White Paper www.vaisala.com

Managing GxP environmental systems to ensure data integrity

In this paper, we provide an overview of data management best practices for life science systems and an overview of regulatory expectations. We offer eight recommendations for establishing and maintaining good practices for data integrity.

More than Bytes management, companies can find Thanks to the publication of and Signatures changes hard to implement in terms enforcement actions such as GMP of updating systems and behavior. non-compliance reports, warning As efforts to ensure the quality and letters, import alerts, and notices, safety of drugs increase, so does it is evident that regulators are the amount of data generated by Data integrity: targeting data integrity failures those efforts. Over the last few Related regulations during inspections. Subsequent years, global regulatory scrutiny has turned to providing guidance Data integrity requirements are enforcement actions have led to on preserving the integrity of data. core elements of basic GMPs, and the withdrawal of supply across Throughout life science industries are broadly addressed in the FDA’s multiple markets, product recalls, — pharmaceutical, medical devices, Title 21 CFR Part 11 and the EU's consent decrees, and reputational and biotechnology research and GMP Eudralex Volume 4, Chapter damage for the firms involved. production — regulatory guidance 4 and Annex 11. However, with With increased targeting of and enforcement strategies are data integrity from regulators, being re-evaluated with a focus increasing automation based on it is now crucial that everyone on data integrity. With increasing computerized systems, as well as involved in GxP-regulated awareness of data collection and the globalization of operations storage, there comes increased and the increasing cost of bringing activities understand correct data awareness of gaps between products to market, new guidance management practices. industry practice and existing was needed to clarify regulatory technology. Although there are expectations around the creation, new strategies available for data handling, and storage of data. Principles and Practice The acronym ALCOA is used by the FDA, MHRA, the World Health Organization and others to outline expectations on records, including In essence, data integrity means paper-based, electronic, and hybrid records (systems that use both paper that data collected and stored and electronic records). ALCOA is a useful guide to remembering key must be original, complete and points of data management for GxP compliance. ALCOA stands for: traceable. Several regulatory agencies have defined the term “data integrity.” In the UK, the Medicines and Healthcare A = Attributable to the person generating the data products Regulatory Agency = Legible and permanent (MHRA) defined data integrity L in their 2015 document: “MHRA C = Contemporaneously recorded GMP Data Integrity Definitions O = Original or a true copy and Guidance for Industry” as the degree to which all collected A = Accurate data are “complete, consistent, accurate, trustworthy, reliable… throughout the data lifecycle.” The WHO added some extra definitions to ALCOA in their working document “Guideline on data integrity” expanding the acronym to For their 2016 draft guidance ALCOA+. In addition to original emphasis of ALCOA principles, the “+” for industry “Data Integrity and highlights the importance of the attributes of being complete, consistent, Compliance with CGMP” the enduring and available.* FDA defines data integrity as: “… the completeness, consistency, Thus, ALCOA+ is now the goal for every piece of GMP information – and accuracy of data. Complete, information that can impact the purity, efficacy, and safety of products. consistent, and accurate data ALCOA+ is the standard by which data integrity is evaluated. In practice, should be attributable, legible, this means that companies must maintain control over all intentional and contemporaneously recorded, unintentional changes to GMP data, including the prevention of data loss original or a true copy, and or corruption. accurate (ALCOA).”

