AI-enabled prediction of clinical outcomes, based on RWD from EHRs

What is Artificial intelligence (AI) in healthcare?

It refers to the ability of a machine to perform tasks commonly associated with intelligent human behavior without direct human input. The use of complex algorithms and software can emulate human logic, recognize patterns, and create logics for the analysis of complicated medical data. Artificial intelligence can outperform traditional statistical models at predicting a range of clinical outcomes based on a patient’s entire raw electronic health record (EHR).


EHRs are a digital version of patients’ paper charts, which if extracted and analyzed accurately and timely would offer a unique perspective into individual and population health, diagnostic evaluations, clinical research, patient recruitment and improving overall decision-making strategies for value-based care. More and more hospitals around the world transition to digitize their patient records as EHRs offer a vast amount of digital data for improving patient care. However, much of this electronic information stored away is in unstructured text. The ability to extract and normalize this information is challenging but crucial for understanding patient care and making (and ultimately predictingpatient outcomes. This demands the development of new age, robust algorithms, and analytics platform technologies to process these kinds of high-volume data.

AI offers a number of advantages

The integration of AI into patient care can enable healthcare professionals to better understand patient health and risks. Machine learning algorithms can be trained to achieve greater accuracy with increasing data. In addition, healthcare data is highly complex with multiple time points and high dimensionality about all aspects of each patient, ranging from drug treatments to biomarker statuses, and imaging results. Machine learning can facilitate the extraction and normalization of these unstructured data and find less obvious patterns.

Machine learning (ML) algorithms to forecast clinical outcomes based on RWD

ML algorithms can uncover hidden patterns as they are exposed to more training data. But even more suited to analyze unstructured data is Deep Learning (DL) algorithms. Their structure of layered artificial neural networks allows the development of sophisticated models with the ability to understand data at different levels of abstraction. Big data simply empowers DL algorithms to train, learn and perform better to optimal features. This ability of deep learning to outperform traditional ML algorithms has already started to generate true and significant recognition in healthcare.

In clinical, artificial intelligence applications are increasing and show promise in predicting disease development across large healthcare systems. Deep learning has become an attractive target to predict clinical outcomes, drug development, and repositioning. Even in its early stages, the performance of AI-enabled applications continues to improve and offer great potential to improve clinical outcomes.

The full potential has not been completely realized to date. The road would always be challenging but not going forward is not an option anymore. Heading into the 2020-year, data analytics and application of emerging technologies such as artificial intelligence will be the backbone for care delivery transformation. Even policymakers are increasingly stressing the need to address challenges and develop tools to digitalize and analyze healthcare data.

Sqilline is committing to continue developing hi-tech innovation solutions for healthcare and is working vigorously to achieve significance. As part of the company’s product pipeline, Danny Predictions of clinical outcomes for targeting therapies will be available in Q4 of 2020.

Danny Prediction screen

Share this article:

More News & Highlights


What if we could change the common approach to clinical trials?

Desislava Mihaylova, CEO and Founder, Sqilline At Sqilline, we listen to the people we work with and often ask ourselves what if… This was the...



Breaking Barriers in Healthcare Data Interoperability: The Danny Platform Solutions

Mihail Jekov, CTO and Partner, Sqilline Electronic Health Records (EHRs) contain a large quantity of machine-readable data. However, institutions choose different EHR vendors, and the...



Sqilline and Fundeni Clinical Institute unite through a strategic partnership for access to high-quality Real-World Data

Sqilline is proud to announce its partnership with The Fundeni Clinical Institute in Romania. The focus of this partnership is the deployment of the analytic...