Early Injury-Related Predictors of Disability 6 Months After Moderate to Severe Trauma: A Longitudinal Study

Abstract

Trauma often causes long-lasting functional impairment, but the extent varies across populations. This study investigated disability six months after moderate to severe injury and identified sociodemographic and early injury-related predictors, including estimated rehabilitation complexity. Further, we assessed the implementation of direct transfer from acute care in the trauma centre to specialised inpatient rehabilitation, bypassing local hospitals. In this prospective study 398 adults, treated from January 2020 to January 2021 at two Norwegian trauma centres, were included. Self-reported disability was measured using the 12-item World Health Organization Disability Assessment Schedule 2.0. Ordinal logistic regression was applied to identify predictors of the 6-month disability outcome. At follow-up, 22% reported no disability, 49% mild/moderate and 29% severe. In multivariable analyses, low education, comorbidities, number of injuries and higher estimated rehabilitation complexity at discharge from acute care were significantly associated with greater disability. Only 20% were transferred directly to specialised inpatient rehabilitation, while 17% went via local hospitals. Participants with orthopaedic injuries and women were less likely to be transferred directly. In summary, most participants experienced some disability at 6 months. Indirect pathways to specialised rehabilitation via local hospitals remained common. Understanding predictors of disability and optimising rehabilitation pathways may improve trauma rehabilitation outcomes, highlighting the need for guidelines that identify patients with high rehabilitation needs.

1. Introduction

Global health has improved over recent decades, but physical trauma is still a major cause of disability [1,2]. Trauma often causes long-lasting functional impairment [3,4]. The proportion of trauma survivors who experience reduced function varies in different populations between 18% [5] and 68% [6]. Patients experience limitations in physical function, participation and mental health [7].

Early rehabilitation improves outcomes for patients with severe traumatic brain injury (TBI) [8] and multi-trauma [9]. A delay in admission to rehabilitation may result in poorer outcomes for patients with severe TBI [10]. Early systematic assessment of rehabilitation needs for primary rehabilitation (estimated rehabilitation complexity) is important to determine the optimal treatment pathway for trauma patients. For this, a structured instrument like the Rehabilitation Complexity Scale (RCS) may be useful [11,12]. Several guidelines emphasise systems to identify patients needing rehabilitation after trauma and seamless pathways to avoid delays in the treatment chain [13,14,15,16,17]. A recent systematic review by Jones et al. [18] stresses the importance of communication pathways, service planning, distribution of resources and coordination within trauma care networks for patients to experience good rehabilitation services. However, adherence to the recommendations for acute rehabilitation in the Norwegian Trauma Plan [19] was found to be limited [20]. Of the patients with severe trauma, only 22% were transferred directly from acute trauma care to specialised inpatient rehabilitation. In patient pathways where the recommendations were followed patients had more severe injuries than other patients [20].

There are several ways to measure outcomes after trauma, and several instruments have been developed [21]. The World Health Organization Disability Assessment Schedule (WHODAS 2.0) is a generic instrument developed by the WHO to provide a standardised method for measuring health and disability based on the International Classification of Functioning, Disability and Health (ICF) [22]. With measures of disability within activity (cognition, mobility, understanding and self-care) and participation (getting along, life activities and social participation), as well as degrees of disability, this instrument may give perspectives other than the commonly used health-related quality of life (QoL) measures [23]. Trauma may affect all dimensions of the WHODAS score. Treatment planning after trauma may also be guided by prognostic factors for disability. In the long-term, patients with TBI are less likely to be employed or in a relationship than patients with traumatic orthopaedic injuries [24]. However, extremity injuries are also shown to predict both lower physical and psychosocial functioning [25]. Higher age, pre-injury comorbidity and some injury types are associated with poorer outcomes [3,4,25], while the data regarding gender are ambiguous, with both worse as well as more favourable outcomes reported for women [5,26,27].

There is a knowledge gap about the occurrence of disability in the Norwegian trauma population and about early rehabilitation after trauma and its impact on outcomes. Findings about factors that predict disability and thus may direct clinical decision-making are inconsistent. The overall aim of this study was to investigate (1) the degree of disability at 6 months post-injury, and (2) sociodemographic and acute injury-related predictors of disability, including estimated rehabilitation complexity. In addition, one recommendation from the Norwegian Trauma Plan, the transfer of patients directly from acute trauma centre care to specialised inpatient rehabilitation and not via acute local hospital departments, was evaluated.

