Organizations can better understand the maturity of AI-driven automation technology across their organization’s IT landscape through effective pharmacovigilance.
A safety surveillance strategy is a requirement for pharmaceutical companies to monitor product safety in the post-marketing environment, as well as to keep physicians informed on product risks and how to prescribe appropriately. This is usually seen as a labor-intensive cost center that often overburdens teams with work, making it difficult to retain talent. As the industry becomes more cost-conscious and is handling more real-world data sources than ever, companies are seeking ways to improve the efficiency and effectiveness of their pharmacovigilance (PV) activities. Automation offers the solution for organizations prepared to embrace it and to do the hard work required to support digital transformation.
Why automation is vital in pharmacovigilance
Many reasons exist to support the adoption of automation in drug safety surveillance. Using artificial intelligence (AI) in safety surveillance can automate many aspects of the process, including non-electronic data collection, processing of inbound data, and triaging data according to its source. With these applications, companies can increase productivity, lower their technology costs, and reduce their reliance on manual labor. At the FT Global Pharmaceutical and Biotechnology Conference in London, which occurred in November 2022, staffing challenges were top of mind for the attendees. During the conference, there was much discussion about the difficulties of retaining talent in life sciences, particularly in PV. With new safety risks constantly emerging and new methods of locating and identifying them, PV positions need candidates with applied experience and the ability to understand the significance of a safety signal amid real-world market feedback. Additionally, even as data volumes are increasing, the offshore cost-savings model has plateaued with companies no longer able to simply outsource their problems and cost increases away. Companies worldwide are looking for new ways to optimize PV processes and achieve scalability and efficiency through innovation. Knowing this, where should companies begin and achieve the journey of rolling out automation in their PV processes? What does success look like, and what factors must be considered from starting the business case to measuring the outputs and, finally, achieving a successful outcome?
How automation in PV environments work
Automation includes everything from basic, rule based processes to more advanced levels of AI, machine learning (ML), deep learning, and natural language processing. In short, it is cognitive computing, and it is changing PV as the industry knows it. This form of “intelligent” automation enables predictive analysis and the capture and translation of adverse event (AE) data to determine significant safety trends. It also gives PV professionals a stronger ability to manage future AEs and a better understanding of safety issues. An advanced PV platform automates intake via web and mobile apps, sending results directly to the safety system. It contains sophisticated AE tracking tools that analyze both structured and unstructured data in real time. The data are gleaned from multiple channels, including social media and chatbots. PV query tools can automate the completion of case management documentation and reports, while natural language processing (NLP) tools analyze complex narratives, including patient charts, social media posts, articles, and other unstructured data. Analytics tools identify safety trends across populations to educate healthcare professionals, regulators, and payers.
Factors to consider in adopting automation
Organizations are looking for a safety surveillance strategy to achieve transformation without disrupting compliance. For companies about to embark on their PV journey, factors to consider include the ease of deployment, the use of technology, the value of that deployment, and the risk. For example, some of the challenges involved in this process include the availability of data, the resources to oversee the transformation project, the impact on current operations, and the changes it will have on company culture. When organizations implement advanced PV platforms successfully, they can generate far more value from PV tasks than was previously possible. Solutions conduct deep analyses of integrated data sets to identify meaningful safety trends, resulting in reduced errors related to manual data entry, strict patient data privacy controls, and minimized pharmacovigilance risks to the business.
Safety surveillance automation
There are actions companies need to take at an enterprise level to be successful and accelerate their journey to automation. It is not unusual for companies to set unrealistic expectations around AI, choose the wrong vendors, launch projects without linking them to measurable performance outcomes, and fail to support change management initiatives. While these are common mistakes on the path to PV automation, they can be avoided by utilizing a tool that can help benchmark the automation adoption process and its results, such as IQVIA’s Safety Surveillance Automation Maturity Matrix (1). The tool walks users through the four phases required to achieve automation and become an automation-first organization. It outlines specific criteria for progress, giving decision-makers clear guidance on how to vet their readiness for automation, choose the best pilot projects, and identify potential obstacles. The phases ensure companies stay on a consistent path, continuously monitor benefits, and avoid pursuing projects that will not deliver value.
