Inverbis recently held a webinar with DataSalus. In this learning space, experts spoke about process mining’s applications in the health field and delved into its advantages, possible areas for improvement, and the ways in which this solution can optimize processes in health management.
To begin, Gonzalo Martín, CEO of Inverbis, and John Wallace, head of business development, briefly explained what process mining is and what advantages it offers in terms of optimizing and improving processes in the healthcare industry.
An introduction to process mining
During his speech, Wallace affirmed that process mining is a new field that opens many possibilities for analyzing aspects of a company in depth.
He began his speech by underscoring that processes are the ways in which companies add value for their customers, their employees, and their shareholders. Processes are dynamic, constantly evolving, and—as they are executed by people—variable.
For Wallace, the pace of process change is, by its very nature, too fast. That speed causes a gap between ideal business processes and the processes executed in practice. This execution gap usually translates into inefficiencies, cost overruns, loss of quality, or increased risk for the business.
What does process mining do, or where does it apply in this context? Basically, it allows for knowing how a process is really executed within a company. Process mining allows us both to visualize processes and to make a representation of their real executions, thus revealing variations in a process and answering questions that may arise.
Process mining extracts the fingerprints a process leaves on a system. It provides an objective view of what is happening within a company. Any activity carried out within a company leaves a trace that, through process mining, can be transformed into a representation of the process with all the relevant variants.
How are these fingerprints collected? A record of events is needed. Most current systems have event records of different activities that are carried out within the system. Process mining requires an identifier for each sequence of actions that make up an entire process, a time stamp, and the different activities that are executed.
Additionally, it is possible to add context to the different events in a process. These data (or attributes) can help to analyze and better understand problems. The more attributes you have, the richer your analysis of what is really happening.
What is process mining used for?
There are four main categories of use cases for which process mining is valuable.
The first is process optimization, which allows for analyzing variants and their duration, deviations’ root causes, people’s workloads, and, of course, trends.
The second is process automation, which enables discovering real activity, making a selection of automatable variants, analyzing the impact of cost and duration of the whole process, and reviewing improvement statistics.
The third is compliance and risk control, which facilitates knowing the degree of compliance; making comparisons between organizations, teams, departments, or regions; assigning a degree of compliance to standards; and identifying the execution of activities and risk situations.
Finally, the fourth category of process mining use cases is prediction and planning. Process mining allows companies to develop models and databases, anticipate and plan resources, and predict the possible impacts that they may have.
Potential use cases of process mining
Manuel Lama, a process mining researcher and professor at Santiago de Compostela University, participated in the second phase of the webinar.
Lama began his presentation by talking about the characteristics of health processes, noting that they are complex, highly variable processes due to the large number of activities they involve, the occurrence of unforeseen events that need to be addressed and adapted for users, and the execution of threads.
In sanitary processes, infrequent behavior is important because it influences inefficiencies in a process. Knowing the most frequent behavior is crucial because we are talking about patient management, an element that has a lot of value.
Another important factor is the possibility of facilitating professionals’ understanding of descriptive and predictive analytics generated through process mining techniques.
In short, according to Lama, it is necessary to make health professionals understand what is happening in a process. Both the techniques that can be used and the ways in which information is transmitted are crucial in health processes. Furthermore, it is important to keep in mind that these data can be displayed visually as well as be complemented by natural language narration.
Likewise, information can also be transmitted to health professionals descriptively, which means that it must be explanatory and predictive.
Taking these characteristics into account, Lama presented four sanitary processes on which they have been working. Within the exposed use cases, the expert presented on the management of patients with aortic stenosis. Fundamentally, these patients must make a big decision about their treatment: to have surgery or not.
This process’s objectives include facilitating medical decisions and comprehensively guiding patients as they decide whether to receive cardiac surgery or another less invasive treatment. The process also aims to reduce the time between a patient’s admission into a treatment program and their decision to undergo surgery or not.
This process involves tests undergone by patients and consultations with health professionals. Of course, the protocol that is followed is affected by unforeseen events, such as hospitalizations, emergency room admissions, and testing delays.
How are doctors informed and how is that information supplemented? It can be complemented by automatically generated descriptions of what happens to patients. A doctor can be presented with information through both graphics and text so as to better understand the process.
A study found that health professionals given a choice between visual and textual presentation of information preferred descriptions over graphs.
From theory to practice: applying process mining to the health field
To end the webinar, Mateo Ramos, co-founder of DataSalus, explained how processes in the health environment can be analyzed with the Inverbis platform.
DataSalus selected the process of preparing parental nutrition, a tremendously important process. Ramos presented a real-world example where parental nutrition bags, designed for intravenous infusion of nutrients for correcting or preventing malnutrition, were used.
As Manuel Lama said, one of the characteristic difficulties of health processes is that they involve a high number of highly variable activities.
Therefore, the example used in this webinar was simplified considerably.
The process of preparing parental nutrition bags, which in practice may involve a hundred activities, especially with regard to adding components to the bags, was simplified down to fourteen. This simplification revealed the ideal succession of activities that would be followed in a perfect world.
Before analyzing the process, Ramos identified three main dimensions: temporal characteristics, including the ideal execution, maximum desired, and expiration times; the discarding of preparations, where reasons and occurrences are analyzed; and nonconformity analysis.
During the exercise, Ramos used different filters to explore the process’s variants to analyze its scope and visualize all the possibilities it offers to health personnel. Ramos also inquired into the infinite possibilities and predictions offered by the tool.
If you want to access the full webinar and other available courses, go to Inverbis Skillshare and register.
If you want to know more about our solution and how we apply process mining to process improvement, you can click on this link to register and request yowur demo.