2. Materials and Methods

2.1. Design and Setting

We conducted a prospective, observational cohort study at Oslo University Hospital (OUH) Ullevål in the city of Oslo and the University Hospital of North Norway (UNN) in the city of Tromsø, two of four Norwegian trauma centres. The centres cover the southeast (approximately 3.1 million people) and northern regions (approximately 480,000 people) of Norway. According to the annual report of the Norwegian Trauma Registry of 2020, 1235 and 177 patients were received by trauma teams at OUH Ullevål and UNN Tromsø in 2020, with 267 and 28 of the admitted patients having sustained serious injuries with an Injury Severity Score (ISS) of >15 (Table 5, [28] p. 37). This study was approved by the Committee for Medical and Health Research Ethics, Health Region Southeast on 22 October 2019 (reference number 31676), and the Institutional Data Protection Officers at OUH on 8 August 2019 and UNN on 27 February 2020 (approval number 19/26515 and 02423). A protocol article for the main study has been published elsewhere [29]. This report adheres to the STROBE checklist for observational studies [30].

2.2. Study Population

The inclusion periods were from 1 January 2020 to 31 December 2020 at OUH and from 1 February 2020 to 31 January 2021 at UNN, comprehensively, one year. As described in our recent article [20], patients from 18 years or older who had sustained injuries were included in the study when residing in Norway and admitted either directly to the regional trauma centres or transferred from local hospitals within 72 h of injury. Eligibility required at least a two-day hospital stay and a New Injury Severity Score (NISS) > 9, indicating moderate to severe trauma. Individuals were excluded if they were younger than 18 years, non-residents, lacking skills to communicate in Norwegian or English or died during the acute hospital stay in the trauma centre. Individuals who did not participate in the 6-month follow-up or failed to complete the WHODAS 2.0 12-item questionnaire were excluded as well. Information about the study was given orally and in writing to the patients or their proxies where applicable. Patients or their relatives gave informed consent to participate in the study.

2.3. Patient Identification and Recruitment

Patients were identified through participation in regular trauma team briefings, admission logs maintained by the trauma department and electronic health record searches using trauma-related diagnostic codes. Most identified patients were met with a trauma team upon hospital admission. Eligibility was assessed based on the inclusion criteria.

2.4. Injury Severity Assessment

The study physicians, certified as AIS registrars, evaluated injury severity using the Abbreviated Injury Scale (AIS) and calculated NISS daily (Monday through Friday), ensuring the inclusion of patients admitted over the weekend. The NISS is a summary measure of anatomic injuries and is computed by summing the squares of the three most severe AIS scores, irrespective of body region [31]. A score of 10–15 indicates moderate injury, while scores > 15 represent severe trauma [32].

The Norwegian Trauma Registry utilises the 2008 update of the 2005 AIS version [33]. The AIS codes describe the body region affected and the severity of each injury. Injuries are categorised on a scale from 1 (minor) to 6 (maximum severity). AIS scores received from the local trauma registers at OUH and UNN, two of 38 hospitals constituting the Norwegian Trauma Registry (NTR) [28], were used for analysis. The registry has been shown to have good validity and reliability [34]. Injury classification was aligned with the system proposed by Anke et al. [35], with the most severely injured body region determined by the highest AIS score. Scores related to the external body region (skin) were excluded. For simplification, injuries in the thorax and the abdomen region were aggregated to one group. Spinal cord injuries identified via ICD codes were treated as a separate category. Deviating from Anke A et al., injuries in the upper and lower extremities and the spine without cord injuries were aggregated as orthopaedic. If multiple regions had identical AIS scores, a predefined hierarchy was used: head > extremities > face > thorax/abdomen.

One patient who had suffocated and had only external AIS scores was categorised as having a head/neck injury due to suspected brain hypoxia.

