Explanation for the choice of comparators

Exercise and early rehabilitation are beneficial for older people in terms of improved physical functioning, shorter hospital stay, and reduced care home admissions [8, 9]. Thus, to evaluate whether robot-assisted physical resistance training can be a good strategy to mitigate the negative outcomes of inactivity during hospitalisation for this population, passive training movements by the robot are considered as a sham control group for comparing the effect of the active resistance training intervention group. Sham controls provide the highest quality and potentially most generalisable clinical trial data. This setup is particularly useful when studying devices controlling for the ancillary effects of a procedure, optimising the ability of the investigator to evaluate for a placebo or procedural effect in an unbiased fashion [15].

Intervention description

The robot-assisted training is performed using an innovative training robot (ROBERT®) (Fig. 1). The robot is able to hold the patient’s leg and perform extension movements of the hip and knee (Fig. 2). The movement can be performed either active or passive. In the active training mode, the patient must use their muscular power to stretch the leg while the robot provides low/moderate resistance. In the passive mode, the robot moves the leg independently without the patient using their own muscle power. Exercise is performed on both legs separately. The robot-assisted training is defined as a minimum of three sessions during the hospital stay. Following each training session, all participants are offered nutritional drinkable supplements each containing a minimum of 18 to 26 g of protein per serving (125–250 ml) [16, 17].

Fig. 1Fig. 2

ROBERT® can hold the patient’s leg and perform extension of hip and knee

Intervention group

The intervention group will receive active robot-assisted resistance training twice a day until the day of discharge in addition to standard individualised care and physiotherapy. A resistance training session consists of three sets of active extensions of the hip and knee under verbal motivation in order to perform as many repetitions as possible. Training intensity is 65–100% of maximum capacity with a break of 60 s between each set. To ensure progression, the level of resistance is assessed at each session. The Borg scale is used for ratings of perceived exertion following every training set [18].

Control group

Participants in the control group will receive passive robot-assisted sham training by the robot twice a day until the day of discharge in addition to standard individualised care and physiotherapy. A sham training session consists of three sets of eight repetitions performed by the training robot, which passively moves the patient’s leg in extension of the hip and knee with a break of 60 s between each set.

Criteria for discontinuing or modifying allocated interventions

Participants will discontinue allocated intervention if there is a deconditioning in their health state or if they develop contradictions for robot training during their in-hospital stay. Also, the intervention will terminate if a participant is referred to another hospital.

Any patient requesting to terminate their study participation can be withdrawn from the study regardless of the stage they have reached in the study process. Patients do not have to provide the reason of their study withdrawal.

Strategies to improve adherence to interventions

Appointed research staff are present in the department every day to carry out screening, inclusion, and robot training with the patients. The research staff will encourage patients to continue daily training (both sham and resistance) to improve adherence and furthermore collect data on the amount of robot training. This will ensure daily motivation for continued participation. Reasons for discontinuing participation will be collected.

Relevant concomitant care permitted or prohibited during the trial

The robot training will not require alteration to usual care pathways for participants during hospitalisation or following discharge (including the use of any medication or physical activity).

Provisions for post-trial care

All patients who participate in this study are covered by Danish Patient Compensation if they suffer harm from trial participation [19].

Outcomes

Outcomes in the current study are selected based on inputs from patient public involvement. During study preparation patients, relatives, and stakeholders including representatives from the municipality’s health care setting participated in a workshop with the aim of designating research focus areas. The group rated functional independence and quality of life as the most important outcomes of interest.

Baseline data

Demographic data will be collected at baseline to describe the study population and assess potential subgroup differences. This includes the following data: age, gender, civil- and living status, body mass index (BMI), use of medications, C-reactive protein (CRP) blood sample, historic Barthel Index (14 days before hospital admission), and reason for hospital admission (ICD-10 diagnose [20]).

Primary outcome

The primary outcome of interest is functional status defined as activities of daily living (ADL) measured by Barthel and sit-to-stand performance measured by 30-s chair stand test from baseline prior to randomisation and to the day of discharge. ADL characterise the capability of a person to do everyday routine activities. The current study uses the Barthel Index-100, which is a recognised and simple scoring instrument used to evaluate basic ADL functions, the level of physical performance, and the intensity of needed care [21]. The Barthel Index is often used as a gold standard comparator in studies addressing ADL [22]. It is a sum score across ten domains of ADL and the total score ranges from 0 (completely dependent) to 100 (completely independent) [23]. Sit-to-stand performance is assessed as the number of times a person is able to rise and sit from a standardised chair within 30 s [24].

Secondary outcomes

Several relevant secondary outcomes are assessed.

Functional status

Functional status as secondary outcome is assessed by Barthel Index-100 and sit-to-stand performance at baseline before randomisation and at 1- and 3-month follow-up.

Discharge destination

Discharge destination will be collected (own home, temporary rehabilitation units, care homes).

Need of care at home

Individual level data from the municipalities will be used to assess amount of home care provided by the municipality during the period of 3 months before admission to 3 months after discharge. Need of care at home is divided into categories of practical help, personnel care, nursing, and training. Amount of care will be presented in hours of provided assistance.

Muscle quantity

Muscle quantity is assessed using bioelectrical impedance by InBodyS10® [25] at baseline before randomisation, at discharge, and at 1- and 3-month follow-up. Bioelectrical impedance analysis (BIA) is a method for estimating body composition, in particular body fat and muscle mass.

