Prospective evaluation of the Quality of Life after Brain Injury (QOLIBRI) score: minor differences in patients with major versus no or mild traumatic brain injury at one-year follow up

Given the increasing interest in the longer-term outcomes following severe trauma, several generic measures of subjective health status, such as the Short-Form Health Survey-36 (SF-36) or the EuroQoL five dimensions questionnaire (EQ-5D), as well as trauma related functional outcome instruments, such as the Trauma Outcome Profile (TOP) were evaluated in recent years [1]. Traumatic brain injury (TBI) accounts for an important percentage of severe trauma [2‐4] and is reported depending partly on the exact definition of TBI [5, 6] and partly on the spectrum and severity of trauma chosen [7, 8]. In addition, the impact of TBI on the traumatized patient may be even more important than other body lesions both in monotrauma and multiple trauma [9, 10]. The newly developed Quality of Life after Brain Injury score (QOLIBRI) [11] was the first instrument designed to assess disease-specific health-related quality of life (HRQoL) following brain injury [12]. We therefore chose this score for our investigation, which still appears to be the most commonly used score for this topic, even though other instruments have been developed in the meantime [13‐15]. Interestingly, work on the QOLIBRI is based fundamentally on original initiatives aimed at assessing non-physical residuals following trauma in general and developing a disease-specific HRQoL tool for multiple trauma [11, 16]. Longer-term follow-up studies in trauma patients both with and without TBI underscored the need to supplement the SF-36, for example, with a measure of cognitive function when evaluating outcome. Few differences were found with regard to scoring of cognitive function in phone interviews between TBI patients (matched according to grade on the Abbreviated Injury Scale, AIS) with and without additional orthopaedic injury [17]. Given the lack of knowledge in the literature on the degree to which cognitive deficits following trauma, whether independently of or dependent upon the TBI sustained, will be reliably identified by scores for HRQoL and functional outcome, we were interested in specifically investigating the QOLIBRI by assessing the longer-term course of patients who had suffered severe trauma (Injury Severity Score (ISS) > 15) and comparing it to other well established outcome scores. Currently, QOLIBRI validation studies have been undertaken only in cohorts with TBI, defined using the Glasgow Coma Scale (GCS), with case selection based on the International Classification of Diseases (ICD) and restricted to patients of working age [12, 18, 19]. In contrast, clinicians are accustomed to grading trauma severity according to the AIS and ISS independently of age. Literature searches did not reveal any investigations of the QOLIBRI that involved unselected, i.e. severely injured patients, including those without head injury, or evaluated possible correlations with trauma severity classified by AIS-grading.

Against this background we undertook a prospective trauma centre evaluation of the QOLIBRI with regard to the longer-term outcomes for the severely injured. The study cohort comprised consecutive patients who had sustained severe trauma to any body region, using the AIS to define the presence or absence of TBI and the severity of trauma. The investigation’s objective was to compare one-year outcomes in the form of QOLIBRI scores for patients with major versus no or only mild TBI by performing detailed analysis of the underlying demographic and trauma characteristics and in relation to established HRQoL and functional outcome instruments.

The investigation took place in a dedicated trauma centre in Switzerland, serving a region of about 750,000 inhabitants. As part of a quality control project (NCT02165137) all major trauma patients (New Injury Severity Score, NISS ≥8) passing through the emergency department from 1.1.2011–31.12.2015 within 24 h of trauma were consecutively evaluated for this prospective investigation with a cross-sectional study design, approved by the local ethics committee. Hospital treatment guidelines followed international standards [20, 21]. This follow-up study included all survivors of trauma admitted to the emergency department of the hospital who were > 15 years of age at the time of the accident and who sustained severe trauma defined as an Injury Severity Score (ISS) > 15 [3, 22]. Major TBI was defined as a trauma severity of > 2 [23, 24] according to the Abbreviated Injury Scale (AIS) [25] of the head (version 2005, update 2008 of the TraumaRegister of the German Trauma Society). The no or mild TBI group contained all injured persons in the study cohort who did not sustain a trauma severity of > 2 according to the AIS head. AIS coding was executed according to the guidelines of the Association for the Advancement of Automotive Medicine (AAAM). For this evaluation of trauma cases all with a minimum ISS > 15, monotrauma was defined as an injury severity of > 3 according to the AIS in one body region and no injuries in other body regions (> 0) [4]. The term multiple trauma was used if at least two Abbreviated Injury Scale (AIS) regions were involved, and the ISS determined at the end of the hospital stay was 16 or above [26]. Given pilot evidence showing only a minimal impact of age on disease-specific HRQoL as measured by the QOLIBRI [27], and, in contrast to original validation studies of the QOLIBRI, we did not exclude retirees from this investigation, but controlled for age in multivariate analysis. Exclusion criteria for this survey were patients under the age of 16, those with an ISS ≤15, those deceased or presenting with a Glasgow Outcome Scale (GOS) [28] of 2 (persistent vegetative state) at hospital discharge or follow-up.

