Disease outbreak and bioterrorism: The ultimate supply chain test

In the event of a disease outbreak or bioterrorist attack, public health officials must make decisions about how to allocate finite medical resources — decisions that impact the spread of the disease and the number of lives lost. Ajay Vinze and Raghu Santanam, both information systems professors at the W. P. Carey School of Business, wondered what might be the best way to allocate critical resources in such scenarios. They realized that, viewed from a business perspective, the public health system is actually a very large and complex supply chain — in many ways even more intricate than the highly developed supply chain of an enterprise like Wal-Mart.

A smallpox outbreak starts in Los Angeles, and before long cases are reported in the Phoenix metro area, and in the small communities that dot the desert across California and Arizona. Public health officials must make decisions about how to allocate a finite number of drug doses and hospital beds — decisions that impact the spread of the disease and the number of lives lost.

Ajay Vinze and Raghu Santanam, both information systems professors at the W. P. Carey School of Business, wondered what might be the best way to allocate critical resources in this scenario.

They realized that, viewed from a business perspective, the public health system is actually a very large and complex supply chain — in many ways even more intricate than the highly developed supply chain of an enterprise like Wal-Mart.

Other software tools, such as Los Alamos National Laboratory's well known Epidemiological Simulation System (EpiSims), model disease outbreak patterns; the researchers' model demonstrates how various schemes for doling out medicines and personnel among fixed assets like hospitals affect the efficiency of quashing an outbreak.

Vinze and Santanam have discovered that optimizing the various facets of emergency preparedness can be the ultimate proving ground for ensuring that the right supplies are available at the right place and at the right time.

Finding the right model

In 2003, Vinze and a team of researchers at the W. P. Carey School of Business obtained a grant from the Centers for Disease Control (CDC) to study how state and local public health organizations could better address a national mandate to prepare for biological terrorism. Rather than looking at the issue from a purely medical perspective, Vinze and Santanam saw an opportunity to approach it as a business process problem.

"The novel aspect in our research is the incorporation of the resource allocation decisions that help us understand how decision makers can impact the outcome," says Santanam. The two created a working computer model to determine how municipalities should best prepare for and respond to public health crises such as smallpox or pandemic flu outbreaks.

Working with a generic model, the researchers were able to plug in different supplies &mdash medicines, hospital beds — and run simulations in order to predict the relative success or failure of each resource allocation scenario. Although the stakes were much higher in Vinze and Santanam's life-or-death calculations, what they were modeling was similar to the planning that businesses do to ensure proper levels of inventory.

It is a specialized supply chain — an information supply chain. Vinze defines an information supply chain as "the collection of information and communication technologies providing a secure integrated decisional environment that enables business partners to collectively sense and respond to opportunities and challenges in a networked eco-system."

"What our simulation allows is for policy makers to look at this and experiment to see, 'Is that the best way of doing this? What is the implication of this? Is there a better way of doing this? Are there other regional cooperation policies that work better?'" says Vinze. Actually running the simulations for the Maricopa County Department of Public Health — which serves the 3.8 million residents who live in and around Phoenix, Arizona — netted results that Santanam says are not necessarily intuitive.

For example, in addressing bioterrorism and epidemics, the CDC has decided to stockpile antiviral medication in a central location rather than disperse it to states in advance of a threat. This is actually in line with what Vinze says is often seen as a best practice in supply chain management: the longer one waits to commit to a decision, the more flexible and responsive it makes the organization.

But the simulations show that in many cases, the idea of a central stockpile is not as efficient in saving lives once the disease spreads. By dispersing resources before they are needed, Vinze and Santanam found that localities could get resources faster by pooling among themselves.

In such cases, hospitals in Phoenix could network with those in Tucson or California, rather than always having to go to a central depot that might be hundreds of miles and days away. Indeed, Vinze says that an optimum dispersion was one where the CDC proactively distributed much of its stockpile throughout the country but held some in reserve.

In effect, this was the best of both worlds: local jurisdictions could help one another out by sharing supplies and resources and doing so quickly, but the CDC would still have some resources that would act like cavalry to bolster areas hit by a health crisis.

