The Early Functional Abilities (EFA) scale to assess neurological and neurosurgical early rehabilitation patients

Background It is difficult to assess neurological and neurosurgical early rehabilitation patients comprehensively. Available scales focus on activities of daily living (Barthel (BI) and Early Rehabilitation Barthel Index (ERBI)) or wakefulness (Glasgow Coma Scale (GCS), Coma Remission Scale (CRS)) while cognitive items are missing. Methods The Early Functional Abilities (EFA) scale comprises 20 items referring to activities of daily living (ADL), wakefulness and cognitive abilities. To evaluate its validity, n = 623 early neurological and neurosurgical rehabilitation patients (most of them after ischemic stroke or cerebral bleeding) were assessed on admission using the EFA, ERBI, GCS, CRS and measures of morbidity (co-diagnoses). Results The more co-diagnoses the lower EFA sum scores were obtained (Spearman-Rho rs = -0.509, p < 0.001). EFA predicted length of stay (LOS, rs = -0.565, p < 0.001) and BI at discharge (rs = 0.571, p < 0.001). Conclusions The results suggest that EFA is a valid instrument to assess critically ill neurological and neurosurgical early rehabilitation patients. It may be used as a measure of morbidity and a predictor of LOS and outcome. Further studies are strongly encouraged.


Background
Patients with severe neurological impairment, e.g. after stroke or brain injury, require specialized early rehabilitation following acute hospital treatment [1,2]. An early initiation of rehabilitation is essential because it may lead to a better neurological outcome [3]. These patients frequently suffer from disorders of consciousness, high morbidity and functional dependence [1,2,4]. In addition, many are dependent on intensive care treatment and mechanical ventilation while undergoing neurological early rehabilitation [5] and may be colonized with multi-drug resistant germs [6].
A valid assessment of neurological early rehabilitation patients is difficult. The Barthel Index (BI), for instance, is a measure of independence in activities of daily living (ADL) and even allows length of stay prediction in subsequent rehabilitation [7]. However, there are ceiling effects and change sensitivity of the BI is low when assessing severely impaired neurological early rehabilitation patients [8]. For that reason, an extension of the BI has been developed, the so-called "Early Rehabilitation Barthel Index" (ERBI) [4,[8][9][10]. The ERBI has even entered the German ICD-10 catalogue [11] and the definition of the early neurological rehabilitation procedure 8-552 in the German DRG system (ERBI ≤ 30 as inclusion criterion) [12]. It includes highly relevant items like tracheostomy, mechanical ventilation or monitoring on a dichotomic scale, but only few studies have proven its validity in neurological early rehabilitation [4]. Further, the ERBI has limitations because it does not allow evaluation of wakefulness and cognitive abilities.
Most other assessments frequently used in early rehabilitation focus on wakefulness/consciousness, like Glasgow Coma Scale (GCS) [13] or Coma Remission Scale (CRS) which is mainly used in Germany [14]. An evidence-based review included 13 more assessment    scales for disturbances of consciousness [15]. Among these, the Coma Recovery Scale -Revised (CRS-R) [16] was recommended with minor reservations to assess consciousness. The Sensory Modality Assessment Technique (SMART) [17], Western Neuro Sensory Stimulation Profile (WNSSP) [18], Sensory Stimulation Assessment Measure (SSAM) [19], Wessex Head Injury Matrix (WHIM) [20] and Disorders of Consciousness Scale (DOCS) [21,22] may be used to assess consciousness with moderate and the Coma/Near-Coma Scale (CNC) [23] with major reservations as far as reliability and diagnostic/prognostic validity are concerned [15]. The German Early Functional Abilities (EFA) scale has been introduced to assess both -ADL and cognitive functions (including wakefulness) of neursosurgical early rehabilitation patients [24,25]. The EFA scale (Table 1) comprises 20 items in 4 categories (autonomic, oro-facial, sensorimotor and cognitive functions/abilities). Each item is rated on a five-point-scale: 1 = "no function", 2 = "severe disturbance", 3 = "moderate disturbance", 4 = "slight disturbance", 5 = "normal" [24,25]. Thus, EFA total scores may range from 20 to 100 [24,25]. Besides vegetative abilities, inter-rater reliability of the EFA scale was found to be moderate to good for most other items [25]. With respect to its validity, however, there is a considerable lack of evidence.
Compared to other scales which are focusing on consciousness [15], the EFA also allows an assessment of vegetative (e.g. tolerance to postural changes and excretion functions) and oro-facial functions (e.g. oral hygiene, tongue movements and chewing) [24,25]. Further, it has to be pointed out that neurological and neurosurgical early rehabilitation patients frequently suffer from disturbances of consciousness at early stages but with improving awareness, sensorimotor and cognitive functions are more and more important for the patients`recovery. Therefore, an evaluation of vegetative, oro-facial, sensorimotor and cognitive functions through only one assessment tool is very useful. The EFA allows a monitoring of the patients`progress throughout the whole early rehabilitation process.  In the present study, the EFA scale was used in a cohort of neurological and neurosurgical early rehabilitation patients. The study`s goals were defined as follows: (1) Examination of concurrent validity The study wanted to explore whether it was associated with expected measures of severity and co-morbidity. The Well established assessments (like Barthel index) have been measured concurrently to study its concurrent validity. (2) Examination of prognostic validity Further, the prognostic validity was explored to see whether EFA scores on admission were predictive of later outcomes and length of stay (LOS).

