Written by: Weslei Gabrillo

What does it look like to reshape the experience of scheduling appointments in a way that enhances the continuity of care? Patients are juggling competing commitments and often face barriers that make it hard to schedule healthcare appointments and follow-ups. Community health center staff members are tasked daily with a large volume of patients and meeting the often insurmountable expectations of matching a given patient to the correct provider in a timely appointment.

Petaluma Health Center took on this challenge, leading with a data-driven and patient-attuned mindset aimed to optimize appointments while also enriching staff capability despite limited human resources and ensuring for overall greater access to health.

Listen and subscribe to the CCI Health Pilots podcast on Apple PodcastsSpotify, and elsewhere. Below is a transcript of the episode, edited for readability.


EPISODE TEASER / Kimberly Keller (guest):

“…The difference was, with this machine-learning, backend and data-driven approach, we could do more unique things with our schedule. We could do a prediction, that this appointment, if scheduled, will be a no-show 40% of the time. We could do predictions based on that. We could do predictions based on the history of the patient: The patient has never completed an appointment in the morning? Probably don’t offer them morning appointments…”

Jessica Ortiz (host):

Hello, everyone. I am Jessica Ortiz with the Center for Care Innovations. And today, we’re talking about one of the exciting projects that Petaluma Health Center has been working on for the past year. Our hope is that, by sharing the highs and lows of this particular digital health solution, safety net organizations can apply the lessons learned to their own challenges.

Jessica Ortiz:

I’m here with Jessicca Moore and Kimberly Keller, members of our Tech Hub Learning Network, which is comprised of 14 tech leaders for California-based community health centers, clinic coalitions and primary care departments in county health systems, that are working to accelerate the adoption of innovative technology. We partner with our tech hub members to vet, pilot, evaluate and spread innovative digital health solutions, targeting Medicaid markets and historically underinvested communities. We are excited to bring you this story today.

Jessica Ortiz:

Kimberly and Jessicca, can you introduce yourselves to our listeners? Maybe we’ll start with you, Kim.

Kimberly Keller (guest):

Sure. I’m Kimberly Keller. I’m the director of business operations at Petaluma Health Center. I’ve been in my role for about four and a half years, new to healthcare.

Jessicca Moore (guest):

And my name’s Jessicca Moore. I’m a family nurse practitioner and our director of innovation at Petaluma. I started here 16 years ago.

Jessica Ortiz:

We’re happy to have you, Jessicca and Kim. So, tell us more about your project. Where did you start?

Jessicca Moore:

Well, I was just going to start by kind of talking about the problem that we identified. So, it was really on two fronts, and those are challenges that are definitely not unique in primary care. I’ve seen these challenges in a lot of different organizations, but the demand for appointments and our no-show rates are really variable on any given date. And, even though the schedules are really full, at the end of the day, there’s almost always some unused capacity. And that really hurts us in primary care, and it hurts our patients, because we know there are other patients who needed and wanted that access that didn’t get it. And we actually did have the capacity, but we couldn’t realize it at the time.

Jessicca Moore:

And then on the other side, our scheduling staff was just spending tons of time combing the schedules on multiple screens, trying to find open slots for patients. Sometimes, it included messaging other staff for permission to double book, and so we really were trying to find, how can we streamline the experience of scheduling for patients and staff, and optimize our appointment utilization, so that we can increase and maintain access and health of the organization and really provide the services that we want to for our patients and community.

Kimberly Keller:

For me, I saw a very similar thing, but more from the employee standpoint. There was just an unrealistic huge expectation of matching patients to providers, in a way that made the patient happy and it made the provider happy. And, reconciling those two things, it seemed very difficult for the staff that was making the appointment, especially with our volume, taking 1,400 calls a day, in a call center, trying to match the correct appointment with the correct provider and the correct patient. So when I first began in Petaluma, I thought about this, and I thought about ways that we could sort of get past that hurdle and how we could right match those appointments.

Jessica Ortiz:

Great. And in thinking about your solution to this challenge, could you share with us a bit more about the digital health solution that you chose?

Kimberly Keller:

So, we evaluated a bunch of different solutions, specifically solutions in healthcare, but we ended up going with a company out of Ohio, Aidan Health. And early on, about three and a half years ago, I was looking through a brochure of a conference in New York or in Las Vegas, and I saw that there was somebody who was working on artificial intelligence and machine learning in a healthcare setting. It was by behavioral health, but it was very similar to our healthcare setting, and they were working on that very problem, right? Sizing appointments, so that patients and providers had the right appointments on their schedule throughout the day. So, I went to Las Vegas, and I met the founder, Quentin Fisher of Aidan Healthcare, and I talked with him about the possibility of doing scheduling in an FQHC setting.

