Good questions (and good answers) come up when sourcing experts get together to talk about reducing waste and variation in the healthcare supply chain. When Tony Gribble and Joe Jackson met up in this webinar, the audience took advantage of their collective expertise to get real answers to real questions.
Tony Gribble (VP, Supply Chain, Vanderbilt Health Purchasing Collaborative) joined Joe Jackson (VP, Customer Experience, Curvo) for an insightful session on how clean, organized data enables smarter, faster, and more cost-effective decisions across large supply chains.
If you have questions about how to leverage data for stronger supply chain performance, this is your blog. See what your peers learned in this Q&A about “Reducing Waste & Variation with Data-Driven Supply Chain Strategies.”
Read for a deeper understanding of how to implement sourcing practices that reduce waste and control costs. Then watch the rest of the webinar for best practices and the VHPC value journey. (Edited for clarity and length.)
TG - There are lots of different things that make it much more difficult if you are still doing data normalizing from a manual perspective. The quick answer is the data sets or data points that come in are so different from every different organization. Sometimes what you get from the hospital doesn't use the same name as the distributor. Sometimes they truncate the item that comes over.
Also the repeatable function of that task. Being able to repeat the same types of steps over and over accurately, and doing the same thing every time is the key.
The AI learning from the Curvo platform also finds something that's off. We can identify that, and then scrub it. It also goes back to the AI to learn, and then we don't see those issues anymore.
So just the amount of disparate data and having different formats and functions and features, being able to normalize that data is so key to what we do from a day-to-day perspective,
The most difficult part as new members come on is there's always some slight variation to their naming conventions and their supplier names. Our human analysts or AI have already been trained to look for and identify a lot of that, but there's always that one variation that might pop through.
So that's the greatest challenge. As more members come on, probably 95% of those data anomalies we've seen before. It's capturing the other 5% and making sure they're being caught and brought up as quickly as we can on those edge-case variations.
TG - It's hard to answer. The way I look at data from an overall operations perspective is different from how our sourcing and contracting team looks at it compared to what our account executives look at. And I think that’s where the Curvo relationship helps us day-to-day.
It's not just a single view that you have. With the visualizations, you can customize those views for what you're looking to achieve, depending on what your role is within the organization. I may just look at it from a high level. I want to know:
I want to build work plans where the sourcing team is going to look more individually at the contract cycle itself and the sourcing itself.
However, the analytics team, they're doing a whole cost study. They're worried about match rates and types of spend that's missing. They're looking at different things. And account executives are looking at how we are performing from a day-to-day perspective.
All of that combined is what makes it a good partnership. It's not just one piece of the puzzle that was solved. It was a large puzzle trying to integrate all this data to make sense of and to visualize, but then to departmentalize that from a data perspective.
We meet with Curvo on a monthly basis as a team and as different departments to continue the training and evolution of what we do from a partnership perspective. Also real time communication channels are open to solve and rectify these types of situations as well.
So it's not just the AI learning that goes on here. It is the personal learning back and forth between both entities that makes the strategic partnership work like it does today.
TG - There are plenty of safeguards in place from a Curvo and the VHPC perspective. Any information that's provided back to a member hospital takes confidentiality into consideration and what they can see from an overall perspective.
True confidentiality goes into what we provide back out to members. We do not share pricing. We do not share hospital names. We do not share physician names. Any type of PHI that we happen to collect is de-identified for those purposes. Within the platform, only the VHBC team has access to all the data sets, with a firewall between members and the data that we share back to them.
So not only within the membership are we making sure we're not sharing confidential information, but even outside of the VHPC family. All that data is blinded, so the team can't see individual hospital prices and start to cherry pick.
All that data is de-identified, and the benchmarks are stratified at the top decile, top quartile, median and lower quartile. We have those four different benchmarks that are still directionally helpful as to what you want to target in your sourcing initiative. But again, all that benchmark information is blinded, and so confidentiality is something that we take very seriously.
TG - That's a million dollar question. You've got to be very strategic and calculated about how you communicate the processes and procedures and what you're doing. You've got to make sure, especially if you're going through a bid, to get the proper stakeholders together up front and on the same page.
No one particular entity within those committees or structures has more importance over the other. When going out to bid for a contracting cycle, they all must have input into that process and making changes. If you make a change, that affects those doing the actual work from a supply chain perspective, because there's a cost associated with that.
You have to get stakeholders up and going as soon as possible before doing a contracting event. Make sure you have measurable data points and communication points to make these types of RFPs or clinical decisions have the output you want.
TG - The rebate tracking piece is something we've just launched over the last month or so. We do load all of our portfolio contracts within the Curvo platform. If we need to load local agreements, we can do that as well. I do know that some of the members track rebates on their own.
Right now, if it's a contract that VHPC holds the paper on or it's one of our contracts that has a rebate, that’s being tracked through the Curvo platform. If a member tells us they have a rebate on CRM, and this is what it is, we've got a way to build that in as well. That's the short answer from a rebate perspective.
TG - Great question, especially in today’s times. This is a real-life situation of going through a larger exercise due to tariff impacts and other market impacts. We’re trying to first understand: did we buy something because of a back order information? That's where utilization comes in and talking to our members.
The second part is understanding from a total membership perspective where we have some standardization and where we can actually further that based on outlying spend.
Other times things come out [of the data] and show if they would move from a particular item, then we could reduce the variation. We could go back to the supplier, and we actually could reduce the cost, because we're adding more spend to that particular category.
So it's an ongoing thing. It's not easy, especially when you're looking at variation from an entire membership perspective, but it's something we do work on with our members specifically.
Does your sourcing team struggle with spreadsheets, emails and manual workflow? And don’t forget all the disjointed data.
If the topics covered in the Q&A caught your attention, you’ll want to view the full webinar for more ideas and advice. Download and watch on demand and start bringing tough sourcing challenges like waste and variation under control.
Reach out with any questions that weren’t covered in the session.