Methods
The BDH Clinic Hessisch Oldendorf, Germany, is a teaching hospital of Hannover Medical School (MHH). It offers acute hospital treatment (including stroke unit and intensive care treatment), neurological and neurosurgical early rehabilitation, as well as subsequent rehabilitation (e.g. medical-occupational rehabilitation [26]). Medical records of n = 623 early rehabilitation patients admitted to the clinic in 2010 have been carefully reviewed with respect to age, morbidity, length of stay (LOS) and functional independence (ADL) measures. On admission, the following scales/assessments have been applied: ERBI [4], GCS [13], CRS [14] and EFA [24,25]. In addition, ERBI (including BI) at discharge was used to evaluate outcome. The ERBI consists of the Early Rehabilitation Index (ERI) and BI (ERBI = ERI + BI) [4]. Each item of the ERI (intensive care treatment, mechanical ventilation, confused patient, behavioural disturbances, impairment of communication, dysphagia) is rated on a dichotomic scale [4]. If an item is applicable, it scores with a minus value (-50 or -25 points) [4]. The ERI sum score (-325 to 0 points) is added to the BI (0 to +100 points). Thus, the total ERBI scores range from -325 to +100 [4]. GCS ranges from 3 to 15 [13], CRS from 0 to 24 [14].
BI served as primary outcome variable. According to previous studies, a poor outcome was defined as BI < 50 at discharge [27,28].
In addition, length of stay (LOS), period of time until first remission signs were observed (remission time in patients with disorders of consciousness, e.g. eye contact), duration of autonomic instability (defined as at least 1 week without any autonomic instability like body temperature, blood pressure or heart rate peaks) and morbidity parameters (number of co-diagnoses, Patient Clinical Complexity Level [2]) have been analyzed. As neurophysiologic measure of wakefulness, EEG rhythms of n = 183 patients have also been included in the analysis.
Only those cases have been added to the database with a complete assessment of at least EFA, BI and ERI. The assessors have been nurses (BI, ERI, EFA vegetative functions), physicians (GCS, CRS, review of BI and ERI), speech therapists (EFA oro-facial), physiotherapists (EFA sensorimotor) and occupational therapists (EFA cognitive) who were well experienced in neurological and neurosurgical early rehabilitation.
In the results section, mean values and standard deviations (in brackets) are displayed. In parametric testing (t-tests and ANOVAs with post-hoc LSD-tests), differences were regarded as significant with p < 0.05. Correlations were computed using the Spearman-Rho correlation coefficient (r s ).
Local ethics committee (BDH-Clinic Hessisch Oldendorf ) determined that the study was exempt from ethics approval and hence waived approval because the study was a retrospective database analysis, only (no intervention).