One hundred ninety-nine severely injured persons responded to the one-year follow-up, i.e. 41.5% of eligible persons (Fig. 1). 53.8% of investigated patients had sustained major TBI. The main patient and trauma characteristics did not differ between respondents and non-respondents (Table 1). Patients with major TBI differed from those without (48 cases with an AIS head = 0, 10 cases with an AIS head = 1 and 34 cases with an AIS head = 2) in most trauma specific characteristics, presenting with more severe trauma in the AIS 1 and 2 regions (head & neck and face) and less injured in all other body regions. No difference was found between groups regarding age, gender or overall ISS (Table 2).

To our knowledge, this is the first evaluation of the QOLIBRI with regard to one-year outcome in a cohort of severely injured patients that includes both major and no or only mild TBI. We found two unexpected major results:

1.), The total QOLIBRI did not correlate at all and the cognitive subscore correlated only poorly with major TBI in this investigation. 2.) The cognitive dimension of the QOLIBRI correlated weakly, but better with TBI than the other mental scales of HRQoL or functional outcome scores tested.

From a historical point of view, although von Steinbüchel et al. in their original evaluation studies correctly describe the broad spectrum of HRQoL to be measured by the QOLIBRI [11, 12, 18, 19], they and subsequent authors almost univocally argue for its use in TBI only. This is already indicated by the fact that its name includes ‘brain injury’. This conclusion is even more astonishing as the founding consensus group explicitly cited the original initiatives, e.g. research by Neugebauer in the 90s, as aiming to develop a disease-specific HRQoL tool for multiple trauma [11]. Their efforts to assess non-physical residuals following trauma resulted in the development of scores such as the QOLIBRI. Reviewing the literature on the QOLIBRI to date, all subsequent studies have focused on TBI cohorts only. Beginning with the first validation studies published by von Steinbüchel et al. [12], the QOLIBRI remained an HRQoL instrument propagated only for traumatic brain injury, even though many of the score questions clearly indicate that a broader spectrum is to be covered.

This prospective investigation is the first to identify in a comparative manner deficits in a mixed cohort of severe trauma patients (ISS > 15) by implementation of the QOLIBRI as a standard questionnaire and including other validated scores on HRQoL and functional outcome such as the GOS, EuroQoL, SF-36 or the TOP for further evaluation.