Data Management A review of enforcement actions ▪ Electronic data could have been Challenges proves that many companies manipulated or deleted without are misinterpreting guidance traceability. Regardless of the methods used to documents. Other industry gather and store data — manual, stakeholders try to help with ▪ Raw data were copied to a CD automatic, or a combination — more explicative documents. For and then deleted from the hard there are always opportunities for example, the European Compliance drive. Data copied were selected failure. Manual processes entail Academy (ECA) published an manually without assurance that obvious points of possible failure: article specifying data integrity all raw data was copied before operators can forget to record failures that caused one company being permanently deleted. information, record incorrect to receive an FDA Warning Letter. values, lose records, or even Observations included: Each of these deviations could intentionally falsify data. The risks have been addressed by systems Failure to exercise sufficient associated with computerized ▪ and methods including: controls over computerized systems are more technical. For systems to prevent unauthorized Unique usernames and passwords both manual and automated ▪ access or changes to data, and methods, regulatory agencies ▪ A durable and inerasable audit to provide controls to prevent have described the regulatory trail or event log omission of data. expectations in their guidelines ▪ Separate administrator and user and draft documents. ▪ The computerized system access rights lacked access controls and audit trail capabilities. ▪ Good standard operating * See also: WHO Technical Report procedures (SOPs) 996, Annex 5, “Guidance on good ▪ All employees had administrator data and record management rights and shared one user name. ▪ Oversight and regular review of practices” processes Key areas of data integrity control There are seven functional areas consistently mentioned in regulations and guidance on data integrity. Here we review these key areas, focusing on how they apply to environmental monitoring applications.

Quality Risk Management1 Personnel2 Documentation ▪ Understand the potential ▪ Document and communicate ▪ Implement and require Good impact of all data on product roles and responsibilities. Documentation Practice quality and patient safety. (GDocP) in all written ▪ Provide technical support for documents and SOPs. ▪ Understand the basic systems administration. technologies used in your Follow relevant regulations data processes, and their Assign responsibility for data ▪ ▪ when creating and reviewing inherent limitations. throughout its entire lifecycle. documents. For example, ▪ Implement systems that ▪ Encourage a workplace culture CFR Title 21, Part 211 “Current provide an acceptable state that supports issue reporting. Good Manufacturing Practice of control that is matched for Finished Pharmaceuticals” to risks and criticality of the ▪ Implement systems that Subpart J - Records & Reports. process in question. can identify and minimize potential risks. ▪ Identify and document points of risk for unauthorized ▪ Create behavioral controls for or untraceable deletion personnel, procedural controls Data Life Cycle or amendment, as well as for processes, and technical opportunities for detection controls for technologies. ▪ Implement change through routine reviews. management and control of Analyze the root causes of Schedule and perform periodic ▪ incidents and deviations. ▪ compliance failures in order to risk assessments as technology fix them systemically. and processes change. ▪ Ensure corrective and preventive action (CAPA) Provide technology training to Authorize appropriate access ▪ ▪ processes and procedures. ensure existing technologies privileges for each system. are used to their full potential.

Training Vendors/Providers Audits & Internal Inspections3 ▪ Provide regular training, and ▪ Ensure providers have qualified ▪ Create detailed review document training completion and trained personnel. processes for inspection including personnel identities findings, non-compliance and dates. ▪ Review providers’ quality management systems. reports, and Warning Letters. ▪ Ensure training is matched ▪ Perform routine in-house data to different roles involved ▪ Note compliance to standards audits, including: audit trails, with data - including quality such as ISO 9001, or ISO 17025. raw data and metadata, and assurance, quality control, ▪ Perform regular checks original records. production and management of providers’ systems ▪ Schedule regular review of - with an emphasis on Good and services; audit where system user access rights. Documentation Practices. necessary and/or allowable. Report audit results to senior ▪ Store training documentation ▪ Review contracts, technical ▪ where it is easily retrievable by management and other agreements, and quality those involved with regulatory relevant stakeholders. agreements. and 3rd party inspections.

