The global digital health market is projected to grow to USD 508.8 billion by 2027, and the World Economic Forum has forecasted that by 2040, annual global healthcare spending will reach approximately USD 25 trillion.

Digital health is transforming healthcare systems in emerging countries. Today, it is no longer difficult to imagine a future where digital tools are fully integrated into healthcare delivery. In fact, they are already changing—and will continue to change—the relationship between healthcare professionals and patients.

While voice recognition has existed for decades, recent advances in computing power and machine learning have made it a clinically useful tool. In the context of a medical consultation, it enables healthcare professionals to record relevant observations—such as foot swelling or skin rashes—via spoken commands, while the system automatically transcribes the information and suggests possible diagnoses and courses of action. The transformative potential of this technology is so significant that Nuance, a leading company in the sector, was recently acquired by Microsoft for nearly USD 20 billion.

Telemedicine and remote consultations have demonstrated that numerous clinical interactions can take place without the need for an in-person physical examination. A notable example is in dermatology, where automated image analysis allows for the classification of skin lesions with high accuracy. In this vein, Google has developed machine learning–based software that, using its vast database of dermatological images, can assist in evaluating suspicious lesions, such as the early detection of potentially cancerous moles.

Digital tools can complement clinical judgment by integrating multiple variables that enhance predictive accuracy. This allows not only for more precise survival estimations but also for anticipating responses to medical procedures and guiding personalized treatments. Machine learning, for example, can already detect left ventricular dysfunction or predict atrial fibrillation from a single electrocardiogram (ECG), identifying patterns imperceptible to the human eye. Meanwhile, electronic health records open new possibilities for applying predictive models, discovering clinical correlations, and improving decision-making—especially when combined with administrative data and robust artificial intelligence techniques.

Patient subtyping is an essential tool in modern medicine, especially in the context of complex or chronic diseases. This process involves categorizing patients into subgroups based on specific characteristics, such as common disease patterns, genetic profiles, and treatment responses. The main goal is to personalize treatments, since patients do not respond uniformly to the same medical approaches. By identifying subgroups with similar traits, it is possible to design more effective therapies and anticipate patient responses to certain treatments.

In complex diseases such as cancer, subtyping plays a crucial role. For example, certain tumor subtypes may show greater efficacy with specific treatments, while others may be more resistant. Similarly, in chronic conditions like diabetes or hypertension, subtyping helps identify patients with rapid disease progression versus those whose condition may be more manageable over the long term.

This approach is not limited to physical illnesses. In fields such as neurology and psychiatry, conditions like depression and schizophrenia also benefit from subtyping models, allowing for a deeper understanding of the various manifestations and treatment responses across patient groups.

A paradigm case illustrates the impact of subtyping in medicine. An oncology drug, initially withdrawn from the market due to adverse effects in a small group of patients, was later reevaluated through a retrospective study. It was discovered that the affected patients shared a specific biomarker, which would allow clinicians to identify those susceptible to side effects beforehand—without restricting access to the therapy for the rest of the patients who could significantly benefit from it.

Patient subtyping employs several advanced methods to identify subgroups with specific characteristics. Some of the most commonly used methods include:

🔹Genetic analysis: Using genetic information to identify subgroups with traits that influence treatment response.
🔹Clinical data: Analyzing medical history, symptoms, and associated conditions to form subgroups.
🔹Biomarkers: Identifying specific biomarkers (biological molecules from the human body) to differentiate and classify patients.
🔹Big data and machine learning technologies: Applying algorithms to detect patterns in large volumes of patient data and help uncover disease subtypes.


Another paradigmatic case of patient subtyping can be found in breast cancer. Although patients may share the same diagnosis, tumors can differ significantly based on the biomarkers present in the tumor cells. Some tumors may be sensitive to specific drugs, while others are not. For instance, specific proteins can be identified on the cell membranes of tumors, which serve as targets for targeted therapies.

Moreover, it contributes to avoiding biases that may impact minority populations, thus ensuring that patients receive the most suitable therapy according to their molecular profile.

The Future of Digital Health Is Personalized Medicine.

Author: Valeria Analia Dávila
‍
References:
Thomason, J. (2021). Big tech, big data and the new world of digital health. Global Health Journal, 5(4), 165-168.
Liu, F., & Panagiotakos, D. (2022). Real-world data: a brief review of the methods, applications, challenges and opportunities. BMC Medical Research Methodology, 22(1), 287.
Butcher, C. J., & Hussain, W. (2022). Digital healthcare: the future. Future healthcare journal, 9(2), 113-117.
Sharma, R., & Kshetri, N. (2020). Digital healthcare: Historical development, applications, and future research directions. International Journal of Information Management, 53, 102105.