Jessica Ortiz:

And what was the vetting process like? I know that you had had this relationship early on, but just the approach in thinking about why this would be a good fit, mission fit with your organization, operational fit, sustainability. Could you tell us more about that?

Kimberly Keller:

Yeah. So there are a lot of organizations out there right now that do scheduling. They’ll interact with your EHR, and they’ll give you scheduling options. Most of those scheduling options though, are based on availability, like the soonest available appointment. Some of them are smart enough to do as soon as available appointment with your PCP, and then if not with your PCP, with a set list. But they’re all very structured and stagnant. They don’t really evolve with the changing needs of your patients.

Kimberly Keller:

The difference was, with this machine-learning, backend and data-driven approach, we could do more unique things with our schedule. We could do a prediction, that this appointment, if scheduled, will be a no-show 40% of the time. We could do predictions based on that. We could do predictions based on the history of the patient. The patient has never completed an appointment in the morning, [so] probably don’t offer them morning appointments, these sort of things. And most of the solutions we evaluated didn’t have that kind of option in the backend to make the appointments being offered a little bit better, than just “this is the first available.”

Jessicca Moore:

And I think that, for us, some of our values, like we are a very data-driven organization and have a really robust informatics department, with programmers on staff, data analysts on staff. And, how could we really use some of this unused data that we have, in a way that’s going to help us move forward and really support our patients and our staff.

Kimberly Keller:

So, for us, it’s really important that patients will be able to schedule their own appointments, and schedule their own appointments with this sort of smart backend, so that we’re not offering patients appointments and they just look at those and they’re like, “I don’t want any of those,” and then pick up the phone and call us. We want to avoid that situation. We want patients to be offered appointments that make sense for them, that are with their PCP, that are at the time of day they like, that are for the thing that they need. So, that’s sort of the end goal.

Kimberly Keller:

But there’s also the bigger problem that we kind of touched on earlier, of our staff, and our staff being able to easily schedule appointments that make sense, and that when they’re offering those appointments to patients, that they have a good interaction with patients, who are like, “Yeah, that’s the perfect appointment. I want that appointment. That’s great.” And they’re able to quickly and easily schedule that, rather than look at a huge panel of open appointments and try to figure out, on the fly, which one matches this person. So, for us initially, we are focusing on a tool that allows staff to schedule those appointments on the phone first. And then the end goal will be to open up that same tool, after we go through iterations with the staff on the phone, to the public, to our patients.

Jessica Ortiz:

And, as you learn from the predictive AI, that certain things will probably happen or not happen with certain patients, how do you go about finding out the reason why these patients are showing up or not showing up at particular times of the day?

Kimberly Keller:

Yeah. So, that’s kind of the goal really far down the road. The idea is, the more data you put into the engine, the more patterns will emerge. First, we’ll use our appointment data, right? The no-show, that’s a pretty easy one. This patient has no-showed this much, this often. There’s a lot of other data we can use. We can use the PCP. We can use how many children are on the account. We can use demographics. We can use all sorts of different data. And, as we just sort of put that into the engine, the engine iterates over and over again, and it pops out correlation.

Kimberly Keller:

So, we would be able to eventually then see, “Hey, it looks like the further away a patient is, the more no-show they have,” which makes sense to us intuitively, but seeing that inside of data is a whole different ball game. That really sort of opens it up to being real. Once you see, “Oh, people who are coming from this far away, in this neighborhood, are no-showing way more than they are over here. Why? Why is that happening?” And, it allows humans to sort of interact with the computer also, and get that data come together and then figure out how we can better serve our community.

Jessicca Moore:

And it’s also important to realize that this is not the only data point that we have, right? So we’re still screening our patients, using the PRAPARE tool, to identify social determinants of health, and do they need transportation? And, I can envision a world in which we see, these are our top no-show patients. And then we do use some of that human power to outreach to those patients specifically and find out, “What are your barriers? It seems like you’re having trouble getting here. How can we better support you in this?” So it’s really a piece of the whole package of what we’re doing, that’s going to enrich our capability to better utilize the limited human resource that we have.

Jessica Ortiz:

And, how does the tool work with new patients?