2.5. Demographic and Clinical Variables

Demographic information (age, gender, living situation, educational level and occupation) was obtained from the patients’ electronic records and supplemented with data from the 6-month follow-up when necessary. Pre-injury health status was assessed using the Norwegian adaptation of the American Society of Anesthesiologists (ASA) Physical Status Classification System defined by the NTR [36]. ASA scores were dichotomised into no disability (scores I–II) and disability (scores III–V) [37]. Clinical data acquired from patients’ records were diagnoses, number of injuries, substance use at the time of the injury and length of stay in the ICU.

2.6. Recommendation for Acute Rehabilitation

In line with the 2016/2017 version of the Norwegian Trauma Plan [13], this study examined discharge pathways from acute care to specialised rehabilitation. Discharging patients directly from acute care in the trauma centre to specialised inpatient rehabilitation is recommended. Acute care was defined as treatment in the (neuro-) ICU or surgical wards (mainly neurosurgery, orthopaedics and thorax surgery) at the trauma centre. Discharge destinations were categorised as either direct transfer to inpatient rehabilitation, indirect transfer via acute departments at local hospitals or no specialised inpatient rehabilitation within six months post-injury. Rehabilitation referred to inpatient treatment in specialised hospital departments or specialised rehabilitation institutions.

2.7. Instruments

WHODAS scores based on questionnaires completed by patients or/and relatives as part of the 6-month telephone-based follow-up were used to assess functioning. The WHODAS 2.0 12-item scale is a generic assessment instrument that can measure disability both at a population level and in clinical practice. As described in the manual, this tool evaluates the ability to be active in six domains (cognition, mobility, self-care, social interaction, life activities and participation) using two questions per domain. Each item is scored from 0 (no disability) to 4 (extreme difficulty or complete inability). The total score, the sum of the sub-scores, ranges from 0 to 48, with lower scores indicating better functioning [22]. Total scores were classified as mild (1–4), moderate (5–9) and severe disability (10–48) [23]. To simplify the model, clinically meaningful groups were considered. In this study, the mild (1–4) and moderate (5–9) groups were combined into a single mild/moderate group (1–9), as their baseline characteristics were similar (age, gender, ASA and NISS). No disability (0) and severe disability (10–48) were retained as distinct categories. In 23 cases which had one item missing, scores were imputed using the mean of the other 11 items, according to the WHODAS manual. A total of 8 cases with more than one missing item were excluded.

The Rehabilitation Complexity Scale–Extended Trauma (RCS-E Trauma) was used to assess the need for specialised rehabilitation [38]. RCS-E Trauma scores were estimated by physical medicine and rehabilitation specialists on the basis of clinical judgement at discharge from the acute department. Inter-rater reliability was tested with intra-class correlation coefficients (ICCs) between the three rehabilitation specialists who contributed scorings based on 11 patients, yielding excellent consistency (ICC = 0.933) and good agreement (ICC = 0.899).

RCS-E Trauma evaluates seven domains: medical needs (M, 0–6); basic care and support (C, 0–4) and risk/behavioural needs (R, 0–4); skilled nursing needs (N, 0–4); therapy needs, including therapy disciplines (TD, 0–4) and therapy intensity (TI, 0–4); and equipment needs (E, 0–3). For scoring, the higher of the C and R scores is used, yielding a total score from 0 to 25 [38]. The RCS-E was found to be a feasible and useful tool for the assessment of rehabilitation complexity in acute trauma care [11] and primary rehabilitation needs in patients with acquired brain injury [39].

2.8. Statistical Analysis

IBM SPSS Statistics Version 28 (IBM Corp., Armonk, NY, USA) was used for statistical analysis. Descriptive data are presented as proportions (percentages), medians with the interquartile range (IQR) and means with the standard deviation (SD). Dropout analysis for patients who had no complete WHODAS score at the 6-month follow-up was performed using the chi-square test. In the current study, we included all participants in the analyses without selecting for injury severity. Univariable and multivariable ordinal logistic regression analyses were used to investigate the association between sociodemographic and injury-related factors and the outcome disability, expressed in three categories (none, mild/moderate and severe) based on WHODAS 2.0 12-item scores at 6 months. To explore factors associated with different pathways to inpatient rehabilitation, groups were compared with the chi-square test, One-way ANOVA or Mann–Whitney U-test.

