The year 2023 will likely be remembered as the year that artificial intelligence (AI) transformed society. And it’ll also likely be remembered as a time when patient trust in the American health care system was at an all-time low. Here we’ll explain AI and patient access more in depth.
Some 43 percent of patients are dissatisfied with their health care, due in large part to tedious and irritating scheduling experiences. Remote patient access – scheduling, in particular – is an endless nightmare for all stakeholders: patients, providers, and telehealth staff.
For those of us in the health care industry, AI has tremendous potential to improve patients’ dismal scheduling experiences. Health care organizations should focus first on transforming their core scheduling processes, because this transformational change will serve as the bedrock upon which effective AI automation can then be built.
AI can automate or semi-automate complex scheduling, freeing up staff to perform more high-touch functions with patients. But if you attempt to implement AI scheduling tools without first transforming your core scheduling processes, these tools likely won’t be effective, and may even be dangerous. Health care is a sector characterized by unique complexity, which means that fundamental alterations to your processes are required for AI and patient access to be successful.
Small steps
To begin your scheduling transformation, first consolidate how appointment calls are handled into one simple workflow. For example, train your staff to log details from each call into a single customer profile or template so anyone can pick up where the last staffer left off. This eliminates the need for transferring patients or repeating information, which in turn lowers hold times and handle times.
Next, standardize as much of your scheduling processes as possible. Create universal, standard instructions that are easy for your staff to read or skim. The goal here is clarity and simplicity. Identify at least four elements: decisions to make, scripting, information to lookup, and services to perform. Think through each element to define how to standardize the format, including the look, feel, colors, links, and more.
Third, eliminate and streamline steps. Anywhere you can remove redundant work will reduce the time it takes your staff to understand and execute the scheduling process. Finally, get feedback from your staff and make appropriate changes. Once you’ve completed these steps, you’re ready to start evaluating AI and patient access tools.
Privacy, trust, and safety
After you’ve consolidated and standardized your scheduling workflows, you should then analyze different AI scheduling tools to determine which ones are a good fit for your organization. Here are some questions you should ask yourself.
Was this AI developed with health care specifically in mind? AI that was built for, say, advertising or marketing purposes likely won’t meet the rigorous privacy and security demands of health care. Systems built from the ground up for health care are more likely to comply with HIPAA laws and other regulations.
Can the AI be trained for your specific workflows and practices? You need to be able to train the AI not just to perform effectively generally, but to perform effectively for your patients within the context of your organization. For example, if a pediatric clinic’s AI fails to account for its 30-minute time slots for routine vaccinations and has double-booked several patients, this will cause long wait times and frustration.
Does the AI have its own independent security and management? If the AI relies on open cloud tools for storage, reporting, and management, that opens up risks around privacy and control of patient data. It’s better for your AI to have its own self-contained database and management portal. As an example, an AI system that leverages existing advertising platforms to track usage or target messages would raise major red flags around privacy. It’s imperative AI and patient access are closely aligned.
What happens to patient data after it’s collected? Understanding the full lifecycle of patient data is important. AI systems that gather data for one purpose but then use it for another without clear consent should be avoided. Again, invest in AI that has built-in compliance mechanisms specifically designed for health care.
How transparent is the AI in how it reaches its conclusions or recommendations about proper scheduling decisions? You should avoid “black box” AI that can’t provide clear explanations for its outputs, because as we know the reasons behind decisions and recommendations in health care really matter. If clinicians and patients can’t understand why the AI made a choice, it will undermine trust and adoption. The AI should be able to offer a logically coherent rationale for any conclusion.
Integrating AI
After you’ve selected your AI tools, you can then begin integrating your AI with your EHR platform, which is a key step in the scheduling optimization process. The AI needs the data from your EHR system to match patients with the optimal provider.
Start by connecting your AI directly to your EHR platform, which allows the AI to access patient health records in real-time. When a patient calls to book an appointment, the AI can then immediately review details – diagnosis codes, medications, recent labs, vitals – to determine the most appropriate type of visit.
For chronic conditions, teach the AI to set up automatic follow-ups by inputting recommended visit intervals for different conditions. For example, patients with diabetes may need visits every three to six months with their endocrinologist and annual foot exams. The AI should automatically prompt patients when they are due for the next appointment and even proactively book the visit.
For preventive screenings, organizations should program schedules based on age and gender to enable automated outreach and reminders for wellness visits, cancer screenings, vaccinations, and other routine care. The AI can then notify patients when they are due for mammograms, colonoscopies, annual physicals, or vaccines.
For example, the AI might send a reminder three months before a man turns 50 that it’s time to schedule his first routine colonoscopy. Or for pediatric patients, the AI could auto-schedule immunizations at the recommended intervals to keep the child’s record up-to-date.
To enable coordinated care, program the AI to review medical records for overdue or recommended follow-ups with previously-seen specialists. When patients call, the AI may ask if they need to schedule a visit with their cardiologist, allergist, or other provider based on recommended guidelines or intervals between past visits.
Trust and safety
What patients really want, more than anything, are simple and streamlined experiences. Simplicity is the primary driver of consumer satisfaction. Simplicity and ease, particularly in regard to scheduling, build trust and safety. Patients don’t care about the latest cutting-edge technology or cool new tools; they care about easily accessing care, minimizing frustration, and achieving excellent health outcomes.
If we lose sight of the patient experience in favor of chasing AI for AI’s sake, we’ll quickly erode any gains we’ve made in patient satisfaction and trust. But if we focus first on streamlining and optimizing our core scheduling processes, and then thoughtfully apply AI to enhance and personalize the patient experience, we’ll be able to provide the simple access to care that our patients want and deserve. AI and patient access need to run hand-in-hand.
