Tele-rehabilitation for cognitive home-based treatment in young traumatic brain injury: a pilot study

Traumatic brain injury (TBI), commonly seen as a sudden, shattering event with short-term sequelae, is instead the most common cause of long-term disability among young adults that poses a substantial social and public health burden.¹

TBI affects around 64-74 million persons each year worldwide and causes a variety of motor, speech, and cognitive impairments which occur immediately after the trauma.² Nevertheless, neurological sequelae can persist up to months, years or for all life.³ In particular, while the limitation in activities due to motor impairment is evident, the persistence of problems in the cognitive sphere is less obvious but heavily interfering with work, relationships, leisure, and daily activities, with a consequent worsening of both patients’ quality of life and caregiver’s burden.^{4, 5}

For these patients who need extended follow-up and monitoring after discharge from a dedicated neurorehabilitation unit, telerehabilitation offers the possibility of long-term treatments and can improve access to neurorehabilitation professionals, especially for people living outside metropolitan areas.

However, the efficacy of telerehabilitation for adults with traumatic brain injury (TBI) has yet to be demonstrated, as only few feasibility and acceptability studies are available.^6-9 A systematic review conducted by Ownsworth et al. in 2017,¹⁰ reported that the most commonly used telerehabilitation modality for TBI is telephone-based intervention, with poor impact on cognitive function; in this review only 3 Internet-based interventions were identified that focused mainly on feasibility.

On this background, we evaluate whether a cognitive treatment, delivered in telerehabilitation mode, can be feasible and applicable in young adults with TBI. The secondary aim was to evaluate the effectiveness of this treatment in terms of improvement of cognitive functions in the patients examined, comparing an experimental group with a control group of outpatients subjected to traditional treatment.

For the experimental group (EG), six patients discharged from the Severe Acquired Brain Injuries Unit of Versilia Hospital were enrolled consecutively. They were 100% males, 21.60±3.19 years of age; education, expressed in years, was on average 13.16±2.41, time since the injury was 4.8±3.62 and the length of stay in the Severe Acquired Brain Injuries Unit was 55.17±26.11 days. The control group (CG) was selected retrospectively using clinical data deriving from the clinical records of patients hospitalized in the same department, who received outpatient cognitive rehabilitation treatment. They were 66.7% males, 33.3% females, 22.41±7.40 years of age; education, expressed in years, was on average 10.83±1.86 and the length of stay in the Severe Acquired Brain Injuries Unit was 26.50±12.04 days (Table I).

Table I. —Demographic and clinical data for experimental group (EG) and control group (GC).

Demographics	EG		CG		P value
	Mean±SD N. (%)	Median (IQR)	Mean±SD N. (%)	Median (IQR)
Participants	6		6
Age (years)	21.60±3.19	20.82 (19.19-23.01)	22.41±7.40	19.51 (16.81-29.06)	0.749
Education (years)	13.16±2.41	13.00 (12.25-13.00)	10.83±1.86	11.00 (9.50-12.50)	0.134
Gender
Male	100%		66.7%
Female	0%		33.3%
LOS (days)	55.17 (26.11)	55.17 (34.50-74.50)	26.50±12.04	21.50 (19.50-34.00)	0.025
LCF discharge	6.83±0.69	7.00 (6.25-7.00)	7.33±0.47	7.00 (7.00-7.75)	0.206
DRS discharge	10.67±4.82	10.00 (7.75-13.00)	6.17±2.34	6.00 (4.25-8.50)	0.076
Time since injury (months)	4.8±3.62	3.00 (2.25-6.00)	5.00±3.47	4.00 (2.25-5.75)	0.934

LOS: long of stay in Rehabilitation Unit; LCF discharge: level of cognitive functioning at discharge from rehabilitation unit; DRS discharge: Disability Rating Scale at discharge from Rehabilitation Unit.

No statistically significant difference are reported between EG and CG except for LOS even if the very small sample size makes the statistical power not very strong

As for the neurological functional status at the admission in Rehabilitation Unit, patients in EG showed mean LCF 4±1.52, DRS 18.83±4.13, whilst patients in CG showed LCF 4±1 and DRS 19.16±3.23. The difference in LOS may not be due to different neurological condition but to other factors (including orthopedic injuries, no family support, etc.).

Inclusion criteria for enrollment in the study were: traumatic etiology of severe brain injury; age between 18 and 35 years old; one or more cognitive impairment in the domains of attention, memory, executive functions and language; time elapsed between TBI ≤12 months. Exclusion criteria were: Level of Cognitive Functioning (LCF) rating scale score at discharge <5; Disability Rating Scale score ≥22; pre-existing cognitive impairment. The study was approved by local Ethical Committee and all patients gave an informed consent.

The device used to perform remote cognitive treatment at home is the HOMEKIT Khymeia^®. The system is divided into numerous rehabilitation modules (cognitive, neurological, speech therapy, orthopedic); all activities can be performed both using the touchscreen and via additional devices such as sensors, footboard, etc.