The following sociodemographic and injury-related factors were included in the univariable ordinal logistic regression analysis based on the literature and expert opinions: patients’ age (continuous), gender, living situation (living alone/not living alone), work situation (working or studying/not working or pensioned), educational level (high education, >13 years/lower education, ≤13 years total, including primary school), pre-injury ASA (dichotomised into score 1–2 versus 3–4) and substance use (influenced by any legal or illegal drug) at the time of injury (yes/no). Also included were the acute injury-related variables severity of injuries (NISS), total number of injuries in all body regions and the categories of body region with the most severe injury (head, orthopaedics, spinal cord or thorax/abdomen). The estimated rehabilitation complexity (RCS-E Trauma) sum score was added as a continuous variable. The discharge direction was grouped in three categories: directly or not directly transfer to specialist rehabilitation or no specialised rehabilitation. Factors with p < 0.10 in the univariable analysis were included in the multivariable analysis. Based on the event-per-parameter criterion, approximately 10 variables could be included. In the present analyses, a few more variables were entered. In consultation with a statistician, model diagnostics indicated an acceptable fit.

The results are given as odds ratios (ORs) with a 95% confidence interval (95% CI). The proportional odds assumption was satisfied. The degree of multicollinearity was checked using Spearman’s correlation coefficient, with ≥0.7 as the cut-off. Explored characteristics of patients with different pathways to inpatient rehabilitation were age, gender, living situation, education, work, drug influence, most severe injury, NISS, ICU days, acute ICU rehabilitation, RCS-E Trauma and WHODAS sum. The level of significance was set to p < 0.05.

3. Results

Of the 527 participants eligible for a follow-up at 6 months post-injury (Figure 1), 452 participants attended the 6-month follow-up, and 398 (75.5% of eligible participants) had complete WHODAS 2.0 scores and were included in this study.

The 129 participants not included in this study were younger (mean 47 vs. 52 years p = 0.004), lived more often alone (54 vs. 34% p < 0.001), were more often not working at the time of the injury (59 vs. 39% p = 0.006), more often had an ASA ≥ 3 (17 vs. 10% p = 0.032) and had a longer average length of stay in the ICU (6.1 vs. 3.7 days, p = 0.006). The 398 participants who completed WHODAS 2.0 questionnaire had a mean age of 52 years (SD 18), and 90 (23%) were women (see

Table 1. Demographic and injury-related characteristics of 398 included adult patients with moderate and severe trauma.

	Participants N = 398
Demographic characteristics
Age, years
Median (range)	55 (18–93)
Mean (SD)	52 (18)
Gender, n (%)
Female	90 (23)
Male	308 (77)
Education, n (%)
Lower education ≤ 13 years	217 (54.5)
High school/university > 13 years	173 (43.5)
Unknown	8 (2)
Work at the time of trauma, n (%)
Working/studying	238 (60)
Without occupation, pensioned	151 (38)
Unknown	9 (2)
Living situation, n (%)
Living with someone	261 (66)
Injury related characteristics
Substance use at time of the accident, n (%)	87 (22)
Pre-injury ASA, n (%)
ASA I–II, no disability	358 (90)
ASA III–V, disability	40 (10)
New Injury Severity Score (NISS)
Median (range)	22 (10–75)
Moderate NISS 10–15, n (%)	91 (23)
Severe NISS ≥ 16, n (%)	307 (77)
Injury Severity Score (ISS)
Median (range)	17 (4–66)
ISS ≥ 16, n (%)	214 (54%)
AIS organ area with highest score, n (%) *
Head	154 (39)
Extremities/spine without spinal cord injury	88 (22)
Spinal cord	36 (9)
Face	9 (2)
Thorax/abdomen	111 (28)

* Participants may have injuries in more than one body region. ASA = American Society of Anaesthesiologists Physical Status Classification System; AIS = Abbreviated Injury Score; ISS = Injury Severity Score; NISS = New Injury Severity Score.

).

