Doctors and nurses in ICUs can benefit greatly from tools that help them choose the right treatment for their patients. Decision support systems can improve the quality of care for patients and even save lives. However, these types of tools have not been adopted widely in healthcare, because they don’t integrate easily in the daily practice of caregivers. Ashley De Bie and Shan Nan, researchers at Eindhoven University of Technology (TU/e) and Catharina Hospital, have developed a new online tool that works significantly better in clinical practice. They will obtain their PhDs on Thursday 19 November 2020.
Tools that help us make decisions have become an integral part of our daily lives. They help us choose the best route when we need to go somewhere or give suggestions for an enjoyable music list that precisely fits the moment. These developments are also occurring in the medical world. There, decision support systems are being developed to help healthcare professionals make the best choices for their patients.
Checklists and guidelines
Critically ill patients are vulnerable because they have little reserve. It is therefore vital that doctors and nurses make the right decisions for them. This prevents errors, so that patients recover more quickly, and are less likely to die. Medical guidelines and checklists should help healthcare providers in this respect, but they often don’t work well in practice. Caregivers at a patient’s bedside often do not have the time to look up the relevant information in these guidelines.
Moreover, this information often does not match the specific problems and requirements of an individual patient. A paper checklist often contains too much general information and is always the same. These limitations prevent healthcare providers from using medical checklists, while a checklist only works well if it is completed in full.
Another problem is that information about patients is often stored in different systems, which barely communicate with one another. In addition, the amount of medical data is increasing, which makes it harder for healthcare providers to keep an overview when they need to make a decision.
TraceBook: dynamic and personalized
Researchers and have developed a novel decision support system that has two major advantages. TraceBook collects data from various systems (e.g. medication data from the pharmacy, lab results, and electronic health record data), and formulates recommendations based on these data that are exactly tailored to the situation and the patient. The tool is therefore not only dynamic, but also personalized.
TraceBook is an online tool, so you can consult it directly on your phone, tablet, or laptop. It works with algorithms that analyze the data of each patient and then gives a scientifically substantiated advice to the healthcare provider. The rules used are completely transparent, so the healthcare provider who uses TraceBook knows exactly why a certain advice is given. Healthcare providers can also adjust the rules themselves if they come to new insights.
Research
To see how TraceBook works in practice, De Bie, who is also a resident internist-intensivist at in Eindhoven, tested the tool at the hospital’s ICU. He followed a total of over 400 patients where doctors could either use a paper checklist or TraceBook’s dynamic checklist.
The research shows that TraceBook works significantly better than an ordinary checklist. For example, healthcare providers were far less likely to overlook crucial patient information. As a result, they were better able to assess whether a patient needed a particular examination or treatment. This had a direct effect on the length of stay (which decreased) and the rating of pain experience (which improved because patients were given painkillers more often). The use of unnecessary antibiotics also decreased.
Innovative
Healthcare providers were generally positive about the new tool, says De Bie. “The participants in the study found TraceBook to be an attractive and innovative tool that is easy to use and certainly has potential for the future. That’s important, because the success of this type of tool depends on the willingness of healthcare providers to use it.”
TraceBook has also been tested in a large hospital in China. Experiences there are also positive, says engineer and researcher Shan Nan. The tool is now being used in China for the treatment of COVID-19 and in percutaneous coronary interventions (PCIs).
During PCIs, the cardiologist uses a contrast medium that is harmful to the kidneys. A special treatment can reduce this damage to the kidneys. Initial research results from China show that thanks to TraceBook almost twice as many patients who need this treatment, actually receive it.
Future
Several companies and hospitals have already shown interest in TraceBook’s innovative checklists. The results of the current studies will be used to further improve the tool. For example, it will be investigated whether it is possible to integrate machine learning and text analysis into the system, so that it automatically learns based on available data, such as patient records or consultation reports. The user-friendliness of the tool can also be further improved.
The research of De Bie and Nan is an initiative of (a collaboration between TU/e, Catharina Hospital, Maxima Medical Centre/Kempenhage, and Philips Research) and Brainbridge 2 (a collaboration between TU/e and in Hangzhou).
“The validation of TraceBook shows the power of e/MTIC well,” says Franklin Schuling, Innovation Program Manager at Philips and member of the supervisory board of e/MTIC. “We are working together on new solutions to improve patient care. The goal of e/MTIC is to speed up the development and implementation of innovative solutions in hospitals. In addition, TraceBook shows the enormous potential that decision support offers to improve clinical outcomes in care. We at Philips support that wholeheartedly.”
Ashley De Bie . First promotor: Prof. dr. Erik Korsten (TU/e)
Shan Nan . First promotor: Prof. dr. Erik Korsten (TU/e). Second promotor: prof.dr. X. Lu (Zhejiang University).
