The Dutch Ministry of Health, Welfare and Sport (VWS) is currently working on the successor to the Integral Care Agreement (IZA), the Supplementary Care and Welfare Agreement (AZWA). At that table, there are extensive discussions about the digital infrastructure needed for the care transformation, and the role that VWS will take up in it. An important issue therein is facilitating efficient data exchange. Because no matter how and where you deliver care, digital or face-to-face, at home or in a care facility, in the primary care or the VVT (nursing home), easy exchange of data has enormous benefits for both care recipients and care providers.
To this end, it is essential that the healthcare sector embrace CumuluZ. The initiative for this platform lies with the umc's, and the project was quickly embraced by the entire healthcare industry. Among others, the Dutch Association of Hospitals (NVZ), the Dutch College of General Practitioners (NHG), the Dutch Mental Health Services and Actiz have since joined CumuluZ. Together, these organizations are thinking about how to make data exchange in healthcare easier; an important and highly topical task. The Ministry of Health, Welfare and Sport also sees the great value of CumuLuz: it was recently announced that they will finance the initiative.
Exchange and interpretation
By making the exchange of medical data more effective, fewer repeat examinations need to take place and the registration burden is reduced. This subsequently benefits the workload and job satisfaction of healthcare professionals. In addition, good data exchange improves the safety of patients and reduces healthcare costs. These are all much needed developments that meet the current needs of the healthcare system.
Yet exchange is not the only important issue in the land of data. The interpretation of medical data must also be standardized between healthcare providers, departments and institutions. Essential to this is making - and adhering to - clear agreements on the language and definitions used. Without equal interpretation, data exchange makes no sense at all, and each healthcare party must play its part in equalizing it. Especially in view of the growing role of AI in healthcare.
Rubbish in, rubbish out
For the use of AI, rubbish in equals rubbish out. If the quality of the data used to run the model is substandard, little can be expected of the results either. This principle must also be observed when applying AI to medical data. If one definition in a digital infrastructure is linked by users to different implementation options, or if information is entered in the wrong places, the AI model is built on loose sand. And this results in unreliable output. While in healthcare in particular, reliability of AI tools is a spearhead.
At MUMC+, we switched to a new electronic patient record at the end of November. A great opportunity to bring about a digital culture change. But this does not only require systemic changes. We really need a change in behavior from our healthcare professionals, prioritizing the correct entry of information into the patient record. If, for example, which medication a patient is taking is entered in the wrong box, the patient record becomes nothing more than a notepad. And in doing so, we complicate future applications of AI based on the data in the patient record. In other words, we need a new digital culture, involving both a systems component and a behavioral component. This culture change is the foundation of MUMC+'s ambition: to be a leader in digital care.
Analyzing voice
But we also look forward to other applications of AI in our hospital. For example, we are increasingly using AI transcription tools for the reporting of meetings, and in our care and innovation lab we are investigating whether AI can help prepare letters for general practitioners. We are also looking forward with great interest to new AI applications within imaging, which can start to help radiologists analyze images, for example, and thus speed up diagnostics.
Finally, several research projects are underway at MUMC+ studying whether AI tools can detect whether patients are ill based on their voice. For example, there are indications that AI can determine, based on (vibrations in) the voice, whether someone is suffering from COPD or heart failure. If AI indeed turns out to be able to make those predictions, it could prevent many acute hospitalizations. And that in turn would be beneficial for reducing the pressure on hospital care, increasing the quality of care, reducing healthcare costs and promoting greater patient autonomy.