As shown in Table 1, the median NISS was 22 (range 10–75), and 77% of the participants were severely injured (NISS > 15). The most common major injuries were head (39%), thorax/abdomen (28%) and orthopaedic (22%) injuries. Reported degrees of disability were none, 22%; mild/moderate, 49%; and severe, 29% (

Table 2. Association between demographic and injury-related factors and the outcome variable disability in 398 adults with moderate to severe trauma.

	Total N = 396 n (%)	No Disability N = 86 n (%)	Mild/Moderate Disability N = 197 n (%)	Severe Disability N = 115 n (%)	Multivariable
					OR	95% CI	p Value **
Age
Mean (SD)	52 (17.7)	51 (17.7)	51 (18.2)	56 (16,4)	1.0	0.99–1.02	0.359
Median (IQR)	55 (38–67)	55 (37–66)	53 (34–66)	56 (45–68)
Gender
Female	90 (23)	15 (17)	42 (21)	33 (29)	0.72	0.45–1.17	0.187
Male	308 (77)	71 (83)	155 (79)	82 (71)
Education < 13 years n = 390	217 (55)	34 (40)	111 (56)	72 (63)	1.65	1.09–2.49	0.018
Working/studying	238 (60)	59 (69)	126 (64)	53 (46)	1.07	0.64–1.79	0.788
Pre-injury ASA ≥ 3	40 (10)	5 (6)	12 (6)	23 (20)	0.34	0.16–0.71	0.004
Register NISS median (IQR)	22 (17–28)	22 (17–27)	22 (14–27)	24 (17–34)	1.0	0.98–1.03	0.842
Number of injuries mean (SD)	6 (3.7)	5.1 (3.1)	5.7 (3.1)	7.1 (4.7)	1.12	1.04–1.2	0.003
Head/neck	154 (39)	41 (48)	74 (38)	39 (34)	1.01	0.25–4.17	0.988
Orthopaedic injuries	88 (22)	10 (12)	46 (23)	32 (28)	0.37	0.09–1.55	0.173
Spinal Cord	36 (9)	2 (2)	16 (8)	18 (16)	0.22	0.05–1.03	0.055
Thorax/abdomen	111 (28)	30 (35)	56 (28)	25 (22)	0.82	0.2–3.32	0.781
No specialised rehabilitation	251 (63)	63 (73)	128 (65)	60 (52)
Direct specialised	80 (20)	14 (16)	38 (19)	28 (24)	1.08 *	0.20–2.06	0.816
Indirect specialised	67 (17)	9 (11)	31 (16)	27 (23)	1.26 *	0.66–2.39	0.485
Sum RCS-E Trauma baseline median (IQR)	9 (1–12)	6 (1–6)	12 (11–15)	13 (10–14)	1.06	1.0–1.12	0.036

Multivariable ordinal logistic regression analysis. Disability was measured with the WHO Disability Assessment Schedule (WHODAS 2.0, 12 item version) expressed in three categories: none (score 0), mild/moderate (score 1–9) and severe (score 10–48). Factors with p < 0.10 in the univariable analysis were included in the multivariable analyses. * No rehabilitation is the reference for direct and indirect rehabilitation in the analysis. ** Significant results are written in bold.

). Age, female gender, working, a higher NISS, dominant orthopaedic injuries, spinal cord injuries and thorax/abdomen injuries were significant predictors for disability in univariable regression, as well as direct and non-direct transfer to specialised rehabilitation. In the multivariable analysis, only low education (OR 1.65, 95% CI 1.09–2.49), comorbidities (OR 0.34, 95% CI 0.16–0.71), number of injuries (OR 1.12, 95% CI 1.04–1.2) and higher estimated rehabilitation complexity at discharge from acute care (RCS-E Trauma needs) (OR 1.06, 95% CI 1.0–1.12) were significantly associated with higher levels of disability at 6 months. In contrast age, living situation, substance use, NISS and direct transfer to specialised inpatient rehabilitation were not significant.

As seen in Table 2 and

Table 3. Personal and clinical characteristics of 398 patients in different care pathways to inpatient rehabilitation.

