S4E5: Improving Throughput & Capacity with DIY Metrics (ft. Matthew Massaro, Northwestern Medicine)

Jordan Cooper   1:46
Yeah, I understand that on on a few different metrics you work through financials, you mentioned staff productivity and patient volumes. Within that you kind of break it out into labs and infusions and visits. Can you talk about how you are parsing out your different?Data sources and how you're able to look at them, present them and analyze them in a meaningful way.

Massaro, Matthew   2:12
Yeah, there's a variety of different sources that we can pull data from in our system. We have data warehouses that exist, some that are directly embedded within our EMR system, and then some that exist outside of the EMR system but are able to tap in and pull information.Out and really the gamut of anything that we can think of as far as data can be helpful to us. Some of the data, depending on how granular we get, can be a little trickier to pull, a little more exhaustive because we want to make sure that what we're getting is valid, is true.We'll use data in terms of our patient volumes, in terms of different diagnoses that we treat, medications that we give. Infusion is a big area for us in oncology. So understanding the volume of patients who necessitate the variety of different medications that we that we give in an infusion space.And also correlating that to how long do these infusions take? Is this an infusion that's 30 to 45 minutes or is this an infusion that's two to five hours? And then understanding the quanti quantifying those those volumes gives us a sense of what capacity we have in our operations.

Jordan Cooper   3:19
Mm-hmm.

Massaro, Matthew   3:29
Also aligning with our staffing models and how many individuals we need on certain teams or what types of roles. Nursing is a big one, of course, and looking at patient hours per nurse. So it's a it's a pretty wide spectrum of data that we look at and it's a little bit more finite, but.Spectrum of sources for the data, where we're pulling those sources and and putting that all together, yeah.

Jordan Cooper   3:57
So Matthew, once you can quantify the volumes, for example, figuring out how long an infusion is taking, how do you take action? What's kind of intervention can follow understanding how long an infusion takes if if it if it actually takes about 40 minutes and you would.It would take 50 minutes. Are you able to add an additional patient each day? Are you able to add more staffing? Just walk us through how exactly once you see what your KPIs are and you define how long an infusion might take, then how does that?Affect operations and allow you to increase throughput and capacity.

Massaro, Matthew   4:40
Sure. It certainly depends on the lens of which we're looking through and why we're looking at the data. What is the problem that we're looking to solve to? What are we trying to understand about ourselves? So for example, with infusion, as we continue to to grow, we have certain areas where.

Jordan Cooper   4:47
Mhm.

Massaro, Matthew   4:59
Recent renovations and expansion projects have happened where we build out the capacity so that we have it and then we look at OK, how much if we have a chair or two that we don't utilize throughout certain days of the week, if we were to open that chair, how do we make sure that as we bring in?

Jordan Cooper   5:06
I.

Massaro, Matthew   5:18
More appointments. If we add more appointments and patients to be seen, how do we ensure that we're not going to overload what we have for for capacity? So we'll look at the average of in a given area, what is the percentage of the patient population that receives XY and Z treatments and then looking at.The length of time that those treatments make. So it's all about forecasting for us to understand what is that future state going to be. Let's go ahead and expand. Let's add a couple more chairs. Now when we do that, what's that going to look like in the clinic for the physicians? How many more?

Jordan Cooper   5:56
Mhm.

Massaro, Matthew   5:58
Questions would be open? How many more slots on a given day would we would we open? And that's kind of what we're piecing those two things together to ensure that we have a, you know, a stable environment.

Jordan Cooper   6:09
So in order to forecast those future estates, I understand you've built methods and models of calculating staffing and capacity needs. Can you speak about how those models have been constructed?

Massaro, Matthew   6:21
Yeah, yeah, it's so in in some of the areas where we more recently constructed some models, it's it's really on the clinic side. So it's outside of the infusion space. So we have a model for navigation for new patient navigation. We use nurse navigators to to help connect with new patients.Who are newly diagnosed and they call our our centers and within 24 hours they're on the phone with a nurse navigator, an embedded ally within our within our practice who provides upfront education because sometimes they might get their diagnosis, they call our facilities.And if their appointment is a few days or a week out leading up to that time, we want to be sure that we're able to connect with the patient, meet them where they're at and give them some support and some education so that when they have that first appointment in the exam room, it is a very efficient and effective visit.

Jordan Cooper   7:04
Mhm.Mhm.

Massaro, Matthew   7:19
And so that that involves getting the patient ready for it as well as our our clinical practice teams. So how we modeled that to your question, how we built quantified the capacity we look at for new patient navigation, we look at how many new appointments are we completing per week. So if the idea is we want.

Jordan Cooper   7:27
OK.

