Authors (including presenting author) :
See KH(1), Hui WH(1), Mak HL(1), Chan TY(1), Yeung KC(1), Tsui YC(1)
Affiliation :
(1)Physiotherapy department, Prince of Wales Hospital
Keyword 1: :
Artificial Intelligence
Keyword 2: :
Scoliosis Assessment
Keyword 3: :
Physiotherapy
Keyword 4: :
Documentation efficiency
Keyword 5: :
Data integration
Keyword 6: :
Clinical Management System
Introduction :
Rising demand for scoliosis physiotherapy services has increased pressure on clinic efficiency and documentation workload. Assessment sessions require detailed history taking, postural evaluation, patient education, and structured clinical recording. In routine practice, therapists reported duplicated documentation processes, including “double entry” of information into the Clinical Management System and Excel, documentation being “too rushed” while performing assessment, and the process being time-consuming. These workflow challenges reduced efficiency and increased administrative burden. A more streamlined documentation approach was therefore needed to better manage growing service demand.
Objectives :
To develop and implement an AI-assisted documentation tool for scoliosis physiotherapy assessment, reduce duplicated documentation and administrative time, and improve service efficiency through structured documentation and automated data capture.
Methodology :
A custom HTML-based documentation tool was developed using AI-assisted code generation. The interface used tick-boxes and structured input fields to capture key scoliosis assessment findings, automatically generate a formatted progress note for Clinical Management System entry, and export quantitative data for Excel-based analysis. Evaluation included 18 randomly selected scoliosis assessment sessions, comprising 10 sessions before implementation and 8 sessions after implementation. For each selected session, total assessment time, documentation time, and data input time were recorded. As patient attendance occasionally varied, all outcomes were normalized as minutes per patient by dividing total session time by the number of patients seen in that session. Group means were compared using an independent-samples t-test.
Result & Outcome :
The AI-assisted tool was associated with marked improvement in workflow efficiency. Mean total time decreased from 22.05 ± 1.13 minutes per patient before implementation to 12.44 ± 1.26 minutes per patient after implementation (p< 0.001), equivalent to approximately 111.5 minutes saved per 12-patient session. Assessment time per patient was similar before and after implementation (10.37 ± 1.00 vs 11.15 ± 1.16). Documentation time decreased from 6.39 ± 0.39 minutes per patient to 0.88 ± 0.07 minutes per patient, and data input time decreased from 5.29 ± 0.53 minutes per patient to 0.33 ± 0.10 minutes per patient. The tool also improved standardization and completeness of documentation, reduced duplicated entry, and lowered therapist time pressure. This AI-assisted documentation tool improved service efficiency by substantially reducing administrative time per patient and streamlining routine scoliosis clinic workflow. The approach is low-cost, practical, and potentially transferable to other allied health settings that require structured assessment, repeated documentation, and parallel service data capture. It demonstrates an innovative strategy for better managing growing demand while maintaining documentation quality and supporting more patient-facing care.
