**3. Results**

A total of 1887 patients ≥18 years, residents of the region and with CT-verified TBI were admitted to OUH from 2015 to 2019. Patients discharged as dead (GOS 1) were excluded (*n* = 163); thus, 1724 patients discharged alive from the acute care units at OUH were included in this study. The mean patient age was 57 years (SD 20), 69% were males, 87% lived independently at home, and 69% had a preinjury ASA-PS score of 1–2. The most frequent trauma mechanism was falls (56%). Head injury severity was categorized as complicated mild in 49%, moderate in 30% and severe in 21%. Further patient and injury characteristics are given in Table 1.


**Table 1.** Patient and injury characteristics.


**Table 1.** *Cont.*

<sup>1</sup> Patient discharged directly to specialized rehabilitation. <sup>2</sup> Patient discharged to home, local hospital, nursing home or other. <sup>3</sup> One patient may have more than one type of traumatic pathology. <sup>4</sup> Includes basilar, linear vault, depressed vault. <sup>5</sup> tSAH: traumatic subarachnoidal hemorrhage.

ICU admission was registered for 64% of the patients, with a median ICU stay of 3 days (IQR 2–10). A neurosurgical procedure was performed in 21% of patients, evacuation of mass lesions was the most frequently performed procedure (13%). An ICP sensor was inserted in 22%. Data on acute treatments provided are presented in Table 2.


**Table 2.** Acute treatment.

<sup>1</sup> Patient discharged directly to specialized rehabilitation. <sup>2</sup> Patient discharged to home, local hospital, nursing home or other. <sup>3</sup> Neurosurgery includes evacuation of mass lesions, decompressive hemicraniectomy, CSF diversion, duraplasty/cranioplasty, and vascular surgery. One patient may undergo several procedures.

Functional outcomes at the time of discharge from the acute care units, measured with the GOS, indicated good recovery in 5%, moderate disability in 46%, severe disability in 46%, and a vegetative state in 3% (Table 3). In patients with good recovery as measured by the GOS, 82/83 (98.8%) were discharged directly home (one patient was discharged to "other"). The two main reasons registered for reduced functional outcome (GOS < 5) were TBI alone and a combination of TBI/extracranial injury/comorbidity. The majority of patients were discharged to their local hospital (43%), followed by home (32%) and specialized rehabilitation (17%) (Table 3). In patients with severe TBI, 39% (139/353) entered direct pathway, and 20% (105/520) of patients with moderate TBI.


**Table 3.** Day of discharge from acute care units at OUH: functional level and destination.

<sup>1</sup> Patient discharged directly to specialized rehabilitation. <sup>2</sup> Patient discharged to home, local hospital, nursing home or other. <sup>3</sup> GOS: Glasgow Outcome Score.

> Tables 1–3 contain a comparison of patients discharged to specialized rehabilitation (direct pathway group) with patients discharged to other destinations (indirect pathway group). Comparing the two groups, the direct pathway group was younger, was more often male, had less comorbidity, had more severe TBI, had more intensive hospital treatment (neurosurgical procedures, ICP monitoring, ICU stay and days on ventilator), and had lower GOSs at the time of discharge.

> The proportion of patients with moderate–severe TBI (N = 873) discharged directly to rehabilitation increased during the period, from 22% (36/166) in 2015 to 35% (63/180) in 2019 (Figure 1). A patient with moderate–severe TBI admitted in 2019 had higher odds for entering a direct pathway than a patient admitted in 2015 (OR 1.17, CI 1.05–2.30, ptrend = 0.004).

> However, the proportions of direct pathways to rehabilitation differed substantially between age strata (moderate–severe TBI). For the youngest patient stratum (18–29 years), the proportion following the direct pathway was 53%, in contrast to patients of retirement age (65–79 years) in which it dropped to 10%, with the majority discharged to local hospitals (74%, 166/225). The distribution of discharge locations within age strata is shown in Figure 2.

> To identify potential predictors for discharge directly to a rehabilitation unit, uniand multi-variate logistic regression was performed. In univariate logistic regression, the

following factors were associated with an increased likelihood of discharge directly to a rehabilitation unit: younger age, male sex, living independently, low preinjury comorbidity (ASA 1–2), increased TBI severity, placement of ICP sensors, neurosurgical procedures, extracranial surgery, lower GOS at the time of discharge, and lower GOS at discharge due to TBI and no concomitant extracranial injury. Substance dependence showed no association with direct transfer to specialized rehabilitation and hence not included in the multivariate models. In multivariate regression, the following factors remained significantly associated with an increased likelihood of discharge directly to a rehabilitation unit: younger age, living independently, low preinjury comorbidity (ASA 1–2), increased TBI severity, lower GOS at the time of discharge, and lower GOS at discharge due to TBI and no concomitant extracranial injury (Table 4). Subgroup analysis for patients with moderate–severe TBI showed similar results, except lower GOS due to TBI was not significantly associated with an increased likelihood of entering a direct pathway (Table 5).

**Figure 1.** Direct pathway by year for patients with moderate–severe TBI (*n* = 873). The percentage of patients following the direct pathway (rehabilitation) increased during the period. "Other" includes general rehabilitation, nursing home and other.

**Figure 2.** Direct pathway by age strata for patients with moderate–severe TBI (*n* = 873). The percentage following the direct pathway (rehabilitation) decreased with increasing age. "Other" includes general rehabilitation, nursing home and other.


**Table 4.** Predictors associated with discharge directly to a rehabilitation unit (*n* = 1724). A total of 1637 patients were included in the model; 87 were excluded (83 patients' GOS 5-GR, 4 patients' GOS not available).

<sup>1</sup> The Hosmer and Lemeshow goodness-of-fit test was not significant, indicating a satisfactory fit of the model (χ 2 = 3.05, df = 8, *p* = 0.93). The area under the ROC curve was 0.86 (95% CI: 0.84–0.88), indicating good discriminative ability.

**Table 5.** Predictors associated with discharge directly to a rehabilitation unit. Subgroup analysis of patients admitted with moderate–severe TBI (*n* = 873): 862 patients were included in the model; 11 were excluded (7 patients' GOS 5-GR, 4 patients' GOS not available).



**Table 5.** *Cont.*

<sup>1</sup> The Hosmer and Lemeshow goodness-of-fit test was not significant, indicating satisfactory fit of the model (χ<sup>2</sup> = 4.14, df = 8, *p* = 0.85). The area under the ROC curve was 0.82 (95% CI: 0.80–0.85), indicating good discriminative ability.

#### **4. Discussion**

*4.1. Main Findings*

In this study, we aimed to provide an overview of discharge to specialized rehabilitation following acute TBI from 2015 to 2019 and to identify factors associated with a direct pathway to rehabilitation from acute care units. We found a significant positive trend in the number of patients who followed the direct pathway during the five-year study period. Patients discharged to the direct pathway typically had the following characteristics: younger age, low preinjury comorbidity, moderate–severe TBI and disability due TBI at the time of discharge. However, the study revealed significant differences in the proportions of patients following a direct pathway among age strata.
