Quality assurance (QA) and quality control (QC) units exist in all product organizations, but are especially important in the pharmaceutical industry. They are necessary to check, measure and verify processes and products such as materials, equipment and experiments – crucial, when the products that are being produced are going to be consumed by the public. However, with new technologies such as AI and the cloud quickly becoming popular in laboratories and research organizations, how might the role of the QA/QC unit change?
A quick reminder for those who forget or don’t know – QA is assuring that a product will meet a given set of quality metrics through the implementation of processes. QC is quality control which is testing at given stage through the manufacturing process to monitor product quality. Both are required but have differing drivers.
In pharmaceutical companies, the QA/QC department are essential to deliver safe medicines and, as such, are deemed mission critical and have significant influence. However, implementing and tracking QA processes is notoriously difficult; collating and tracking QC data in real time and reporting “out of specification” results is complicated and often time consuming. Informatics tools have been developed and used in most pharma organizations to reduce this burden. Often sitting under the umbrella of manufacturing execution systems (MES) and enterprise resource planning platforms (ERP), pharma has leveraged IT and informatics to help address QA and QC.
However, there is an increasing acceptance of cloud computing and digital strategies that could mark a seismic shift in what is possible and what the future of how QA/QC works. Digital strategies exist now at the board level of many organizations and often link to an IoT-related strategy as well. IoT can be considered at the highest level as the connection of “things” (all instruments, analytical systems and process sensors) via the internet (of sorts – likely not public in this context) that allows “connected things” to be discoverable and controllable to facilitate easier access to the data.
Considering the implications of just these three things – discoverable, controllable and accessible data – we can see some immediate and valid concerns, security being the highest on people’s minds. Who and what can do things that affect the environment will be at the forefront of everyone’s minds. But for the purposes of this article let’s assume that their issues will be solved and proper control systems and process are put in place to meet the rules and provide a secure environment – what does that mean for QA/QC departments and their thinking?
One significant potential development area is the use of AI. Deep learning, machine learning and tools that sit under this umbrella, all have potential application in this space. Imagine the future state where all the data that a given process is producing is being aggregated, analyzed and extrapolated to either help make informed decisions to support the business or, even further, make automated decisions and change how a product is being produced to ensure the QA metrics are met. This kind of self-managing pharmaceutical production process is obviously conceptual now but there are examples of parts of it happening – just in isolation and waiting for technology advances to extend further.
Production lines are already connected, data is often captured and managed centrally and analytics is performed on these data to help inform decisions. But the analysis is done on “known knowns” – trends that are already known to impact QA and QC. Data is not aggregated at a macro level, often due to an absence of somewhere to put it. Data is defined by those attributes that are known as important, so are collected and aggregated. Each of these aspects impacts the connectivity and therefore the amount of derivative data that can be created.
Each of these current “restrictions” are only temporary if we look at the trajectory of cloud and AI technologies. Data can be aggregated at a macro level easily in cloud “scalable” storage – i.e. storage that increases automatically as you need it. Enrichment can be applied to this data to make it more consumable, which can be a continual process – as you learn more about what is important for QA/QC you can enrich the historical data if possible. The enriched data can be made available to AI tools – deep learning (automation of a known decision process) and machine learning (looking for new relationships and patterns that can be converted into new knowledge or understanding).
This is a very forward-looking view of the pharmaceutical QA and QC but it is likely to arrive faster than anticipated if other technology adoption trends are mirrored: ten years ago, we all though that music combined with a mobile phone and Filofax was absurd. Look where we are now – we can now do our banking on these same devices and control most of our homes environmental systems (heating, lighting, doors and appliances) with them as well!
Pharma manufacturing has concepts like this already – quality by design encompasses some of these aspects. But, in the future, QA and QC must leverage these technology trends, including the IoT and AI and we will see increases in product quality control to a point where manufacturing deviations are not show stoppers in a production run but are managed within the plant itself. The result is a “self-adjusting” process that maintains product quality inherently and can adjust itself within a given set of variables without human intervention.
Scary thought, I know, but likely better that relying on us fallible humans to dial valves, press buttons and conduct analysis.