NHP Synthetic Data Project: Scoping & Presentation Document

Project Overview

This repository contains the ongoing work for an internship project focused on exploring and developing synthetic data generation techniques for the New Hospital Programme (NHP) demand prediction model. The primary goal is to address the challenge of data accessibility and privacy, enabling the NHP's open-source model to be truly reproducible, shareable, and evaluable without compromising sensitive real patient data.

The project's key findings and methodological insights are documented through Quarto, utilising the NHS Strategy Unit (SU) Presentation Theme for a structured and visually consistent presentation of the work.

Problem Statement

The NHP demand prediction model is a tool for strategic healthcare planning. While its source code aims to be open, its usability for external researchers, collaborators, and even internal testing is limited by the inability to share the sensitive real patient data. Existing artificial data solutions provided by NHS England currently lack the necessary statistical relationships and complexity to accurately represent real world patient cohorts and their interactions. As a result, it is difficult to meaningfully evaluate the model or perform a thorough analysis.

Objectives

This project aims to:

1. Evaluate existing synthetic/artificial data approaches, specifically assessing their ability to preserve complex multivariate relationships that are important for healthcare demand modelling.
2. Enhance the existing artificial HES dataset by creating relationships, so it can be used on the demand model to effectively assign mitigators.
3. Produce a comprehensive scoping document/presentation detailing the findings, comparing different approaches, outlining their pros and cons, and providing clear recommendations for future full scale implementation.
4. Enable true open-source collaboration for the NHP model by facilitating reproducible testing and analysis without privacy concerns.

Repository Structure

.
├── notebooks/                                  # Folder containing Quarto files
├── scripts/                                    # Reusable Python scripts
├── _extensions/                                # Quarto theme
├── data/                                       # Placeholder for data files
├── .gitignore                                  # Git ignore rules
├── _quarto.yml                                 # Main Quarto project configuration
├── README.md                                   # This overview document
└── synthetic-data-rough-note.qmd               # The primary Quarto document/report

Getting Started

To explore this project locally:

1. Clone the repository:

   git clone https://github.com/AshAninze/synthetic-data-rough-note.git
   cd synthetic-data-rough-note

2. Install Quarto: Ensure Quarto is installed by following the instructions on the official Quarto website.
3. Render the scoping document:

   quarto render synthetic-data-rough-note.qmd

   This will generate the final output, an HTML file.