RxDataScience Blog

Machine Learning and AI in Healthcare: Practical Applications (Part 2)

How can AI technology improve the healthcare industry?

Artificial Intelligence (AI) in healthcareOur last blog post explored the implications of a recent study on artificial intelligence (AI) and machine learning business application. While the healthcare industry can expect incredible benefits from the adoption of AI technology, it remains under-adopted and receives below-average attention when compared to other industries.

This post aims to highlight the areas within healthcare where AI has potential. This technology can improve operations, patient experiences, medical procedures, and solve regulatory challenges.

Using AI to improve operations management

Virtual Assistants/Clinical Documentation

You've heard of Siri and Alexa, two of the most well-known and widely-used virtual assistants. Both are software entities that provide a variety of services which include performing tasks, offering information, assisting with purchases, playing music and videos, and more.

In September 2017, Nuance Healthcare launched an AI-powered virtual healthcare assistant designed to improve clinical workflow efficiency for healthcare professionals. Voice recognition allows the clinician to interact with the software as they would with any employee. The AI presents the clinician with a patient's lab results and any abnormalities, current medications, possible drug interactions when a doctor suggests a new prescription, and alternative medication suggestions. When the doctor has determined the correct drug, the AI system sends the prescription directly to the patient's preferred pharmacy.

The designers of this healthcare-specific AI technology aim for a solution that is secure, intuitive, conversational, and that will reduce the time a doctor spends at a computer.

Cybersecurity Boost

The potential of AI in healthcare extends to data security. AI and machine learning can supplement existing security systems to prevent data breaches. By using algorithms to identify and react to actionable threats, the software is able to both respond to (and learn from) potential data leak patterns.

Another method of security enhancement through AI adoption is to introduce a machine learning system that assesses risk and acts as another level of protection. The health insurance giant, Aetna, recently unveiled a new security system for its mobile apps that did away with password requirements, of all things. This behavior-based solution instead monitors a user's devices and learns both the way a consumer interacts with their phone or computer as well as where they access their device.

Using AI to improve the patient experience


Chatbots have enormous potential. Healthcare providers spend an inordinate amount of money on customer service and appointment scheduling that is done via phone, email, and live chat. The annual projected cost savings from the adoption of chatbots in healthcare is expected to exceed $3.6 billion in the next five years (up from $2.8 million in 2017). Chatbots can be used with remote patients to simulate a physician or nurse experience. The AI bot connects with data on conditions, drug interactions, symptoms, and more, allowing the patient to both interact with and get treatment advice from a machine. The benefits are not just financial; chatbots allow for 24/7 global support that facilitates data accessibility for the patient. The technology taps into natural language processing, knowledge management, and sentiment analysis capabilities.


Wearables are smart electronic devices that are either worn directly on the body or by incorporating the technology into clothing or accessories. Activity trackers monitor vital signs while smartwatches relay and communicate important health information. AI can look for leading indicators that would suggest a patient needs to see a physician. For example, an increased resting heart rate, a change in blood pressure, or a lack of sufficient mobility could trigger alerts that notify a patient proactively. The goal with wearables in healthcare is three-fold: better patient well-being, reduced stress, and a lower cost for both patients and the healthcare system by preemptive measures to avoid hospitalization.

Clearer explanations of patient results

Another application for AI in healthcare involves improvement on explanations of diagnoses and lab results for patients. One such advancement comes in the form of an advanced natural language processing platform (affectionately referred to as a "robodoctor" by its inventor). The platform plans to interpret lab results, starting with blood tests and gradually expanding to other types of tests. It uses natural language processing to explain lab results to patients within an app conversation in simple language that eases the concern some have with robots replacing face-to-face interactions with clinicians. While robots (or in this case, apps) lack the bedside manner so crucial to the patient's journey, natural language AI simulates the experience better than a pre-programmed response system.

Improving medical procedures

Artificial intelligence can augment medical procedures. Examples of this technology include AI-based spinal surgery, cerebral bleeding detectors, and one-second colorectal cancer detection. Another potential application uses emotional intelligence to read into a patient's vocal tone, so that AI platforms could have the ability to detect depression or underlying chronic conditions such as heart disease.

Using AI to solve regulatory challenges

Data insight can improve healthcare costs and performance, but the industry is highly-regulated and difficult to access. Regulations like the pending GDPR compliance reduce the ability of pharmaceutical manufacturers to get high-quality patient level data. AI algorithms and machine learning allow manufacturers to use various data points to accurately derive desired (but legally inaccessible) information. Of course, there are difficulties bringing together AI-derived data into a single view and extracting insights from it. The sheer volume of data points, coupled with the vast command of knowledge required makes it a complicated problem.


AI technology and machine learning have a multitude of potential applications, ranging from virtual assistants that can improve operations, to chatbots that walk patients through their lab results, and even as an answer to regulatory challenges.

In order to realize the full potential of AI in healthcare, the solution requires the unique combination of data science and pharma-specific data domain knowledge. That’s why, in order to take advantage of this rapidly emerging space, more and more organizations are turning to technology and outsourced data science expertise. Learn more about RxDataScience’s analytics as a service solution for the pharmaceutical industry.


Learn More

Related Posts