Health-related quality of life in children and adolescents with paediatric acquired brain injury: Secondary data analysis from a randomised controlled trial

Any injury to the brain that occurs after birth, whether it be traumatic brain injuries (TBI) or non-traumatic injuries (e.g., brain tumour, cerebrovascular accidents, and infection) is categorised as a paediatric acquired brain injury (pABI). The consequences of pABI include long-term cognitive, behavioural and psychosocial problems and fatigue, that may restrict children’s opportunities for educational achievement, participation and quality of life [1‐9]. However, the subjective impact of pABI on health-related quality of life (HRQOL) is still not fully recognised.

HRQOL is a multi-dimensional construct including physical, social, and psychological functioning, all relevant to overall health [10]. It is often approached in terms of how well individuals participate in a lifestyle typical for their age [11] and taking into consideration the impact their disease or illness might impose [12]. Individuals with pABI have been found to demonstrate similar levels of HRQOL to that of the general population [13, 14], but more often they seem to have much poorer scores on HRQOL measures compared to healthy peers or even other paediatric clinical populations [15‐19]. Notably, findings on the associations between self-reported HRQOL and objective measures of health (e.g., injury severity indices and functional outcomes), are found to be weak, and rather more strongly related to reports of executive deficits, behavioural problems, fatigue, and limitations in participation [18, 20, 21]. Particularly a significant association between executive dysfunction and poor quality of life has been confirmed in children with pABI as well as other chronic paediatric conditions, such as epilepsy, developmental coordination disorder, and autism spectrum disorder [22‐24].

Higher levels of HRQOL have been associated with better subjective executive function (EF) in children and adolescents with pABI [25]. From the perspective of clinicians, this could represent a new avenue for improving the lives of these children. Executive dysfunction and its effect on participation and overall quality of life is of great concern. Notably, the young brain is especially vulnerable to any disruption or insult that can cause deviations from the typical developmental trajectory of EF [9, 26]. EF is a concept that refers to a set of higher-level controlled cognitive processes, such as inhibition, mental flexibility, and working memory. These functions subtend appropriate goal-directed behaviour and are therefore critical in managing the multiple and complex tasks of daily life [27]. To date, evidence-based interventions to improve EF as well as overall health in children and youth in the chronic phase of pABI is limited compared to the adult ABI literature [28, 29].

In 2018 our group conducted a multicentre double-blinded, parallel group randomised controlled trial (RCT) on children in the chronic phase of pABI (> 12 months since injury/illness or completion of cancer therapy), with parent-reported executive complaints [30, 31]. Here, we investigated the efficacy of a paediatric adaptation of Goal Management Training (pGMT) relative to a psychoeducational program (paediatric Brain Health Workshop, pBHW). An integral part of the RCT was to explore the experiences made by the children and their families throughout the intervention as the new skills, strategies, and knowledge were absorbed and consolidated, and how they perceived the effectiveness of the intervention. After all, the perspective of the patient—in this case the child and his/her caregivers—could be the most potent indicator of the quality of treatment outcome [32]. Therefore, self- and proxy-reports were conducted at all time-points. In the questionnaire package provided, the EuroQol 5D children and adolescent versions (EQ-5D-Y-3L) [33] were included as a quality indicator of the intervention. The EQ-5D-Y-3L is part of the EQ-5D family of instruments and it is recognised being a relatively simple, yet a good tool for self-assessment of HRQOL in healthy populations as well as children and adolescents with disabilities [34, 35]. The various EQ-5D instruments can in many instances be more favourable compared to the more disease-specific measures within rehabilitation, as they in addition to low respondent burden include a health profile with broad relevance across health problems and/or diagnoses [32].

