James F. Jordan, a professor of healthcare and biotechnology management at Heinz College, a division of Pittsburgh’s Carnegie Mellon University, shares his perspectives on the near-term imperatives facing CFOs and materials managers in healthcare, and how advanced data analytics and artificial intelligence will need to be used to drive forward key operational advances.
Even as digitization has dramatically advanced the operational management of hospitals, clinics, and health systems, the potential remains vast for the further advancement of patient care organization operations. And one area in which a huge amount of untapped potential remains is in the supply chain and materials management arena. One expert who has been thinking a great deal about that potential is James F. Jordan, a distinguished service professor of healthcare and biotechnology management at Heinz College, a division of Pittsburgh’s Carnegie Mellon University.
Professor Jordan is also the president of StraTactic, a Pittsburgh-based strategic decision-making consulting firm, is the national co-chairman of the BIO Bootcamp (the Biotechnology Entrepreneurship Boot Camp), and the Founder of the Healthcare Data Center. He has published numerous articles and books on innovation, startups, intellectual property, and health systems. He has been a senior executive for Fortune 100 companies and has participated in and led angel and venture capital investments.
Jordan sat down recently with Healthcare Innovation Editor-in-Chief Mark Hagland to discuss the challenges facing the healthcare supply chain, and the potential offered by technology and data analytics to dramatically advance the efficiency and cost-effectiveness of hospital and health system operations. Below are excerpts from that interview.
When you think about the potential for digital transformation in healthcare over the next several years, what are your thoughts?
If you think about public health and healthcare delivery, they have two different objectives, but those objectives are aligned. Public health is something like the national defense of our healthcare system. And there’s no formal hierarchy. Ironically, the public health network is very loose. Where the hospital systems and suppliers were surprised during the COVID-19 epidemic, is that they had no sense of where the supply chain was in healthcare. And prior to COVID, there was a lack of connections. For future suppliers, we’d love to have the optics—the concept of Oracle and SAP—healthcare system-wide.
So if you go back to 2005, you saw that medical device diversion was going on, and how that was caught was that the J&Js of the world looked in and saw the number of devices being sold, and wondered where they were being used. So the concept is that you’ve got acute care and chronic care going on all the time, and the health systems don’t have projections for what is coming. The goal would be to improve processes.
You’re talking about improving the predictability of supply chain processes for care delivery purposes, correct?
Yes, and we’re talking about this in terms of products and acute care processes, but we also need to do this for clinical trials. And we did an analysis of clinical trial processes, and clinical trials were failing because they couldn’t recruit enough patients into the trials. So this ties also into drug discovery and innovation and being able to optimize that, too.
What needs to happen in the hospital and health system supply chain in the next few years?
When we think about fraud and deviation and similar issues, we have the opportunity to take blockchain into supply chain and validating things like drug administration. We need to make sure that transactions are traceable and that we know where things are at any time. That’s the first issue. The other issue has to do with contracting, such as drug price negotiations. A colleague, Jeffrey Lewis, had done an analysis and found that there was a lack of transparency around supply costs in HC. He concluded that governments and states and large health systems coming together can overcome those issues. We have group purchasing organizations and pharmacy benefit managers involved, but they’re also taking profit from the system. And that profit is happening because of lack of clarity and the ability to see prices. If you go to Carvana to buy a car, you can see prices, but it’s all hidden in healthcare.
And for example, if I’m J&J or Boston Scientific, if there’s a recall, I have to pull back product. But if I’m McKesson, I might put a McKesson label on a J&J product for my purposes. And so I could end up with four different barcodes on a product. And it’s all legitimately done for supply chain security, but in doing so, we’ve hidden the pricing. So with every good intention, there are unintended consequences.
And we had a competitor of ours who were making higher margins than we were, and we couldn’t figure out what they were doing. They would bid a contract and then change the stocking number units. So they would change the labels and charge more. This was a major publicly traded distributor—to be clear, not McKesson.
How can senior purchasing leaders make the system work better?
