Chapter 06: Directing Strategic Business Initiatives at UPMC
Files
Description
In this chapter, Ms. Kaul describes her work as Senior Director of Strategic Business Initiatives in the International Commercial Services Division at UPMC in Pittsburgh, Pennsylvania. This was a new, entrepreneurial venture for UPMC, and it involved creating opportunities to work with other companies to develop and sell technology. She explains one project that involved developing computer-assisted medical coding. She explains why this is a challenge and why the technology is valuable in healthcare. She managed the eventual joint venture between UPMC and A-Life Medical and she tells the story of how the project evolved and explains the lessons learned.
Identifier
KaulR_02_20160425_C06
Publication Date
4-25-2016
Publisher
The Making Cancer History® Voices Oral History Collection, The University of Texas MD Anderson Cancer Center
City
Houston, Texas
Interview Session
Topics Covered
The Interview Subject's Story - Professional Path; Finance, Entrepreneur, Biotechnology; Professional Practice; The Professional at Work; Discovery, Creativity and Innovation; Discovery and Success; Technology and R&D; Fiscal Realities in Healthcare; The Healthcare Industry
Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 License.
Disciplines
History of Science, Technology, and Medicine | Oncology | Oral History
Transcript
T. A. Rosolowski, PhD:
OK, and the counter is moving. And it is about 11 minutes after 2:00, and today is the 25th of April, 2016. And I'm in the mid-campus building on the ninth floor, interviewing Rebecca Kaul. This is our second session together. So thank you for making the time. I know that we had to bump this, because you've been so, so busy, which is good news.
R. Kaul, MBA:
Yep.
T. A. Rosolowski, PhD:
And we strategized just a little bit before, just a reminder that we had ended up our conversation last time talking about your international experience, and mentioned the fact that you had moved back to Pittsburgh for some family reasons, but began on kind of a new phase in your career, of working with UPMC. So I wanted to kind of hit the high points of that work, and how it prepared you, really, for the next step, which was to come here.
R. Kaul, MBA:
OK. So when I started at UPMC, I was working in a, what was a new division at the time. It was called Strategic Business Initiatives, and it eventually got renamed to be International Commercial Services Division. It was born out of a past experience, UPMC had, prior to me arriving, whereby UPMC had developed a PACS system.
T. A. Rosolowski, PhD:
And what is that?
R. Kaul, MBA:
Picture archiving and communication system, basically, for medical images, and sold that technology, or that company, got investor funding, formed a company. The company got moved to California from Pittsburgh, and eventually it was sold to Phillips. And it is what it is today, the Phillips iSite system. That experience for UPMC kind of gave UPMC a taste of entrepreneurship, but in an unsatisfying way. And the reason is, is it was technology that was solely developed at UPMC, using the expertise and knowledge. But by the time there was an exit, a financial exit, UPMC had been diluted down to, a minority share of the company. So though there was a large financial exit, it could have been a lot larger. So the notion was --
T. A. Rosolowski, PhD:
0:02.56.4 So excuse me for sec, so was the concern the financial outcome? Or was there also a reputation and international property stamp thing?
R. Kaul, MBA:
There was -- the concern was financial and community, which was, we let the company get moved from Pittsburgh, and UPMC has somewhat of a community mission to build the Pittsburgh community and bring jobs to Pittsburgh. And so when you have a successful startup company, getting moved to California, it's not satisfying, it's sort of anti-mission. But UPMC had lost control of the company, being a minority shareholder, and that was not in their control. And secondly, the financial return, being a minority shareholder when there was a large exit, when it was something that was solely our creation, the question was, if venture can bring money to the table, and that's what they're bringing -- they're not bringing necessarily the intellectual asset -- why wouldn't we bring money to the table and wholly own it, keep it in Pittsburgh and still have the same outcome? So that was kind of the hypothesis, going into Strategic Business Initiatives, which was, can we look at opportunities, technology opportunities inside the house and spin them out using our own funds and be successful and create a diversified revenue stream, and/or can we take technologies form the outside that are emerging markets and leverage our clinical and intellectual capabilities, as both a subject matter expertise, test bed and reference site to accelerate the development and success of these companies that we would be bringing in, and providing the kind of co-development contribution and test bed and reference that contribution, could we, share in the economic returns of those companies, whether it would be forming joint ventures, whether it would be taking a minority share or a majority share in a company, or just participating in a shared royalty stream, right? Thirdly, it would offer us an opportunity as a customer to be leveraging technologies for free. You know, from not necessarily paying for them, because we're part of the creation, or part of the co-creation of the technology. So it again creates a different kind of economic model for the organization, and an diverse revenue stream.
T. A. Rosolowski, PhD:
Does it also do something to the culture?
R. Kaul, MBA:
I think it creates more of an innovative and entrepreneurial culture to the institution.
T. A. Rosolowski, PhD:
Yeah, I was curious about that.
