Skip to content

resplab/acceptpexa

BranchesTags

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

26 Commits
R
R
 
 
man
man
 
 
.Rbuildignore
.Rbuildignore
 
 
.gitignore
.gitignore
 
 
DESCRIPTION
DESCRIPTION
 
 
NAMESPACE
NAMESPACE
 
 
README.md
README.md
 
 
acceptpexa.Rproj
acceptpexa.Rproj
 
 

Repository files navigation

acceptpexa

A ModelsCloud wrapper for the ACCEPT package — ACute COPD Exacerbation Prediction Tool.

Overview

This package wraps the resplab/accept package for deployment on the ModelsCloud platform. It supports all versions of ACCEPT including ACCEPT 3.0-CPRD, a UK primary-care specific recalibration derived from the Clinical Practice Research Datalink (CPRD).

Design philosophy

model_run() takes a single model_input argument — a named list containing:

  • patients_data — the array of patient records
  • version — which ACCEPT model to run
  • country — country code for ACCEPT 3.0

This follows the ModelsCloud uniform API contract where every model on the platform looks identical to the client.

Installation

Direct R usage

remotes::install_github("resplab/acceptpexa")

Via ModelsCloud

remotes::install_github("resplab/modelscloud")

Quick Start

Direct R usage (without ModelsCloud)

remotes::install_github("resplab/acceptpexa")
library(acceptpexa)

# Default — accept2, single patient
model_run()

# UK primary care
model_run(list(
  patients_data = accept::samplePatients,
  version = "accept3",
  country = "GBR-primary"
))

# Canada
model_run(list(
  patients_data = accept::samplePatients,
  version = "accept3",
  country = "CAN"
))

Via ModelsCloud

remotes::install_github("resplab/modelscloud")
library(modelscloud)
connect_to_model("resplab/accept", access_key = "YOUR_KEY")

# Default — accept2
model_run(get_default_input())

# UK primary care
model_run(list(
  patients_data = accept::samplePatients,
  version = "accept3",
  country = "GBR-primary"
))

Examples by Version

Note: ID can be any string — "P002", "10001", "patient_1", "ABC123" — it is just a label to identify the patient in the output. The model does not use it for any calculation; it simply echoes it back in the results so you can match predictions to patients.

ACCEPT 1.0

Web interface (Lite tab):

{
  "model_input": {
    "patients_data": [
      {
        "ID": "10001", "male": true, "age": 70, "smoker": true,
        "oxygen": true, "statin": true, "LAMA": true, "LABA": true,
        "ICS": true, "FEV1": 33, "BMI": 25, "SGRQ": 50,
        "LastYrExacCount": 2, "LastYrSevExacCount": 1
      }
    ],
    "version": "accept1"
  }
}

R (modelscloud):

library(modelscloud)
connect_to_model("resplab/accept", access_key = "YOUR_KEY")

model_run(list(
  patients_data = data.frame(
    ID = "10001", male = TRUE, age = 70, smoker = TRUE, oxygen = TRUE,
    statin = TRUE, LAMA = TRUE, LABA = TRUE, ICS = TRUE, FEV1 = 33,
    BMI = 25, SGRQ = 50, LastYrExacCount = 2, LastYrSevExacCount = 1
  ),
  version = "accept1"
))

ACCEPT 2.0

Web interface (Lite tab):

{
  "model_input": {
    "patients_data": [
      {
        "ID": "10001", "male": true, "age": 70, "smoker": true,
        "oxygen": true, "statin": true, "LAMA": true, "LABA": true,
        "ICS": true, "FEV1": 33, "BMI": 25, "SGRQ": 50,
        "LastYrExacCount": 2, "LastYrSevExacCount": 1
      }
    ],
    "version": "accept2"
  }
}

Note: "country": null also works for accept2 — country is not used.

