AI-Supported Physiotherapy After Hip Fracture in Frail Older Adults

This randomized, parallel-group, controlled clinical trial is conducted in the Orthogeriatrics Unit of Ramón y Cajal University Hospital (Madrid, Spain) and follows the SPIRIT 2025 and CONSORT 2025 guidelines. It evaluates whether adding an automated motivational support and movement-capture system to standard hospital physiotherapy improves short- and medium-term functional recovery in frail older adults after hip fracture surgery, compared with standard physiotherapy alone.

Eligible participants are randomly assigned 1:1 to the experimental or control arm using concealed allocation (sequentially prepared, opaque envelopes drawn by the principal investigator). Outcome assessors and the statistician are blinded to group assignment.

Both arms receive one daily session of standard physiotherapy (Monday through Friday) during the hospital stay, following a structured two-phase program (seated exercises, then sit-to-stand and gait training). Participants in the experimental arm additionally complete one daily session with an automated motivational support and movement-capture system. The system uses a standard RGB camera and a body-pose estimation model (MediaPipe, inference mode) to record movement during the prescribed exercises. It performs no facial recognition or automated biometric identification and makes no clinical decisions. During the session it provides only pre-scheduled, generic auditory and visual motivational feedback (for example, session start/end cues and encouragement phrases), independent of how the patient performs the exercise; it gives no real-time technical correction. Video and derived movement data are analyzed afterward, off-line, for research purposes (for example, joint-angle estimation), and this analysis does not influence the feedback the patient receives. A study team member observes each session without intervening, except to stop it immediately in case of an adverse event, alarm, or other safety concern. Control-arm participants complete the same structured exercise program autonomously, per the unit's current protocol, without the motivational/capture system.

The intervention under study is a support and data-capture tool, not an investigational medicinal product or medical device: it provides no diagnostic or therapeutic function and does not adapt its output to exercise quality. The trial is therefore not subject to Regulation (EU) 2017/745, Royal Decree 1090/2015, or AEMPS authorization requirements.

Outcomes are assessed at three time points: baseline, once weight-bearing is authorized (T0); hospital discharge (T1); and 1-month follow-up (T2). The primary statistical analysis follows the intention-to-treat principle, with multiple imputation for missing data. Between-group adherence is compared using a symmetric, REDCap-based metric collected identically in both arms (structured observation by the study team in the experimental arm; a brief structured interview conducted by the treating physiotherapist in the control arm, given the autonomous nature of that arm's additional session).

• Conventional hospital physiotherapy (Other): Standard structured physiotherapy program delivered once daily, Monday through Friday, during the hospital stay, following a two-phase protocol (seated exercises; sit-to-stand and gait training).
• Automated motivational support and movement-capture system (Device): RGB camera-based system using a body-pose estimation model (MediaPipe) to record movement during exercise sessions for deferred, off-line biomechanical analysis. Provides only pre-scheduled, generic auditory and visual motivational feedback, independent of how the patient performs the exercise; performs no facial recognition, no automated biometric identification, and no real-time technical correction.

Primary outcomes

• Physical performance - Short Physical Performance Battery (SPPB) [Baseline (T0, day weight-bearing is authorized after surgery), at hospital discharge (T1, up to 30 days after baseline), and 1 month after discharge (T2)]: The Short Physical Performance Battery (SPPB) is a standardized, validated measure of lower-extremity physical function in older adults. It combines three timed tests: standing balance (feet together, semi-tandem, and tandem positions), gait speed over a 4-meter walk, and lower-limb strength (five repeated chair stands, performed without using the arms). Each component is scored from 0 to 4, for a total score ranging from 0 (worst) to 12 (best). Lower scores indicate greater functional impairment and a higher risk of disability and falls.

