ACT-PharmacyFirst-Protocol2-healthcare-usage

View on OpenSAFELY

Details of the purpose and any published outputs from this project can be found at the link above.

The contents of this repository MUST NOT be considered an accurate or valid representation of the study or its purpose. This repository may reflect an incomplete or incorrect analysis with no further ongoing work. The content has ONLY been made public to support the OpenSAFELY open science and transparency principles and to support the sharing of re-usable code for other subsequent users. No clinical, policy or safety conclusions must be drawn from the contents of this repository.

Pipeline Overview and Data Flow

Last update: 26 May 2026

Files prefixed with archive_... represent older or deprecated scripts that are currently not used in the active analysis pipeline, but are retained for reference and development history.

The repository is organised around several core components:

• patient-level dataset generation
• practice-level aggregation and summary (STP-level outputs)
• validation workflows, including
  • pregnancy variable checking and validation
  • several patient-level measures
  • snomed code occurrancence counting

These are reflected in the code blocks specified in the current workflow.

The diagram below illustrates the overall data flow and dependencies between steps.

graph TD
%% Pregnancy pipeline
A1[Pregnancy dataset Nov 2024] --> B[Generate pregnancy measures]
A2[Pregnancy dataset Nov 2025] --> B
B --> C[Pregnancy checks and charts]

%% Patient pipeline for aggregation
D1[Patient dataset deinition] --> E1
D1[Patient dataset deinition] --> E2
E1[Patient dataset Feb 2024] --> F1[Combine monthly datasets]
E2[Patient dataset Mar 2024] --> F1

%% Patient pipeline for measures
D2[Patient dataset definition for measures] --> E3[Generate patient measures] --> F3[Processed outputs]

%% Practice pipeline
D2[Patient dataset definition for measures] --> E4
E4[Generate practice measures] --> F2[Aggregate to practice-level dataset]
F2 --> G[STP-level summary output]

Loading

Core patient-level dataset definitions

• dataset_definition_patients: Main patient-level dataset definition used to generate monthly datasets for downstream analyses. Monthly datasets are generated separately for each study month.
• dataset_definition_patients_measures: Separate patient-level dataset definition used specifically for generating measures and validation outputs. This dataset is primarily used for measure generation, exploratory summaries and validation, and practice-level aggregation.

Monthly data generation and aggregation

Alongside the two main files,

• analysis/generate_project_action.py: Because monthly dataset generation cannot easily be parameterised within project.yaml, this script automatically creates the repetitive monthly actions in batch.
• analysis/preprocess_combine_gz.py: Utility script used to combine monthly compressed patient-level CSV outputs across the study period.

All monthly dataset generation steps rely on dataset_definition_patients.

Because these actions cannot easily be parameterised within project.yaml, separate actions must be created for each study month.

Steps:

1. In config.py, define the start and end dates.

2. Run the following command to generate the monthly actions:

   python analysis/generate_project_action.py > project_test.yaml

3. Copy the generated monthly actions from project_test.yaml into project.yaml.

4. Update the combine_monthly_patient_gz step:

   • Modify the needs field so that it depends on all generated monthly datasets
   • The current configuration includes only two months as an example
   • When using config.py, this should be updated to include all months in the specified range

Practice-level aggregation and summary

Folder: practice_variables

Measures are generated at patient level and grouped by practice, STP and region over monthly intervals.

The resulting outputs are then aggregated to:

• practice-level datasets;
• STP-level summaries;
• regional summaries.

These outputs support descriptive comparisons across practices and regions.

Pregnancy dataset and measures generation

Folder: check_pregnancy_variables

Please refer to this issue for implementation details.

Patient-level measures (for validation)

Folder: validation

This folder contains several validation workflows used to check:

• population structure;
• consultation counting;
• consultation mode classification;
• eligibility definitions.

Each validation workflow typically includes a measure definition file and a corresponding process_... Python script for summarising outputs. Some processing scripts also generate CSV summaries and plots for validation purposes.

The current project.yaml configuration mainly includes validation actions for October and/or November 2025.

SNOMED code counting (for validation)

Folder: validation_snomed

This folder contains workflows used to examine SNOMED coding patterns within GP consultations related to PF conditions. Details are noted in the README file.

About the OpenSAFELY framework

The OpenSAFELY framework is a Trusted Research Environment (TRE) for electronic health records research in the NHS, with a focus on public accountability and research quality.

Read more at OpenSAFELY.org.

Licences

As standard, research projects have a MIT license.