Expert Insights: In Conversation with Dr. Gabriel Cuatrecasas

At UNLOKall, we bring together leading voices shaping the future of obesity and cardiometabolic care. Through our expert interviews, we explore real-world perspectives on AI, early intervention, clinical innovation, and patient-centered care, helping healthcare professionals translate emerging technologies into meaningful practice.

Meet the Expert

Dr. Gabriel Cuatrecasas

Dr. Cuatrecasas recently led the UNLOKall session Why Early Matters: AI Revolutionizing the Future of Obesity Care. Watch the recording here.

Dr. Gabriel Cuatrecasas is a specialist in Family and Community Medicine based in Barcelona and a member of the Endocrinology, Diabetology, and Nutrition team at Hospital Clínic Barcelona. He coordinates the Primary Care section within the hospital’s Endocrinology and Obesity Service and serves as the Obesity Representative for Primary Care Diabetes Europe (PCDE). He is also the founder of the Obesity Study Group of the Catalan Family Medicine Society and a national leader in integrating obesity care into primary care practice.

Do you see early identification and intervention changing outcomes for patients with obesity and cardiometabolic risk in real world clinical practice?

Yes, early intervention is transformative because the length of time a person lives with obesity matters as much as the severity of the condition. Research shows that obesity duration is independently associated with worse cardiometabolic outcomes, particularly for diabetes risk, and that more than 80% of children with obesity will continue to live with it into adulthood. AI models are now capable of identifying high-risk trajectories one to five years in advance, allowing clinicians to trigger preventive measures before complications become irreversible. Importantly achieving even a 5-10% weight loss in these early stages can significantly improve insulin resistance, blood pressure, and lipid profiles.

The webinar featured an AI driven patient simulation. Based on your experience, how close are we to integrating these types of tools into routine clinical workflows, and what practical challenges still need to be addressed?

We are already seeing the first phase of integration, with some clinics piloting AI-supported risk assessment tools and visualized prediction systems using algorithms. However, several practical challenges remain. These include data heterogeneity, the risk of algorithmic bias (especially in models trained on non-representative populations), and the lack of standardized technical infrastructure in resource-constrained settings. Additionally, there is a significant need for comprehensive training programs to ensure healthcare providers can effectively interpret and integrate AI-generated insights into their daily practice.

AI can generate large volumes of data and predictive insights. In your view, how can clinicians translate these insights into clear, actionable decisions that improve patient care without adding complexity to their workflow?

The key lies in using “explainable AI” methods, which provide a clear breakdown of which specific factors (e.g., sedentary time or fasting glucose) are driving a patient’s risk. This allows clinicians to translate complex data into personalized, target-setting advice for lifestyle counseling. Furthermore, integrating these insights into Clinical Decision Support Systems (CDSS) that follow established guidelines—like the 2025 Obesity Algorithm—helps automate the prioritization of interventions, ensuring AI acts as a helpful “clinical co-pilot” rather than an added administrative burden.

As digital tools and AI become more prominent, how can clinicians ensure that care remains personalized, empathetic, and centered around the patient rather than the technology?

Clinicians must adopt a “human-in-the-loop” (HITL) governance framework, where professional accountability remains non-delegable and all AI-generated content is human-verified before being delivered to the patient. AI can actually support empathy by using large language models (LLMs) to craft more encouraging, personalized behavioral nudges that respond to a patient’s real-time “streaming data” from wearables. By removing barriers like travel and weight stigma through virtual platforms, technology can actually make care more accessible and individualized to a patient’s daily living situation.

Reflecting on your discussion, what is the single most important message you would like clinicians to take away regarding the role of AI in obesity care?

The most important message is that AI has moved from being an experimental tool to a core enabler of personalized obesity care pathways. It offers a scalable solution to bridge the massive gap between the growing clinical need and the limited availability of specialized obesity care. From a primary care perspective, the single most important takeaway is that AI is not a replacement for the clinician, but a transformative “clinical co-pilot” that allows us to move from a reactive model of treating complications to a proactive model of early, personalized intervention. Ultimately, AI empowers the primary care clinician to fulfill their role as a pivotal coordinator of long-term health, providing the scalable infrastructure needed to treat obesity with the same precision and continuity we apply to other chronic conditions.

This session introduced an interactive format using an AI avatar and clinical simulation. How do you see these types of educational approaches supporting clinicians in building confidence and applying new knowledge in practice?

These interactive formats can be interesting for translating complex, abstract ideas into practical clinical reasoning. By working through real-world patient cases like “Alex,” clinicians can practice making live decisions and see the impact of early risk changes in a safe environment. This approach aligns with cognitive load theory by providing curated, digestible content that helps clinicians build confidence and retain information more effectively than traditional lectures.

For clinicians who may feel uncertain or hesitant about adopting AI in their practice, what would be your advice on where to start and how to approach this transition?

For clinicians feeling uncertain or hesitant, the best approach is to view AI not as a replacement for clinical judgment, but as a transformative “clinical co-pilot” that enhances your ability to provide proactive, personalized care.

The lowest barrier to entry is using AI to reduce administrative burden. Additionally, AI can be used to draft patient education materials that are tailored to different health literacy levels and languages, improving patient engagement before they even enter your office.

Start with EHR-based risk prediction tools and AI-enabled behavioral coaching (AIBC) platforms, as these currently have the strongest evidence for clinical utility and patient benefit. These tools allow you to identify upstream high-risk trajectories and provide patients with the continuous support that brief, episodic clinic visits cannot offer. To build trust, choose systems that use “Explainable AI” methods that allow you to translate complex data into clear, actionable counseling for the patient. Furthermore, every AI-generated insight or patient-facing message must be human-verified before delivery, ensuring that care remains safe and guideline-concordant. The goal is to move from a reactive model of treating complications in a siloed fashion, to a proactive, precision-based model in a more comprehensive and integrated way.

Learn from Dr. Gabriel Cuatrecasas