Secondary outcomes

• Pain - Visual Analog Scale (VAS) [Baseline (T0, day weight-bearing is authorized after surgery), at hospital discharge (T1, up to 30 days after baseline), and 1 month after discharge (T2)]: The Visual Analog Scale (VAS) is a self-reported instrument for measuring subjective pain intensity using a 10-centimeter line, where 0 represents no pain and 10 represents the worst pain imaginable. The participant marks the point on the line that best reflects their pain; the resulting score, measured in centimeters, classifies pain as mild (1-3), moderate (4-7), or severe (8-10).
• Muscle strength - Medical Research Council (MRC) Scale [Baseline (T0, day weight-bearing is authorized after surgery), at hospital discharge (T1, up to 30 days after baseline), and 1 month after discharge (T2)]: The Medical Research Council (MRC) Scale is a manual muscle-testing method used to assess the strength of the knee flexor/extensor, hip flexor/extensor, and hip abductor muscle groups. Each muscle group is scored from 0 to 5, where 0 indicates no visible or palpable contraction and 5 indicates normal strength against maximal resistance.
• Joint range of motion - Goniometry [Baseline (T0, day weight-bearing is authorized after surgery), at hospital discharge (T1, up to 30 days after baseline), and 1 month after discharge (T2)]: Active joint range of motion is measured with a goniometer for knee flexion-extension and hip flexion-extension and abduction.
• Basic activities of daily living - Barthel Index [Baseline (T0, day weight-bearing is authorized after surgery), at hospital discharge (T1, up to 30 days after baseline), and 1 month after discharge (T2)]: The Barthel Index assesses a person's level of independence in ten basic activities of daily living, such as eating, bathing, dressing, and toileting. Scores from each activity are summed for a total ranging from 0 to 100, with higher scores indicating greater independence (100 = total independence; 0-20 = total dependence).
• Ambulation capacity - Functional Ambulation Categories (FAC) [Baseline (T0, day weight-bearing is authorized after surgery), at hospital discharge (T1, up to 30 days after baseline), and 1 month after discharge (T2)]: The Functional Ambulation Categories (FAC) is a 6-level scale that assesses a patient's walking ability, ranging from total inability to walk (level 0) to independent ambulation on any surface (level 5), based on the amount of physical or device-based support required.
• Quality of life - SF-36 Health Survey [Baseline (T0, day weight-bearing is authorized after surgery), at hospital discharge (T1, up to 30 days after baseline), and 1 month after discharge (T2)]: The 36-item Short Form Health Survey (SF-36) assesses health-related quality of life across eight domains: physical functioning, physical role, bodily pain, general health, vitality, social functioning, emotional role, and mental health. Each domain is scored from 0 to 100, with higher scores indicating better health status.
• Treatment adherence - structured REDCap logs [Recorded daily (Monday through Friday) from baseline (T0) to hospital discharge (T1), up to 30 days]: Adherence to the prescribed exercise program is recorded for both study arms using identical structured REDCap forms. It is broken down into three components: session adherence (percentage of additional sessions completed relative to those prescribed during the hospital stay), intra-session adherence (percentage of prescribed exercises completed per session, averaged across sessions), and weekly persistence (number of days per week with a completed session relative to the planned schedule).
• Usability - System Usability Scale (SUS) and adapted survey [At the end of the intervention period, at hospital discharge (T1), up to 30 days after baseline (experimental group only)]: Usability of the automated motivational support and movement-capture system is assessed in the experimental group using the System Usability Scale (SUS), a standardized 10-item questionnaire scored from 0 to 100, with higher scores indicating better perceived usability. It is complemented by a specific survey adapted for older adults, covering clarity, ease of use, sense of safety, and perception of the motivational feedback.
• Operational safety of the automated system [Recorded during each session with the system, from baseline (T0) to hospital discharge (T1), up to 30 days (experimental group only)]: Operational safety of the automated motivational support and movement-capture system is assessed by recording any incidents or adverse events occurring during sessions, including the level of clinical supervision required.

Sex: All, ages: 70 Years to any

Inclusion Criteria:

* Age ≥ 70 years
* Diagnosis of hip fracture due to bone fragility
* Hospital admission to the Orthogeriatrics Unit, Ramón y Cajal University Hospital (Ward 3DA, rooms 301-314)
* Signed informed consent
* Frailty, assessed with the FRAIL-Spain scale (score 3 to 5)
* Weight-bearing authorized by the medical team (Trauma Surgery Department)

Exclusion Criteria:

* Hip fracture due to high-energy trauma
* Pathological fracture (bone metastasis or tumor)
* Pre-fracture ambulation capacity (FAC) of 0 to 2
* Other coexisting fractures
* Inability to complete clinical follow-up
* Lack of cooperation with the physiotherapy program
* Neurodegenerative disease at a moderate stage or beyond (cognitive impairment GDS ≥ 6, or Parkinson's disease Hoehn-Yahr stage 3-5)
* Institutionalized patients (due to earlier hospital discharge)
* Post-surgical complications (e.g., infection, poor surgical wound healing, suspected deep vein thrombosis)
* History of underlying conditions affecting balance or posture (cancer, dizziness, vertigo, or stroke)
* Physical disability that could compromise sample homogeneity
* Sensory impairment preventing use of digital tools
* Severe cognitive impairment (Lobo's Mini-Mental State Examination score ≤ 9, equivalent to GDS ≥ 6)

