PhysioRehab

ICF-native, single-case clinical-reasoning tools for rehabilitation.

Rehabilitation reasons differently from disease medicine: it targets disability and life reconstruction, evaluates the individual rather than a randomised group, combines qualitative judgement with quantitative signals, and is delivered by allied-health professionals (PT / OT / ST) with varied statistical backgrounds. PhysioRehab re-centres physiological measurement around that way of reasoning: it anchors measures and goals to the International Classification of Functioning (ICF), detects change with single-case methods, keeps the clinician in the loop, and drafts an ICF-structured progress note.

Installation

# the containers build on Bioconductor, so its repositories are needed too
install.packages("BiocManager", repos = "https://cloud.r-project.org")
install.packages("PhysioRehab",
  repos = c("https://x-biosignal.r-universe.dev", BiocManager::repositories()))

Quick start

The core is a single guided workflow:

measure the individual over time against personal goals
  -> detect change defensibly with single-case methods (NAP + MCID + 2-SD band)
  -> visualise against ICF and the trajectory
  -> auto-draft an ICF-structured progress note

library(PhysioRehab)

ep  <- simulate_stroke_gait_case()          # a synthetic single case
hyp <- rehab_hypothesis("task-oriented training improves gait",
                        expect_speed_mcid = TRUE, expect_icf = c(d450 = 1))
res <- rehab_workflow(ep, hyp)              # analyse -> reason -> plot -> draft

cat(res$report)                             # the ICF-structured progress note
res$plot_file                               # the single-case trajectory plot

The source repository bundles a fuller end-to-end demo script, run from a checkout of the package root:

Rscript inst/examples/run_demo.R

It simulates a chronic-stroke, left-hemiplegia single case (AB design, 12 sessions), runs the workflow, and writes two clinician-facing artefacts — an auto-drafted, ICF-structured progress note and a single-case trajectory plot — both to a temporary directory.

Design

Component	Description	Code
ICF-native semantics	Measures, goals and observations are anchored to WHO ICF codes with 0–4 qualifiers — the domain data model of rehabilitation.	R/icf.R, R/goals.R
Quantitative–qualitative fusion with an honest reasoning scaffold	Single-case change detection fuses signals with ICF qualifiers; a hypothesis-to-evidence check keeps the clinician in the loop and explicitly reports what the evidence does not support.	R/sced.R, R/reasoning.R
Guided workflow and clinician GUI	One call, rehab_workflow(), runs the whole path; a Shiny interface, rehab_app(), exposes it to non-programmers, while experts call the components directly.	R/workflow.R, R/app.R

The clinician-facing output (R/report.R) drafts the progress note the therapist already has to write, so measurement and analysis accompany routine documentation.

Ecosystem integration

PhysioRehab builds on the x-biosignal ecosystem rather than reimplementing it:

• Single-case statistics — PhysioAppKit (NAP effect size, threshold-and-band analysis; reconciled with PhysioClinStats::scedNAP).
• ICF ontology — PhysioAnnotationHub (icf_for_measure() resolves a measure to ICF codes from the maintained ontology).
• Real signals — PhysioMoCap (markerless gait via add_session_from_mocap()) and PhysioEMG (muscle fatigue via add_session_from_emg()).
• Ordinal-to-interval measurement — Rasch calibration via PhysioAppKit.
• Reporting — PhysioReport.

Scope and limitations

• Early-stage (0.1.0). The bundled examples use synthetic data; the worked instance is stroke hemiplegic gait, and the framework is condition-agnostic.
• PhysioRehab is a research tool for measurement, documentation, and clinician-in-the-loop reasoning support. It is not a medical device and is not intended for diagnosis; final clinical judgement rests with the clinician.

Citation

See CITATION.cff. Licensed under the MIT License.