Introduction
Trauma is the most common cause of death and functional impairments among children and adolescents [
1‐
3]. The currently held view is that traditional outcome measures, such as survival rates or presence of physical symptoms, are inadequate and do not capture the range of ways in which a patient may be affected by injury, treatments, and sequelae [
4]. In the last decade, several authors have highlighted the importance of measuring health-related quality of life (HRQL) as an essential aspect of assessing outcome after injury [
4‐
8].
HRQL instruments need to be multidimensional, consisting of physical, emotional, and social health dimensions based on the World Health Organizations (WHO) definition of each concept [
4‐
9]. Because injury characteristics are heterogeneous, generic instruments are preferred and enable comparisons across multiple groups to facilitate understanding of how demographic variables or clinical groups differ in their reported HRQL scores [
7,
9]. Disease-specific instruments can complement generic measures focusing on specific aspects of health with respect to particular disease or organ systems [
7,
9].
Studies of children’s and adolescents’ HRQL face many challenges. One is that different researchers use different measuring instruments, which raises the question of whether the same health dimensions have been measured [
9]. Another challenge arises from when and how the information was obtained. Most studies of HRQL of children after injury have been carried out within 2 years after injury and have relied on parents’ proxy reports. These studies have focused on different age ranges and injuries and have revealed rapid recovery during the first year after moderate to severe injuries, followed by a plateau phase during which any remaining disabilities remain more or less unchanged [
10‐
17]. The few existing long-term follow-up studies have found that children continue to recover 5–10 years after moderate to severe injuries and a majority of them report HRQL scores similar to those of healthy peers [
18‐
21]. Those studies have, however, either focused on specific injuries such as traumatic brain injuries [
19,
21] and or had a specific focus on children with the most severe injuries [
18,
20]. No long-term studies have been found representing the full spectra of injuries and injury severities found in a general pediatric population.
HRQL measures are by definition an individual’s perception of the effects of disease and treatment on their well-being [
22]. The gold standard for measuring pediatric HRQL is self-report, as children have a unique awareness of their own health and earlier research has revealed that children as young as 5 years can self-report their HRQL [
23,
24]. It is well documented in the literature that there are discrepancies between children’s self-report and parents’ proxy reports, where lower agreement have been found in subjective domains such as emotional and social functioning and higher agreement for objective domains such as physical functioning [
25‐
27]. In studies where differences have been investigated it has been suggested that parents rate their children’s HRQL worse than the children themselves [
27,
28]. There is also concern regarding the influence of parental distress and other related factors on parents’ perception of their child’s HRQL [
14,
19,
28]. Most authors agree that it is important to include parents’ proxy report as a complement to child self-report as a secondary outcome measure. Moreover, there are situations where the child is unable to provide a self-report and parent proxy report is the only source of information [
5‐
7,
27,
29]. A number of studies have examined children with traumatic brain injuries (TBI) and found that caring for children with TBI may have adverse effects on the home environment, potentially leading to parental mental health pathology, family dysfunction, and changes in the parent–child relationship [
30‐
32]. There is, however, a knowledge gap regarding the situation in families after other types of injuries. Research is needed investigating the agreement between child self-report and parent proxy report of children’s HRQL and parental factors that may influence ratings of children’s HRQL.
In this study, we decided to use the Pediatric Quality of Life Inventory (PedsQL 4.0) since it assesses the domains outlined by the WHO, includes both child and proxy versions, has good psychometric properties, is widespread and easily interpretable, and recommended as a generic instrument for measuring children’s and adolescents’ HRQL after injury [
4‐
8]. The purpose in the present study is to examine the relationship between child self-report and parent proxy report of children’s HRQL and how parents’ mental health status relates to ratings of child HRQL 6 years after the child had sustained a minor to severe injury.