1 A key document in this area 2 Personnel management directs 3 If a hybrid system is in use is ICH Q9. This guideline from and controls how companies (both paper and electronic data the ICH Expert Working Group function to achieve business goals. are generated), the original data provides a methodology for Focusing on personnel ensures should also be checked routinely a risk-based approach to that resources are allocated in addition to trend data, reported data management, including to the functions that support documents, or PDF files. recommendations. recommended practices and promotes accountability among all levels of management and staff. Eight Ways to Ensure Data Integrity in monitoring systems Qualify IT & Validate Systems The following recommendations give an overview of how to maintain data ▪ Validated systems require an integrity for computerized systems. IT environment that has been fully qualified. Perform Risk-based Validation Audit your Audit Trails ▪ Validate only systems that ▪ An audit trail must be an Plan for Business Continuity are part of GxP-compliance. inerasable record of all Ensure disaster recovery Ensure protocols address data changes made to data in a ▪ planning is in place. quality and reliability. system. To be useful in GxP ▪ In some cases, it’s cost- compliance an audit trail must ▪ Your plan should state how effective to have the system answer: Who? What? When? quickly functions can be vendor perform qualification And Why? restored, as well as the probable and validation of the systems. impact of any data lost. Define the data relevant to To help decide between in- ▪ Look for software and systems GxP and ensure any changes ▪ house or purchased validation that can record and store data to this data will be recorded service, use the ISPE’s redundantly to protect it from by an audit trail. GAMP5 (Good Automated loss during power outages or Manufacturing Practice) Assign roles and schedules network downtime. categorizations to determine ▪ for testing the audit trail Employ solutions such as UPS the validation complexity of functionality. ▪ your system. (Uninterrupted Power Source), battery-powered, standalone The depth and frequency of Account for all electronic data ▪ recorders or devices that can ▪ an audit trail review should be storage locations, including switch to an alternate power based on the complexity of the printouts and PDF reports source when required. E.g. system and its intended use. during validation. data loggers that can also be ▪ Ensure your quality ▪ Understand what audit trails battery powered. management system defines comprise: discrete event logs, the frequency, roles and history files, database queries, responsibilities in system reports or other mechanisms Be Accurate validation. that display events related to ▪ Verify system inputs. For Your validation master plan the system, electronic records example, an environmental ▪ or raw data contained within must outline the approach monitoring system requires the record. you will use to review regularly calibrated sensors. meangingful metadata, ▪ For networked systems, test including audit trails, etc. Select Appropriate System that data are coming from the ▪ Your validation master plan and Service Providers correct location. should require periodic ▪ Select systems that provide re-evaluations of every ▪ Ensure your providers are alarm messages in case of validated system. familiar and fluent with the communication failure, device relevant regulations.* problems, or data tampering. Change Control ▪ Systems must be fit-for- purpose. Compare your User Archive Regularly ▪ Ensure system software Requirements to system updates are designed to Functional Specifications ▪ Backup and save electronic comply with changing prior to acquisition to prove data, including metadata, to regulations, especially when the suitability of a candidate a secure location on a regular implementing new features. software for any GxP schedule. ▪ Collaborate with providers to application. ▪ Verify relevant GxP Data can stay informed about changes be readily retrieved during Learn about your suppliers’ and update your systems ▪ internal audits. organizational culture and accordingly. maturity relating to quality and ▪ Electronic archives should ▪ Select systems that are easy data management. Ask what be validated, secured and to update and validate upon systems are in place to ensure maintained in a state of the addition of new hardware data integrity and audit those control throughout the data or other system inputs. systems if possible. life cycle.

* See also: EU GMP EudraLex Annex 15, Section 2, Documentation including VMP 2.6: “Where validation protocols and other documentation are supplied by a third party providing validation services, appropriate personnel at the manufacturing site should confirm suitability and compliance with internal procedures before approval.” viewLinc & VaiNet Features ▪ Access to the system is controlled by individual login IDs, user names and passwords. ▪ User-specific rights and access control permissions create different authority levels, fulfilling the regulatory requirement for segregation of duties. ▪ viewLinc includes device checks to guarantee the origin of the data and validation alarms to guarantee the validity of data. ▪ Only viewLinc, not users, can create data records, and these are uneditable and inerasable. ▪ Creation and modification of data and system parameters is recorded by an audit trail shown in viewLinc’s “Event” view.