	Total n (%)	No Inpatient Rehabilitation N = 251 (63%) n (%)	Direct Specialised Rehabilitation N = 80 (20%) n (%)	Indirect Specialised Rehabilitation N = 67 (17%) n (%)	p Value
Age, mean (SD)	52 (18)	52 (18)	50 (17)	54 (16)	0.904
Women	90 (23)	54 (22)	13 (16) a	23 (34) a	0.026
Men	308 (77)	197(78)	67 (84)	44 (66)
Living alone	137 (34)	84 (34)	32 (40)	21 (31)	ns
Education < 13 years	217 (55)	131 (53)	45 (56)	41 (61)	ns
Working/studying	238 (60)	150 (60)	55 (69) a	33 (49) a	0.042
ASA ≥ 3	40 (10)	26 (10)	5 (0 6)	9 (13)	ns
Drug influence at time of injury	87 (22)	53 (21)	26 (33) b	8 (12) b	0.010
Head	154 (39)	76 (30)	51 (64) b	27 (40) b	<0.001
Orthopaedic injuries	88 (22)	60 (24)	5 (6) b	23 (34) b	<0.001
Spinal Cord	36 (9)	12 (5)	19 (24)a	5 (8) a	<0.001
Face	9 (2)	9 (4)	0	0
Thorax/abdomen	111 (28)	94 (38)	5 (6) a	12 (18) a	<0.001
NISS mean (SD)	24.3 (11.9)	20.6 (8.7)	32.1 (14.5)	29 (12.7)	<0.001
ICU days mean (SD)	3.7 (5.6)	2.2 (2.4)	7.8 (7.5)	6.7 (8.1)	<0.001
Sum RCS-E Trauma baseline median (IQR)	9 (1–12)	6 (1–6)	12 (11–15)	13 (10–14)	<0.001
WHODAS 6 month sum score median (IQR)	5 (1–11)	3 (0–9)	5 (1–14)	8 (4–14)	0.012

^a Difference between direct and indirect specialised inpatient rehabilitation, p < 0.05. ^b Difference between direct and indirect specialised inpatient rehabilitation, p < 0.01.

, the recommendation to transfer patients directly from acute trauma centre care to specialised inpatient rehabilitation was fulfilled for 20% of the participants, 17% were transferred indirectly via a local hospital, and 63% had no specialised inpatient rehabilitation after the acute care hospital stay. Table 3 explores characteristics of patients with different pathways to inpatient rehabilitation. Despite similar demographic and injury-related variables analysed in this study, women were transferred less often directly to specialised inpatient rehabilitation than men (14% vs. 22%, p = 0.026) and more often indirectly via local hospitals (26 vs. 14%, p = 0.012). Significantly fewer participants with major orthopaedic injuries were transferred directly than non-directly (6% vs. 26%, p < 0.001) to specialised rehabilitation. There was no statistically significant difference in median WHODAS sum scores between directly and indirectly transferred patients (5 vs. 8, p = 0.175).

4. Discussion

This prospective multicentre study investigated outcomes and rehabilitation pathways for adults during the first 6 months after moderate to severe traumatic injuries. Nearly half of the participants reported mild to moderate disability after 6 months, and about 30% reported severe disability after 6 months. In the multivariable analysis, factors associated with disability at 6 months post-injury were low level of education, as well as acute injury-related factors, including occurrence of comorbidity, number of injuries and rehabilitation complexity at discharge from acute trauma care. Having followed the recommendation to transfer patients directly to specialised rehabilitation and not via a local hospital was not associated with disability in this observational study of a general trauma centre population. Among the approximately 40% who received specialised inpatient rehabilitation, half were transferred via a local hospital and not directly from acute care. In the exploratory analysis, women were under-represented among the direct-transfer patients, an unexpected finding. Participants who were transferred directly more often had head and spinal cord injuries and less often, orthopaedic injuries.