Kimberly Keller:

Yeah. So, right now, we’ve been focused primarily on our current patients, and on being able to schedule those appointments, people we’ve seen. We have a little tool that allows us to have people pre-register through our EHR, and the pre-registration goes in. The hope is that eventually, we’ll be able to take that pre-registration, make some automation through that, and then sort of prompt people to pre-register and then be put into automation. If there’s no data on the history of somebody, or there’s no data set that we’re using, then obviously, we wouldn’t be able to really pop out a great appointment for them. But hopefully, over time, we would get better and better with those appointments. So that’s later, later.

Jessica Ortiz:

Great. Thanks for sharing, Kim. And, could you share to listeners what the top highlight of the project is thus far?

Kimberly Keller:

For me, the highlight of the project has been just sort of learning about this technology and how we can use it to better serve our patients. We’ve been able to look at continuity of care a little bit through this, and the idea that we might be able to improve our continuity of care, while at the same time, getting people in the appointment they want to be in. Those two things have been difficult for me, because they’ve been sort of at odds with each other. Because a lot of the time, a patient just wants the appointment on this day or the soonest appointment. And then, we all know that it’s so important that they see their PCP. So, having those two things work together, I think, is the biggest highlight. I can really see where we can make those things work together, and we can get better continuity of care while still getting the exact appointment that a patient wants.

Jessicca Moore:

And, I’ll just add to that. In addition, I think something that’s really exciting for me, and I know it’s really exciting for Kim, is just the idea that we’re going to be able to give our call center a really powerful tool that is going to help them to feel good about their job, and that we can support them in a really meaningful way. It is a terribly difficult job that we are asking them to do, with like a million different permutation and rules and changes all the time. And, they’re getting it from the patients, because the patient isn’t happy with the appointment that they’re offered, or they waited on hold for a long time. And then, the clinical staff is unhappy, because they wanted it scheduled here and not there. So, it’s really, really a tough job, and to be able to prioritize this staff, and to give them a tool that’s really going to help them in their job, I think I’m really excited about.

Jessica Ortiz:

And, how has acceptance been for a non-call center, operational and clinical staff?

Jessicca Moore:

So, I’ll speak to clinical staff. I mean, because we’re still kind of in the development phase of this, we haven’t had large scale conversation around it with the clinical staff. We’ve had some kind of subsets of clinical folks. And, I mean, nobody’s thrilled with the current state. It’s just, anything that can make the day more predictable, increased continuity and really minimize the friction in the system is welcome change to the clinical staff that we’ve engaged so far. So, I think people are, across the board, pretty excited. And I think Operations is obviously very excited that we might be able to reduce some of this unused capacity, right? Because nobody likes to see that. We have the capability to see a hundred appointments, and we had 105 booked, but we only saw 85 or 90. We just missed. It was a mismatch, and that’s unused resource. And so, I think Operations is obviously really excited to see this move forward.

Jessica Ortiz:

Do you have any thoughts about implementing this type of technology with conversational voice AI or text SMS messaging?

Kimberly Keller:

Along with the web-based tool, we also will be able to launch at the same time text rescheduling. So, we’re not going to launch text scheduling right off the bat, but we are going to launch text rescheduling. So, those confirmation texts that we send out will have a little interactive menu that allows you to press 2 to reschedule, and it’ll give you appointment options. Do you want this one, this one, this one, or this one? You choose a number, and it’ll come back through. And that’s using the same backend AI machine learning engine.

Jessica Ortiz:

That’s really exciting though.

Kimberly Keller:

Yeah. That’s really cool. We’re pretty excited about being able to turn those confirmations into an opportunity to fix your appointment if you wanted to.

Jessica Ortiz:

I’m even thinking that I would love that, as a patient. Very convenient.

Jessicca Moore:

Right. Not to have to go through that whole, “Oh, I have to schedule it again.” Yeah.

Kimberly Keller:

My favorite part about it is, there’ll be a last option on it. It’s like option four that says, “No, I changed my mind, just keep the same appointment.” Because we’ve all been there, where we’ve tried to reschedule something and been like, “You know what?-”

Jessica Ortiz:

It doesn’t work.

Kimberly Keller:

“I made a horrible mistake.”

Jessica Ortiz:

Yep. Been there too.  Not every patient gets a 15-minute slot, right? With a clinician. Some are scheduled for more depending on their needs. How does this platform handle that?

