as possible mediation effects. H3: Based on the findings so far, it is expected that all predictors will significantly contribute to the explanation of criteria variance, but it is not possible to predict the share of individual predictors. H 3.1: Measures of current cognitive functioning and functional status will be directly related to measures of participation. H 3.2: Cognitive reserve and severity of the injury will be indirectly related to participation measures, through current cognitive functioning and functional status. 4. To examine the possibility of predicting productivity one year after TBI based on initial injury severity, measures of current cognitive functioning, functional recovery at the end of rehabilitation and cognitive reserve. H 4: Based on existing findings, it is expected that all predictors will significantly contribute to the explanation of productivity variance one year after TBI, but it is not possible to predict the share of individual predictors. Methods: Participants: A total of 104 patients whose primary diagnosis was TBI and who underwent inpatient rehabilitation from 2015 to 2017 were included in the first part of the study. In order to be included, the participants had to be: aged 18–65 at the time of assessment and diagnosed with TBI by a physician according to ICD-10. Excluded were patients older than 65, patients with previous TBI or neurological illness, and patients who did not emerge from posttraumatic amnesia (PTA) during inpatient rehabilitation. The participants, 87 men and 17 women, had an average age of 32.1 and received an average of 11.9 years of education. The average participant Glasgow coma score (GCS) was 7.4, and the average duration of PTA was 9.6 weeks. The second part of the study analyses the Community Integration Questionnaire (N=91) results one year after TBI. There were no statistically significant differences in the aforementioned demographic and injury-related variables between the group who answered the CIQ and those who did not. In the analysis of the fourth research question, the results of 83 previously productive participants were included in the analysis. Those participants were significantly younger than the whole group and also had lower GCS scores. Instruments: Various neuropsychological tests and scales were used in the study, to determine: A) severity of brain injury, B) cognitive reserve, C) current cognitive impairments and D) functional outcomes and participation. The severity of the injury was assessed using the Glasgow coma scale, the length of posttraumatic amnesia and Marshall's computerized tomography classification. Measures of cognitive reserve were: 1. Premorbid intelligence, estimated using Peabody Picture Vocabulary Test results, as a continuous variable; 2. Education, dichotomized according to a median of the years of education into the low cognitive reserve (8-11 years) and high cognitive reserve (12 and more years), and 3. Occupational attainment, also categorized, using participants’ preinjury occupation according to the International Standard Classification of Occupations (ISCO-08), into the low cognitive reserve and high cognitive reserve. The outcome measures included: results of the neuropsychological tests and Functional independence measure which were assessed at the end of inpatient rehabilitation, as shortterm outcome measures, and the Community Integration Questionnaire and return to productive status, which were assessed one year after TBI, as long-term participation outcomes. Results and discussion: In the first part of the research, the results demonstrate a statistically significant low to moderate correlation of all three cognitive reserve measures with the raw results of NP tests, which correspond to conclusions found in literature. After correcting the results of the NP tests for the effects of education and age, using standardized results from the manuals, the correlations remained statistically significant mainly for the tests that assessed the perceptual speed and executive function, with the highest correlations measured by the TMT-B test. The use of standardized results as a correction due to ascertainment bias was discussed. The analysis of the relationship of cognitive reserve measures with short-term functional outcomes shows a statistically significant low correlation of all three cognitive reserve measures with the cognitive functional outcome. On the other hand, the motor functional outcome did not show significant correlation with any of the cognitive reserve measures, which corresponds to the results of previous researches. In the prediction of shortterm functional outcomes, we did not find a statistically significant unique contribution of the cognitive reserve measures in the prediction of either cognitive or motor functional outcomes. As opposed to that both functional outcomes were moderately to highly associated with injury severity measures, as well as with NP tests results. The relationship between cognitive reserve and participation outcomes measured by the Community Integration Questionnaire one year after TBI shows that the correlation between the two cognitive reserve indicators (education and occupation) is moderately high. The vocabulary test, as a measure of premorbid IQ shows low, but statistically significant correlation. Three hierarchical multiple regression analyses were employed to predict community integration. Education and occupational attainment had a significant unique effect in the full model, but estimated premorbid IQ did not. Mediation analysis shows statistically significant direct effects of education and occupation on one-year community integration. The estimated premorbid intelligence shows an indirect effect, mediated by the length of PTA and results of NP test assessed at the end of the rehabilitation. Results of the NP tests and Functional independence measure also show a statistically significant direct effect on community integration. The relationship between cognitive reserve and return to productivity, as the second participation outcome, is statistically significant for all the three of the applied cognitive reserve measures. These correlations were also higher than correlations with the Community Integration Questionnaire results. Cognitive reserve significantly contributes to the prediction of the likelihood of return to productivity, above and beyond the factors related to the severity of TBI, cognitive impairment and limitations in activities of daily living. Conclusion All three of the cognitive reserve measures show significant correlation with shortterm cognitive outcomes measured by the NP tests results, which corresponds in direction and magnitude to the findings from the previous literature. However, due to methodological limitations, these findings cannot be unambiguously interpreted as a contribution to cognitive reserve theory, because of the possible ascertainment bias. In the prediction of short-term functional outcomes, we did not find a statistically significant unique contribution of the cognitive reserve in the prediction of either cognitive or motor functional outcomes. On the other hand, these outcomes were moderate to highly correlated with injury severity measures, functional independence assessed at the beginning of the rehabilitation, as well as with NP tests results. The association of cognitive reserve with community integration outcomes indicates a significant role of cognitive reserve for long-term outcomes of TBI. Nevertheless, the association of demographic variables and the Community Integration Questionnaire results in the healthy population suggests caution in the conclusions. The relationship between cognitive reserve and return to productivity provides further evidence of the importance of cognitive reserve measures in predicting long-term TBI participation outcomes. These results could have their application in vocational rehabilitation, where groups of patients who need help most when returning to work or school would be detected. The findings of the high correlation of NP tests with short-term and long-term outcomes after TBI should be viewed in the light of multifactor cognitive reserve models (McGarrigle et al., 2019; Satz et al.,2011), a topic which requires further academic research.