As the medical field starts to adapt IoT technologies and other smart solutions, several domains become susceptible to cyberattacks, due to the crucial nature of their work and the high amount of sensitive data they handle. The research reported here focuses on evaluating fault detection in medical IoT devices in Intensive Care exposed to cyberattacks through computational analysis. The computational method used is adaptive dynamical system modeling via adaptive self-modeling temporal-causal networks. The paper aims to answer the following main research question: How can intrusion detection on IoT devices in Intensive care be improved to better protect against ransomware? The research question is answered through the following sub research questions: What is the current process of fault detection in IoT devices? What are the weaknesses of this process? How can these weaknesses be improved?