R (modelscloud):

model_run(list(
  patients_data = data.frame(
    ID = "10001", male = TRUE, age = 70, smoker = TRUE, oxygen = TRUE,
    statin = TRUE, LAMA = TRUE, LABA = TRUE, ICS = TRUE, FEV1 = 33,
    BMI = 25, SGRQ = 50, LastYrExacCount = 2, LastYrSevExacCount = 1
  ),
  version = "accept2"
))

ACCEPT 3.0 — Country-specific (e.g. Germany)

Web interface (Lite tab):

{
  "model_input": {
    "patients_data": [
      {
        "ID": "10001", "male": true, "age": 70, "smoker": true,
        "oxygen": false, "statin": true, "LAMA": true, "LABA": true,
        "ICS": false, "FEV1": 33, "BMI": 25, "SGRQ": 50,
        "LastYrExacCount": 2, "LastYrSevExacCount": 1
      }
    ],
    "version": "accept3",
    "country": "DEU"
  }
}

R (modelscloud):

model_run(list(
  patients_data = data.frame(
    ID = "10001", male = TRUE, age = 70, smoker = TRUE, oxygen = FALSE,
    statin = TRUE, LAMA = TRUE, LABA = TRUE, ICS = FALSE, FEV1 = 33,
    BMI = 25, SGRQ = 50, LastYrExacCount = 2, LastYrSevExacCount = 1
  ),
  version = "accept3",
  country = "DEU"
))

ACCEPT 3.0-CPRD — UK Primary Care (two patients)

Web interface (Lite tab):

{
  "model_input": {
    "patients_data": [
      {
        "ID": "10001", "male": true, "age": 70, "smoker": true,
        "oxygen": true, "LAMA": true, "ICS": true, "FEV1": 33,
        "SGRQ": 50, "LastYrExacCount": 2, "LastYrSevExacCount": 1
      },
      {
        "ID": "10002", "male": false, "age": 42, "smoker": false,
        "oxygen": true, "LAMA": true, "ICS": false, "FEV1": 40,
        "SGRQ": 40, "LastYrExacCount": 0, "LastYrSevExacCount": 0
      }
    ],
    "version": "accept3",
    "country": "GBR-primary"
  }
}

R (modelscloud):

model_run(list(
  patients_data = data.frame(
    ID    = c("10001", "10002"),
    male  = c(TRUE, FALSE),
    age   = c(70, 42),
    smoker = c(TRUE, FALSE),
    oxygen = c(TRUE, TRUE),
    LAMA  = c(TRUE, TRUE),
    ICS   = c(TRUE, FALSE),
    FEV1  = c(33, 40),
    SGRQ  = c(50, 40),
    LastYrExacCount    = c(2, 0),
    LastYrSevExacCount = c(1, 0)
  ),
  version = "accept3",
  country = "GBR-primary"
))

ACCEPT 3.0-CPRD — Missing optional predictors (auto-imputed)

Missing statin, LABA, BMI are automatically imputed using the CPRD model.

Web interface (Lite tab):

{
  "model_input": {
    "patients_data": [
      {
        "ID": "10001", "male": true, "age": 70, "smoker": true,
        "oxygen": true, "LAMA": true, "ICS": true, "FEV1": 33,
        "SGRQ": 50, "LastYrExacCount": 2, "LastYrSevExacCount": 1
      }
    ],
    "version": "accept3",
    "country": "GBR-primary"
  }
}

R (modelscloud):

model_run(list(
  patients_data = data.frame(
    ID = "10001", male = TRUE, age = 70, smoker = TRUE, oxygen = TRUE,
    LAMA = TRUE, ICS = TRUE, FEV1 = 33, SGRQ = 50,
    LastYrExacCount = 2, LastYrSevExacCount = 1
  ),
  version = "accept3",
  country = "GBR-primary"
))

Custom patient

Web interface (Lite tab):

{
  "model_input": {
    "patients_data": [
      {
        "ID": "P002", "male": true, "age": 80, "smoker": true,
        "oxygen": true, "statin": true, "LAMA": true, "LABA": true,
        "ICS": true, "FEV1": 20, "BMI": 18, "SGRQ": 80,
        "LastYrExacCount": 5, "LastYrSevExacCount": 3
      }
    ],
    "version": "accept2"
  }
}