R. Kaul, MBA:
You know. So it started small. There was just a couple of us looking at opportunities, essentially creating your pipeline. People internally will come to us with ideas, people externally, we would look at the external markets, we would study the markets, figure out what are kind of interesting places to invest and talk to external people, right? And so I was on the team, there was maybe three or four of us. And --
T. A. Rosolowski, PhD:
0:06.48.8 Did you all have similar backgrounds? Or what was the profile of the group?
R. Kaul, MBA:
We had different backgrounds. I was sort of the person who understood kind of the technology and the technology market, and the other individuals were more finance or investment banking types. So, I was sort of looking more at the operation business and technology side of these opportunities, not necessarily doing the financial diligence. And we had, of course, attorneys that were doing the deals.
T. A. Rosolowski, PhD:
Now, I'm sorry, I just want to make sure I mention for the record that you took on that role starting in 2006, is that correct?
R. Kaul, MBA:
Yes.
T. A. Rosolowski, PhD:
OK, OK, just to place it for the record.
R. Kaul, MBA:
So as I was looking at opportunities, I came across one that I thought was really interesting. And a company had come to us, and I was looking at it, it was in the computer-assisted medical coding space. They had developed a technology, leveraging natural language processing to automate, in part, the coding, the medical coding process. So basically, using technology to read the dictated notes that physicians create, and codify them into the ICD-9 coding system, which has traditionally been a manual effort by a set of medical coders. So I thought that the use of natural language processing in this space was interesting, and had promise for the future. So it wasn't necessarily coding that interests me, but it was the use of this natural language processing technology, coding being a very good space to start. If you look at the market of natural language processing, you can see that the companies that try to be kind of generic and kind of everything, kind of do everything medical failed, because natural language processing really is a technology that requires training.
T. A. Rosolowski, PhD:
Now, when you say "natural language processing," what are you referring to, exactly?
R. Kaul, MBA:
It's the ability for the computer to read blobs of text, meaning sentences, and structure it into data. So basically think about it reading a paragraph and being able to pull out facts and structure facts out of that.
T. A. Rosolowski, PhD:
So you're not talking about voice recognition, you're talking about text?
R. Kaul, MBA:
No. I'm talking about text. So I think 80 percent of medical data is actually unstructured, it's in these blobs of text that physicians dictate, right? And it intrigued me, the notion of how do we get meaningful information from these blobs of text and be able to do large-scale analytics in the future? But you need to start somewhere, and you need to start in a place that's low-risk; you don't want to start with trying to do that at a point of care standpoint, because you don't want to put patients at risk. So the best way to start to test and train these type of technologies, because these technologies are built around --the more they're used, the more data that it processes, the smarter they get, right? So the best way to kind of start to introduce medical vocabulary, medical terminology, medical data and training these systems is just starting a space that's relatively structure, like coding, so there's a finite set of codes and in a space that's low-risk, where it's not going to impact patient care, it affects billing essentially, which, we're not going to hurt anybody by that, and as long as you're having a human being in the mix to verify it before it goes out the door, then you're not going to do anything illegal, either. So you would never let the technology do everything on its own. You would always have someone review the results of the technology, like an auditor. So but interestingly, I wasn't interested in the company that came to us. I was interested in the technology and the concept, but I didn't think that the company was that great, and the technology that this specific company had to bear was interesting. So I went in on my own, kind of analyzed the market and found a couple of different alternative companies that I thought were viable. So I liked the problem space, and I liked the concept of using natural language processing for computer-assisted coding. But I was kind of looking more broadly in the marketing at the different companies that could solve this problem. And I found a company and ultimately vetted it, and we decided to do a 50-50 joint venture with the company whereby we would take their core technology, which was used for ambulatory coding and apply it to the hospital space, kind of doing your in-patient and hospital out-patient coding, which is a significantly more complex field, because you're going from a single document, single specialty in a single point in time, a single episode, to an in-patient which is many documents, many physicians, many diagnoses over a period of time where the prognosis could change. So when someone's admitted to a hospital, they could be in the hospital for 30 days, be interfaced with multiple specialties, as you're kind of figuring out what the problem is, and there may be multiple -- kind of comorbidities and things like that happening to this patient. The clinical perspective on the patient could change, it could be different on Day One than it is on Day Fifteen. So asking a computer to reconcile all of that is significantly more complex than the ambulatory space and asking it to reconcile a single note for a single specialty, with a single doctor. So for me, the extending this technology a little further was sort of a logical next step for the technology, to see how far we could start to push it.
T. A. Rosolowski, PhD:
What did you envision in the future for this? I mean, because I have a feeling this was very much part of a strategic plan for how this would grow.