Sarcopenia

Sarcopenia is assessed based on the 2019 European guidelines [26] at baseline before randomisation, at discharge, and at 1- and 3-month follow-up.

Clinical frailty scale

The 9-point Clinical Frailty Scale (CFS) with pictograms is used at baseline before randomisation, at discharge, and at 1- and 3-month follow-up [27] to examine the impact of the exercise intervention on frailty and to determine the impact of baseline frailty on the effectiveness of the intervention [28].

Quality of life

Quality of life is assessed using the questionnaire “Quality of life EuroQol-5 dimension (EQ-5D)” (patient-reported outcome (PRO)) at baseline before randomisation, at discharge, and at 1- and 3-month follow-up.

The EQ-5D comprises five dimensions (mobility, self-care, usual activities, pain/discomfort, and anxiety/depression). The answers given to EQ-5D generate 243 unique health states or can be converted into an EQ-5D index, a utility score anchored at 0 for death and 1 for perfect health. The EQ-5D questionnaire also includes a visual analogue scale (VAS), by which respondents can report their perceived health status with a grade ranging from 0 (the worst possible health status) to 100 (the best possible health status) [29, 30].

Concern about falling

Concern about falling is assessed using the 16-item Short Falls Efficacy Scale International (Short FES-I) questionnaire (patient-reported outcome (PRO)) [31, 32] including information on the actual number of falls. Assessments are performed at baseline before randomisation, at discharge, and at 1- and 3-month follow-up.

Cognition function

Cognitive function is assessed by the Mini-Mental State Examination (MMSE) at baseline before randomisation, at discharge, and at 1- and 3-month follow-up. The MMSE is a set of 11 questions used to assess potential cognitive impairment (problems with thinking, communication, understanding, and memory). The maximum score for the MMSE is 30. A score of 25 or higher is classed as normal. If the score is ≤ 24, the result is usually considered to be abnormal, indicating possible cognitive impairment.

Mood

Mood status is assessed by the 15-item Geriatric Depression Scale (GDS) at baseline before randomisation, at discharge, and at 1- and 3-month follow-up. The short form GDS consists of 15 questions. Of the 15 items, 10 indicate the presence of depressive symptoms when answered positively, while the rest indicate depressive symptoms when answered negatively. Scores of 0–4 are considered normal, depending on age, education, and complaints; 5–8 indicate mild depression; 9–11 moderate depression; and 12–15 indicate severe depression [33].

Hospitalisation

Length of hospital stay (LOS) will be defined as the number of days in the geriatric department. Readmission is defined as any unplanned hospital contact with a duration of 12 + h, occurring between 4 h and 30 days following discharge from the geriatric department [34]. Data will be collected using a medical records review.

Mortality

Mortality is assessed during hospital stay and at 1- and 3-month follow-up.

Patient perspective

Participant observation and qualitative semi-structured interviews will be performed with 10–12 patients during the trial period to explore patient perspectives and experiences (patient-reported experience (PRE)). The qualitative interviews will provide additional information to the results from the quantitative parts of the study. By using this mixed methods approach the research results will be strengthened by the complementary findings. The analysis of the qualitative data will be completed in line with manifest content analysis by Graneheim and Lundman [35].

Health care cost evaluation

A researcher in health care economics will perform a health care cost evaluation addressing running costs using the training robot, discharge destination, hospital readmissions, need of care at home, and visits to the general practitioner after 3 months.

Participant timeline

Patients will be enrolled and allocated within a maximum of 48 h from the time of admission to the geriatric department. Each day patients are listed in order according to most recently admitted and screened according to this list. When finding an eligible patient, the research assistant immediately asks the patient about participation and starts the baseline tests if the patient has agreed and signed the consent form. After baseline test completion, the screening continues. Daily screening ends when logistic capacity does not allow further daily participation. After baseline tests, patients will be allocated randomly to either active robot training or passive robot training. At discharge, the dates for 1- and 3-month follow-up are arranged with the patient. The schedule of enrolment, interventions, and assessments are shown schematically in Table 1.

Table 1 SPIRIT—schedule of enrolment, interventions, and assessmentsSample size

The minimum clinical important change for the 30-s chair stand test is 2.6 repetitions [36] and the mean reference value for people in the relevant age groups is 13 according to a recent Danish study [37]. With a significance level of 5%, 80% power, and an expected drop-out rate of 20%, 74 participants in each group (148 in total) are needed to detect an inter-group difference of 2.6 repetitions. The magnitude of clinical meaningful change of the Barthel Index is 5 points and a recent study in geriatric patients reported a change of 6.9 points during exercise [6]. The mean (SD) Barthel Index score of patients at the geriatric department OUH is 59.5 (24.3). To achieve 80% power for demonstrating the same mean difference, this study would require 244 participants per group (488 in total) with a significance level of 5% and an expected drop-out rate of 20%.

Thus, a total of 488 participants is required for our primary outcome.

Recruitment

A total of 74% of the approached eligible older patients agreed to participate in our previously performed pilot and feasibility study and expressed that they would recommend the training robot [13]. Thus, our study group already has experience using a training robot in geriatric patients. Our department is a 32-bed unit with ~ 2000 patients admitted each year. If we assume only 25% of the patients will be eligible for the trial and accept participation a realistic inclusion period to obtain 488 participants is 16 months.