Patient cohort
Data of n = 623 (n = 283 female, n = 340 male) early neurological and neurosurgical rehabilitation patients have been analyzed. Mean age was 64.9 (15.4) years. Main diagnoses are displayed in Table 2. 36.5 (142.0) days since onset of neurological/neurosurgical disease had passed. Results of the scales on admission are displayed in Table 3. There were no significant differences of EFA scores (subscores and total) between stroke and head injury patients (p > 0.05).
(1) Concurrent validity A mean of 16.4 (5.4) co-diagnoses per patient was documented. The more co-diagnoses, the lower EFA total scores were found (r s = -0.509, p < 0.001, Fig. 1). 91.0 % of all patients (n = 567) had the highest Patient Clinical Complexity Level (PCCL) of 3 or 4. The higher the PCCL, the lower EFA total scores were found (Fig. 2). Patients with a PCCL of 4 had a significantly lower EFA total score than any other group (F = 22.82, all LSD-tests p < 0.001). For those patients with disorders of consciousness like coma or unresponsive wakefulness syndrome (n = 238), first remission signs were observed after a mean of 13.6 (16.4) days. The longer the period until remission signs were observed, the lower the EFA total r s = -0.192, p < 0.01, Fig. 3), oro-facial (r s = -0.179, p < 0.01), sensorimotor (r s = -0.153, p < 0.05) and cognitive score on admission (r s = -0.172, p < 0.01). There was, however, no significant correlation between remission time and EFA vegetative. Autonomic instability was observed for a mean of 25.8 (28.7) days after admission. EFA vegetative correlated significantly and negatively with duration of vegetative problems (r s = -0.341, p < 0.001). EFA total score correlated significantly with BI (r s = 0.570, p < 0.001, Fig. 4), ERI (r s = 0.505, p < 0.001), CRS (r s = 0.732, p < 0.001) and GCS (r s = 0.751, p < 0.001, Fig. 5) on admission. EFA vegetative and cognitive abilities domains as a measure of wakefulness correlated significantly and positively with GCS (r s = 0.560 resp. r s = 0.727, p < 0.001). EFA oro-facial and cognitive as a measure of communication and facial motor abilities correlated with CRS subscale talking (r s = 0.370 resp. r s = 0.299, p < 0.001). Further, EFA cognitive was significantly lower when a disturbance of communication was documented in  Fig. 2 The higher PCCL, the lower the EFA total scores were observed. Mean EFA values (bars) and mean standard deviation is indicated (on top of bars) Fig. 1 Correlation between EFA total score and number of co-diagnoses. The more co-diagnoses the smaller EFA scores were recorded Fig. 4 Correlation between EFA total score and BI on admission EEG patterns of n = 183 patients were analyzed, Alpha rhythm (8-13 hz) was found in n = 124, theta (4-7 Hz) in n = 53 and delta (0.5-3 Hz) in only n = 6 patients. EFA total scores differed highly significantly between the three EEG rhythms (ANOVA, F = 19.8, p < 0.001). In post-hoc LSD-tests, EFA total scores were lower among patients with theta (p < 0.001) and delta (p < 0.001) than alpha rhythms. In addition, EFA scores of patients with delta and theta rhythms differed significantly (p < 0.05), Fig. 6. (2) Prognostic validity Patients with poor outcome (BI at discharge < 50) had significantly lower EFA total, vegetative, oro-facial, sensorimotor and cognitive scores than patients with better outcome (p < 0.001),

Discussion
The ERBI is frequently used in Germany to assess the progress of neurological and neurosurgical early rehabilitation patients [4]. However, the ERBI has some limitations. It focuses on activities of daily living (ADL) and some items relevant to this group of patients, such as tracheotomy or mechanical ventilation [4]. Wakefulness and cognitive items are missing in the ERBI. Well established scales like GCS [13] or CRS [14] measure wakefulness but do not comprise ADL or cognitive items. The EFA assessment has been developed to evaluate cognitive abilities (including wakefulness) as well as items of ADL among neurosurgical early rehabilitation patients [24,25]. While inter-rater reliability was found to be moderate to good [25], no studies on its validity are available, yet. The rationale of the present study was to contribute further knowledge to the question whether the EFA scale is a valid instrument to assess progress of early rehabilitation patients as well as its prognostic value.
While neurological early rehabilitation patients suffer from different disorders, no significant differences of EFA scores on admission could be detected when comparing stroke and head injury patients suggesting that EFA scale may be useful for a broad spectre of diagnoses. We found that EFA total scores correlated very well with measures of morbidity, such as PCCL or number of co-diagnoses. In addition, EFA on admission predicted LOS, remission time and duration of autonomic instability. These findings indicate that patients with higher morbidity had lower EFA values on admission. It seems reasonable to believe, that critically ill patients have worse functional abilities than healthier subjects.
EFA also predicted outcome. There was a highly significant correlation between EFA on admission and BI at discharge. In addition, patients with a poor outcome (BI < 50) had significantly lower EFA total and category scores. In an ANOVA, EFA sum had the highest significant influence on BI at discharge (p < 0.001).
Comparing EFA with well established scales like ERBI [4], BI, GCS or CRS [13,14], there were also highly significant correlations. The higher EFA scores were observed, the higher GCS and CRS values could be obtained.  As a neurophysiologic measure of wakefulness, faster EEG rhythms could be observed among patients with higher EFA scores. This finding indicates that the EFA scale seems to be of some value when assessing patientsẁ akefulness. Some limitations of the study need to be addressed. First of all, it has to be pointed out that it was a retrospective data analysis. In general, the quality of the database is better in prospective controlled studies. Secondly, the study contributes some knowledge to the concurrent and prognostic validity of the EFA scale, but inter-rater-reliability has not been examined at all. Only well experienced (in early rehabilitation) staff has been involved in the study, but rating of BI, ERI and EFA might differ considerably between different professions (e.g. nurses and physicians). Thirdly, more sophisticated statistical analyses like an Item Response Theory (IRT) analysis could be of some value.

Conclusions
The concurrent and prognostic validity of the EFA scale is supported by the finding that it correlates with morbidity, LOS, established scales of wakefulness, ADL and outcome. The EFA scale may be used to evaluate progress of vegetative, oro-facial, sensorimotor and cognitive functions of critically ill neurological and neurosurgical early rehabilitation patients. Prospective and controlled studies on reliability, in particular inter-rater reliability, and validity of the EFA assessment are strongly encouraged.