Within the cognitive module, the exercises involve different neuropsychological and cognitive function such as: attention, memory, executive functions, praxis, spatial orientation, logic and mathematics; some of these activities can be present in multiple domains as they integrate multiple cognitive functions. For each exercise there are settings that can be modified and customized basing on patients’ characteristics (e.g. difficulty of the exercises, the sensitivity, etc.) Exercise modality can be modified remotely at any time by clinicians with the aim of personalizing the rehabilitation session based on the patient’s functional abilities. The device automatically records every patient activity, thus generating a complete and objective reporting system containing the analysis and result of each performed exercise. There are two treatment methods: online and offline telerehabilitation. In the first case, the therapist connects via videoconference and takes remote control of the patient’s device by interacting with him in real time and he is able to modify the exercises; in the second case the patient performs the personalized exercise scheduled by clinical staff, guided by the Smart Virtual Assistant, which accompanies him in real-time interactive mode throughout the entire duration of the rehabilitation session.

According to the results, the exercises can be modified by the clinician during the online session.

At the time of discharge from the Neurorehabilitation Unit (T0), all patients underwent functional and cognitive assessment by means of: Level of Cognitive Functioning (LCF),¹¹ Disability Rating Scale (DRS)¹² and Brief Neuropsychological Examination (ENB2).¹³

The experimental group (EG), appropriately trained, took at home the HOMEKIT Khymeia^® for the training.

The rehabilitation session consisted of 30-minute sessions each 3 times/week for a total of 11 weeks of treatment. Every week an online connection was provided. At the end of the treatment (T1) all patients were evaluated with ENB2.

Adherence to the program was assessed in both groups as the percentage of completed rehabilitation sessions. The feasibility of the protocol was defined as the adherence of 80% of participants to the program (a subject is defined adhering to the protocol if completing at least 80% of the rehabilitation exercise sessions). In EG patient satisfaction was examined through the CSQ-8 questionnaire (Client Satisfaction Questionnaire)¹⁴ and the usability of the application through the SUS scale (System Usability Scale).¹⁵

The treatment of the control group (CG) was matched to EG consisting of 30 minutes each session, 3 times/week for a total of 11 weeks provided by face to face treatment in hospital. At the end of the treatment (T1) all patients were evaluated with ENB2.

As regards adherence to the protocol, EG completed 93.42% of sessions and CG completed 83.85% of the total sessions; therefore, for what concerns feasibility, all subjects of EG exceeded the threshold of 80% of sessions completed and 4 subjects of the CG exceeded the threshold.

The Client Satisfaction Questionnaire CSQ8 showed an average score of 25.00±3.70, with a minimum score of 21 and a maximum of 32. Overall, patients expressed a good level of satisfaction in both the qualitative and quantitative aspects of the service received.

The SUS score was 77.08±12.45 with a minimum score of 62.5 and a maximum of 92.5; in this scale scores of 90 indicate excellent usability, while scores below 50 indicate usability difficulties.¹⁵ The data therefore highlighted an overall acceptability of the device with values that indicate good and excellent usability of the system, revealing that the patients would recommend this program to others with similar impairment.

As for the outcome, a Mann-Whitney Test on the ENB2 scores showed an improvement in cognitive performance in all patients without statistically significant differences between the two groups both for the absolute score at T0 (P=0.330) and T1 (P=0.687) and when comparing the variation (Δ) of the outcome scale (Δ ENB P=1.000) (Table II). As for the intra-group analysis a Wilcoxon Signed Rank Test shows a statistically significant differences in both groups (P=0.031 for CG and P=0.028 in EG) (Table III).

Table II. —Comparison between cognitive assessment of EG and CG.

Parameter	EG	CG	P value
ENB2 T0	51.67±23.91; 66	66.67±6.82; 70.00	0.330
ENB2 T1	66.50±20.72; 74.00	75.33±4.46; 74.00	0.687
Δ (T1-T0)	14.83±19.12; 8.00	8.67±5.71; 7.00	1.000

Data expressed as mean±SD and median. ENB2: Brief Neuropsychological Examination.

Table III. —Intra-groups analysis.

Group	T0		T1		P value
	Mean±SD	Median (IQR)	Mean±SD	Median (IQR)
EG	51.67±23.91	66 (30.50-69.25)	66.50±20.72	74 (71-76.25)	0.028
CG	66.67±6.82	70 (64.75-71.50)	75.33±4.46	74 (72-76.75)	0.031

In conclusion telerehabilitation for cognitive function in TBI showed an excellent feasibility and the system seems to be easy and intuitive to use.

Overall, the score obtained on the satisfaction questionnaire in EG highlighted a good level of appreciation both in the qualitative and quantitative aspects of the services received with a higher adherence in the EG compared to the CG.

As regards the effectiveness of the treatment, the results highlight an improvement of cognitive performance in both the experimental group and the control group, even if without any difference in inter-groups analysis, suggesting that telerehabilitation is not inferior to standard treatment in this population.

In conclusion we believe that the use of a cognitive telerehabilitation system could be integrated into clinical practice to ensure continuity of home care, giving the possibility of extending rehabilitative treatments in frequency and intensity. This method could represent an opportunity to guarantee equal and better access to care and a stronger integration between hospital and territory, even if larger samples, multicenter studies are needed to further investigate the effectiveness of treatments provided in telerehabilitation in TBI.

The most problematic aspect, however, remains the efficiency of the internet line: paradoxically, the more peripheral areas, where remote treatment would be more advantageous, are those often worst served, with consequent difficulty in completing the sessions.

Furthermore, if the system requires a connection between the hospitals server and an external server, it could jeopardize the safety of internet connection.

Finally, the actual economic benefit that the healthcare system can derive from telerehabilitation, as well as some legal aspects particularly relating to the liability profiles in case of accidents, remain to be investigated.