R (modelscloud):

model_run(list(
  patients_data = data.frame(
    ID = "P002", male = TRUE, age = 80, smoker = TRUE, oxygen = TRUE,
    statin = TRUE, LAMA = TRUE, LABA = TRUE, ICS = TRUE, FEV1 = 20,
    BMI = 18, SGRQ = 80, LastYrExacCount = 5, LastYrSevExacCount = 3
  ),
  version = "accept2"
))

How it works

model_run() receives a single named list (model_input) structured as:

list(
  patients_data = <data frame or list of patient records>,
  version       = "accept3",
  country       = "GBR-primary"
)

It extracts version and country, passes patients_data to accept::accept().

Versions

Version version country needed? Missing optionals imputed?
ACCEPT 1.0 "accept1" No No
ACCEPT 2.0 "accept2" No No
ACCEPT 3.0-CPRD "accept3" Yes — "GBR-primary" Yes
ACCEPT 3.0 UK specialty "accept3" Yes — "GBR-specialty" No
ACCEPT 3.0 other countries "accept3" Yes — e.g. "CAN" No

Supported countries for ACCEPT 3.0

ARG, AUS, BRA, CAN, COL, DEU, DNK, ESP, FRA, ITA, JPN, KOR, MEX, NLD, NOR, SWE, USA

For UK: use "GBR-primary" (primary care) or "GBR-specialty" (specialty care).

Required Identifier

Variable Description
ID Unique patient identifier — required to label output rows, not a predictor

Mandatory Predictors

Variable Description
age Age in years
male TRUE/FALSE
FEV1 FEV1 % predicted (10–120)
LastYrExacCount Total exacerbations last year
LastYrSevExacCount Severe exacerbations last year
mMRC or SGRQ Symptom score — at least one required. SGRQ is converted to mMRC internally if mMRC is missing

Optional Predictors

LABA, oxygen, ICS, LAMA, statin, BMI, smoker

These are optional when using version = "accept3" with country = "GBR-primary" — missing values are automatically imputed using a UK-specific sequential regression model derived from CPRD data. For all other versions, all predictors should be provided.

Functions

Function Description
model_run(model_input) Run ACCEPT predictions
get_sample_input(n) Get sample input list with patients_data, version and country
get_default_input() Get default input list with one patient, version and country
echo(...) Echo input back — for testing API connectivity

References

If you use this package, please cite the relevant ACCEPT paper(s):

ACCEPT 1.0

Adibi A, Sin DD, Safari A, Johnson KM, Aaron SD, FitzGerald JM, Sadatsafavi M. The Acute COPD Exacerbation Prediction Tool (ACCEPT): a modelling study. The Lancet Respiratory Medicine. 2020; 8(10): 1013–1021. https://doi.org/10.1016/S2213-2600(19)30397-2

ACCEPT 2.0

Safari A, Adibi A, Sin DD, Lee TY, Ho JK, Sadatsafavi M and IMPACT study team. ACCEPT 2.0: Recalibrating and externally validating the Acute COPD Exacerbation Prediction Tool (ACCEPT). EClinicalMedicine. 2022; 51: 101574. https://doi.org/10.1016/j.eclinm.2022.101574

ACCEPT 3.0-CPRD (UK primary care)

Mehareen J, Lim LHM, Adibi A, Amegadzie JE, Xia Y, De Vera MA, Law MR, Sin DD, Bhatt SP, Quint JK, Sadatsafavi M. Performance of multivariable risk prediction algorithms in predicting COPD exacerbations: a population-based study. Thorax. 2026. https://doi.org/10.1136/thorax-2026-224814

Related

About

ModelsCloud wrapper for the ACCEPT

Resources

Readme
Activity
Custom properties

Stars

0 stars

Watchers

0 watching

Forks

0 forks
Report repository

Releases

No releases published

Packages

 
 
 

Contributors

Languages