5 Lessons for Predictive Procurement in Healthcare

For years, healthcare procurement was largely measured by one question: how cheaply can an organisation secure the products it needs?

However, a series of global disruptions has reshaped perceptions around procurement’s core principles.

Shortages, shifting demand, transportation disruption and geopolitical volatility have exposed the risks of relying on procurement models that are primarily reactive – particularly when the products being sourced can determine whether a patient receives critical medical treatment.

This was the central theme of a webinar hosted by Healthcare Digital in partnership with Amazon Business, exploring how predictive procurement and data-driven decision-making can help healthcare organisations strengthen supply continuity.

You can watch the full webinar on demand here.

During the session, Dylan Bartlett, President and CEO at Supply Chain Canada, Joseph Shobe, Senior Managing Consultant at IBM and Kenny Otto, Head of Supply Chain at the NHS’ Royal Papworth Hospital, discussed how healthcare procurement is evolving from a transactional function into a strategic pillar of operational resilience.

The panel explored the role of AI and predictive analytics, but also highlighted some of the more prosaic foundations required before healthcare organisations can make effective use of these technologies – from data quality and interoperability to workforce training and supplier collaboration.

Here are five key lessons from the session.

1. Healthcare procurement must move beyond cost

The biggest shift in procurement has been a change in what organisations are actually trying to optimise.

While cost remains important, the panel argued that healthcare organisations can no longer afford to make price the overriding consideration when procuring critical medical supplies.

“It shifted from a cost alone equation to…availability and resilience….Supply chains want to be able to absorb the next impact,” explains Joseph Shobe, Senior Managing Consultant at IBM.

“People are starting to pay much more attention to procurement as a…strategic tool inside their organisations,” he continues. 

The COVID pandemic provided a stark demonstration of the problem. Healthcare organisations rushed to stockpile critical supplies, but some of those products ultimately went unused and expired.

At the same time, organisations are now expected to respond faster to disruption while maintaining availability, controlling waste and managing supplier risk.

Due to the volatile nature of the global landscape, gone are the days when procurement leaders could follow a “set it and forget it” approach, says Dylan.

For healthcare, the consequences of getting the procurement balance wrong can be particularly severe.

“The worst moment to discover a shortage is when you’re in one,” says Joseph. “From that point…your options are limited. They’re expensive. You risk delivery performance. You risk the patient.”

For the panel, resilience therefore needs to sit alongside cost, with procurement teams considering supplier capability, delivery risks, category risks and the likelihood of future disruption.

2. Data quality comes before AI

Predictive procurement depends on being able to understand what an organisation has, what it is using and what it is likely to need next.

Yet healthcare supply chains remain characterised by fragmented systems and inconsistent data.

Hospitals can have separate definitions for products and suppliers, while procurement systems may not be connected to electronic health records or other operational systems.

That means supply chain teams may not have visibility of critical information such as hospital census, scheduled procedures or when inventory is actually being consumed.

When asked how healthcare organisations can overcome these barriers, Joseph put it succinctly. “It starts with data quality. It’s not glamorous,” says Joseph. 

The panel pointed to standards and integrated systems as an important foundation for predictive procurement.

Joseph also described how better visibility into inventory consumption had previously enabled one healthcare organisation he worked with to reduce inventory held on the floor by 50%, after teams gained greater confidence in the data showing how much stock was actually being used.

The lesson is clear: before asking AI to predict what happens next, organisations need to establish a reliable picture of what is happening now.

3. Predictive procurement isn't necessarily AI – yet

One of the more striking moments of the webinar came when the panel challenged the idea that existing predictive procurement tools should automatically be described as AI.

While advanced forecasting and analytics can help organisations anticipate demand, these technologies are not necessarily examples of full AI integration.

“When we start to bring new tools together that can bring disparate data together, unstructured and disparate data, and start to provide advice,” Joseph says, organisations can begin to move towards AI systems that make recommendations based on a much broader pool of information. 

The distinction is important.

Healthcare organisations are already using forecasting, ERP and inventory-management technologies to analyse demand. The next stage will involve bringing together far more information – including hospital census, operating-room schedules, actual demand, supplier risk, geopolitical developments and cost pressures.

More advanced AI systems could then assimilate that information and provide recommendations about where risks are emerging and what action should be taken.

But the panel stressed that this is still an emerging capability rather than a universally established reality.

4. AI should augment procurement teams, not replace them

Even as AI becomes more capable, the panel was clear that healthcare procurement remains a human-led function.

AI can analyse large amounts of information, identify patterns and recommend actions. But procurement professionals still need to assess those recommendations, particularly where decisions involve expensive products or patient safety.

“AI tools are to help provide recommendations, and the action is taken by humans. So you use that data for decision-making,” explains Kenny.

AI could potentially progress from an advisory role to taking on lower-level tasks below oversight, and eventually evolve towards more autonomous systems.

However, high-risk categories will continue to require human intervention.

Governance will therefore be critical to defining where AI can act independently, where it can make recommendations and where it should simply flag a potential problem.

The panel also highlighted workforce capability as a major consideration. Procurement professionals will need the skills to understand, train and trust new tools – rather than simply being expected to adopt them.

5. The future of procurement is connected

The next three to five years could see healthcare supply chains become much more closely connected to clinical operations and suppliers.

For Joseph, the direction of travel is clear: supply chain and clinical systems need to be able to “see” one another, allowing organisations to understand demand and supply in the same context.

He sees “supply chain and clinical systems coming together…Organisations know that they have to create visibility both from demand to supply, supply to demand.”

That connectivity could also extend beyond the hospital.

By sharing forecasts and real-time information with suppliers, healthcare organisations could give manufacturers greater visibility into when products are likely to be needed, rather than relying on suppliers simply to replenish stock.

Kenny cited one example of connectivity in action. In the NHS, the use of a central hub for 95% of product sourcing demonstrates progress in addressing systemic fragmentation.

The longer-term opportunity extends beyond resilience and cost. Better forecasting could help organisations reduce excess inventory, improve supplier relationships and potentially support sustainability goals.

But technology alone will not deliver this transformation.

As the panel repeatedly stressed, healthcare organisations will need to invest in data standards, integrated systems, collaboration and training alongside AI.

Ultimately, predictive procurement is less about replacing the procurement professional with an algorithm and more about giving that professional a clearer view of what is coming next.

What is Amazon Business?

Amazon Business is a global platform that helps organisations make purchasing easier, control spending and improve supply chain efficiency.

The platform makes complex buying processes simpler and offers tools to help everyday purchases support sustainability goals.

With Guided Buying, administrators can set central rules, highlight preferred items, or block purchases that fail to meet an organisation's ESG standards.

Amazon Business also shows supplier sustainability ratings, like EcoVadis scores, right on the supplier listings when you make a purchase.

To hear the full discussion hosted by Healthcare Digital and Amazon Business on integrating predictive procurement into healthcare supply chains, watch the webinar on demand.