Larger systems are starting to create their own purchasing groups, in metro areas like Boston, for example. And they’re putting resources and analytics to it, and that’s where AI comes in. Instead of monitoring pricing and availability, AI can learn patterns. For example, that in situation with that company that was trading things out and back-ordering to make profit, AI could see that kind of move. So applying AI is key.
And I don’t know what they call it today, but in the past, they called it a back-flush system. The thought was cutting deals with major vendors to have them bid on a part of the DRG system in the hospital and risk-share. So say one organization looks at stents, which can consume up to 30-40 percent of the DRG; but the DRG is different by payer. So they’re trying to get people to price based on the DRG they’re billing, to get profits to be more consistent. So imagine if you’re the CFO of a hospital, you don’t actually know on a daily basis the true costs and prices. You’re predicting based on past reimbursement patterns. What CFOs and big supply chain people are trying to do is to get distributors to bid based on reimbursement patterns. So Patrick Flannery at UPMC told us, we’re looking to de-risk our part of it. But no one took them up on that, because the manufacturers aren’t yet under pressure to do that. But you might see that change as the government gets involved in drug price negotiations.
That will push the industry to look at costs in a predictive way?
Yes, it will. For example, lets’ say I have a $10,000 DRG for Patient 1; and for Patient 2, I have $7,000; but for the Medicare patient, I get only $5,000. For the entire procedure. So you have to figure out your margins. And my labor, nurses, doctors, operating room; those won’t change. So they’re trying to get the flexibility to say to the distributor, you will get a percentage rather than a fixed amount, for the supply items. And they will give the manufacturer 100 percent of the volume for a supply item. Initially, it’s products from Medtronic and J&J—stents, cardiac rhythm management devices, da Vinci robots, high-end items.
Could such agreements eventually be applied to the purchase of surgical supplies as well?
I think so; but I would be cautious, in that it’s complicated. You would need pretty sophisticated artificial intelligence to do that with rubber gloves, catheters, etc. And who’s doing the negotiation? The manufacturer, the distributor, both? Who negotiates that? I think that’s a second-stage execution.
But will CFOs and materials managers try to apply AI [artificial intelligence] to these calculations?
Yes, absolutely. But it’s also low-risk. And you’ve got augmentative intelligence, generative AI, and other types of flavors of AI. There’s clinical risk: did I do the right thing for the patient? The lower risk is in the administrative areas, including supply chain. If I buy too many of a supply chain item, I won’t hurt people.
How should CFOs and materials management leaders plan for this?
They should strategize around purchase-price variance, and the measurement of executing to a plan: you want to begin to track how much you’re paying for specific items. That’s the basis of purchasing today. And how do I plan better to get better ways of buying? And how do I have less committed capital to inventory? So the DuPont model is finance model that involves net income over sales times sales over assets, times assets over equity. And the sales cancel out the assets, and you get net income over equity. And for example, a drug company will have very high expenses. But McKesson will have very high net income over assets, based on high turnaround. So when you’re a supply chain manager, you want to improve net income over sales, planning and buying, and then you want to reduce the amount of product you have waiting to be used, which is sales over assets.
So I can use planning techniques and AI to get a sense of the flow of planning of my patient volumes, so I have just enough materials to handle the flow. That’s what we saw as turning out to be problematic during the pandemic. So per net income over sales: they have to improve net income through planning and acquisition techniques. And by managing my inventory more closely through planning, I can reduce my assets and improve my sales over asset ratio, and thus improve my profitability. So for example, a lot of times, inventory expires. And a lot of manufacturers might take back big-ticket items, but they won’t for smaller items. So you’re stuck with a central supply area, and now your nurses are walking a mile to get their supplies.
Will a large number of CFOs and materials managers be doing this in a few years?
One of my former students is at Mass General, and is putting data into a system called Tableau. And she and her colleagues can see the daily movements of patients and volumes and start times, and they can start looking at scheduling and planning issues. So I have planning issues, and patient flow issues. I have deviation issues, if something unforeseen happened, like the power going out. And I have variability. And if people can knock out planning, flow, and deviation, they can reach predictability. And the bigger institutions are hiring data analytics folks around this.