CLiTICAL for iOS

日本語版: README.ja.md

An iOS app that predicts post-operative risks of revascularisation for chronic limb-threatening ischaemia (CLTI) from patient data.

It implements the prediction models by Miyata et al., built on the JCLIMB registry of the Japanese Society for Vascular Surgery (JSVS), and is a native SwiftUI rewrite of the Flutter/Android version of CLiTICAL.

Disclaimer This app is intended to support the clinical judgement of healthcare professionals. The values it calculates are estimates from statistical models and do not determine diagnosis or treatment. The final decision remains the responsibility of the attending physician.

Features

• Patient data entry — six sections: basic information, social history, clinical information, arterial lesion sites, other vascular lesions, and comorbidities
• Risk prediction — five indices calculated from the entered data
• References — the source papers opened in an SFSafariViewController
• Settings — Japanese/English switching applied instantly in-app, the Terms of Use, Privacy Policy and Support pages, and app information
• Accessibility — VoiceOver support, Dynamic Type, and risk levels never conveyed by colour alone

Predicted indices

Index	Description
30-day death or major amputation	Risk of death and/or major amputation within 30 days
30-day MALE	Major adverse limb events within 30 days
2-year OS	Two-year overall survival, with a low/medium/high risk classification
2-year AFS	Two-year amputation-free survival
GNRI	Geriatric Nutritional Risk Index, with a four-level classification

GNRI is calculated as 14.89 × Alb + 41.7 × min(weight / (22 × height²), 1.0) and classified as no malnutrition risk (≥98), mild (92–98), moderate (82–92), or major (<82). The 30-day risks use logistic regression; 2-year OS/AFS use Cox proportional hazards models (baseline survival 0.922 for OS and 0.876 for AFS).

At least one arterial lesion site must be selected — AI (aortoiliac), FP (femoropopliteal), or BK (below the knee) — and the most proximal selected site determines the lesion classification.

Entry validation

Age, sex, height, weight and albumin are all required, and sex has no default. The numeric entries are range checked so that a unit mix-up — a height typed in metres, albumin in g/L — is refused rather than turned into a plausible looking result: age 18–120 years, height 100–250 cm, weight 20–300 kg, albumin 1.0–6.0 g/dL. An out-of-range entry shows an alert quoting the accepted range and no prediction is made.

Requirements

• iOS 16.0 or later (iPhone and iPad)
• Xcode 16 or later / Swift 5 language mode
• No third-party dependencies (the only networking is opening the reference papers and the legal pages)

Project layout

clitical-ios/
├── clitical/                     # App target (SwiftUI)
│   ├── CliticalApp.swift         # Entry point
│   ├── MainTabView.swift         # Tabs, references, settings and about
│   ├── ContentView.swift         # Patient data form
│   ├── PredictedRiskView.swift   # Results screen
│   ├── ChoiceListView.swift      # ToggleRow / SegmentedRow / MenuChoiceRow
│   ├── AgeFormView.swift, …      # Numeric fields (age, height, weight, Alb)
│   ├── Labels.swift              # Domain enums → localization keys
│   ├── LocalizationManager.swift # In-app language switching
│   ├── QuestionError.swift       # Input error definitions
│   └── CLPatientData/            # Local Swift package (domain logic)
│       ├── Sources/CLPatientData/
│       │   ├── PatientData.swift # Patient data model (value type)
│       │   ├── PatientRisk.swift # Risk calculation and classification
│       │   └── Questions.swift   # Predictors and regression coefficients
│       └── Tests/CLPatientDataTests/
├── cliticalUITests/              # XCUITest
├── ja.lproj / en.lproj           # Localizable.strings
└── clitical-ios.xcodeproj

Architecture

• Domain logic kept separate — risk calculation lives in CLPatientData, a local Swift package independent of the UI, so it is testable without SwiftUI. The regression coefficients are collected in Questions.swift as one enum case per predictor.
• Patient data as a value type — PatientData is a struct passed to the forms through @State / @Binding, and is never persisted.
• In-app language switching — LocalizationManager swaps the class of Bundle.main so that the selected .lproj wins, letting Text("key") re-resolve without a restart. Only navigation titles, which UIKit caches, are resolved explicitly via localization.string(forKey:).
• HIG-conformant UI — settings-like items are grouped in a single Settings tab rather than holding tabs of their own, and choices use inline Picker / Toggle controls instead of pushed screens.

Build and run

Open clitical-ios.xcodeproj in Xcode and run the clitical-ios scheme.

From the command line:

xcodebuild -project clitical-ios.xcodeproj -scheme clitical-ios -destination 'platform=iOS Simulator,name=iPhone 16' build

Tests

Development follows red/green TDD. The domain unit tests are written with Swift Testing.

Package tests only:

swift test --package-path clitical/CLPatientData

All tests including the UI tests:

xcodebuild -project clitical-ios.xcodeproj -scheme clitical-ios -destination 'platform=iOS Simulator,name=iPhone 16' test

Localization

Japanese (default) and English are supported. Strings live in ja.lproj/Localizable.strings and en.lproj/Localizable.strings, and the key sets are kept identical between the two. The selected language is stored in UserDefaults under app_language; on first launch the app follows the device language.

Versioning

MARKETING_VERSION (display version) and CURRENT_PROJECT_VERSION (build number) are managed in the Xcode project's build settings. The display version is shown under Settings > About.

References

1. Miyata T. et al, Risk prediction model for early outcomes of revascularization for chronic limb-threatening ischaemia. Br J Surg. 2022 Oct 14;109(11):1123. https://doi.org/10.1093/bjs/znab036
2. Miyata T. et al, Prediction Models for Two Year Overall Survival and Amputation Free Survival After Revascularisation for Chronic Limb Threatening Ischaemia. Eur J Vasc Endovasc Surg. 2022 Jun 7;S1078-5884(22)00340-9. https://doi.org/10.1016/j.ejvs.2022.05.038

Terms of use

The legal documents are hosted outside the app and opened in an SFSafariViewController from Settings, in whichever language is selected there:

• Terms of Use — Japanese / English
• Privacy Policy — Japanese / English
• Support — Japanese / English

The URLs are built by AppInfo.legalURL(for:language:) in MainTabView.swift.

On first launch an intended-use notice is shown in place of the app, with the Terms of Use reachable from it; tapping "I understand" records agreement to both. The acknowledgement is stored in UserDefaults under intended_use_disclaimer_version as the version of the notice that was accepted (IntendedUseDisclaimer.currentVersion), so a material change to the wording asks again.

Privacy

Patient data is processed on this device only and is never sent or stored elsewhere. The app saves the selected language and the acknowledged version of the intended-use notice, and nothing else.

License

Released under the MIT License.

The regression coefficients and classification thresholds come from the papers listed under References; the MIT License applies to the software implementation only.

Credits

• Published by: Japanese Society for Vascular Surgery / JCLIMB Committee (2022)
• Developed by: Kazuhiro Miyahara
• Terms of Use: https://studiome.github.io/clitical-legal/terms/ja/