The present study presents the subjective experiences made by the participants throughout the above-mentioned RCT in regard to self-reported health. Specifically, the aims are to explore the characteristic health profiles and patterns of HRQOL across the five health states of EQ-5D-Y-3L for children and adolescents who undergo cognitive rehabilitation as perceived by both the youths and the parents, and the degree of agreement. Furthermore, we aim to investigate whether improvement in EF following cognitive rehabilitation is linked to perceived improvement in HRQOL. Considering the promising results of improved EF in pABI following intervention [31], it could be expected that HRQOL may improve equally. However, as the improvement of cognitive function seems to be mainly involving the metacognitive facet of EF and not the behavioural, whether HRQOL shows distinct patterns in relation to the two accordingly is undecided.

The present study aimed to explore the characteristic quality of health profiles across the five health states of EQ-5D-Y-3L for children and adolescents who undergo cognitive rehabilitation as perceived by both youths and their parents, and to investigate whether reliable clinical change in EF following cognitive rehabilitation is linked to perceived improvement in HRQOL. Despite our study not finding a significant change in HRQOL following the interventions, attention should be drawn towards the tendency of a differentiated association between HRQOL and distinct behavioural and metacognitive aspects of EF. Moreover, the HRQOL profile formed by the health dimensions provides important information considering how the participants appear to have only minor problems across HRQOL dimensions, and seemingly evaluate their overall health as being relatively good.

There was no significant improvement in VAS from baseline to 6-month follow-up in either of the two RCI (no/negative vs. positive clinical change) groups. This was the case for both BRI and MI, which was different to what we had expected, although a positive trend was more noticeable in the positive RCI groups. Potentially, these results could be attributed to our choice of measure in terms of it not being sensitive enough to detect significant changes that are clinically relevant in pABI. The choice of using EQ-5D-Y-3L as our generic measure was a pragmatic decision in terms of it being a good patient-oriented questionnaire frequently used in health economics research, enabling us to compare our results to others if deemed appropriate. Disease-specific measures that include items that are considered likely to be affected by the specific condition or treatment in question could be more suitable for the evaluation of clinical trials and their influence on QOL [43]. Nonetheless, the calculated RCIs of the metacognitive facet of EF (MI) stands out in terms of those reporting a positive clinical change, showing a significantly higher VAS compared to those not reporting a clinical change on metacognition already at baseline and throughout the intervention. One interpretation could be how the metacognitive aspect of EF plays a key role in predicting participation and regulation of activities or vice versa, subsequently contributing to the corresponding physical components within HRQOL [44]. One of the dimensions reported as being most problematic was that of feeling pain and/or discomfort. We venture this to be a possible indicator of the EQ-5D-Y-3L capability to capture an important nuance of the association between HRQOL and the metacognitive aspect of EF.

Some may question the relevance of applying RCI (clinical change) as grouping variables rather than differentiating by intervention group (in this case pGMT vs. pBHW). The intervention groups have been similar without any significant differences on any of the primary and secondary outcomes. We argue that for the present study in which the purpose was to investigate HRQOL as perceived by the participants, RCI is a means to provide us with information that would have been lost otherwise. For instance, it is interesting to note that even though not significant, those who had a positive clinical change in EF following the intervention, had higher VAS already at baseline and throughout the intervention. Whether this is an expression of the children being more satisfied with their HRQOL because of their capabilities in EF [18], or if it rather demonstrates how children reporting poorer HRQOL are struggling with specific aspects of EF remains to be decided. Whatever the direction, HRQOL could be a more reliable point of reference for predicting treatment outcome than neuropsychological assessments. Compared to adult ABI, the extent and pervasiveness of problems are less likely to be identified until years after the acute period and initial recovery [45], as deficits may not become fully apparent until the cognitive processes are expected to be fully developed [46], or until the cognitive and social demands and expectations to the child reach a critical level [8]. Many of these children will perform within the normal cognitive range, yet they undeniably can display problems with EF, resulting in impaired adaptive and social functioning [47]. This illustrates how determining a child’s functioning level or even treatment outcome solely on cognitive measures can have its limitations in this clinical population. How treatment and outcome is perceived by both the child and their family, might offer a better way to evaluate the treatment’s effectiveness and determine whether a neurocognitive intervention should be implemented any further. Considering the association between EF and HRQOL, improvements in EF could potentially improve HRQOL accordingly. Hence, if we rethink how the success of a neurocognitive interventions that targets EF should be measured, more children could benefit.