Event Details Add Comment ▪ Calibration data is stored in each device, and in the software, ensuring accuracy Monitoring software should show all events within the system, including: specifications of devices are threshold and device alarms, messages sent (Emails or SMS), User also tracked. login/out, automated report generation, devices added, etc. ▪ Reports are created in secured PDF files that cannot be Data Integrity in Environmental Monitoring modified. ▪ All graphs, system reports Vaisala understands the diligence and attention that is required to ensure and environmental reports data integrity. As a manufacturer of environmental measurement and are easy to read, fulfilling the monitoring systems used in GxP applications, our customers turn to us requirement of human readable for assistance to protect their data. Vaisala is invested in understanding copies of data. the relationship between computerized systems, network functionality, ▪ All measurements are device efficacy and data integrity. We continuously develop our system synchronized against the software with the goal of ensuring data integrity for our customers. Here system server clock so it’s easy we outline several features of viewLinc that guarantee reliable, complete, to compare data sets across and accurate data. time zones. ▪ The viewLinc software can be used in multiple time New Generation, Same Data Integrity zones simultaneously without Vaisala’s proprietary VaiNet wireless technology is a wireless device compromising the data because connectivity option for the viewLinc Continuous Monitoring System. The all records are based on UTC VaiNet data loggers system include several security features, which are (Coordinated Universal Time). designed to ensure data integrity in GxP-regulated applications. VaiNet ▪ Thorough system provides secure connectivity between data loggers and access points documentation helps with with a licensed ISM (Industrial, Scientific and Medical) protocol. With qualification, validation radio band frequencies that vary depending on global location, VaiNet and future usage of the system (User Requirement relieves already overburdened Wi-Fi networks that are often difficult to Specification, Functional secure. Vaisala used the LoRa™ modulation technique to create wireless Specification, Design data loggers with wired reliability. VaiNet’s highly modulated CSS (Chirp Qualification, Traceability Matrix, Spread Spectrum) signal achieves ranges of 100 meters or more in typical Risk Assessment, Validation manufacturing environments. The unique modulation is highly reliable, yet Protocols and Reports). requires less power for data transmission. The result is a long-range signal ▪ Metadata is easy to find that is readable only by Vaisala network access devices within a VaiNet and provides contextual network, and superior protection of data integrity. information on all data. Two additional security features Data integrity by design Data integrity is about more than further enhance data integrity: data compliance with regulations; encryption and data authentication. Risks to data integrity are reduced it is about protecting research Data encryption means that by implementing correct data and products for human use. specific code is required to read management practices that In GxP applications, data and understand transmitted include behavioral, procedural, often represents a significant information. In VaiNet, the original and technological controls. investment in development, data is transmitted between data In environmental monitoring clinical trials, donated tissue, loggers and the network access applications, there are common and the hopes of patients for a point (VaiNet AP10) and cannot scenarios that entail expensive new therapy or drug. The data be intercepted by a non-VaiNet risks: an undetected compressor represent assets that require device. Data loggers encrypt the failure could destroy the entire fail-safe, trustworthy systems and data before transmission, and only contents of a fridge or freezer. practices that ensure product the access point can decrypt this These chambers may be storing safety. The devices, software, data. Encryption is performed with irreplaceable samples from infrastructure, processes, and proven AES-128 technology (AES research in a crucial stage of operating procedures must = Advanced Encryption Standard) development. If the monitoring all be aligned to ensure that and data authentication uses system includes remote alarming, data are complete, consistent, CMAC technology (Cipher-based the costs of an equipment or accurate, and exemplifying the Message Authentication Code). process failure are mitigated. Even characteristics of ALCOA+. Authentication ensures that data when equipment failure is not is coming from the correct source immediately catastrophic, accurate To learn more, see our webinar on and the origin of the sent message and reliable data is sent in an alert data integrity. is always identified and tracked. through email or SMS will indicate a problem.

Ref. B211613EN-B ©Vaisala 2021 Please contact us at This material is subject to copyright protection, with all copyrights www.vaisala.com/contactus retained by Vaisala and its individual partners. All rights reserved. Any logos and/or product names are trademarks of Vaisala or its individual partners. The reproduction, transfer, distribution or storage of information contained in this brochure in any form without the prior written consent of Vaisala is strictly prohibited. Scan the code for All specifications — technical included — are subject to change www.vaisala.com more information without notice.