Corporate supplies vs. emergency situations

On the surface, getting the right amount of medicines to the right place seems little different than shipping the right inventory of shoes and shirts from a warehouse to retail locations. However, trying to model a supply chain for public health has been an eye-opening exercise for Vinze and Santanam.

While large corporations that are heavily dependent on their logistics — Wal-Mart or FedEx, for example — must have incredibly complex supply chain models and simulations, the logistics of a public health catastrophe are far more complex, says Santanam.

Individual companies can operate within a system that is more or less closed; while outside forces like traffic or weather can affect shipments, the company is largely concerned with its own resources in a warehouse, on trucks or on store shelves. Also, a typical corporate supply chain is based on long-term relationships and is nurtured and managed on a continuing basis with the thought that it will perpetuate and improve gradually over the course of years.

"Emergency response, on the other hand, requires collaboration to happen 'on the fly' and on an 'on-demand' basis. This is a much more difficult proposition," says Santanam. Emergency supply chains that must come into being all of a sudden to handle a threat are, by their very nature, disconnected and open to many more permutations than individual corporate supply chains.

As opposed to having solid trading relationships with customers and suppliers, the CDC exerts significantly less control over state and local health departments, police precincts, fire and EMS crews, hospitals, medical labs and pharmaceutical companies.

So-called "bundling issues" also come into play as a disease outbreak or terrorism requires not just having enough medicine but also having enough medical personnel to properly administer it. On top of those issues, the CDC would still face the same run-of-the-mill hiccups such as weather and traffic that FedEx faces.

And, of course, the whole equation is made more difficult as the disaster itself would compound the chaos. "When you take all of them and start piecing them together, what we found was this was a very complex problem, not something you can just analyze in a single dimension," says Vinze.

Ramifications of multidimensional supply chains

Due to the collaboration needs and the greater level of interconnectedness, Santanam and Vinze's supply chain model begins to look less linear and more like a supply ecosystem. Although the two have only developed an early stage engine to better allocate resources, it could be a first step in adapting the supply chain to incorporate more complexities and better respond to unexpected demand surges and supply disruptions.

That prospect has IBM interested in how this nascent engine could help the company adapt its products to better fight viruses — the computer kind. Vinze and Santanam are now working with researchers at the company's Thomas J. Watson Research Center to apply their findings about real virus outbreaks to autonomic computing.

With autonomic computing, computers take on the job of keeping themselves up and running by fixing viruses, adding storage capacity and balancing which servers handle traffic during normal operations and during spikes in demand.

It turns out that trying to effectively share the burden of a dramatic demand surge between neighboring hospitals is applicable to sharing demand when, for example, a web site's visitation spikes and the servers become overwhelmed.

The business models Vinze and Santanam developed for emergency preparedness are going back to work in the corporate world. "It's far more complicated than what typical business problems are: from a collaboration perspective, from a logistics perspective, from an information-exchange perspective, from multiple stakeholders," says Vinze.

"It's a really messy problem and one of the advantages is that if you can play with this problem, two things happen: public health benefits from the application of business practices in driving policy and, on the other side, business understands that there are other complications that we typically don't take into account that we might start taking into account."

Bottom Line:

• Business best practices can benefit society outside of the for-profit world including applications in public health and emergency preparation.
• Getting the right resources to the right place and doing so expeditiously is far more complex when dealing with a public health emergency like a smallpox outbreak than companies face when trying to get their goods or services to the right place on time.
• Due to the multiple dimensions that go into an emergency response supply chain, it appears more like a "supply environment" than a linear "chain."
• Although best practices in supply chain management dictate that it is best to hold off as long as possible in committing to a decision — such as where to send supplies — dispersing supplies before there is a disease outbreak or bioterrorism and then advocating that localities collaborate may be more efficient, coupled with a small centralized stockpile.
• How people respond to the need to suddenly and unexpectedly address public health demands has parallels to how computer networks are trained to respond to viruses and traffic spikes. Better understanding the former may improve the latter.