Massaro, Matthew   7:39
Want to get on the phone and have interactions at least a 60 minute call with every new patient before they even touch down in the clinic. We need to understand how many new patients are coming in per week and how long do they exist in navigation, meaning how long is it from the first point of contact to when they first see the physician.That period of time is easy to quantify because we understand all those metrics. Beyond that, there's a little bit of work that needs to be done to get deeper into the granular, like a time study where we'll have teams, we'll sit with our teams and.Quantify how much work is being done on each patient. They get off the phone. We understand that it's a telemedicine appointment. It's easy to quantify that that's a 40 minute call or a 60 minute call. But beyond that, does the patient reach back out over that next week? What are they reaching out about? Are there records that need to be gotten from the administrative team? Do they call back?With symptoms that have come up that are necessitating further phone calls. So we do time studies where we'll sit in with our teams and quantify how many extra minutes beyond the initial call is every individual patient having the nurse be accountable toward so that we can take that if every single patient before they have their first appointment.Requires what we find is about 120 minutes, two hours of time for the nurse, for the nurse navigator. What we do is then multiply that by the amount of new patients that we see every week and that gives us a sense of how much work is to be done in every week for a nurse navigator. How many minutes, how many?

Jordan Cooper   9:07
C.Mm-hmm.

Massaro, Matthew   9:16
How many hours in a week are they actively navigating, not going to a nurse meeting, not going to an education session, not meeting with their managers, not doing these other ancillary, but actively navigating our patients. And when we look at our model to say, well, how much do we need, we use these metrics.

Jordan Cooper   9:28
3.Mm-hmm.

Massaro, Matthew   9:36
It's a few different sources of metrics and how we put them together to understand what's the workload if we're going to say.

Jordan Cooper   9:42
Sir, are you, Matthew, are you and your team actively building your own new de Novo process measures? Or are these things that you're lifting from the National Quality Forum or from the American Medical Association or other kind of associations that standardize metrics? Or are you, are you developing your own measures?2.

Massaro, Matthew   10:01
You know, it's a little bit of both. It's almost like sourcing different recipes and figuring out what are the ingredients that kind of work for our area. There are certainly standards. Nurse navigation isn't something new that we've that we've developed in the idea of a nurse navigator connecting with a patient prior to their care.Isn't isn't this new novel concept that exists only in our area, but the way we quantify and leverage it is is unique to us. So how we come up with our calculations is.Is the part that I would say is specific and unique and homegrown where we are, but the concept of what we're trying to deliver is more of a kind of a best practice and I would say nationally recognized standard of practice.

Jordan Cooper   10:42
OK.I see you mentioned that you alluded to the topic of data integrity. You mentioned that you want to ensure that your data sources for these measures are valid and true. You mentioned can you can you delve into how you're solving for how to create accessible data and reliable data?

Massaro, Matthew   11:03
Mhm.Yeah.Sure. Yeah, that's that's a very good question. So a lot of it, what I've found is how accurate the data is from the system that's pulling it is often dependent on the users and how they're interacting with the platform that the data is being pulled from. So for example.

Jordan Cooper   11:10
Yeah.

Massaro, Matthew   11:29As practices exist for a while in the same EMR system, there'll be different codes that mean the same thing. Let's take a new patient appointment over time. As practices evolve, let's say they move physically from one space to another, which is not uncommon. That happens a lot, especially here in Chicago, where I'm at, where we have a finite amount of.

Jordan Cooper   11:35
Mhm.Hmm.

Massaro, Matthew   11:49
And we're always re tinkering existing space and moving it. When that happens over time, there might be 3 different new patient appointments that exist and they're coded differently, but they mean the same thing. They might be worded slightly differently.Or they use a different code and so some schedulers might come in and schedule a new patient appointment under one particular code, same slot as any other new patient would use, whereas other schedulers happen to pick a different code that means the same thing to the templates and getting scheduled in. But when you go on the.

Jordan Cooper   12:14
Mhm.

Massaro, Matthew   12:26
Back end and you pull the data and you ask the data and you filter down the search criteria of what they're pulling. You need to make sure that all different visit types are being accounted for. And if there's three different types of new visits, you need to make sure all three are being accounted for. What we do on the back end then is work with our teams to standardize and say.

Jordan Cooper   12:28
Mhm.

Massaro, Matthew   12:46
We're going to go with one. We know that some of you use this one, some use this one. We're going to define it, winnow it down to one existing. We're going to remove the other ones from the system and then we're just going to have the one. So when we pull the data, we know it's true. Let's say a new if I if I'm the one pulling the data and I and I'm no longer in my position and somebody else.Into this position, they're likely not going to have that deep rooted context that, Oh yeah, let me pull up all of those visit types that correlate with the new appointment. Let me make sure I'm accounting for all of those. So we try and winnow that down, but I scrutinize the data. We have a couple different databases here and I like to pull the same.

Jordan Cooper   13:11
Hmm.

Massaro, Matthew   13:25
Same information from both of them to make sure are they telling me the same thing?

Jordan Cooper   13:29
So Matthew, I understand from what you just said that there there a lot of your solutions to data integrity are governing standards, meaning that you're going to say at Northwestern, though there may be 3 codes for doing X, we're going to eliminate two of those codes that there's only one to simplify data analytics on the back end.I'm wondering if there are any technological challenges to normalizing, aggregating and deduplicating data in order to ensure that you have reliable data when you're trying to develop your own process metrics.