Original validation studies report an average total QOLIBRI score in TBI patients of about 65 one year following trauma [11, 12, 18, 19]. A few studies showed mean values of about 5 points lower [39] or higher, even if the complete cohort under evaluation was limited to severe TBI only (defined as a GCS < 9) [40]. The mean QOLIBRI in our study was found to be 75, independently of whether patients sustained major TBI or not, with lowest values in patients with no or mild TBI. Wilson et al. suggested a grading for recovery after TBI for the QOLIBRI by stratification for the GOSE, whereby an average score of 76 meant good recovery, 62 moderate disability and 56 severe disability [38].Following this stratification, our patients presented with good recovery on average at the one-year follow-up. To explain this difference between the literature and our data, further comparative analysis of study cohorts and procedures in the literature has to be undertaken. Surprisingly, to date, apart from the isolated exception found in the publication by Soberg et al. [41], all investigations on the QOLIBRI only gave a precise definition of their study cohorts with regard to TBI, but not regarding possible additional body injuries or overall trauma severity of patients [11, 12, 18, 19, 38, 42]. Almost all investigations of the QOLIBRI selected their TBI patients based on the ICD-classification and defined the severity of TBI by the worst GCS within 24 h following trauma. Therefore, most studies report on cohorts that are comprised to about 55–60% of patients with a GCS < 8 or investigate severe TBI (GCS < 9) only [40, 41]. In contrast and similar to other trauma centre evaluations [4, 6], our definition of TBI was based on the prospective consecutive assessment of unselected trauma patients with trauma severity graded according to the AIS. Our study cohort comprised 19% patients with a GCS < 8, i.e. 31% of patients in the major TBI-group. At first glance, this important difference in the percentage of severe TBI patients identified by GCS (3–8) in our cohort compared to previous studies might explain the difference in outcomes as stated above and quantified with the QOLIBRI. Nevertheless, the analysis of only severe TBI-cases with a GCS 3–8 in our investigation (as used by QOLIBRI validation groups) did reveal a mean total QOLIBRI of 74, which was comparable to the whole study group and/or unselected TBI-patients in terms of GCS. In our study cohort the lowest total QOLIBRI values for patients with a GCS 3–8 (mean 54) were found in the small subgroup of no or mild TBI (n = 5), i.e. for patients with an AIS head 0–2. The literature regarding a possible association between the severity of TBI and resulting HRQoL appears conflicting with both better and worse outcomes reported for more severe trauma [43, 44]. As in our analysis, most international studies reported no association of the QOLIBRI with the GCS, with some exceptions such as the latest Finnish evaluation of TBI patients undergoing intensive residential rehabilitation that reported a low negative correlation (Spearman r = − 0.21) [39]. One possible explanation for such divergent findings may be the selection of the study samples under investigation. For example, in a large Trauma Audit Research Network database analysis of over 25,000 patients with isolated TBI (AIS head > 2) for an equivalent severity of intracranial injury (defined according to the AIS), GCS was found to be higher in older patients than in young ones, an observation unlikely to be explained by differences in mechanism of injury or types of intracranial injury according to the analysis of authors [45]. In addition, a general lack of standardization in assessment and reporting of the GCS was reported in recent surveys and reviews [46, 47] and may be, at least partially, responsible for the observed variation of outcome data.

With regard to the severity of head injury (AIS), comparison with a simpler outcome score such as the EuroQoL in our investigation surprisingly showed correlation values not inferior to the QOLIBRI cognition (Pearson r = 0.16). In addition, the magnitude of the correlations for the cognitive QOLIBRI with the AIS of the head were not higher than for the total QOLIBRI with the AIS of the extremities, underlining the finding that the QOLIBRI is not injury specific. Interestingly, higher correlations (r about 0.4) were found for the EuroQoL regarding the grade of injury of the extremities. Overall, more somatic oriented outcome (sub-) scores correlated better with somatic lesions than did the QOLIBRI and its subscores or the mental and cognitive dimensions of the SF-36 or the TOP with brain injury. According to the literature the SF-36 may not be sensitive enough to detect key issues in patients with a TBI, such as cognitive dysfunction, severe physical restrictions or patients with psychological problems. In consequence, to assess the sequelae of TBI the use of both a generic (e.g. the SF-36) and a disease-specific measure (e.g. the QOLIBRI) of HRQoL is recommended [10, 14], even though no standardised interpretation aids for such a combined use are yet available. One example is the POLO chart, representing a battery of scores developed for the multiply injured including TBI and comprising the GOS, the EQ-5D, the SF-36 and the TOP [36, 48]. Unfortunately, such extensive scoring appears difficult to handle for the standard evaluation of single patients, even though the trauma specific TOP in the few studies published to date appeared to be a reliable and well discriminating score, covering both relevant general dimensions of HRQoL and trauma-specific aspects of longer-term outcome [36, 49]. Currently, no reports on any comparison of the TOP with the QOLIBRI are available, the present investigation being the first to compare both scores in a trauma centre setting. With regard to the clinically relevant question as to which patients are identified as conspicuous based on their boundary score values and, therefore, needing further individual examination or even therapy, we found that in 31% one, i.e. either the SF-36 or the QOLIBRI, and in 16% both indicated an impaired HRQoL. Such findings additionally support the conclusion that scoring for detection primarily of non-somatic residuals following severe trauma still has to be improved [14]. A recent detailed correlation analysis of QOLIBRI and SF-36 in the original international validation cohort of the QOLIBRI demonstrated a higher discriminative power of the QOLIBRI and all its subscales in comparison to the SF-36. At the same time, the informative value of subscales did differentiate between moderate and good recovery categories [42].

