Issue link: https://www.e-digitaleditions.com/i/1494960
16 Pharmaceutical Technology ® Quality and Regulatory Sourcebook March eBook 2023 PharmTech.com Pharmacovigil ance prise level and cannot be done only for a single busi- ness process such as safety. Building a strong foundation through the right technology partner When implementing data-driven technologies in pharmacovigilance, it is crucial to work with a tech- nology partner who has expertise in both the field of pharmacovigilance and data-driven technologies. Such a partner will have a deep understanding of the unique challenges and requirements of the pharma- covigilance industr y, including knowledge of the regulatory requirements and best practices for data management and governance. They will also have a strong track record of successfully implementing da- ta-driven technologies in various industries, including experience with data mining, ML, and artificial intel- ligence. This expertise will ensure that the organiza- tion's data-driven technology implementation will be done in compliance with the industry standards and regulations. A technology partner should also be able to provide solutions that can scale with the organiza- tion's needs as it grows and evolves. The right tech- nology partner will also provide ongoing support and training to ensure that the organization can contin- uously monitor, evaluate, and improve their models. This will ensure that the organization can stay ahead of the curve and continue to improve its data-driven technology implementation over time. One of the key areas where technology partners can help is in data governance and management. This in- volves ensuring that the data are accurate, complete, and consistent, and that it is used in compliance with regulations and ethical guidelines. Technology part- ners can provide solutions for data quality assessment, data integration, and data warehousing that can help to improve the reliability and trustworthiness of the data. Another area where technology partners can help is in the development and deployment of ML models. ML models can be used to identify potential safety signals and predict adverse events, but they require large amounts of data and can be affected by biases and errors. Technology partners can provide expertise and tools for data pre-processing, feature engineer- ing, model selection, and evaluation that can help to achieve reliable and robust models. Furthermore, technology partners can provide sup- port for the implementation and maintenance of the ML models, including monitoring and troubleshoot- ing, and help in creating the interface for the end-us- ers. Also, they can provide expertise in creating and maintaining the regulations, ethical guidelines, and compliance for the data usage. Technology partners can play a vital role in max- imizing the benefits of data-driven technologies in pharmacovigilance by providing expertise and solu- tions for data governance and management, and for the development and deployment of ML models. By working together, pharmacovigilance practitioners and technology partners can ensure that the data are used to improve patient safety and support the devel- opment of new and better treatments. Conclusion Data gaps can pose significant challenges to the de- velopment of reliable ML models in pharmacovigi- lance. To overcome these challenges, it is important to consider a range of strategies, including active surveillance, using diverse and representative data sources, employing data augmentation techniques, and utilizing in-stream supervised learning. Addi- tionally, it is essential to implement data governance and management practices to ensure the quality and integrity of the data being used for ML. By leveraging these strategies, organizations can effectively close data gaps and achieve more accurate and effective ML models in pharmacovigilance. It's worth noting that these strategies should be tai- lored to the organization's specific needs, as it's im- portant to find the balance between data privacy, data security, and data quality. It's also crucial for organizations to choose the right technology partner when implementing data-driven technologies in pharmacovigilance. This partner should have expertise in both pharmacovigilance and data-driven technologies, as well as a solid under- standing of the regulatory requirements in the indus- try. The right partner will be able to provide guidance on how to effectively implement data-driven technol- ogies, including how to provide pharmacovigilance professionals with the skills needed to understand, use, and maintain the ML algorithms. Furthermore, ongoing support and training should be provided to ensure that the organization can continuously moni- tor, evaluate, and improve their models. References 1. Sch m idt, L .; Olor isade, BK; McGui n ness, L .A.; Thomas, J.; Higgins, J.P.T. Data Extraction Meth- ods for Systematic Review (Semi)automation: A Living Review Protocol. F1000Res. 2020 Mar 25; 9, 210. DOI:10.12688/f1000research.22781.2. PMID: 32724560; PMCID: PMC7338918. 2. Pugh, S.L.; Brown, P.D.; Enserro, D. Missing Re- peated Measures Data in Clinical Trials. Neuroon- col Pract. 2021 Jul 16; 9 (1), 35–42. DOI: 10.1093/nop/ npab043. PMID: 35096402; PMCID: PMC8789297. 3. Wieder, P.; Nolte, H. Toward Data Lakes as Cen- tral Building Blocks for Data Management and Analysis. Front Big Data. 2022 Aug 19; 5, 945720. DOI: 10.3389/fdata.2022.945720. PMID: 36072823; PMCID: PMC9442782. ■