Step-by-step guide to FHIR implementation with Microsoft Fabric for healthcare

Designed for the electronic exchange of health information, FHIR (Fast Healthcare Interoperability Resources) is a standard that promotes interoperability between various systems and applications using health data. It is widely adopted in the healthcare industry and supported by major cloud providers like Microsoft Azure.

Microsoft Fabric provides powerful tools and frameworks for developing and managing healthcare solutions utilizing FHIR standard. This FHIR implementation guide will walk you through the process of creating a healthcare solution with Microsoft Fabric, from setting up your environment to configuring and running data pipelines.

Steps Involved in Creating a Healthcare Solution

Obtaining a Microsoft Fabric License

To get started, ensure you have a Fabric or a trial license. This is essential for creating and exploring Microsoft Fabric items. 

Creating a Workspace

1. Open the Fabric Service by visiting Microsoft Fabric. 
2. Create a Workspace by selecting the Fabric trial license.

Configuring Runtime Version

Set the runtime version to Runtime 1.1, as other versions currently do not support FHIR solutions. Navigate to the following path to configure the runtime version: 

Workspace settings > Data Engineering/Science > Spark Settings > Environment > Runtime Version > 1.1 (Spark 3.3, Delta 2.2)

Selecting Industry Solutions

1. Select the Experience tab at the bottom left corner of the Fabric portal. 
2. From the available options, choose Industry Solutions. 
3. Microsoft currently offers three data solutions; select Healthcare data solutions.

Deploying Modules

Deploy all the different modules within the healthcare data solutions into your workspace. This process might take some time. Once completed, verify the deployment by viewing the files inside the workspace. 

Configuring Files for the Notebooks

Configuring Notebooks

• Open the file healthcare_msft_config_notebook, containing all the global configurations needed for the healthcare solution. 
• Set the kv_name to blank to import the sample data. 

• Close the file to autosave the changes. 
• Do not run this configuration file. 

Ingestion of Data from Raw to Gold

The data ingestion process involves moving data through various stages: Raw, Bronze, Silver, and Gold.

Ingestion from Raw to Bronze Layer

1. Open the notebook healthcare_msft_raw_bronze_ingestion. 
2. Run the file. 

Ingestion from Bronze to Silver Layer

1. Open the notebook healthcare_msft_bronze_silver_flatten. 
2. Run the file. 

Ingestion from Silver to Gold Layer

1. Open the notebook healthcare_msft_silver_omop. 
2. Run the file. 

Once these notebooks successfully run, the data will be ingested through all three medallion layers (Bronze, Silver, and Gold Lakehouse), transforming raw sample data in the Bronze Lakehouse to structured enriched data in the Gold Lakehouse.

Creating Pipeline with FHIR Sample Data

Setting up Pipeline in Workspace

• Navigate to the Data Engineering experience option from the bottom left corner. 

• Select the Data Pipeline option. 

Creating the Pipeline

• Open the pipeline canvas. 
• Under the Activities tab, select Notebook and create three different notebook activities, connecting them sequentially upon completion. 
• For each notebook, go to the Settings tab and select the workspace with the FHIR sample dataset. Assign the following notebooks respectively: 
  • healthcare_msft_raw_bronze_ingestion 
  • healthcare2_msft_bronze_silver_flatten 
  • healthcare2_msft_silver_omop
• Add a base parameter under the settings tab of all three notebooks, named _inlineInstallationEnabled, and set it to True. This enables running notebooks with the %pip magic command.

• Set up Office 365 Outlook to receive email notifications on the success or failure of the pipeline.

• Schedule the pipeline by clicking on the Schedule option and then applying the settings. 

• Validate the pipeline for any errors before running it. 
• Click on the Run option to execute the pipeline. 

By following these steps, you can effectively leverage Microsoft Fabric for healthcare to create a robust healthcare solution using FHIR standard, enabling efficient data management and interoperability in the healthcare sector.

Conclusion

Microsoft Fabric offers a streamlined approach to building healthcare solutions using the FHIR standard. This guide allows healthcare organizations to efficiently set up and configure their environments, deploy necessary modules, and create robust data pipelines. This ensures effective data management and interoperability, enhancing healthcare delivery and patient care.