How Technology and Trade Data Solve Supply Chain Disruptions

A solution to mitigating and resolving supply chain disruptions would be to fuse retrosynthesis and global trade data from a variety of reliable sources, using advanced data science techniques to identify vulnerabilities in the supply chain.

Leigh Prather
Leigh Prather

As the Coronavirus disease (COVID-19) pandemic ravaged the world, it brought life as we knew it to a halt. In particular, many industries experienced significant and unexpected supply chain issues—difficulty sourcing necessary ingredients for vital drugs or finding workers to fill relevant positions—holding up production lines and imposing a considerable financial burden. The absence of just one raw ingredient can halt development of a product; unfortunately, this threat has not lessened with the introduction of vaccines. A pandemic on this scale is rare, but natural disasters, political boundaries and route blockages are common and can be just as devastating to production.

Pharmaceutical supply chains are complex, and it can be difficult to discover or resolve the root cause of disruptions. A deep understanding of the underlying supply chain can enable analysts and organizations to identify and react to shortages quickly or even proactively, ensuring that delivery of critical medications is not interrupted.

A solution to mitigating and resolving supply chain disruptions would be to fuse retrosynthesis and global trade data from a variety of reliable sources, processing the data using advanced data science techniques to identify vulnerabilities in the supply chain and displaying results using interactive visualizations. Many existing solutions can track the distribution of end products; however, few can trace the supply of raw materials upstream, a more difficult process due to the incomplete and uncertain data.

For example, the processes pharmaceutical companies use to manufacture drugs are proprietary. Some ingredients must be divulged, but the specific method used to create the product is rarely shared. An analysis of retrosynthesis information from libraries of data on drug development may reveal several methods on how a drug is manufactured. Machine learning can then probabilistically assess the likelihood that any given method might be available to a given manufacturer and identify where in the world the materials might be sourced from to support potential processes for generating the drug.

Further complicating this analysis, shipping data is also held close to the vest in many countries outside the United States. Data in these countries can be completely withheld; for example, there may not be any trade data that accurately locates the specific location within China where a particular ingredient is sourced. This can make it difficult to determine if that ingredient will be disrupted by an event that only impacts certain regions of the country. Facility information is complex, too—even if the facility has the machinery to make the source component, they may not be able to do so with approved ingredients and legal labor.

Many people rely on a steady supply of certain drugs to survive. If the supply chain of these key drugs is disrupted, people can die. Technology that can both predict when a disruption might occur and analyze existing supply chains to detect where a chain is brittle, will be essential both to provide awareness of potential problems and let distributors look to other sources to keep life-sustaining drugs in place for those who need them.

It will be important to develop solutions that can handle incomplete and uncertain data with deeply enmeshed complexities. Incorporating open-source probabilistic programming language can help.

Presenting data in a usable fashion is almost as complicated—and important—as compiling it. Having a user-centered, interactive dashboard employs visualization techniques, including supply graphs and maps, would enable a rapid understanding of the structure and risks within possible pharmaceutical supply chains. Pharmacists and other users in the pharmaceutical industry could use this dashboard to diagnose and assess issues quickly and easily.

Incorporating a visualization approach grounded in scientifically proven design methodologies to combine data visualizations and various cueing and highlighting techniques will help users to predict, identify and understand critical gaps in the supply chain and the non-obvious impacts of disruptions on final products or compliance issues.

The need to solve and prevent supply chain issues is important for many industries, not just pharmaceuticals. It is necessary to map global supply chains for industrial partners, public agencies, global governance organizations and non-governmental organizations.

Flexible technology that can draw information from a variety of sources, including trade data, commercially available data networks and the user’s proprietary data will make it possible to pivot the technology to fuse a different set of data. For example, if a shipping blockage occurs, as it did in March in the Suez Canal, suppliers could have a greater set of scenarios to select from to keep assembly lines running.

In addition to manufacturing-related industries, trends indicate increasing interest of placing information about the supply chain into the hands of product distributers. For example, hospitals could use the technology to manage its supply of medicines and vaccines to ensure that the needs of the local population can be met.

This type of technology has the potential to prevent devastating issues caused by disruptions to the supply chain. It can provide a proactive method to identify and remedy issues before they affect the availability of end products.