*3.3. Relationships between Neuropsychological Screening and MPAI-4*

There were significant correlations between both BNIS and Cognitive Flexibility(TMTB-TMTA) and BNIS and total score on MPAI-4 (Table 3). Dividing MPAI-4 into subscales revealed that it was primarily the participation subscale that was related to BNIS score, while for Cognitive Flexibility(TMTB-TMTA), this correlation was driven by the adjustment and ability subscales.

**Table 3.** Correlation between MPAI and neuropsychological screening instruments.


\* = *p* < 0.05, § = *p* = 0.06.

#### *3.4. Regression Analysis*

Linear regression was used to assess the independent effect of BNIS and Cognitive Flexibility(TMTB-TMTA) on functional outcome. When adjusting for acute injury severity and educational level (Table 4), linear regression analysis showed that BNIS was significantly and independently related to outcome according to the participation subscale on MPAI-4; Cognitive Flexibility(TMTB-TMTA), however, was not. For the MPAI-4 ability and adjustment subscales, Cognitive Flexibility(TMTB-TMTA) but not BNIS was significantly and independently related to outcome (Table 4). For the adjustment subscale, acute injury severity was also independently and significantly related to outcome in the model both with BNIS (estimate = −0.200, *p* = *0*.020) and with Cognitive Flexibility(TMTB-TMTA) (estimate = −0.2246, *p* = *0*.006). Educational level was not significantly and independently related to outcome in any of the models.

**Table 4.** Linear regression investigating the relationship between neuropsychological screening instruments and MPAI, adjusting for acute injury severity (including age) and educational level.


\* = *p* < 0.05, *\*\* = p <* 0.01. Separate linear regression models for BNIS and cognitive flexibility for each MPAI-4 outcome scale. In the adjusted models, CRASH and educational level are used as adjusting variables.

When Cognitive Flexibility(TMTB-TMTA), CRASH, and educational level were included in the model, the MPAI-4 adjustment subscale model was found to be statistically significant (F(3,25) = 5.73, *p* = *0*.004). No other models reached statistical significance.

#### **4. Discussion**

The present study investigated the relationship between findings from early cognitive screening using BNIS and TMT and long-term outcome assessed with MPAI-4 in sTBI by considering whether this is influenced by cognitive reserve, age, and injury severity. We found that both BNIS and Cognitive Flexibility(TMTB-TMTA) correlated with long-term outcome after sTBI measured with a MPAI-4 total score. When considering the MPAI-4 subscales, we found that BNIS correlated with the participation subscale of MPAI-4, while Cognitive Flexibility(TMTB-TMTA) mainly correlated with the adjustment and ability subscale. Adjusting for cognitive reserve and acute injury severity did not significantly alter these relationships. The results also show, except for the model regarding the adjustment scale containing the variables Cognitive Flexibility(TMTB-TMTA), CRASH, and educational level, that very little of the variance was explained by these variables. This strongly suggests that there are other variables, besides results on cognitive screening, age, and acute injury severity, that influence outcome after sTBI.

Nonetheless, these results suggest that early screening with TMT and BNIS can be used as a tool to help predict later outcomes for patients with sTBI and that these two tests seem to measure different aspects of outcome. Scores on BNIS are related to participation, both independently and when adjusting for confounders. The participation subscale mainly reflects community integration, including a return to work, and our results are in line with previous research emphasizing the link between cognition and return to work [26]. Cognitive Flexibility(TMTB-TMTA) did not correlate significantly with the participation subscale.

The ability subscale in MPAI-4 measures problems with both motor and cognitive abilities and did not correlate with the BNIS score, neither on its own nor when adjusting for controlling variables. The ability scale does, however, correlate with Cognitive

Flexibility(TMTB-TMTA). A possible explanation is that Trail Making requires more motor skills compared with BNIS, thereby requiring more of the abilities that the ability scale is measuring. Still, one might have expected a correlation between BNIS and the ability scale since both measure cognitive abilities. The lack of relationship might be due to the fact that rating problems with cognitive abilities are not the same as measuring cognitive function. Prior research has found that subjective measures of cognitive problems have a stronger relationship with a concurrent emotional state than with objective cognitive test measures [27]. An alternative explanation is that individuals with better cognitive function participate more in daily life, according to the correlation between scores on BNIS and the MPAI participation subscale. Greater participation, for instance, more leisure activities, and higher rate of employment likely lead to greater cognitive demands and, therefore, possibly the same amount of experienced cognitive problems in spite of having higher ability.

There was no relationship between BNIS and the MPAI adjustment index, which rates problems with psychological well-being and social interaction. However, acute severity of injury (CRASH) and the score for Cognitive Flexibility(TMTB-TMTA) were both moderately related to outcome on the adjustment scale, acute injury severity slightly stronger than Cognitive Flexibility(TMTB-TMTA). The relationship between the CRASH index and the MPAI adjustment index indicates that CRASH not only can predict mortality but is also related to functional outcome, as have been seen in previous studies [28]. Impaired executive functions have also previously been demonstrated to affect functional outcome negatively, including social functions [14], and difficulties seem to increase with more severe forms of TBI [29]. Executive function is a broad concept that also encompasses aspects such as impulse control, emotion regulation, and motivational drive [12]. In this study, only one aspect of executive function was examined, namely, cognitive flexibility. In order to learn more about the relationship between executive impairments and outcome on MPAI-4 subscales, a more detailed assessment of the executive function would be needed.

We found that as early as 3 months after sTBI, a large effect, according to Cohen's *d*, could be seen between education groups on both BNIS score and Cognitive Flexibility(TMTB-TMTA). These results were not statistically significant, however, and probably due to a lack of power. This finding is in line with previous research, which has pointed out that the effect of education on cognitive measures is strong even in the early stages after a brain injury, increasing the possibility of misclassifying patients with longer education as having lesser consequences of their brain injury [9,10,30,31]. In our study, 28% of the patients with longer education were classified as having no obvious cognitive impairments from their brain injury 3 months after sTBI, compared with 8% (one patient) in the low education group. In the present study, there was also an age difference between the educational groups, and some of the difference in BNIS scores might be due to the high education group being younger, as there was a correlation between BNIS score and age. Taken together, our findings support the risk of misclassifying patients with longer education as having no cognitive impairments using BNIS, especially if they are also younger.

In the present study, there was no significant difference on outcome according to MPAI-4 based on the level of cognitive reserve, which differs from that of previous studies. However, several previous studies have primarily used cognitive measures as outcome variables, which are more directly influenced by cognitive reserve [32,33]. Nonetheless, other studies have found that cognitive reserve also influences long-term functional outcome after TBI, although these studies have used less detailed outcome assessments [6,34].
