OR WAIT 15 SECS
Jennifer Markarian is manufacturing editor of Pharmaceutical Technology.
Eli Lilly and Company experts share the vision and the value of the digital plant for pharmaceutical manufacturing.
The “digital plant” and technologies that go along with digital transformation-such as robotics, data analytics, artificial intelligence (AI), and the industrial Internet of things (IIoT)-promise greater efficiency. Eli Lilly and Company is advancing in applying these technologies to its pharmaceutical manufacturing organization. Pharmaceutical Technology spoke with Jim Weber, advisor for Manufacturing and Quality IT [Information Technology] Digital Manufacturing, and Wilfred Mascarenhas, advisor for Data and Analytics, Manufacturing, and Quality IT, both at Lilly, about some of the key issues.
PharmTech: What do you see as the most significant drivers for making investments in digital manufacturing technologies?
Weber (Lilly):Our Lilly Manufacturing organization is united around a common purpose-to make medicines that make life better for people around the world, and to do that with safety first and quality always. Therefore, our biggest motivation is to achieve this purpose as best we can and to continuously strive to do even better. The digital plant can accelerate improvements. For example, we can reduce ergonomic risks through robotics lifting boxes and ensure quality through real-time analytics rather than after-the-fact testing. These technologies also drive cost efficiencies.
PharmTech:What do you see as the most valuable benefits of digital transformation technologies for pharma manufacturing?
Weber (Lilly):For us, ‘value’ means anything that enables us to better make medicines. We start by asking our people who make the medicine: what are the problems; where is the waste; and what are the most frustrating things they have to do? Then, we think creatively about solutions. It’s at this ‘creative thinking’ step that the technology comes into play. We have to educate ourselves on emerging digital capabilities, so we can see the full set of possibilities to find a good opportunity. Before we start a project, we require a business case review and approval to keep value at the center, and not technology. Transformation will happen through time, replication and accumulated learning. We believe the most valuable benefits of digital transformation come from the new capabilities that our people develop.
PharmTech: What are some examples of ‘low hanging fruit’/initial projects for digital transformation?
Weber (Lilly): A tree has to be somewhat mature before it has any fruit, much less low-hanging fruit. Right now, technologies like robotics and advanced analytics tend to be a bit more mature, and thus these are the technologies of choice for our initial projects. For robotics, we’ve got projects to automate tedious material handling steps that are done manually today. Some examples include using a six-axis robotic arm to load material into packaging lines, transporting these materials to the line from the warehouse by autonomous guided vehicles (AGV), and even things as basic as picking up the trash with mobile robots. We’ve got projects that are using advanced analytics to help our people identify trends and patterns more quickly and effectively. Some examples of those include using natural language processing to find patterns in complaints, applying data models to predict process failures, and using artificial intelligence techniques to identify and fix errors in supply chain data. We’re doing these things in our existing facilities and leveraging data from current IT and automation systems. At the same time, we’re trying to anticipate equipment and system lifecycle timelines, so that we can build digital plant capabilities into new installations from the start.
PharmTech: What is the role of Big Data/analytics?
Mascarenhas (Lilly): Data and analytics play a critical role in realizing the Digital Plant vision. It is recognized as one of the nine key technologies that is transforming manufacturing and industrial production (1). Not only is Big Data/analytics a critical capability in itself, but it is also foundational to many of the remaining eight technologies such as IIoT, robotics, etc. Our Digital Plant vision, also called Smart Manufacturing (whereby ‘smart machines’ and robots are inter-operating and self-adapting based on communication with one another and with humans) is largely driven by analysis of data and the communication of what adjustments need to be made based on the results of the data analysis. In many cases, this analysis will be done at the ‘edge’ (i.e., at the machine itself) and then communicated to other relevant machines/processes and humans.
Over the past 10 years or more, Lilly Manufacturing has been investing in building a strong data foundation. Due to this long-term investment, Lilly has a mature centralized global data store, which was recently modernized to a Big Data data lake. Data from all of the critical IT and OT systems flow into the centralized data lake, which allows the business to connect to a single data source to get all of the data required for analytics and perform a holistic end-to-end analysis. Time to find and integrate the data is minimized, which allows our people to focus on analysis of the data. The data lake is a fully validated system, allowing its use for both good manufacturing practices (GMP) and non-GMP usage. Lilly is considered one of the leaders in centralizing all manufacturing data for analytics and in having a GMP-validated data lake.
PharmTech: What are the next steps for data and analytics and what are the challenges that need to be addressed?
Mascarenhas (Lilly): There are four different types of analytics typically conducted on data. These are broadly referred to as descriptive, diagnostic, predictive, and adaptive. Descriptive tells us what is happening. Diagnostic provides information on why it happened. Predictive helps predict what will happen so we can be proactive to avoid issues or optimize processes, and adaptive is where processes/machines are using analytics to self-adjust with minimal to no human intervention.
Due to Lilly Manufacturing's strong data foundation, Lilly Manufacturing has mature descriptive and diagnostic analytics. We have automated dashboards such as quality and maintenance, periodic reports, monthly metrics dashboards and process monitoring dashboards to help business users perform descriptive and diagnostic analytics on a daily basis. In the past two years, with the advancement of Big Data and analytics, Lilly Manufacturing has invested in predictive analytics. We have successfully completed several proof-of-concept projects that are now being piloted. A few examples of projects in advanced analytics include predictive and condition-based maintenance, natural language processing of unstructured data, and smart search. We are also evaluating Cloud services for data storage, as well as AI platforms for advanced analytics and machine learning.
We already have vast amounts of disparate data in our data lake, and it’s increasing on a daily basis. Some of the challenges in the data and analytics space are to adapt our information management practices to this huge scale. Data governance, data stewardship, automated data quality, data integrity, data catalog maintenance, and master data management are examples of those challenges. Lilly is proactively addressing these challenges to ensure the long-term maturity and sustainability of the data lake. These challenges require a strong partnership between business and IT for long-term success.
1. Boston Consulting Group, “Embracing Industry 4.0 and Rediscovering Growth,” www.bcg.com/en-us/capabilities/operations/embracing-industry-4.0-rediscovering-growth.aspx, accessed Jan. 31, 2019.