• Sevilla-Gonzalez MDR, Moreno Loaeza L, Lazaro-Carrera LS, Bourguet Ramirez B, Vazquez Rodriguez A, Peralta-Pedrero ML, Almeda-Valdes P. Spanish Version of the System Usability Scale for the Assessment of Electronic Tools: Development and Validation. JMIR Hum Factors. 2020 Dec 16;7(4):e21161. doi: 10.2196/21161. PMID 33325828
• Oh YJ, Zhang J, Fang ML, Fukuoka Y. A systematic review of artificial intelligence chatbots for promoting physical activity, healthy diet, and weight loss. Int J Behav Nutr Phys Act. 2021 Dec 11;18(1):160. doi: 10.1186/s12966-021-01224-6. PMID 34895247
• Kwon S, Perera S, Pahor M, Katula JA, King AC, Groessl EJ, Studenski SA. What is a meaningful change in physical performance? Findings from a clinical trial in older adults (the LIFE-P study). J Nutr Health Aging. 2009 Jun;13(6):538-44. doi: 10.1007/s12603-009-0104-z. PMID 19536422
• Nuevo M, Rodriguez-Rodriguez D, Jauregui R, Fabrellas N, Zabalegui A, Conti M, Prat-Fabregat S. Telerehabilitation following fast-track total knee arthroplasty is effective and safe: a randomized controlled trial with the ReHub(R) platform. Disabil Rehabil. 2024 Jun;46(12):2629-2639. doi: 10.1080/09638288.2023.2228689. Epub 2023 Jul 5. PMID 37403684
• Hopewell S, Chan AW, Collins GS, Hrobjartsson A, Moher D, Schulz KF, Tunn R, Aggarwal R, Berkwits M, Berlin JA, Bhandari N, Butcher NJ, Campbell MK, Chidebe RCW, Elbourne D, Farmer A, Fergusson DA, Golub RM, Goodman SN, Hoffmann TC, Ioannidis JPA, Kahan BC, Knowles RL, Lamb SE, Lewis S, Loder E, Offringa M, Ravaud P, Richards DP, Rockhold FW, Schriger DL, Siegfried NL, Staniszewska S, Taylor RS, Thabane L, Torgerson D, Vohra S, White IR, Boutron I. CONSORT 2025 statement: updated guideline for reporting randomised trials. BMJ. 2025 Apr 14;389:e081123. doi: 10.1136/bmj-2024-081123. PMID 40228833
• Chan AW, Boutron I, Hopewell S, Moher D, Schulz KF, Collins GS, Tunn R, Aggarwal R, Berkwits M, Berlin JA, Bhandari N, Butcher NJ, Campbell MK, Chidebe RCW, Elbourne DR, Farmer AJ, Fergusson DA, Golub RM, Goodman SN, Hoffmann TC, Ioannidis JPA, Kahan BC, Knowles RL, Lamb SE, Lewis S, Loder E, Offringa M, Ravaud P, Richards DP, Rockhold FW, Schriger DL, Siegfried NL, Staniszewska S, Taylor RS, Thabane L, Torgerson DJ, Vohra S, White IR, Hrobjartsson A. SPIRIT 2025 statement: Updated guideline for protocols of randomised trials. PLoS Med. 2025 Apr 28;22(4):e1004589. doi: 10.1371/journal.pmed.1004589. eCollection 2025 Apr. PMID 40294521
• Perera S, Mody SH, Woodman RC, Studenski SA. Meaningful change and responsiveness in common physical performance measures in older adults. J Am Geriatr Soc. 2006 May;54(5):743-9. doi: 10.1111/j.1532-5415.2006.00701.x. PMID 16696738
• Kim S, Kim S, Kim H, Kim H, Yoo JI. Cross-sectional study comparing smart insoles and manual methods for short physical performance battery in hip fracture patients. Aging Clin Exp Res. 2025 Mar 1;37(1):61. doi: 10.1007/s40520-025-02960-6. PMID 40021544