Discussion
In this study, we used the PedsQL 4.0 instrument to determine the relationship between child and parent proxy ratings of children’s and adolescents’ HRQL as assessed by 177 child-parent 6 years after an injury to the child. We also used the SF-36 health survey instrument to explore the parents’ mental health status. Hierarchical multiple regression analyses were used to investigate the correlation of the parent’s mental health status to both the child’s and the parent’s rating of the child’s HRQL. To our knowledge, this is the first study that has investigated the relation of parent’s mental health status on child and parent ratings of children’s HRQL after injury.
The main finding of the present study is that a low score for parent’s mental health status was the strongest predictor of poorer children’s HRQL in all domains in parent proxy reports. It was also the strongest predictor of poorer HRQL as reported by children themselves. However, the relationship may be either way or bidirectional; parental mental health may influence children’s HRQL as well as children’s HRQL influencing parental mental health. Two earlier studies have explored parental mental health and the relation between child and parent ratings of children’s HRQL. Panepinto et al. [
44] in a study using PedsQL 4.0 to determine the HRQL of children with sickle cell disease found that parents with lower mental health status proxy rated HRQL scores that were significantly lower than their children’s self-reported HRQL. The authors suggested that the children may have adjusted to their level of functioning and therefore reported better HRQL compared to their parents ratings. In contrast, Vance et al. [
45] in a study of children with acute lymphoblastic leukemia found that parents who were more depressed had children that self-reported poorer HRQL, the parent’s depression was not related to the proxy report of the child’s HRQL. In the same study, illness stressors and perceived vulnerability were correlated with significantly poorer parents’ proxy ratings. [
45] Vance and colleagues suggested that parents are better able to hide stress, but unable to hide more pervasive feelings of depression. The differences between our results and the results of these two studies may be related to time of follow-up, method of data collection, different diagnosis, and cultural differences. Our results could also reflect parents’ knowledge concerning their child’s experiences, health and well-being. This has been showed by Varni et al. [
46] in a study where there were higher agreements in domains that were of clinical importance to the child’s health problem. The possible bidirectional relation in our result could also be related to well-known research findings that parents are affected by children’s exposure to traumatic events and that their responses impact children’s reaction to trauma. [
47] Traumatic events such as a high-speed vehicle crash can cause instability in a child’s life, which has been found to be associated with a range of outcomes impacting development and affecting cognitive abilities, school achievements, social skills, and behavior. [
47] Earlier follow-up studies of injured children have demonstrated that caregiver distress, socioeconomic difficulties, and family burdens are associated with lower parent proxy report scores of the child’s HRQL after injury. [
14,
19,
28,
48‐
50] Wade and colleagues [
48] found in a study of pediatric trauma that social relationships are important for parents’ psychological adjustment regardless of injury. A study by McCarthy et al. [
14] found that unhealthy family functioning prior to the child’s injury, single-parent household, and not being covered by insurance were associated with worse parent proxy reports of children’s HRQL. In addition, Aitkin et al. [
49] found that burden in families after pediatric trauma was greater when health care need was unmet. This has also been found in a study by Limond et al. [
50] on children with spinal cord injury (SCI) where 45 % of the parents perceived that they did not receive enough support after discharge from acute care hospital after their child’s injury. The parents in Limond and colleagues [
50] study reported significantly lower scores on their child’s HRQL. These findings suggest that children’s HRQL may be better in families that have better economic and psychosocial conditions and that such conditions facilitate adjustment after pediatric injuries. Further research is necessary to reveal causal relationship between parents’ mental health and child and parents ratings of children’s quality of life.
