Hi, my name is Tom Clancy and I serve as a Clinical Professor and Assistant Dean at the University of Minnesota School of Nursing. And today, our presentation is going to focus on a new and exciting area of nursing informatics named Computational Modeling and Simulation. Computational modeling and simulation provides advanced methods to facilitate health care. And it's a combination of different tools that we use to improve processes in healthcare. To help with facility design. And in many cases, to help design healthcare products, and how well they fit into the healthcare workflow. And the Toolbox that we use, some of which are advanced statistical methods. some of them are known as knowledge management, discovery and data mining. where we create predictive models from statistical analysis. we use advanced workflow modeling where we can actually model a specific processes. Computer simulation where we can actually bring that workflow alive on the computer. And a new area called network analysis which helps us really understand how information flows throughout an organization. And together, these advanced methods can be easily integrated to help with your traditional performance improvement efforts. such as Six Sigma, Lean, FOCUS PDCA, and help determine return on investment for all sorts of new processes equipment and facility redesign. Computer modeling and simulation really provides a measure, a measure to impact new technology, facility redesign or adoption of best practices. And I have an example of what we call a computational model on the right, which you should be familiar with, if you've ever looked at a flowchart symbols. So, it doesn't really look much different, but actually it's a powerful statistical package that allows us to actually exercise what in the past has been sort of a static flowchart. It actually brings it alive by letting us run data through it. Modeling and simulation actually improves project cycle time by providing instantaneous feedback through computerized scenario analysis. We can actually develop a simulation of something. We can then compare it to the actual process and determine what the difference is. So, it can really help us to problem solve before we actually have to implement something. And there's all different areas that you can do projects both clinical and administrative. For example, in the emergency room, you may want to try to improve how quickly a patient moves through the emergency room. In the operating room, you might want to look at what impact adding two additional rooms will have on throughput. on outpatient clinics and nursing units, you might want to look at what is the impact on workflow for nurses if you add an electronic medical record. So, there's all sorts of different areas you can focus on with these new tools. Let's take one of them, computational modeling and simulation and briefly look at how we develop a model. We start with a simple flowchart. And this is just a, a series of three steps, it could be any process, it could be the medication process, it could be the process for admitting a patient. But for each of those individual steps, we collect some data. typically we collect data for how long it takes for each individual step, and we develop, rather than just averages, we, we take real data for every single event. And we develop a, what we call a probability curve. that curve might look like a bell curve or it might look skewed based upon how long and the inter-arrival rates of patients coming through and those tasks being activated. We then put those together in a model, and that becomes what we call a surrogate system. In other words, it represents the real system put on a computer. And this then allows us to take a flowchart and actually transfer it into the modeling software. As you see here, and recreate that same flowchart, but now in modeling software. And within that then, we can actually run patients through a simulation model, as you see here. And as each of those patients goes through the simulation model, we're able to actually collect data on those individual events and be able to predict if we have a change in a process, add different equipment, what the outcome is going to be, as compared to what we currently do today. So, it becomes very valuable for what we call predictive modeling. And here's the different distribution curves in each of the individual tasks that we do. And this is an example of some of the outputs from the model. And as you can see here it provides us opportunities to look at what we're currently doing and combine processes, or eliminate certain steps, eliminate rework and so forth. I want to focus on one more tool what we call Dynamic Network Analysis. And this really looks at how does information flow in your organization. And to do that, we develop a simple matrix, where we identify every person or thing that's involved with the flow of information. So, for example, let's say that we want to understand how information about quality metrics flows throughout the organization. So, we would list, for example, all of the different people involved, nurses, data analysts and so forth, physicians. And where does that information flow, when we create a matrix, as you see on this slide here. And as the information flows from one person to another, we put a 1 there. And if information doesn't flow, we put a 0 there. And from this matrix then, we can transform this information into a network. And that network then allows us to see how information flows from the bedside here, where performance metrics might be picked up through medical records, through patient interviews, and so forth. It then flows through the various nodes, these are individuals in which that information flows. And as you can see, whoops, it flows all the way up to the chief executive officer and the board of the hospital. And along the way, you can see the bigger nodes, mean that more information flows into those nodes. And that's where potential bottlenecks are. You can also see where information gets clustered, where you see a lot of nodes coming together at the same time. This information can be very helpful, then, as you try to understand how, for example, an electronic health record might help improve information flow. This diagram here shows all of the different places that a nurse has to go to collect information for a patient. And those different areas might be the medical record, the lab result, the radiology result having to go the nurse station, and each of those clusters of information are the different pieces of information the nurse has to get. And from this then, we can then see how we can try to centralize information. And we can actually measure through reports that the software provides, the degree of connectivity within the the system. and where vulnerabilities may be. Where bottlenecks are, or information is i, isolated. And we can also determine how people get information often times through workarounds. and other ways that information can be found. So, this provides you with a brief summary of two tools that we use for advanced modeling and simulation. computer simulation and network analysis.
