Automation – Way Forward to Achieve Data Integrity
- 1. A Global Laboratory Informatics Company Making Labs Proficient Data Integrity Conference - Sept 1, 2016 Mumbai Presented by : Mukunth Venkatesan (CEO, Agaram Technologies) Automation – Way Forward to Achieve Data Integrity
- 2. 2 Why is Data Integrity Important? “Data Integrity is considered as the first and foremost requirement in a pharmaceutical quality system to ensure that the medicines are of the required quality”
- 3. 3 What is Data Integrity Data Integrity is defined as the “the completeness, consistency, and accuracy of data. Complete, consistent, and accurate data should be Attributable, Legible, Contemporaneously recorded, Original or a true copy, and Accurate.” This is also known as the ALCOA principles guiding Data Integrity.
- 4. 4 Contemporaneous Recording Date & Time Accurate Data with no errors or editing Original Justifying data is true copy Legible Permanent Recording Consistent Consistent application of date & time stamps in the expected sequence Enduring Recorded in enduring media ORIGINAL 75+ FTEs supporting customers globally Available Available , accessible for review/audit for lifetime of record Complete All data including repeat or re-analysis Consistent Consistent application of date & time stamps in the expected sequence Attributable Source of Data & Who ALCOA De-mystified
- 5. 5 ALCOA – How to achieve? Attributable Source of data identified with a process and person Legible Permanently recorded. Human Readable. Contemporaneous Data Identified with a date and time. Original or True Copy Data certified as correct by an authorized person Accurate Data should not have any errors Data stored electronically should achieve this Built in Time Stamp at the time of creation or modification of data Electronic Signature should satisfy this need Data captured directly without human intervention should ensure this Built in Audit Trail identifies the data to a source, person & process
- 6. 6 Data Generation & Recording Processing Method Accuracy Result Data Application Data System Authenticated, Secure & Protected Central Storage Automation Automation Architecture for Data Integrity Assurance
- 7. 7 • An environment for controlled access to any application • Automatic pushing of data and meta data generated or modified to server- without user intervention • Users can still continue to modify or update data • Modified data is pushed to server automatically • Full audit trail of all activities to be available at server • Review & approval based on server data (true copy) • Result or decision should be taken from the server Data Integrity- Functional Requirements
- 8. 8 Risk Mitigation- Data Generation & Recording • How and where is original data created? • Created in the local hdd. With a copy in the server (controlled area) • How do you ensure that the data is complete, accurate and traceable to meet ALCOA? • Automation ensures that (raw, meta, human readable) are all moved to server • Is it possible to recreate, amend or delete original data and metadata? • Automation should help in identifying amendments and NO possibility to delete or obscure data • How data is transferred to other locations or systems for processing or storage • Automation can help in transfer of data in a controlled manner for processing or storage. Any change due to processing needs to be handled by automation solution.
- 9. 9 Risk Mitigation- Data Accessing & Processing • How is data processed? • Method used for processing to be identified as metadata for capture • Where no external metadata is available for processing, data should contain relevant metadata or manually record conditions under which data was created or modified • How is data processing recorded? • Any change in data due to processing should always be captured by the automation • Does the person processing the data have the ability to influence what data is reported, or how it is presented? • Make automation server as the primary source of data (to prevent influence) • Even if a person does trials all trials should be captured as independent versions
- 10. 10 Risk Mitigation- Completeness, Accuracy of reported data • Is original data (including the original data format) available for checking? • Always original data is available in original format at server • Accuracy or reported data can be cross checked based on original data for data integrity at any point and time • Does the data reviewer have visibility and access to all data generated & processed? • Reviewer can check all data generated in one go and from a single point (server) without having to look at individual systems. • Cross-checking can be done by download and re-creation of output using the original application
- 11. Data Integrity – Solution
- 12. 12 AUTOMATION – Way Forward to Achieve Data Integrity By Implementing Scientific Data Management System (SDMS), Electronic Lab Notebook (ELN) & Document Management System (DMS)
- 13. 13 •Solution for instrument generated data b •Control over access & audit trail c •Control over generated & processed data d •Automatic data version control e •Electronic Review & Approval Scientific Data Management System (SDMS)
- 14. 14 a •Template Based approach b •Deploy in QC & Production c •Integrate via RS232 or TCP/IP d •Data Versioning e •Audit trail & Electronic Review Electronic Lab Notebook (ELN)
- 15. 15 Electronic Document Management & Issuance control a • Manage organisation wide • Document preparation • Review, Approval & Release c • Document Request & Issuance d • Auto filling of Tags (batch#) e • PDF Document with Electronic signature f • Print Control
- 16. 16 Agaram Technologies 76 Nelson Road, Aminjikarai Chennai 600 029 info@agaramtech.com www.agaramtech.com +91-44-4208-2005 Making Labs Proficient Get in Touch
Editor's Notes
- #7: Users are provided controlled access to application Automatic pushing of data and meta data to server No user intervention required Users can still continue to modify or update data (results or method or reports) Modified data is pushed to server automatically Accuracy, review and approval should happen based on server data Result or decision data should be taken from the server