Patient Twinning and Its Impact on Clinical Decision-Making

In an exclusive interview with India Pharma Outlook, Dileep Mangsuli, Senior Vice President of Global Access to Care at Siemens Healthineers, shared his views on how AI can transform personalized healthcare. He emphasizes a dual-layer approach, first drawing on extensive disease data and then using individual patient insights to create personalized treatment plans specific to a patient’s health profile.  

How do you think AI can be beneficial for personalized healthcare?

There are two main ways AI impacts healthcare: generalization and personalization. On a larger scale, AI can allow healthcare to be more predictable by analyzing large amounts of data to improve diagnostics and treatment possibilities for a variety of diseases. For example, AI can analyze thousands of cancer cases, looking for patterns associated with things like race, gender, and age, building what’s known as a “horizontal” layer of disease data.

On the vertical personalized level, AI collects “vertical” data on patients — genetic information, lifestyle, habits — on the individual level. Once these two data layers are intersected, AI can develop more personalized treatment approaches. For instance, chemotherapy may be helpful for one patient but not for another because of genetic and lifestyle differences. Combining horizontal disease data and vertical patient data, the intersection of horizontal disease data and vertical patient data allows AI to provide customized healthcare solutions for more precise, individualized medicine.

How do you perceive the evolving landscape of patient twinning within clinical practice, and how have these trends changed over recent years?

Patient twinning is a novel idea whose roots are in mechanical engineering, particularly in aviation. Engineers originally captured live data from aircraft engines to monitor real-time performance and determine if an engine could keep flying or needed to be serviced. Since then, this approach has been commonly called “digital twinning” in different industries, such as healthcare, where it is highly transformational.

Patient twinning is a futuristic model for personal health management in healthcare. What if you had a digital replica of your health data on a handheld device, allowing you to be proactive with your health? The data collected from wearable devices such as smartwatches that track heart rate or metabolism are used by this model. This data is combined with your routine medical checkups and continuously updates your health profile. It can give early alerts of potential health risks that can be acted upon in time.

Digital twins are today revolutionizing therapies and clinical treatments. A digital twin of a patient’s heart can be used to simulate different electrode placements for resynchronization or rehearse surgical procedures, all helping clinicians improve accuracy and patient outcomes.

Patient twinning promises to change personal health management enabled by advances in computational power and data transmission. The potential of this technology is to reshape wellness, enabling the proactive and effective monitoring and care of one’s health.

In what ways would patient twinning integrate multi-source health data (genetic, clinical, lifestyle) to improve clinical decision-making, including precision medicine? What challenges do you perceive this integration to present?

Data from multiple sources has to be integrated to create digital twins. For example, Siemens is renowned for its state-of-the-art diagnostic and imaging technology and its contribution to sustainable innovations for various populations. Tests like blood and urine analyses generate vast amounts of health data, which this technology uses to reveal the details of an individual’s condition. However, this data is only a part of a larger health story.

To get a complete picture, data from different devices (wearables, for example) must be combined. In terms of handling all types of health data, such as blood transfusion records, the future of digital twins is to seamlessly standardize data from all sources. This standardization is complicated and it needs to be aligned with regulatory frameworks and build consistent data models.

The National Digital Health Mission is a prime example of this as it seeks to bring together health records from a wide variety of medical professionals. Can we store all of a patient’s health data in a unified format? In that case, we would be able to not only understand their current health but also be able to predict future outcomes.

One of the most important things is privacy. As data goes from one platform to another, data anonymity is essential, but patient confidentiality must also be protected. This data may be the responsibility of private entities, governments, or both, and it is an evolving discussion worldwide.

What ethical considerations and patient privacy issues arise with the use of patient twinning, and how can these be managed effectively?

The biggest problems with AI are privacy, ethics, and bias. Let’s say an AI algorithm has been trained on some patient data from hospitals in Chennai. If this algorithm works well locally, it may not work as well for data from other regions like China, Europe, and South America. These differences are physiological and behavioral and lead to this discrepancy. To combat such bias, AI models have to be trained on diverse, representative datasets that encompass a wide demographic range.

In addition, privacy is an important concern, regulated by strict rules such as HIPAA, which requires anonymization and controlled data access. As AI developers, we have to strike a balance between data privacy and an algorithm that works across different demographics. To realize this balance, we will need ongoing work to improve data management practices and develop algorithms that work well in a variety of populations while remaining flexible enough to adapt to evolving data.

Looking ahead, how might advancements in artificial intelligence and machine learning further transform the concept of patient twinning and its impact on preventive care and precision medicine?

Healthcare is embracing artificial intelligence more and more, and one area where it is being used is to improve education for patients and personalized medicine. Take tuberculosis (TB) for example, one of the oldest and most contagious diseases, which affects about 11 million people worldwide, including India contributing to close to 25 percent of the cases. Governments and global health organizations spend billions of dollars each year on TB treatment and prevention.

AI is recognized by the World Health Organization to be a good option for TB screening, particularly in areas without radiologists. Chest X-rays can be analyzed by AI algorithms to detect TB accurately, and faster. Patients then undergo lengthy treatment regimens lasting more than a year, and patient data must be well-managed post-diagnosis.

Digital twinning technology can build a digital replica of a patient’s health journey, provide insights into recovery, and also enhance patient engagement. Companies in India are innovating, and developing algorithms to detect TB from cough sounds, which could one day be used through smartphone recordings. This is a trend that shows how quickly AI-driven healthcare solutions are growing and how personalized medicine is changing.