We found no discrepancies between the parents’ proxy report and the children’s self-report of the child’s HRQL. The only PedsQL scale that showed a tendency to significant difference in ratings was emotional functioning, where parents tended to rate their children’s function worse than the children themselves. The level of agreement between child and parent proxy ratings of children’s HRQL report was strong in all scales with the exception of emotional functioning which was also the scale with the lowest internal consistency. In a study by Glaser et al. [
51], the authors claimed a higher level of agreement (however, not statistically different) in child and parent ratings of children’s HRQL when the questionnaire was mailed and completed at home compared to completion at a clinical setting. The authors state that this might be explainable by collusion between children and their parents, but another factor that they also mention is that the completion of a questionnaire at home in familiar surroundings may provide a more accurate reflection of the child’s HRQL. The authors also found higher agreement between child and parent proxy ratings compared to child and physician proxy ratings and child and physiotherapist proxy ratings, suggesting that proxies who have the greatest contact with the child respond more comparably with the child [
51]. In a review study by Upton et al. [
52], the authors found no differences in parent proxy reports and children’s self-report agreement depending on method and place of data collection. In the present study, the questionnaires were sent to the children’s and parents’ home addresses with instructions to avoid collusion and so enhance agreement, but there was no control over how the questionnaires were filled out. Therefore, we cannot rule out that the relationship between child and parent reports could have been affected by the method of data collection, both positively and negatively. Further research is needed to identify factors that may influence levels of agreement in child and proxy ratings of children’s HRQL.
In our multiple regression models, older children were found to report higher HRQL in emotional and social functioning. Conversely, parents of older children reported lower scores in emotional functioning, psychosocial health and physical health. These findings are somewhat in line with several previous investigations of child and parent reports on HRQL. For example, Achenbach et al. [
25] found that parents are more adept at assessing a child’s externalizing problems (e.g., aggression and conduct) compared to internalizing problems (e.g., anxiety and depression). Eiser and Morse [
27] have suggested that this could be applied to parents being more prone to rate the child’s HRQL on the basis of visible domains such as physical functioning than on less visible domains such as emotional or social functioning. We also found that parents of female children reported higher scores in psychosocial health.
Children with more severe injuries reported better social functioning in their HRQL. To our knowledge, this finding has not been described earlier. In a Swedish qualitative study of adolescents with spinal cord injury (SCI), Augutis et al. [
53] parents and peers were found to have formed an important support network around the injured child. Parents acted as advocates and containers for sadness, frustration and anger, and friends acted as promoters of activities and identity development. It was perceived that healthcare providers did not make sufficient use of this network [
53]. It is possible that children with more serious injuries receive better support from their social network. Further studies are needed in this area to investigate the impact of social support from family, friends and others regarding help to cope and adjust after different injuries.
Mothers as proxy reporters dominate most studies. In a study by Waters et al. [
54] of healthy school children, the mother’s self-reported HRQL significantly influenced the proxy report on their children’s HRQL. The author did not find this association with fathers [
54]. In the study by Vance et al. [
45] of children with cancer, it was found that children who self-reported poorer HRQL had mothers who were more depressed. In the present study, 77 % of the parent responders were females, and if the proxy reporter was female, this predicted an increase in both child and parent reports of social functioning and in parent reports of physical health, but the strongest predictor of parents’ ratings of their children’s HRQL was the parents’ mental health status.
Strengths and limitations
One strength of the present study is the long-term follow-up. Earlier studies have shown that children’s recovery trajectory continues 5–10 years after injury, indicating that follow-up investigations should go beyond 5 years [
18‐
21]. Another strength is that the population derives from a complete cohort from a well-defined population and geographical area (Stockholm region).
Some limitations of this study should be noted. First, the cross-sectional design does not allow controlling for pre-injury HRQL and other confounding variables such as recurrent injuries or other health problems. Furthermore, we did not have access to data to control for personality characteristics, family dynamics, and resources. Additional exploration of these issues is clearly merited. Secondly, 42 % of the cohort was lost to follow-up causing selection bias which potentially limits the generalizability of the findings. We recommend the readers to interpret the results with caution. Responders and non-responders were comparable with regard to demographic characteristics, but non-responders had less severe injuries and were more often discharged home from the emergency department than the responders. These factors may have influenced the recall of the injury event and the interest in participating, as reported elsewhere (in manuscript). A reminder to non-responders would probably have helped achieve a higher response rate, but such procedures were not permitted by the ethical review board.