Phase 1: reaching readiness
This first phase to reach automation focuses on education about current technology and what it can achieve. Companies can take steps to prepare the organization for transformation, such as hiring people who understand the technology, consulting multiple AI vendors to explore a diverse range of solutions, and finding partners who align with their culture and technology needs. Investing time toward attending demos can help an organization determine whether a prospective solution could be viable. Defining the business case for their move to automation requires determining use cases and setting achievable targets. The technology is not as critical a factor as the benefits it can achieve, and unless those align with the business case, the solution chosen will not be appropriate. When these steps have been completed, companies are ready to engage with vendors who will validate their ideas and targets. They can then embark on a request for proposal (RFP) process and select a vendor. Quality vendors will raise questions to help define achievable targets and work in unison with the company to develop a request for quotation (RFQ).
Phase 2: investment initiatives
This phase of the process contains no shortcuts. Companies looking to automate PV will have to invest in data, regardless of the technology they choose. Most data vendors will not be able to supply data in the desired format, so the organization will need to clean and standardize it according to the software requirements. Technology will fail if the data are not correctly prepared and available in the proper format. A company must also invest in flexibility and create an organization willing to accept change. While technology can adapt, vendors have a specific view of the world that does not necessarily include knowledge of an individual organization’s culture. Be prepared to adjust the company goals during the process. This could mean bringing in different groups or people as implementation progresses and the focus changes to alternative areas. Preparing the groundwork for automation includes mapping how it moves through the system, determining the impact on roles and responsibilities, and preparing employees to understand and accept the changes. Once the data and workflow are ready, the team can roll out its pilot project and begin measuring performance. The outcome of this phase will determine the company’s way forward.
Phase 3: enabling automation
Reaching this phase means the company is ready to begin reaping the benefits of AI through automation of the PV activities. This requires a heightened focus on high-priority workflows that deliver time, cost, and quality benefits through automation. One should begin building a project framework and governance model for all future AI deployments that captures best practices related to timelines, stakeholder support, vendor management, and change efforts. This process will help streamline projects, reduce the time to deployment, and prevent common mistakes from recurring. Change the company’s business model to embrace safety as an automation-driven process, rather than simply a way to ease the human resource burden. Consider the entire workflow through a tech-enabled lens to identify real obstacles and determine how to use technology to eliminate problems. During this phase, companies typically reach the point of being ready to use AI to support all key decisions in the safety workflow. AI is embedded across departments, with fully automated data entry, built-in reporting rules governed by machine learning algorithms, and a suite of applications that enable automated case translation, data analysis, and decision-making as part of the safety surveillance workflow.
Phase 4: innovation and improvement
Companies that reach this Matrix stage have achieved full automation of their PV activities and are ready to consider future possibilities. These include:
• Establishing an AI-first culture that actively seeks new ways to use technology to enhance, accelerate, and act on safety data.
• Focusing on supporting a continuous state of improvement by appointing a dedicated team to monitor digital innovations and constantly question how, when, and where new technologies could add value for the organization, even if they are not yet available for use.
• Pushing vendors to create new solutions and to reimagine their own technology roadmaps to improve processes within the pharmacovigilance environment. The pharmaceutical industry has seen an explosion of data, much of it from devices and sources that did not impact pharmacovigilance in the past. Data flowing from wearable devices, social media, audio recordings, and shared images can all now be used to monitor the safety of pharmacologic products. This transformation will only continue, which means safety surveillance must continue to push the envelope on data relevance and the use of AI to capture, translate, and interpret it.
Results of automating safety surveillance processes
Companies that have introduced automation in their PV processes have seen numerous improvements in efficiency. AI adds value to data and creates new insights and knowledge. It replaces multiple processes and moves the organization from a transactional view of the world to a data-driven view of the world. It challenges the entire business model and the business case that underpins it, freeing up data and enabling the use of knowledge in new ways. Moreover, AI will create new data that we do not know today. We are not going to ask for it; it will happen because that is what AI does. In the longer term, AI will start to answer its own questions. The possibilities for creating new value for companies are endless and will force them to rethink many of their current operation methods.
Moving to an AI-driven organization
To become a fully AI-enabled organization, organizations will need to spend a lot of time looking at their processes, reviewing their business models, and really understanding how AI ideally supports all the critical decisions they make. For changemakers, the key is to figure out where your company is today and where it wants to go.