Kimberly Keller:

Yeah. So, well, I guess four years ago, when I started at Petaluma Health Center, we had an Excel spreadsheet. And the Excel spreadsheet had all of the providers on it, and then it had all of the procedures on it, the procedures that generally take longer than 15 minutes. Some of them are still 15 minutes, but it sort of showed who was able to perform what procedure, because there are some times when your PCP isn’t the right fit for the procedure that you need. This spreadsheet, we would download whenever it was updated by credentialing. So everybody had a different spreadsheet on their computer. Some of them that had older ones. It was very difficult for us to make sure that everybody had the updated version of the spreadsheet.

Kimberly Keller:

I had a consultant come in, a programmer, and we took that spreadsheet and we made into an online GUI. We call it the procedure matrix. And it allows you to choose the age of the patient, the location of the patient, and then to type in a procedure or a provider. And it’ll pop up whether the procedure or the provider is within scope of practice to do that, and how long that procedure or provider takes to perform that procedure.

Kimberly Keller:

So, you take this procedure matrix and the backend of it is all a database. It’s an SQL database. So we were able to provide that database to Aidan Health and have them incorporate it in the backend of the engine. So, part of the initial GUI, the initial graphical user interface, for the scheduling software, asks you, “What do you need done?” And if you type in any of those or you choose any of those procedure options, it’ll automatically book you for two office visits, which would be like a 30-minute procedure, or three office visits for a 45-minute procedure. So most of the work for that in the backend was already done in a database. We were very lucky to have that and be able to incorporate that in our packet.

Jessicca Moore:

And, Kim, this is something I don’t know, because I haven’t been part of this conversation yet, but for the patients in our system that have a flag that they need a 30-minute appointment, I’m assuming that also goes into this tool.

Kimberly Keller:

Yeah. 30-minute alerts are also able to be just put in there. Also translation services, things like that. Which are the people with the 30-minute alerts, but yeah. Any reason for a 30-minute alert. And then, the way that we’ve done it is, we’ve already rules, and, I call them “reservations,” because I think it’s a nicer word than rules, but we already put reservations through our schedule for 15-minute visits. So, the tool will just take two or take three. So, it doesn’t actually have to think about finding a procedure slot or something like that. It just says, “I need two office visits.”

Jessica Ortiz:

It reminds me of when I have people schedule with me, and I have my 30-minute time slot on Calendly, and I have people double book – it’s the same kind of mechanism. Makes things way easier. On the other side of that, you’ve got your highlights and your challenges. Can you share, so far, what’s been the biggest challenge with this project?

Kimberly Keller:

Well, at the same time of doing the scheduling software implementation, we’re also implementing WELL Health for a communication platform. And the goal is to have those two work together, so that every time somebody schedules an appointment using this interface on the web, that a record of that will log actually goes into WELL Health, along with all the other communications. So for me, the biggest challenge that I’ve had with doing this is to have those two projects work simultaneously, and have them work together, so that we don’t end up in a spot where both of them are implemented, but they can’t connect ever. They have to be able to connect at the end. And so, that’s just been a lot of difficulty.

Kimberly Keller:

It’s also been hard to think about simplicity. Simplicity has been so important in this project, because I think it can get really complicated, really fast. And we can start to throw a lot of data at it, and we can start to have a lot of different rules and the engine can go and it can kind of get out of hand, where we don’t really have a handle on the reasoning or the “why” that this is happening. So, making sure that we sort of keep our finger on it and sort of keep it simple has been a challenge.

Jessicca Moore:

And I think if you look at the lifespan of this project, Kim said she first had this conversation three and a half years ago. It’s been challenging over time to match kind of our internal capacity capability, readiness, resources, with kind of the right tool and people to help us support and implement that. And so, it’s kind of crazy that it turns out that we’re just going to do this in the pandemic, but why not? It just kind of happened that way. Kim and I were talking last week about just how we, and everyone in healthcare over the last two years, has just had to be constantly innovating, and changing, and doing things differently, just to meet our basic operations. And so, things are not stopping. And so, we continue to need tools and resources to help us leverage the human capacity that we have, to support our patients and our staff.

Jessicca Moore:

And, thanks to support from CCI and the Virtual Care Innovation Network, we did have some funds for the startup of this, which was really helpful. And, we had our IT support, which had been lacking for a long time. And so, I think it’s challenging if you don’t have all of the right people at the table and convinced that this is a priority and ready to do it. Kim and I have been around long enough to see that it’s easy for things to get constantly “back-burnered.” And so, really getting it to the front of the line and making the case and saying like, “Okay, we have the right people. This is a real problem, and we have a real opportunity to solve it and make a real difference, not just for our staff, but for our patients and for our operations.” So, I think that was a challenge, and then also kind of, “why now?”

