INSPIRE PEACH: USING OHDSI FOR HEALTH INFORMATICS IN MALAWI

Developing an ETL for the Malawi Integrated Disease Surveillance for COVID-19 in Malawi

Amelia Taylor, Gonjetso Chinyama, Moses Gwaza�Inspire PEACH project�Malawi University of Business and Applied Sciences, KuyeseraAI Lab

1

Find this presentation at: https://shorturl.at/jxEIJ

PROJECT AIM

To apply the OHDSI technologies to COVID-19 data in Malawi and Kenya and build an integrated ecosystem for sharing and analysing the data.

2

Find this presentation at: https://shorturl.at/jxEIJ

What does OHDSI offer?

• A Common Data model and methodology to standardise medical data

​

• Open source tools for analysing and visualising observational health data

​

• Possibility to conduct studies of patient characterisation and prediction across multiple datasets, hospitals and countries.

​

• A community of practitioners/developers

3

INSPIRE PEACH

• Project site: https://inspiredata.network
• MUBAS is a partner with others: APHRC, LSHTM, CODAT, SAPRIN

​

4

INSPIRE PEACH cont’d

• Setup stage involved using IDSR (Integrated Disease Surveillance and Response) data for COVID-19

​

• Initially we did not have the data. We generated synthetic data to which we applied individual level fidelity metrics.

​

• We then developed an ETL pipeline for standardising this data.

5

Synthetic Malawi IDSR Data

The synthetic data was generated for the IDSR Case Based form for collecting C19 data used in Malawi.

​

https://github.com/Inspire-Mubas/Malawi-IDSR-COVID-19-Synthetic-DataSet

6

Synthetic Malawi IDSR Data cont’d

7

Developing an ETL Pipeline

Transforming the Malawi IDSR data into a data standard called the OMOP Common Data Model.

​

8

Developing an ETL Pipeline cont’d

This required:

Concepts and terms standardisation using medical vocabularies

Transforming data to the OMOP Common Data Model

Loading data to a database for analysis

All processes were developed with scalability and replicability principles in mind.

9

The OMOP CDM

10

Concepts Standardization

11

Pentaho: ETL Tool

• Open-source Business Intelligence (BI) suite
• Provides a variety of tools for data integration, data analysis, and reporting
• Pentaho is an ETL tool (Extract, Transform, and Load).

12

Applying Pentaho to OMOP

13

The Demographics group

​

​

14

Lessons learned

Concept mapping is an essential process in healthcare data standardization because it helps to establish a common vocabulary and understanding of medical concepts across different healthcare systems and providers.

​

Challenge - identifying the right concepts

​

15

Lessons learned cont’d

The usage of Pentaho for performing the ETL is a bit flexible and offers visual drag and drop capabilities for building complex pipelines unlike just using like SQL only

​

16

Lessons learned cont’d

From using Pentaho for the ETLs

​

Overall, Pentaho can help healthcare organizations to standardize their data, improve data quality, promote interoperability, and comply with regulatory requirements.

​

This, in turn, can lead to better decision-making, improved patient outcomes, and cost savings for healthcare organizations.

​

​

17

Lessons learned cont’d

​

Standardization in healthcare data is important because:

• It ensures consistency, accuracy, and reliability of healthcare-related data across different healthcare systems and providers.

​

• End result:
• healthcare professionals make reliable informed decisions
• Also helps save costs to process and analyze data

18

Further Work

• The ETL pipeline is extended to real C19 data in Malawi and Kenya.

​

• Deploying a containerised OHDI infrastructure.

​

• Working on analysis using OHDSI tools.

19