Companies Focused on Using AI and NLP for Patient Education, Behavior Modification: Part 1

As mentioned in the previous article there are only a handful of companies who are committing a significant proportion of their resources to creating virtual health assistants for direct-to-person coaching, lifestyle change, and adherence. Here are 2 who chose to submit material on what their agents can do and what their approach is. There are many common approaches but each offers a slightly different approach and different technology. But, to be sure, the use of virtual health assistants is quickly becoming not just an approach to help people, but a new healthcare industry! (In alphabetical order)

CodeBaby

intelligent virtual assistant (

CodeBaby feels that the IVA) experience, also known as the customer engagement experience, is just that: it’s about the customer. The user experience design considers the audience, goals, content, and organizational culture. An IVA often starts with an avatar as the character representation creating an emotional engagement with the customer.

Academic studies have demonstrated that people are more likely to reveal information they may find embarrassing to an IVA than they are to another human being. They feel that an IVA can allow patients to provide medical information in a manner that allows physicians to update discharge instructions to enable providers to meet meaningful use and improve collaborative care

CodeBaby operates from a position that the complete IVA experience is a combination of avatars, conversational scripts, wizards, props, interactions, dialog boxes, calls-to-action, natural language query knowledge base capabilities, voice-to-text capabilities, text-to-voice capabilities, and navigational maps to effectively engage, service, and drive the desired user behaviors. These tools can be mixed and matched to achieve multiple objectives, such as education and training, guidance on forms and applications, tech support, branding/loyalty, and transaction completion.

Their IVA draws its answers from many resources, including real-time customer data, knowledge bases, and various customer relationship management/enterprise resource planning systems. In addition, CodeBaby uses predictive modeling based on past behavior and other user preferences from similar situations as well as learns from the totality of users dependent on its help. Over time, its IVA(s) literally learns how to be a better assistant—to anticipate needs versus reacting to them.

CodeBaby also believes that having the right answer alone doesn’t make it completely effective. It has to sound conversational, whether through text, audio, or visual, by using human-like traits, such as gestures, inflections, tone, and colloquial language or natural language response. If the specific implementation involves a natural language, intelligent component, the IVA will use integrated data and socially intelligent voice, text, and/or gestures.

Some CodeBaby virtual assistants are built without the natural language engine. They use conversational scripts, which are pre-programmed to direct users through pathways, with some user flexibility, to achieve specific key performance objectives. However, they use the same human-like social intelligence, human voice, text and/or gestures to achieve a successful customer experience.

CodeBaby has a specific focus on characters because it believes:

• People trust information more with characters.
• Interactive characters can have social roles every bit as influential as those assigned to real people. These roles, when made obvious, can be used strategically to simplify automated transactions and maximize their effectiveness.
• Characters provide an opportunity for user-experience designers of online interactions to deal explicitly with emotional experience exchange (eg, a smile when encouragement is required, an expression of uncertainty when information is lacking).

CodeBaby feels that an exceptional customer experience includes:

• Customer journey mapping
• Improving application accuracy and completion rates
• Reducing contact center and agent usage
• Improved education, navigation and advisement along the way
• Ultimately driving to a desired behavior such as signing up for health insurance or fully understanding patient discharge instructions such that they do not return to the hospital for the same issue

Intelligent Digital Avatars

natural language processing

Intelligent Digital Avatars (f/k/a Geppetto Avatars), another company focusing on the use of (NLP), calls its agent an Intelligent Digital Assistant, or “IDA” for short. IDA combines speech recognition, NLP, and intent-based reasoning. Like CodeBaby, Intelligent Digital Avatars also makes use of 3D animation to provide an experience that it feels makes patients want to engage in on a daily basis. As is common with all of the companies highlighted in this series, IDA does not judge, has infinite patience, and is available when patients want to talk.

IDA is available as a downloadable app on an Apple or Android mobile device. When launched, IDA appears on the screen, asks for your name, and has the ability to communicate via speech, touch, or text. It recognizes you from previous encounters so that you no longer need to enter your user ID and password, but keeps the encounter HIPAA compliant. The device's camera and microphone capture the entire encounter, which is made part of the patient’s personal health record.

IDA exchanges pleasantries and then walks through a series of questions to capture a daily assessment of the patient’s diet, exercise, medication, and mood. It does a pain assessment and allows the patient to point out where the pain occurs, using an outline of a person. It can perform tests such as the “pinch test” for Parkinson’s disease and record a video for comparison with past and future performances. It can also capture still images of a skin condition or any physical anomaly the patient wishes to pass on to the care manager. It captures a writing sample to see any degradation in dexterity.

Next, IDA invites the patient to ask questions about his or her condition. Answers are provided through a database that is continually updated based on the current state of medical knowledge. The entire encounter is stored on a HIPAA-compliant server in the cloud, and analytics are performed not only on the data of the particular patient but also on patients with similar characteristics. This allows identification of patterns and trends for the patient, and eventually predictive analytics based on the collective analysis of hundreds of similar patients across the database.

A daily report is made available to individuals and organizations authorized by the patient including caregivers or clinicians. The report is structured to highlight results on an exception basis (including lack of use of IDA by the patient).

Finally, IDA has a HELP button, which connects the user to the managed care call center at any time, allowing person-to-person communication with a skilled professional to occur as needed.

The next version of Intelligent Digital Avatar's IDA will contain cognitive games, interfaces to medical devices and personal fitness devices, and connectivity to clinical records of the patient, with the ultimate goal of complete integration with electronic health records.

Next, I'll take a look at MedRespond and Next IT.