R. Kaul, MBA:
Right, exactly. To me, this was just a stepping stone, that if you could get technology like this to truly parse the medical record, parse all of these notes, you can start to do analytics that start to inform us on how to drive better patient outcomes. And potentially how to better treat patients. There's a lot of different things you could do with this that are kind of broader and more impactful, but, with innovation, you have to start somewhere. If we had started there, that would have been very risky, number one. And it would have been hard to measure whether it was even working, because an outcome can only be measured over watching a patient over a prolonged period of time, to some extent, to see how they, you know. And collecting the baselines is very difficult as well. Something like coding is very tangible, very defined. So when you start to test a technology that's imperfect, starting in a space that you can define is helpful. Also, from a business perspective, being able to create a business out of this, you need a defined space. You need a space where you can define a value proposition that is quantifiable, and that will speak to an organization, which it's very difficult to sell into the healthcare market. You need to be able to be planned as part of the budget process, which means you need to show that this is something people want to put in their budget and is worth buying. Something like medical coding is really easy to justify this technology, because you can show that it's going to create revenue that drops right to the bottom line of the organization. And we were ultimately able to show when we developed this product a $21 million increase in revenue over the course of the year for the institution, through use of this product. Something like that makes it really easy to sell it to the market, and makes a successful business. And what's great is if you start with a simple use case like this, that creates a value proposition that's very tangible, makes it easy for a CFO to sign the check, once the infrastructure is in the institution and you're feeding data into this infrastructure, it's significantly easier to upsell on the next and the next and next, because you already are processing the data, and you're already in. So strategically, when you go to market with a solution in the technology space, the hardest part is getting in. Once you're in, you can continue to do a lot more. So getting in with a tangible value proposition and a tangible product is easier. It's hard to sell a big vision. It's easy to sell a tangle set of value.
T. A. Rosolowski, PhD:
Can I interrupt you just for a second, just to pick up two details. One is the name of the company that you went into this 50-50 partnership with, can you --
R. Kaul, MBA:
It was called A-Life Medical.
T. A. Rosolowski, PhD:
A-Life Medical? OK. And the second one, second question I have is, you're using this term, "value proposition." Is that a term in the field, or is that your term?
R. Kaul, MBA:
Yeah, it's a term in the field.
T. A. Rosolowski, PhD:
OK. Yeah.
R. Kaul, MBA:
It's a typical business term.
T. A. Rosolowski, PhD:
OK, I just never --
R. Kaul, MBA:
Creating -- yeah.
T. A. Rosolowski, PhD:
I just wanted to make sure. I'd never heard that before, so...
R. Kaul, MBA:
Yeah. So we -- I was so interested in this that when we formed this joint venture, I was provided with the opportunity to go manage the joint venture, to run the joint venture. So I went ahead and did that. I built a team, a product development team.
T. A. Rosolowski, PhD:
How did you do that? What were your thoughts when you put together that team?
R. Kaul, MBA:
Essentially a team that could do commercial software development, product development, leveraging the skillset, having NLP -- natural language processing -- engineers, linguists, because a lot of that is having to do with how the English language is processed. Coders who understood the coding space, operations people who understand how to deploy these products into operations and healthcare, and commercial kind of product managers who know how to position it into the market. So, in my past, did consulting and through my education, I knew in principle how to do this, I had help from the other side of the fence at the A-Life Medical Company, who was already doing that in the ambulatory space, so there was somewhat of a framework for how to build a successful team. So I built that presence, and kind of to our lessons learned about not letting things -- this was a California-based organization, but we put the joint venture in Pittsburgh, so we built a Pittsburgh team so that we could bring jobs to Pittsburgh. And that was one of our stipulations as part of the joint venture. Again, a lesson learned from that past, you know. So I built that team, and we developed the product in 12 months, deployed it across the health system. I think we deployed it in the main -- initially on the main campus, and proved out the value proposition. So we proved that we were able to deliver $21 million of value, so we did that in a couple of different ways. We ran the system against cases that were already coded manually to see if the system coded it differently and then had the coders audit what the system did, versus what was previously done to make sure that it was legit. And we were able to show the difference, so kind of irrefutable. Then we literally rebuilt and collected the money to make sure that this followed that we could unequivocally say that it delivered this value, we took a statistically significant sample and went through this process, and then extrapolated to get the $21 million value proposition based on our volume and our payer mix so that we could go to the market with a case study that says, this is the value proposition in this kind of an institution. It was such a powerful argument. There were many suiters coming to us in the coding space that wanted to just acquire the company. But it was really a joint venture with this other company that had the ambulatory side. That company wanted to be sold, so we kind of had to have an intermediary step where we sold our joint venture back to the company, so that they could package themselves up to be sold, because it's difficult to sell a company when you have a joint venture hanging out there. So it was kind of, for them, the two-part transaction; to kind of us to sell our interests back to A-Life Medical, and then A-Life Medical to sell the entire product suite to a company. So at the 18-month mark, that transaction occurred. And we sold -- so essentially, in the total of 18 months, we developed a product, proved $21 million annually, operationally to our institution and then doubled our money in the transaction itself, and we sold our interests back.
Recommended Citation
Kaul, Rebecca MBA and Rosolowski, Tacey A. PhD, "Chapter 06: Directing Strategic Business Initiatives at UPMC" (2016). Interview Chapters. 1265.
https://openworks.mdanderson.org/mchv_interviewchapters/1265
Conditions Governing Access
Open