Massaro, Matthew   14:02
Yeah, there's a few. When you're in a system as large as we are, you might locally determine that how the system is set up, the data analytic system, the data poll system, how it's set up might need to be modified to suit what we locally within.Our disease group need. However, if we want to make a change, one of the tricky things is if this change is embedded within the system, within the data warehouse system, is that going to then create a ripple effect and impact other areas in the way that they're set up? Is the change that we're making?

Jordan Cooper   14:23
Mm-hmm.Yes.

Massaro, Matthew   14:41Not possible to be made just for our team and is it if that gets retooled, is that going to change it for everyone? Um, so that can be something that we always, you know, uh, that could be a roadblock and something we always want to be mindful for. Uh, the other is.

Jordan Cooper   14:42
2.Mhm.Mhm.

Massaro, Matthew   14:57
There's always there's there's constantly new systems that are being developed and as a new platform comes to fruition, if it looks promising and we believe in it, say locally, but it doesn't exist yet in our system because.The information and the health medical records are so secured in order to get a new system to have access to our medical record system, say like a data system to be able to plug into it. We got to make sure that everything's vetted if we're going to have a new vendor.So to speak, um, it's a it's a little more difficult to just kind of get that set up because we're such a large organization.

Jordan Cooper   15:37
So Matthew, a lot of our listeners are, you know, I think a lot of what you're saying will resonate with them. Many of them are actively migrating to the cloud. I believe this is something in Northwestern Medicine has also begun doing and many of them are involved in active merger and acquisitions activity. I'm wondering.As we.We kind of approach the end of this podcast episode. Just a few more questions, but can you delve into the implications of migrating to the cloud or the implications of acquiring other practices, other hospitals, other entities on data reliability? How do you handle those data issues?That arise from cloud migrations and M and A activity in order to ensure that your calculations are based on accurate data when you're calculating those metrics.

Massaro, Matthew   16:28
Yeah, I mean the integrity of the accuracy, if you're going to be pulling in data and hosting it to say this is what that number is, whatever, whatever it is, you're going to be migrating that to the cloud, I think before it moves to the cloud.There needs to be kind of a team centered around scrutinizing that data to make sure it's accurate because once it moves to the cloud, then the assumption is it's going to be there. It's going to be there for a while. It's going to be accessible for to people from different avenues and they'll want to believe that it's true. So I think it's really important to just make sure that.Everything is scrutinized before it migrates over to the cloud. You know, as far as the cloud, I think there's a lot of cyber security is obviously very important so that when that information goes into the cloud, one, it can't just be viewed by anyone, but just as importantly, it can't be manipulated.Once it's there, it should stay there. That's a, you know, in my experiences, I think healthcare has been as far as the technology of data and sourcing data and quantifying it. Healthcare has been a difficult sector to stay up to date and stay current with what the.

Jordan Cooper   17:26
Mhm.

Massaro, Matthew   17:45
At most advanced height of data analytics is, and I think that has to do with, you know, largely has to do with how secure the information needs to be, yeah.

Jordan Cooper   17:57
Yeah. So as we approach the end of this episode, I'd like to pose this final question to you. You know, you've been working for quite some time. You know, you're the program director of oncology and you've been working on throughput and capacity within the outpatient oncology setting.You've been working on developing these different measures, building models, calculating staffing capacity needs. What advice would you give either to yourself a few years ago as you were earlier in the process of developing and managing these models and measures?Or to a listener who hasn't yet begun to really go down this path. What advice would you give?

Massaro, Matthew   18:38
Um.I would say as best as you can work to standardize how you're organizing your data. So what I do is I'll pull information, but then I like to store a lot of it in my secured files that I have.

Jordan Cooper   18:46
And.

Massaro, Matthew   18:58
Say in in in my drive and it's important that I'm consistent with how I present the data to myself and to do that consistently over time and making sure that there's a system for it. Sometimes I can get lost in it and I swim in it and then I reinvent how I present this data to myself or I reinvent how I analyze it.And I think staying consistent with with what your approach is and how you wrap your mind around it is very important. Sometimes you don't look at data, certain data as frequently as others. So let's say there's data that I might pull once or twice a year. It's really important that I can go back to the last time I I pulled that data and look at.OK, this is how I organized it. Here's the pieces that when I pulled the report that I parsed out. I took these columns out because we don't need them. They're not as relevant. They don't move the needle in our decision making and then keeping everything consistent and organized. I think an organized file system is massively important for data. And how do you label your files and how do you?You go back and know right away when you're looking into a into a folder, OK, that's what that was. That's that time period. That's what that was telling me. Early on, I would just kind of label things a lot very similarly and then I'd open up a file and over time, over a couple of years, it would just be this massive list and I'm clicking in three different.Excel documents, four different Excel documents before I actually find the one that I needed and my head is swimming in all of these numbers and all of these cells and columns. So I'd say like your own organization system, figure that out, define it rigidly and stick with it.

Jordan Cooper   20:28
It may not be sexy, but in the pursuit of profit, vis-a-vis improved throughput and capacity in the outpatient oncology setting at Northwestern Medicine, it sounds like a a efficient and standardized method of organizing data is the best path forward. So Matthew.I'd like to thank you for joining us today.

Massaro, Matthew   20:51
Thank you, Jordan. Thank you so much for having me.