As U.S. healthcare systems face nearly unprecedented shortages of hundreds of drugs and cancer therapeutics, pharmaceutical manufacturers face growing pressure to maximize their productivity and supply reliability.  

Every pharmaceutical product begins as a series of supply chain inputs that come together in a manufacturing plant and then go to market. By applying Internet of Things (IoT) technology, data analytics, and artificial intelligence (AI) in conjunction with lean manufacturing techniques to these processes, manufacturers can improve their operations and do more with the raw materials they already have.  

These solutions can also help pharma and other medical manufacturers continuously optimize their supply chain management and manufacturing processes for efficiency gains and quality improvements over the long term. 

Making Supply Chain Forecasts Smarter with Factory Data 

Accurately forecasting supply chain needs is a classic challenge for manufacturers, particularly in pharma. Market conditions, emerging diseases, new therapies, seasonal demand differences, and more all complicate predictions about how much of a given product to make — as well as the inputs required. For example, if demand for a daily medication is constant, forecasting supply chain needs is straightforward. But for a new medication or a seasonal vaccine, sourcing the right quantities of raw materials can be more difficult.  

Accurate historical data and real-time demand inputs make it easier to forecast more precisely, especially in cases where demand for a particular product is highly variable. Touchless forecasting uses advanced analytics based on this data to identify clearer prediction signals. This data is collected from a combination of historical records, installed ERP and historian systems, and IoT sensors with associated analytical tools. This data provides real-time information on product runs, quality defects and trends in a number of parameters critical to accurate forecasting. 

Making this data available for analysis and forecasting isn’t always easy. However building what’s called a “data fabric” can help. The data fabric makes unified data from the ERP and other sources available in a clean, easy to use platform for analytics and forecasting. More accurate forecasts allow for more precise supply chain planning, which can help prevent shortages based on prediction errors. Better forecasts also support better allocation of resources within the smart factory after the raw materials arrive.

Making the Factory Smarter with IoT-Generated Data

The foundation for the smart factory rests on IoT sensors that track equipment function, employee location, and other key elements of the production process. The IoT sensor data then informs several smart practices, starting with the ability to track factory functions in real time. Rather than watching the activity on the factory floor and reacting when they see or hear something unusual, managers can monitor the data, integrate it into daily sessions, and move faster to resolve problems. 

For example, if the data indicates that a piece of equipment is starting to malfunction, the manager can promptly call for service, and perhaps start setting up another production line to minimize interruptions. 

Over time, as the IoT system gathers more data, it’s possible to perform historical analytics to identify longer-term trends, such as maintenance intervals for specific pieces of equipment. These insights allow for proactive predictive maintenance (PdM) that avoids unplanned downtime due to equipment failure, and reduces costs related to unnecessary interval-based maintenance. Maintenance can be scheduled between production runs to minimize interruptions. PdM can also help manufacturers optimize their yield by reducing the number and size of batches lost to unplanned downtime — a critical consideration for products with hard-to-obtain raw ingredients.  

Another way to leverage historical factory data is to create a digital twin that simulates a particular piece of equipment, a production line, or an entire factory’s operations. Digital twins allow manufacturing engineers and planners to experiment with different elements of the production process to find new ways to optimize it, without disrupting real-world production and risking hard-to-source inputs.

Improving Product Quality for Patients and Providers

Internal and external data can also help pharmaceutical manufacturers improve the quality of their products for better patient and healthcare provider experiences. This helps eliminate costly investigation and rework cycles and has the potential to improve the quality of life for patients who are relying on new medications or experimental treatment regimens, by reducing the probability of undetected defects. For example, for patients following a new cancer regimen under special FDA approval, there might not be much (or any) data on potential side effects. When patients report side effects, that data can quickly be analyzed and viewed against smart factory data to identify areas where the product can be improved to reduce those side effects going forward. 

Better quality products also make it easier for providers to do their jobs. For example, a provider of an injectable biologic cancer therapy will need to scrub, gown, and glove before administering the medication to the patient. If at that point, the product packaging makes it hard to open, that creates a “fit, feel, finish” issue that can be frustrating and waste the practitioner’s time. That product feedback data can help the manufacturer make quick adjustments that eliminate the packaging problem for easier use. 

Starting the Smart Supply Chain and Manufacturing Journey

Laying the groundwork for smart factory and supply chain technology requires three steps.  

The first step is to look at the factory as it is and then imagine what it could become with real-time and historical data to support analytics. This step matters because most supply chains and factory floors evolved long before today’s smart technology, so merging the two requires some conscious brainstorming. A thorough “as is” and “to be” also allows manufacturers to document baseline performance and set reasonable goals for improvement. 

With a vision of what’s possible in terms of optimization and forecasting, the second step is to identify use cases and priorities in order to achieve the pre-determined goals. This might include increasing the rate of on-time-in-full deliveries, optimizing batch yields, improving supply chain forecast accuracy, and beyond.  

Once use cases are identified, the third preparatory step is to identify the resources required for that use case. This is the time to audit the IT/OT (operational technology) landscape to determine whether the use case calls for more or different IoT equipment, more data, better access and analytics through a data fabric, or some combination of those resources. This is also the time to begin shaping the business case in terms of investment, financial and non-financial benefits, and return on investment. 

As goals are formulated, and use cases identified, engagement with talent across various levels of the organization is essential. Leaders must be able to communicate a consolidated, aligned vision of their goals, while the broader workforce must understand the impact on roles and responsibilities. Engaging people early will accelerate the adoption of new tools and processes and help them to thrive. 

Investing in and implementing smart sourcing and manufacturing solutions is not a quick and complete fix for today’s drug and device shortages, but it can help manufacturers bring more product to market and reduce the risk of future shortfalls, while reducing costs and improving product quality.  

Sheetal Chawla is head of life sciences, head of Northeast region, Capgemini Americas; Brian Eden is vice president, global life sciences, Capgemini Group.