Giorgio Quer of Scripps Institute Reveals How Digital Health Is Transforming the Healthcare Landscape

By Lori Cameron

Digital health promises to transform traditional care by increasing its scalability, reducing its cost, and, most importantly, improving the quality of care through the use of real-time health and behavior data.

We found an expert to give us some insight into how digital health is transforming the healthcare landscape.

Giorgio Quer

Accompanying the release of the IEEE Computer Society’s Computer magazine’s special issue  “Digital Health: Active and Healthy Living,” we conducted an interview with one of the issue’s authors, Giorgio Quer of Scripps.

Quer is a senior staff scientist and director of artificial intelligence at the Scripps Research Translational Institute. His research interests include wireless sensor networks, wearable sensors, probabilistic models, AI for time series data, deep learning, and digital medicine.

Related: Check out the newly released digital health special issue of Computer.

At Scripps Translational Institute, Quer works on the data analytics side of the All of Us Research Program, adopting probabilistic models and predictive analytics to extract information from large health datasets available through the program, as well as from other industrial collaborations.

His goal is to extract and present this information in a useful way to medical clinicians and other users.

Here’s what he had to say.

Computer Society: How will “digital health” transform the healthcare landscape?

Quer: Digital health means using digital tools to improve health and quality of life for everybody. The use of digital tools is expected to improve the practice of medicine towards one that is high-definition and more individualized (patient-centric). Currently, a huge amount of data is collected by healthcare centers from their patients, and by single individuals from their wearable devices. This rich and heterogeneous data is the building block for the development of advanced learning techniques that will help us unveil the unique characteristics of each individual, allowing the development of personalized care, including improved diagnosis, monitoring outpatients affected by chronic diseases, personalized drug dosage, and personalized diet.

Computer Society: Over the next five years, what developments in deep learning do you see having the strongest impact on the quality of healthcare?

Related: Like what you’re reading? Explore our collection of more than 50 magazines and journals. 

Quer: Deep learning has already shown its decision-making skills in the recognition and interpretation of patterns in clinical images. Two preeminent examples are the detection of diabetic retinopathy from retinal fundus images, and the detection of skin cancer, where deep learning is performing on par with the best clinicians. Indeed, an important unresolved challenge in the application of deep learning is that it cannot autonomously search for causes, nor provide explanatory power. It has great detection performance starting from a large dataset of labeled data, but it falls short in explaining the reasons behind any decision, thus highly limiting its applicability in a complex scenario as inside a clinic. Novel work in the interpretability of deep learning will have a strong impact, especially in healthcare.

Computer Society: How does deep learning improve healthcare for those suffering from chronic illness?

digital health countries
China and Saudi Arabia are consistent forerunners when it comes to the adoption and use of all new technologies. Some other emerging countries, including India and Russia, are also excelling in specific areas. (Source: Philips Future Health Index.)

Quer: Deep learning may significantly improve monitoring the status and progression of a chronic illness. Currently, the patient’s condition is assessed based on short examinations, done during sporadic visits to the clinic, and important therapeutic decisions may be affected by the uncertainty of the measurements. Indeed, digital tools now allow for the continuous acquisition of several vital signs in free-living conditions. Deep Learning for the automatic interpretation of long time series may ingest this large amount of data, providing to the clinician only the most informative data, while screening the patient on her everyday life, allowing for more personalized and effective treatment. Moreover, medical emergencies or unexpected changes in the patient’s condition may be automatically detected and immediately reported.

Computer Society: How effective are current monitoring methods in detecting clinical problems like atrial fibrillation?

Quer: The state-of-the-art is based on old rule-based algorithms, which present some well-known inaccuracies. The problem is far more complicated when the data analyzed comes from wearable devices, where noise and motion artifacts may play a significant role. In this case, the use of deep learning to identify patterns and clear out noisy segments based on very large datasets can largely outperform current methods and make this data a reliable tool for the clinician.

Related: For a limited time only, claim your free download of Quer’s article, “On the Effectiveness of Deep Representation Learning: The Atrial Fibrillation Case.”

Computer Society: How is “representation learning” an improvement over current deep learning processing methods?

Quer: Representation learning encompasses all the methods capable of an automatic interpretation of data, and deep learning is one of the most promising approaches within this area. By leveraging high-level information, typically provided by human experts, it enables a compact representation of raw data and the automatic detection of insults, even minor insults that are hard to recognize by the human eye.

Computer Society: What will the deep learning experience be like on the consumer’s end? How will it change our way of life?

Quer: Deep learning techniques will allow the automatic analysis of large datasets in a more efficient way and without human intervention in normal conditions. The widespread adoption of wearable and continuous monitoring devices allows the collection of long-time series, but the analysis of such a huge amount of data is impractical for clinical personnel. Deep learning will be of great support in this regard, providing users with simple and effective tools to easily analyze their continuously acquired data. This process is suited for screening purposes, alerting the user if a risky situation, which requires a medical consultation, is detected.

You can contact Giorgio Quer at



About Lori Cameron

Lori Cameron is Senior Writer for IEEE Computer Society publications and digital media platforms with over 20 years of technical writing experience. She is a part-time English professor and winner of two 2018 LA Press Club Awards. Contact her at Follow her on LinkedIn.