4.1. Disability in General

This study of a general trauma population measuring disability with the self-reported complete WHODAS 2.0 12-item score found a higher rate of disability (80%) than other studies (50–68%), although some studies had more severely injured participants. In a study by Gabbe et al., 68% of the participants with severe injuries reported disability after 6 months (measured with GOSE-E, with levels 2–6 representing disability) [40]. In a multicentre study from three hospitals in Hong Kong, Rainer et al. found that 50% of the surviving patients had disability 6 months post-injury, measured with GOSE-E [41]. The study included 400 patients with moderate to severe injuries (ISS ≥ 9); 35% of the patients had severe injuries (ISS ≥ 16), which is 42% lower than in our cohort. In a study by Haider et al., 60% of the patients had physical limitations 6–12 months post-injury, measured with the Trauma Quality of Life (TQoL) Instrument (1537 participants, ISS ≥ 9) [41]. In the same project as the current study but including all age groups, Moksnes et al. [42] found that 58% of their participants had disability after 6 months, measured with Glasgow Outcome Scale—Extended (GOSE-E) (score 2–6). The different instruments used to measure disability are not equal, and use of WHODAS 2.0 as a self-reporting instrument with more detailed measuring of disability compared to GOSE-E may explain the higher rate of participants with disability in the current study. In this study, we did not adjust for pre-injury disability or disability that occurred post-injury independently from the injury; this may have also contributed to a high rate of disability. Time of measurement also influences disability grade because some patients will gain better function after 6 months. Additional measuring points, for example, 12 months post-injury, would give a clearer view of how disability evolves over time.

In this study, participants with spinal cord injuries or major orthopaedic injuries more frequently reported severe disability after six months than patients with a predominant head injury or injuries in the thorax/abdomen. This is in line with the study of multiple trauma patients by Anke et al. [35], where significantly fewer patients with predominant thorax or abdominal injuries had impairments at a 3-year follow-up. The lower rate of participants with a dominating head injury who reported severe disability may be caused partly by an underestimation of problems by the participants, something that is common in patients with brain injury.

4.2. Predictors for Disability

Lower education, comorbidities, number of injuries and higher estimated rehabilitation complexity at discharge from acute care were adjusted predictors for disability in this study. RCS-E Trauma scores, which indicate rehabilitation complexity, have been shown to be correlated to injury severity (measured by ISS) and hospital length of stay [11], which is in line with expectations. A lower level of education is known to be a factor with a generally negative influence on health; it has been shown to have a detrimental effect on disability after traumatic injuries, including restrictions on participation due to work disability [3,7,25,43]. A higher grade of comorbidities as well as number of injuries and spinal cord injuries are also known factors in poorer outcome [3,25]. Older age [3,35,43] and gender [7] are other known factors, but in our study, they were significant only in the univariate analysis. This may be due to the relatively limited number of participants in our study compared to some of the other studies, which recruited participants from registers.

Fifty percent of the participants in this study with spinal cord injuries had severe disability after six months, and 36% of the participants with major orthopaedic injuries had severe disability. Well-organised rehabilitation pathways exist for persons with spinal cord injuries, and their transfer rate to specialised inpatient rehabilitation was high, but participants with predominantly orthopaedic injuries received specialised rehabilitation less often. More systematic assessment of patients with dominant orthopaedic injuries may contribute to their identification and initiation of rehabilitation according to their needs.

4.3. Disability in Relation to Guidelines for Direct Transfer to Specialised Rehabilitation

Direct transfer from acute trauma centre departments to specialised inpatient rehabilitation was not associated with participants’ grade of disability in the present study. Also, in the exploratory analysis, this study showed no statistically significant difference in disability between participants transferred directly from acute departments to specialised inpatient rehabilitation vs. participants transferred indirectly via acute wards in local hospitals. On the contrary, a retrospective study of 6484 patients with mild to severe trauma (patients with ISS ≥ 9 =82%) by Scott et al. found lower mortality and a better chance for good recovery (Glasgow Outcome Scale) in patients following an embedded rehabilitation program, including a customised rehabilitation prescription after assessing patients’ rehabilitation needs within two to four days of admission [44]. Bouman et al. found faster recovery after 6 months in a study of multi-traumatised patients included in a “Fast track” rehabilitation service [9]. The current study used the NISS as the only inclusion factor related to injury severity and only included patients with moderate and severe injuries, while Scott et al. [44] also included patients with mild injuries, and Bouman et al. [9] used more specific definitions of trauma types. We have no data on how long patients who were transferred to local hospitals stayed in the acute wards before being transferred further to a specialised rehabilitation unit. Short delays may have less influence on outcomes, especially if transfer to specialised inpatient rehabilitation is planned and initiated from the trauma centre. In addition, other factors, like the amount of rehabilitative community-based treatment, may be important for the grade of disability after six months. In orthopaedic injuries, weightbearing and other restrictions may interfere with rehabilitation, and planned and individually adjusted delays may be necessary for some of these patients.