Jessica Ortiz:

Yeah, I’m glad that you brought up the pandemic, because that was my next question is just thinking broadly about, what has it been like to innovate during a pandemic, and if there’s anything more you want to share on that?

Jessicca Moore:

I think exhausting is probably the word. I don’t know.

Kimberly Keller:

I would second that, yeah.

Jessicca Moore:

Yeah. Exhausting. And, it has definitely had a way of sort of… We thought we were constrained before. And we weren’t at all, and why can’t we just do some things really differently? And, it’s definitely given us, in a strange way, some freedom that we didn’t feel like we had before, because just everything had to be changed and needed to be different. And so, there have been some energizing moments in that also. But I think, to have to constantly change things is exhausting, and to keep up, and monitor, and support, and train… and in a time of a lot of turnover with staff for various reasons, to kind of keep that going has definitely been challenging.

Jessica Ortiz:

I’ve heard that from a lot of your colleagues so far, that have shared about innovating during the pandemic and speaking of your colleagues and community, and how important it is during this time of change to kind of lean on each other – you are part of our learning network. And we’re big believers in collaboration, not reinventing the wheel, especially during a time like this, with so much change. How have your peers, in the programs that you’ve been a part of, supported you in moving this project forward?

Jessicca Moore:

So I’ll just say that kind of broadly over time, being part of this network, and seeing and sharing various successes and challenges as we try to implement these tech-based solutions and watch other folks, it’s helped me to be more realistic about timelines, for sure. And, just getting encouragement and support from peers in the network– having a sounding board when we feel stuck — has really, really been helpful. Not just with this project, but with all of our kind of innovations and learnings. It’s great to have other folks who — maybe they haven’t implemented this specific solution — but they know kind of what this process is and can help us to ask the right questions and be an ear and support when things are challenging. And, it’s great to be excited with each other when things move forward, and the thing that you’ve been thinking about and wanting to do for four years – it finally is moving forward. To have people there to share that excitement, too – it’s really great.

Jessica Ortiz:

Is there any advice that you have for listeners, in thinking about implementing a similar technology?

Jessicca Moore:

So, I’ll start. I know Kim probably has some good technical advice. But I will say, be patient and make sure that you have all the right people at the table, that you’re really clear about the resources, both financial resource but also person power, that’s going to be needed to successfully implement. And, don’t get discouraged if you’re not ready. It’s okay not to be ready, and it’s worth waiting until you are, because you’re going to save yourself a lot of frustration. And I think in that waiting time, Kim has done an amazing job of really maintaining this relationship with Quentin at Aidan [Health]. And, there are still pieces of kind of co-design. And, if you have a willing partner, you can still be moving forward incrementally, in meaningful ways that are going to make your ultimate implementation much more successful. So, I think that’s what I would say.

Kimberly Keller:

That’s exactly what I would’ve said, but not as well as you did. I would say, yeah, that exact thing. Don’t ignore the little improvements you can make. Always keep your eye on the main goal, but if you’re not ready to implement that bigger project yet, there’s so many little things you can do that sort of step you towards it. The procedure matrix is an example of that, reservations on our schedule and sort of thinking ahead. Just always kind of keep your eye on that bigger goal. And then as you make improvements, think ahead towards that.

Jessica Ortiz:

Thanks for sharing that advice. I’m sure that’s very valuable for others in the broader safety net community that are considering implementing this technology. So, we’re curious, what can we look forward to in the future from you? What’s next?

Kimberly Keller:

Well, we talked about WELL Health, and what’s next is to take the WELL Health platform, which is sort of a communication aggregator, and to combine that with Aidan Health, so that we have an interaction history for all of our patients in one place, a place where we can look back and we can say, “Oh, it looks like you scheduled an appointment yourself via SMS on this date and this time. Great.” We have a record of what’s going on. As we implement both of those together, combining them and making them one big resource management system is the end goal.

Jessicca Moore:

This is what Kim and her staff are getting for Christmas. (laughs) It’s—

Kimberly Keller:

(laughs) Yes. We’re getting a CRM with schedules.

Jessicca Moore:

—what she wanted for so long.

Jessica Ortiz:

A wonderful Christmas gift, right? Great. Well, thank you so much, Jessicca Moore and Kim Keller, for sharing your expertise and your experience with the broader learning community. We’re so happy to be able to speak with you today.

Jessicca Moore:

Thanks for having us.

                          

                           

Find this useful or interesting? We’re constantly sharing stuff like this. Sign up to receive our newsletter to stay in the loop.