Ad 2.) The cognitive component was the only QOLIBRI-score that differentiated between patients with versus without major TBI in univariate analysis. Recent studies reported that HRQoL and recovery patterns differ for mild, moderate and severe TBI [50] and, for the QOLIBRI, it was even shown that severe TBI patients may report better cognitive functioning on the QOLIBRI subscale than did mild TBI patients [51]. Surprisingly for a brain injury score, in multivariate analysis of our data the cognitive QOLIBRI subscore was found to be weakly, but equally, associated (R² = 0.03) with both head trauma and injuries to other body regions (each classified according to the AIS). Nevertheless, in comparison with the mental subscales of the TOP and the SF-36, we found the cognitive QOLIBRI to be the only one to correlate, weakly, but still significantly with the AIS of the head. These findings provide further evidence that the cognitive QOLIBRI indeed elicits specific information on the mental outcome of patients. But, in contrast to the recommendations for its use only in TBI [14], the information on mental residuals following trauma found in our investigation was identified equally for patients with and without major TBI. In our study the cognitive component of the QOLIBRI showed the best association with the mental components of the TOP (r = 0.70), followed by the mental sum component of the SF-36 (r = 0.60). The association of the cognitive QOLIBRI with the physical sum component of the SF-36 was much lower (r = 0.31) – a finding that appears logical. Even though, for example, in the Australian validation study of the QOLIBRI [52], as part of the international QOLIBRI project using the identical inclusion criteria to those of von Steinbüchel et al. [18, 19], the cognitive QOLIBRI correlated less with the mental sum component of the SF-36 (r = 0.44), apparently having more in common with the physical sum component of the SF-36 (r = 0.53). This finding would not be expected for a TBI-specific score but, unfortunately, this discrepancy was not discussed further by the authors. One possible explanation for this result could be that only TBI-patients were included in their study.

Originally, von Steinbüchel et al. [18] stated that the items and therefore the total QOLIBRI score predominantly concentrate on emotional, cognitive, and psychosocial aspects and to a lesser extent on physical changes. The authors concluded that the questionnaire thus measures satisfaction and distress in areas of life typically affected by brain injury. Given our findings in unselected patients following severe trauma of all body regions, we are of the opinion that the use of the QOLIBRI should not be restricted to TBI patients only, but should also be undertaken in more severe non-TBI patients for whom such deficits are rarely expected and therefore not screened for on a routine basis. We are aware that such argumentation opposes the mainstream and would involve shifting focus to the development of defect-specific HRQoL scores that would also apply to TBI [14]. In agreement with investigators such as Dijkers [53] we are of the opinion that developing modules that quantify quality of life in specific functional areas that are underserved given the generic quality of life measures such as cognitive functioning may be a very reasonable approach. Due to the fact that no reports on the use of the QOLIBRI in no or mild TBI patients have been published so far, the pilot findings in this field as presented here have to be carefully weighed up and await further evaluation in cohorts other than ours.

Our results revealed cognitive deficits following major trauma independently of whether patients sustained major brain injury or not. With the objective of detecting possible mental deficits, this finding supports the general application of the QOLIBRI and, above all, its cognitive subscore for outcome measurement of severely injured patients. In a next step, objective neuropsychological tests should further validate our approach using patients’ self rating data. Future investigations may additionally reveal the potential benefits of an instrument that could be used directly after the accident, prior to hospital discharge, and at the beginning and end of the rehabilitation programme. This instrument might include more objective questions in the mental dimensions such as ability to think, reason and concentrate, would enhance long-term comparison capabilities, and provide documentation as a progress indicator for both physician and patient [39].