Koios Care is an Antwerp-based health-tech company. It seeks to transform the monitoring and management of neurological and chronic conditions such as Parkinson’s Disease, through continuous, meaningful insights that are gathered unobtrusively by everyday devices, such as smartphones and smartwatches.
Thank you to Dr. Konstantinos Kyritsis for his time and insights.
——–
** How is artificial intelligence transforming clinical care for Parkinson’s patients?
AI is revolutionising Parkinson’s care by providing continuous real-world data that helps clinicians make better decisions. We’re moving beyond the traditional model where patients see their doctor just twice a year and struggle to accurately recall their symptoms.
Evidence from our patient support programs has shown that providing patients with accurate, continuous data can reduce symptom duration by up to 60% — without requiring new medication or major treatment changes. Doctors receive these insights via their patients at the time of consultation. This transfer happens through passive data collection from everyday devices like smartwatches and smartphones, without requiring active participation or disrupting patients’ daily lives.
** What specific patient information can AI analyze?
Modern AI systems can track and analyse motor symptoms like tremors and bradykinesia, non-motor symptoms, medication effectiveness and adherence, sleep patterns, eating behaviors, physical activity levels, and social engagement.
One important capability is monitoring functional activities – measuring the time required to transfer food from the plate to the mouth, which serves as a proxy for bradykinesia. These detailed measurements create valuable datasets for researchers studying disease progression and developing new interventions. This is exactly what regulatory bodies want to see for evaluating new medications – improvements in everyday life activities rather than just clinical measurements.
** Why is continuous monitoring better than traditional assessments?
Traditional Parkinson’s care relies heavily on patient interviews during consultations and on questionnaires like UPDRS Part II and PDQ-39, completed during infrequent doctor visits. However, there’s a pretty chaotic limbo between visits. Quality of life, tremor, and other symptoms fluctuate greatly because it’s a neurodegenerative condition. It will progress – fast or slow – but unfortunately, it will progress.
AI-powered continuous monitoring reveals what’s happening beneath the tip of the iceberg. For example, in our patient support programs, AI has detected deteriorating sleep quality as the root cause of declining overall function, despite stable motor symptoms – insights that wouldn’t be captured in traditional clinical assessments.
** How accurate are these AI systems?
The technology has been validated through rigorous clinical studies. In our most recent Belgian trial with 130 Parkinson’s patients, our researchers found strong correlations between AI-based measurements and how patients reported their quality of life. When correlating digital measurements with standard quality of life measures, we see excellent alignment between how Parkinson’s disease affects patients and what the technology can detect through passive monitoring. These findings are increasingly being published in peer-reviewed journals, providing scientific validation of AI’s capability to accurately track Parkinson’s symptoms and their impact on daily life.
** What are the key benefits for patients?
First, patients are empowered. Patients want to be in the driving seat of their disease. You want to feel that you are in control. This greatly improves quality of life.
Second, AI enables earlier interventions. There’s absolutely no need to have a troubled few months until your next visit to understand what the issue is. Patients can see their neurologist sooner when data indicates a problem, potentially avoiding adverse events like falls.
Third, patients gain better understanding of their condition. Knowledge is strength. Knowing that sleep is the issue instead of thinking medication isn’t working or something is going wrong because of an emotionally difficult day provides actionable insights.
Finally, clinical data shows AI-informed care helps patients achieve better sleep, improved eating behaviors, and increased social activities.
** What are the main challenges in implementing AI for Parkinson’s care?
There are two major clusters of challenges. First is data privacy and security. People need to be aware about the information they’re sharing. The healthcare industry needs ethical solutions using data for the right purpose. This requires strong encryption, potentially decentralized AI models that keep data on patients’ devices, and strict regulatory compliance.
The second challenge is trust and adoption. AI is highly effective, but patients and healthcare providers must trust it before it becomes widely adopted. The main reason is a lack of understanding about how AI works.
Ten years ago, we wouldn’t put our credit card online. Now we do. Health data is equally precious, but what happens if there’s a breach? We need to educate the whole ecosystem – patients, doctors, families, healthcare providers, insurers – on what AI is, what it produces, and what it requires to use.
** What are the limitations of AI in Parkinson’s care?
AI cannot replace doctors – not for the time being. It lacks human judgment, empathy, and intuition. It processes data based on parameters it’s been trained on. It can provide recommendations, flag potential risks, and streamline administrative tasks – but it’s not a replacement for clinical expertise.
AI is not a magical solution. It’s a powerful tool that can help to manage symptoms, detect patterns, and improve care. It can support drug development and enhance patient quality of life, but it’s not going to cure Parkinson’s disease.
** What does the future hold for AI in Parkinson’s care?
Right now, treating Parkinson’s is largely a trial-and-error process. Patients start a medication, adjust the dosage based on how they feel, but this adjustment period can take months or even years. And because the disease is progressive, the effectiveness of a given medication or dosage will eventually decline – forcing patients to start the process over again from scratch.
What we are moving toward with the help of AI is being able to make better and more timely interventions. AI can detect when medication effectiveness changes and suggest adjustments to treatment regimens before problems become severe.
In 5 to 10 years, we can expect real-time monitoring and intervention, hyper-personalised treatment plans, fewer side effects through better medication management, enhanced deep brain stimulation optimisation, and more targeted therapies.
** Finally, what is the key thing that you want people to understand about the work you are doing?
AI is not a magical solution for Parkinson’s Disease, but it is definitely transforming Parkinson’s care. The goal isn’t technological advancement for its own sake but positively impacting patients’ lives through clinically meaningful monitoring that enables better treatment adjustments and, ultimately, improved quality of life.