A majority of practices should at least be thinking about how artificial intelligence (AI) might affect their businesses during the next couple of years. According to a health services innovation company, 94% of survey respondents agree that AI is the most reliable path toward equitable, accessible, and affordable health care, and almost as many believe there will be a return on investment for the technology.1
The question is, how will AI play a role? While many conversations revolve around the potential for clinical improvements, most of the solutions that show promise are related to administrative tasks.
A recent webinar noted the expectation that most AI solutions will be from existing companies that are looking at ways to further automate and improve mundane repetitive tasks related to patient communication, billing, and more.2 This makes sense because according to Accenture, “AI will adapt and flex to automate tasks and self-learn for repeatability at scale. It will help humans focus on adding value and complement, enhance human ability, and improve capital efficiency.”3
The reason that this is where AI can have the most impact today is that most of today’s AI is a specific type—narrow AI. This is largely related to advanced analytics, natural language processing (NLP), and machine learning (ML). Narrow AI is good for recognizing patterns and analyzing data to understand responses or predict actions. It takes quick action for specific tasks or in response to specific actions. For example, think about the chatbot that helps direct you on the phone to your bank.
Machine learning is also good at digging through data to gather insights. Recently, ML has been used to analyze points of communication—such as appointment confirmations, recall notifications, and patient reviews—to look for trends and best practices. Here are some of the insights.
Giving a reminder at the time of scheduling, when the appointment is more than a month away, has no impact on patient appointment confirmations.
The ideal timing for reminders is three weeks before, three to five days before, and about three hours before the appointment.
Recall reminders sent during the week are far more successful at getting appointments scheduled than those sent during the weekend.
Recall reminders that include a link to schedule online are more effective than those that don’t have an option to schedule online.
Several words appear in online reviews that can make the difference between a top-tier and mid-tier review. These words show patterns around issues that practices should be monitoring. Some are words you’d expect, such as “hold” or “privacy,” but others are less expected, such as “cattle” and “assembly.” These are related to being herded like cattle or treated like part of an assembly line.
These AI-driven insights can help practices improve patient experiences and establish more efficient workflows. Beyond these insights, narrow AI can help patients navigate their health-care experience and receive personalized interactions, while saving time for staff. These include tools such as:
• Automated SMS—Because narrow AI is good at interacting in specific ways, it is great for automating SMS interactions. It can be used to provide an interactive experience for patients where they can ask questions and receive an automated response or be directed to the appropriate person. Therefore, if a patient texts the practice at midnight for directions, the system can respond.
• Identifying potential no-shows and filling appointments—AI can analyze patient behavior to predict the possibility of no-shows so that the schedule can be adjusted accordingly. AI can predict things such as the likelihood a patient will pay bills on time or follow through with treatment. It can also automatically reach out to patients who are waiting for specific appointment types when those times are available.
• Post-appointment care plans—AI can establish the right cadence and content for postcare plans based on prescribed treatment and patient behavior.
Many of these cases can be set up to have as much or as little staff monitoring and interaction as the practice feels is needed. Not everyone is comfortable handing their patient experience over to AI-driven technology. So, the staff can start with a high level of monitoring and work their way to handing over certain tasks as they feel ready to do so. AI only gets better with time and continued interaction.
This approach to growing AI use in administrative areas also makes sense for patients. Patients are interested in using AI to save time and experience more convenience, but they still want to get to a person when they need one.4
Today’s narrow AI options help practices remove mundane tasks while providing a better experience for patients. When done the right way, it can provide the perfect balance between technology and human interaction.
References
1. OptumIQ annual survey on AI in health care. Optum website. https://www.optum.com/resources/library/ai-survey.html.
2. Lynn J, Chatham L. Learn how AI will change healthcare. Solutionreach website. https://solutionreach.wistia.com/medias/w3pmki5dr4.
3. Purdy M, Daugherty P. Why artificial intelligence is the future of growth. LinkedIn SlideShare website. https://www.slideshare.net/. AccentureTechnology/ai-and-the-economy/1. Published September 27, 2016.
4. Accenture 2018 consumer survey on digital health. Accenture website. https://www.accenture.com/us-en/insight-new-2018-consumer-survey-digital-health.
NAGI PRABHU is the chief product officer at Solutionreach, where he oversees product strategy, product management, and product marketing functions. He has 19 years of experience leading software as a service (SaaS) product management, as well as engineering teams building innovative, market-leading SaaS enterprise software products that serve large and small enterprises in various domains. Follow him at @NagiPrabhu.