In this study, women were transferred less often directly to specialised rehabilitation significantly than men. Some studies have shown that women have worse physical and mental outcomes [7,26,45], but this has not been shown in all studies [27]. Nevertheless, there is no indication in the data of this study that women have less need for direct transfer for specialised rehabilitation. However, as stated by Mikolic et al. [46], residual confounding remains possible, or the observed differences could also result from insufficient adjustments for differences in clinical needs. In one study, after traumatic brain injury, patients who were transferred directly to inpatient rehabilitation were younger than patients transferred via a local hospital [47]. Our study population had different inclusion criteria and was older on average, and only a minority had severe traumatic brain injuries, and we did not find the same selection of rehabilitation pathway based on age. Even if this was only found in unadjusted analyses, and thereby not controlled for possible influencing factors like age and injury types, a possible selection of working patients for direct transfer to specialised rehabilitation prevails. Taken together, multiple and highly individual factors are influencing the grade of disability, and given the rather low number of patients transferred directly to rehabilitation, caution in interpretation of the results is needed. However, recommendations for acute rehabilitation and direct transfer to specialised inpatient rehabilitation are based on documentation for several patient groups [8,9,44]. Therefore, more studies with a general trauma population and consequent use of guidelines are needed to evaluate and improve guidelines.

4.4. Strengths and Limitations

We conducted a prospective cohort study from two trauma centres covering about half of the Norwegian population. Although the study population covers a whole year, the number of participants in the different groups is relatively small. Therefore, and to show the whole spectrum of disability, we included all participants in the analyses in the current study, even if the acute rehabilitation recommendation applies only to patients with severe injuries. This leads, at the same time, to a lower percentage of patients transferred directly, as shown in a previous published article [20], giving a worse picture as if only the participants with severe injuries would have been included in this part of the analysis, even if the differences were relatively small. To evaluate the long-term effect of adherence to the rehabilitation guidelines, one ideally should conduct controlled studies. The lack of pre-injury disability data beyond ASA scores may lead to bias as this relative unspecific instrument may not be able to show the whole picture for each patient. WHODAS 2.0 covers mental health problems, which are common after trauma, relatively poorly. This may lead to an underestimation of disability caused by mental health problems that are triggered by trauma. Some selection bias occurred, as the drop-out group was significantly younger and had a longer stay in the ICU. RCS-E Trauma is not validated in a population with general trauma in the Norwegian health care system, and, therefore, it is uncertain to what degree it can predict the level of rehabilitation treatment necessary to fulfil rehabilitation needs. The finding of the under-representation of women in direct transfer to specialised rehabilitation was made in the unadjusted analyses, but age- and injury-related factors were checked and found to be similar to those in men. The design of the study did not include a registration of motivations for referral or discharge decisions made by clinicians and therefore cannot describe if pathways were chosen based on patients’ needs or for pragmatic reasons, like the availability of rehabilitation places. An observation period of 6 months is not appropriate to assess long-term disability for all patients, since some patients will experience further gain in function, while others may develop complications later. However, we consider a 6-month observation point to be relevant to identify patients with further rehabilitation needs after the subacute rehabilitation period, which often is completed at this time.

5. Conclusions

Most patients with moderate to severe traumatic injuries had some grade of disability 6 months post-injury. As expected, low education and the acute injury-related predictors, total number of injuries and estimated rehabilitation complexity, were found to predict higher levels of disability. The recommendation of direct transfer to specialised rehabilitation was not associated with disability in this general trauma centre population. Few patients with orthopaedic injuries and less women than men were transferred directly to specialised rehabilitation.

Understanding the predictors of disability and the characteristics of patients following different rehabilitation pathways can help improve trauma rehabilitation practices and guidelines. Recommendations for acute rehabilitation are based on evidence for several trauma groups. To ensure optimal rehabilitation for all patients, it is important to have specific guidelines that accurately identify target groups.