Assort Health

So today we’re chatting about Assort Health and how it’s used at your company. Before we jump into that, could you give a brief overview of the company and your role there?

We are a large orthopedic practice, and I’m the managing partner of a division of our practice. Our practice has physician partners and employed physicians as well as all the associated ancillary personnel. Our division is the test site for Assort Health.

What was the need that drove you to look for a solution like this?

Orthopedics is a high-volume specialty, and our call center plays a vital role in managing patient appointments. Demand for immediate service is intense due to the competitive nature of the healthcare market and patients’ urgency when injured. We faced a significant challenge with call abandonment, particularly on Mondays, where around 40% to 45% of callers would hang up out of frustration. This issue was compounded by patient complaints about difficulty scheduling appointments and navigating our phone system.

The pandemic exacerbated staffing issues, as we struggled to fill call center operator positions. Our goal was to harness artificial intelligence to reduce call wait times and efficiently manage patient interactions, ultimately enhancing our practice’s service quality. We needed a dependable solution to improve our call center’s operations, since traditional methods were failing.

What use cases were you looking for the AI solution to address?

Right now, we’re focused on appointment scheduling. This includes managing bookings for new patients, established patients, and established patients with new issues. We’re also integrating the capacity for patients to cancel, confirm, and reschedule appointments through the system. We’ve worked with Assort as a design partner to tailor it to the needs of our healthcare practice.

What were the specific requirements you wanted the solution you were developing to handle?

Orthopedics is highly specialized. In our practice, each partner focuses on their area of expertise. For instance, we have dedicated professionals for spine issues, hand problems, knee and shoulder sports injuries, joint replacements, and foot and ankle conditions. This specialization meant the system had to accurately direct patients to the correct provider based on their needs when they called in.

Additionally, the system needed to discern urgent cases that required immediate attention. We integrated signs for emergencies, like infections, dislocated joints, and postoperative complications, into our rules algorithm to prompt patients to seek emergency care instead of scheduling an appointment.

To cater to doctors’ needs for efficiency, we needed an interactive voice response system for patients to describe their issue and be scheduled appropriately. We complemented this with a mobile form sent via text for patients who prefer to input their information directly. By offering both options, we accommodate different preferences, ensuring a seamless experience for everyone. And the patient can still choose to wait on hold for a human call center operator.

What is the pricing model you aligned on?

We decided on a payment model based on tasks completed, which both parties found fair. Unlike a flat fee or tiered system, this approach seemed more equitable. It’s essential for buyers to know their current call center costs to evaluate if a new solution is more efficient and cost-effective. Calculating expenses goes beyond just multiplying the number of operators by hours worked; it’s a complex process. We have a good grasp of our current expenditures, which helped in determining an appropriate payment rate.

What did the implementation and onboarding process look like?

For our practice, accurately matching patients with the appropriate physicians is essential. Our physicians completed a detailed questionnaire, which Assort used to inform their rules engine, ensuring the right patient-doctor assignment. Active engagement and a collaborative approach were key to our success, involving regular dialogue between the healthcare providers and Assort. After our team filled out the questionnaires, Assort developed the rules engine, which we then closely reviewed together. Our physicians tested it by calling in as patients. We started with about five per day for a week and then increased to ten. As with any new technology, there were issues to address, which have been fixed. The system has significantly improved since inception and is now functioning well.

Can you describe the patient experience?

When patients visit us, their face sheet includes demographics and a brief description of their health issue. The language used indicates if the appointment was scheduled via our AI system, Assort, or by a human operator. Lately, we’re seeing more appointments made through Assort and fewer through the call center. I often ask patients about their experience scheduling with AI. Generally, they find it straightforward and unremarkable, which is our goal. They’re aware that they’ve interacted with a computer, but had no issues, which is a sign the system is working well. We receive daily metrics on how many people call, use the system, and complete the scheduling successfully through AI, and these numbers are rising each week.

When patients call, they’re given options: wait for a human operator, schedule via AI on the call, or receive a text to schedule on their phone. We’ve maintained a computer-like voice for the AI on purpose. Despite considering more human-like options, research and advice from linguists suggest a mechanical voice is less likely to unsettle users. This approach seems to be effective, based on the data we have.

Has the product addressed your primary problem of abandoned calls?

Our abandonment rate has significantly decreased, now sitting in the high teens or low twenties. This reduction is primarily due to the AI handling a sizable volume of inquiries, which lessens the load on our call center. Agents are addressing the same number of calls, but fewer calls overall are reaching the human system, since the AI is resolving many queries. This has led to an improved patient experience, evidenced by fewer people hanging up.

What do you see as the strengths of the product, and are there any weaknesses or areas you see room for improvement?

A primary strength is the enhanced patient experience. It’s essential for patients to be able to smoothly schedule appointments, akin to expecting seamless mail delivery. Our system now contributes to improving this experience. With Assort managing a large portion of appointment scheduling, this alleviates pressure on the call center, allowing staff to address other patient needs like secretarial inquiries or requests for documentation.

Another advantage is the extension of service hours beyond traditional call center operations. Typically, our call center is available from 8 a.m. to 4:30 p.m. on weekdays, but with Assort, patients can schedule appointments at their convenience, including on weekends and holidays, which is a tremendous plus.

One area where we still face challenges is linking insurance to the patient, though this has significantly improved since the integration with Epic.

Have you found it to be relatively stable and bug free so far?

We haven’t experienced any system crashes or problems. We rely heavily on computers for our operations these days, and while crashes can be expected from time to time, we’ve been fortunate to avoid them with Assort.

How have you found the integration with Epic to work as you need it to?

Our organization is big enough that we have a robust tech team that collaborated effectively with Assort and other third-party vendors to integrate with Epic. It took a lot of work, but the result has been seamless.

How has the support been?

The customer service has been excellent. They’re highly responsive—whether you email or call, they’re quick to pick up the phone or get back to you, and they promptly address and resolve any concerns.

Thus far, do you feel like you’ve made the right decision in going with Assort?

100%

Do you see any growth areas for the product?

Since April, we’ve made significant progress, achieving an 80% success rate when individuals engage with it to schedule appointments. Much like the evolution of cell phones from the early ’90s to the sophisticated devices we have today, I think the potential for growth is substantial. Plans are underway to integrate revenue cycle management and expand services that can be offered to existing patients, things like automating routine tasks like answering patient queries or completing forms, tasks that are costly when done manually. The AI’s capability to operate 24/7, 365 days a year, compared to our current office hours, presents a significant opportunity to enhance efficiency and accessibility.

Do you have any advice for someone who is thinking about incorporating something like Assort?

They have to thoroughly analyze the cost of their current call center. Costs include not only salaries and benefits of employees but also the high turnover rate often seen in call center staff within physician practices. Opportunities for advancement are limited, exacerbating staff retention issues in a tight labor market.

Furthermore, significant expenses are associated with the recruitment process, such as background checks and comprehensive training. Operators need to be competent in using electronic health record systems like Epic, and this training requires time and resources. And employees take sick days or need to handle personal emergencies, operational setbacks that would not affect automated solutions.

Administrators must also evaluate operational data, such as average hold times and the number of calls managed by each operator. Are current call centers handled by remote workers or in-person call handling, and what are the advantages and disadvantages of each? It is essential that the call center, the foundation of a practice’s infrastructure, operates efficiently, because any shortcomings directly impact the practice’s overall performance.