Life sciences companies are under enormous pressure to drive clinical and economic success through new research, drug development, and better specialized therapeutic solutions. In addition, they are increasingly investing in patient-centric solutions. Due to the COVID-19 pandemic, many pharma companies set up cross-functional teams to ensure continued access to treatment to patients when physical contact wasn’t possible. With clinical trials shifting to a virtual environment, remote monitoring has become even more important to support the patient’s health and safety. This shift has not only taught us that life sciences organizations should become more patient-centric, but also that they can do it in a very short term.
As a technology provider, we want to contribute to building impactful solutions for patients. In this post, we will discuss five areas in which AI can contribute to patient engagement.
What is patient engagement?
Patient engagement is an umbrella term, describing different platforms and solutions that have a common goal to improve the patient’s participation in their own health and wellness. As a result, patients more involved in their health tend to experience better health outcomes.
How can AI help?
At Radix, we believe in building AI solutions that help humans do their work better. Whether that’s a researcher working in a lab spending less time on administrative tasks to focus on discovering new insights, or a patient being more in charge of their own health, AI solutions give the best results if they are designed with end-user input from the start.
Where are the challenges?
According to a survey by Accenture, 76% of patients believe that pharmaceutical companies have a responsibility to provide services that complement their products. However, only 1 in 5 patients are aware of the services the pharmaceutical companies already offer. As for healthcare practitioners, 40% reports being aware of the services that pharmaceutical companies offer.
5 ways in which your life sciences organization can bring value to patients with AI
1) Measure patient awareness and sentiment
Social listening or social media monitoring is a method used by brands to measure and track the social media impact of their products and services. By automatically extracting and analyzing posts from blogs, forums and social networking sites, life sciences companies can track what patients and citizens say about certain disease areas, treatments, or global awareness days.
Using Topic Detection and Entity Extraction, we can extract the topics users are discussing on social media. A technique called Sentiment Analysis allows to track how a certain topic is received by the population, by assigning a positive, neutral, or negative score to a post.
2) AI in remote patient monitoring
Due to the COVID-19 pandemic, the importance of non-contact patient monitoring systems has grown over the last few months. With remote monitoring, patients can continuously monitor vital signs like body temperature, pulse rate, blood pressure without the need of specialized equipment or medical training. By including AI in such remote monitoring solutions, we can add more value to the solution, for example, by sending reminders if data points fluctuate or spike, or by automatically informing emergency services if the patient’s health condition deteriorates.
A machine learning technique that can be used to identify when an observation is abnormal, is Anomaly Detection. With this method, we can detect and identify observations that differ significantly from the rest of the data.
3) Patient outreach, reminders, and appointment scheduling
Digital solutions can help keep patients informed of appointments or treatment schedules. Through a mobile app, the patient can automatically receive a reminder when an appointment is due. In addition, virtual assistant technology can be used here to seamlessly allow patients to make or modify appointments with lower barrier to access.
A virtual assistant or chatbot captures the user’s intent and key information, so that an action can be triggered. This way, a user can request information or set up an appointment in natural language without the interpretation or intervention of a human.
4) Accessible treatment information through Q&A
Treatment information, such as drug information leaflets in packages of medicine are often not user-friendly. A survey demonstrates that many patients don’t usually read the drug information leaflets that come with their medicine. AI can play a great role in making the leaflet information more user-friendly and accessible. The content lends itself perfectly to create a Q&A chatbot, that’s able to answer specific questions from patients on the drug.
Chatbots use Intent Classification and Entity Extraction to capture key information provided by the user. Based on this information, an action is triggered.
In Natural Language Processing, there are also techniques called Text Summarization and Sentence Simplification, that allow to make the information present in leaflets more readable.
5) Lay summaries of clinical trials
A Lay Summary or Patient Summary is a document that intends to share findings of clinical researchers with the general public. The document is written in easy-to-understand language, and makes sure to add definitions of domain-specific terms when these can’t be avoided. Traditionally, these documents are created manually by experts or editorial teams. However, with the use of AI techniques like Text Summarization or Sentence Simplification, part of this work can be automated.
Similar to commercial applications like Grammarly, an automatically generated lay summary uses Natural Language Processing to summarize key findings, simplify sentences, and calculate a readability score for a document.
There are numerous ways in which AI can be used to build patient-centered solutions. In this article, we described five ways in which AI can already be used to bring value to patients.
At Radix, we help organizations build AI solutions that users trust and love to use. In a healthcare context, explainable AI is crucial. In one of our recent posts, we share our tips on explaining AI results.