Our results corroborate the important role cognition has to life quality. Specifically, our study suggests that improvement in metacognitive rather than the behavioural regulation aspect of EF is likely to positively influence HRQOL. However, as our sample was within the clinically normal cognitive range according to BRIEF, their cognitive problems in some areas could be too subtle to be detected by the EQ-5D-Y-3L. A 3-level value system for reported problems on the HRQOL dimensions may not provide a good enough differentiation of perceived problems, resulting in, on a day-to-day basis, the children’s performance therefore being perceived as acceptable enough for them to pass as relatively well-functioning. Changes in behavioural regulative processes could potentially be obscured. It is noteworthy that in adult ABI, facets to EF like inhibition, emotion control, and self-monitoring are found to contribute quite strongly to the mental components of HRQOL [44]. Considering the second most problematic dimension reported on the EQ-5D-Y-3L was that of “feeling worried, sad, or unhappy”, this is clearly an important domain to investigate further. A body of research demonstrates that there are characteristic cognitive growth curves for various EF subdomains that are seemingly in play in pABI, signifying the importance of time when interpreting level of function and disability in this clinical population [9, 48, 49]. In any case, these children and adolescents may have reduced metacognitive skills and therefore have limited insight into the impact of their reduced executive skills on their performance [50]. Additionally, how parents and children report HRQOL are most likely influenced both by measure used and the age of the child [51]. In the end, however, whether a change in HRQOL would be defined as clinically meaningful is set by the premise of who’s experiencing it [52].

Regarding the estimation of HRQOL, very few participants reported the worst severity level (level 3) on the various EQ-5D-Y-3L dimensions. The health profile of 11111 was for both the parent- and self-reports the most common profile at all time points. Notably, “feeling pain and discomfort”, and “being worried”, seem to be the dimensions that are most problematic as reported by both the youths and parents. Potentially we could have obtained a more nuanced picture of their health profiles if we had chosen the 5-level version of the EQ-5D-Y. However, as mainly the first and second levels were used by most participants, it is not unreasonable to believe that 3 levels are sufficient. Moreover, in a systematic review, a similar trend was noted, especially in the general population of school pupils, but also among some patient populations [33]. The authors highlight that this skewness is not unexpected as the EQ-5D is constructed to measure deviations in health, rather than positive health. Arguably, extending the number of levels does not necessarily give clear superiority to the EQ-5D-Y-5L over EQ-5D-Y-3L. In fact, relevant to our study, the selection of EQ-5D-Y-3L may be more suitable than the 5-level, when the population of the study is younger or where a lower literacy level is anticipated [53].

Parents and youths were mostly in agreement when estimating their HRQOL. Notably, there was a moderate discrepancy in child-parent reports on the two health dimensions “feeling pain and discomfort”, and “being worried”. Over time, the disagreement was reduced for the overall group. The problem of proxy-child disagreement is well-known across various populations. Children and parent ratings of HRQOL dimensions that are less observable (e.g., emotional and social functioning) tend to be in less agreement than for dimensions related to e.g., physical functioning and participation in general [23, 54‐56]. Specifically, in pABI populations, studies have found evidence for the parents to report less favourable outcomes on HRQOL [14]. Other studies find parents to rate the HRQOL of their children higher than the children do themselves across HRQOL dimensions [18, 19]. Indeed, when exploring HRQOL of young individuals in the chronic phase of pABI, parent-reports could be sufficient when looking at the overall clinical group. However, we know children with pABI do tend to report a broader range of negative emotions such as worry, sadness, shame, and guilt as being much more problematic than physical limitations [57]. How much children and parents focus on psychosocial consequences compared to physical concerns in relation to having a chronic health condition may be influenced by the severity of injury symptoms and how it is perceived [3, 58]. Thus, parents may not be able to fully understand the experiences of their children. This has clinical relevance considering parents, based upon their perceptions of their children’s health, largely will function as the clinical decision-makers [55].