Background: An explosion in a Chinese factory in 2016 caused a global shortage of essential broad-spectrum antibiotic piperacillin-tazobactam. Hitherto, no detailed, policy-relevant analysis has been conducted on this major shortage event. Thus, we aimed to (1) investigate causes; (2) describe supply chain challenges; and (3) uncover policy gaps to support possible mitigation actions.

Methods: Applying an analytical framework for security of medical supply chains, we investigated the changing roles of Pfizer-led and Chinese API suppliers. We identified demand surge, capacity reduction and co-ordination failures. Triangulating between scientific literature, corporate, and regulatory documents, we analysed the impact of Western and Chinese policy contexts on supply chain resilience.

Results: We uncovered 'red flags': geographically dispersed manufacturing failures due to complexity of sterile production; undetected supply chain concentration and interlinkages; and Chinese policy-led API supplier consolidation. We found these warning signals were ignored in the absence of a co-ordinated policy framework to identify and mitigate emerging global supply risks. Firstly, policy makers lacked visibility on growing 'volume dependency' in the chain. Secondly, national policy makers lacked a global view of supply risk. Thirdly, we show antibiotic API manufacturing economics were impacted by a number of non-pharmaceutical policy decisions (e.g. state aid, environmental standards, procurement rules) which contributed to supply chain vulnerability.

Conclusions: Our findings suggest possible policy gaps in governance of supply chain resilience. Firstly, disclosure of API suppliers including degree of dependency may better pre-empt bottlenecks, facilitating priority setting for public investments in re-shoring where global API supply currently relies on few, or single plants; secondly, a whole-of-government approach may counter the potential impact of non-pharmaceutical policies on supply chain resilience. Our findings confirm suggestions from previous studies that international data sharing would be beneficial considering the global shortage effects which can emerge from a single point of failure.