Modeling care processes in GZA hospital (Kick-off)

Friday the 30th of March I been to Antwerp together with Pieter Van Gorp to talk to H. Van Der Mussele and J. De Sitter. They are policy makers in the GasthuisZusters Antwerpen hospital (GZA) and are working on a new system. In this post I will tell you everything I’ve noticed and what I’ve learned. First I will inform you about the goal of the project form the university’s point of view and what is in it for the GZA. After that I will discuss the meeting on contents and my observations about the current way of working within the hospital and how they would like to work in the near future.

 The first thing I would like to mention is the goal of the project from the university’s point of view. First I will observe how and what there is being modeled within the hospital. I will observe the whether they use advanced constructs within their models and what tools they use to model their pathways. I also will observe what they model, and what they don’t model. For example; I can imagine they model the pathway for a broken hip, but they will not model the pathway for a broken hip and a specific type of infection, because this combination will be very rare. When I have observed the how and what, I will study the effect of a reference model. A reference model is an existing model, which the GZA (or any other hospital) can adapt to its needs. A reference model should increase the quality of the model, improve the use of the syntax, and increase the comparability between different models.
In return I will support the GZA with their migration of some/most/all of their processes to models. I will give my opinion about the models and discuss different options of modeling.

The second thing I would like to mention is that the GZA has been working on and with pathways the last 10 to 15 years. The evaluation of the pathways did not happen through a formal process, but more via a trial and error method. Even to the people who have worked a lot with pathways think the definition of pathways is unclear, so my literature study about this terminology makes sense. Maybe unclear is not the correct word, but it has many forms. But for what I know is that the three layer taxonomy comes close to all these various definitions. The two definitions that came forward during the meeting this morning were the clinical pathway as a model description and as an individual treatment plan. Where it is seen as a model description it describes the process for a general patient requesting for only one specific treatment (e.g. when a healthy and fit person broke his hip). The other definition can be seen as an individual treatment plan where the clinical pathway is fitted to one specific patient. When there is deviated from a pathway it is often the case that the patient needs more than just one treatment (e.g. an older, less healthier patient that has broken his hip, but also needs a treatment for an infection). In the latter case the patient needs to be treated by several medical professionals and therefore a single clinical pathway is not possible.

Another general note is that the people in the healthcare ‘industry’ say that it cannot be compared to other industries, like the manufacturing industry. They claim that other industries only have 1 or 2 processes and have no interference with their customers, while the healthcare ‘industry’ has lots of processes and have patients to deal with. These patients are very demanding because their health is at stake. This is a vision often heard by the people from the healthcare ‘industry’. On the other side, all industries can be compared, but the emphasis is slightly different.

Next, I will describe how they work now, and how they would like to work. Currently, all processes are sequential without any workflow support or support from an information system (IS). Sequential is defined as first do task A, when done, then do task B, etc. The current way of working is depicted in figure 1. The clinicians start a process, they have some output measurements and evaluate the process, and adopt where necessary.

With the implementation of the care pathway paradigm the GZA already gained huge improvements on the time aspect, which leads to less costs and less occupied resources (like beds, nurses, etc.). The downside of this pathway implementation is that the human aspect is becoming less important. A commonly heard critic is that “the patient does not have the time to be sick” or “where previously the patient was waiting in the hospital for the clinicians to be ready to start the treatment, now a days the process is waiting for the patient”. Reducing the time in the hospital has its benefits, but it should not be exaggerated. The patient should be able to rest and to be sick in a hospital.
What they would like to add at the GZA, is a WorkFlow Management System (WFMS). The goal of this WFMS is to support the flow of an entire care process. An example of decision given by Mr. Van Der Mussele is the insertion of a catheter. A catheter in a vein must be replaced after maximum four days. The WFMS will give a nurse a reminder that the catheter for a special patient has to be checked and maybe replaced because it is in a patient’s arm for certain amount of time. This is a very simple example, but it makes clear what a WFMS does. A WFMS does not generate an action for a clinician. So to use the same example again; the WFMS will not say to the clinician to replace the catheter, but only to check the patient’s catheter.

This WFMS is graphically shown in figure 2, where the red parts are existing functional silos, and the blue “tasks” bar is the new system. These functional silos are systems within a system. A care process started within a silo results in a treatment, including important decision that need to be made. For example, in the functional silo “medication” a common treatment is the switch from medicine in a fluid form through a catheter to a medicine which comes is pills. This treatment should be started when a patient is able to swallow. The decision to switch from a fluid medicine to a solid medicine is done, for every patient, at the “tasks” level, but a doctor is able to decide otherwise on a lower level (within the functional silo).