Making the Most of Your Timestamp Data Webinar Transcript
LUYI ZHANG: For today, we will be talking through the following topics. Why are timestamps so important? Which operational metrics should be checked, and how often? How can real-time data help with day to day operations and future scheduling? So as a person who really cares about infusion operations, you know it is important to always keep a finger on the pulse on the day to day activities. As well as look retrospectively to see what has occurred in your center. Analyzing timestamps is a great way to really dive into the details of your operations.
On a daily basis, you should be asking some, if not all of these questions regarding to the patients, the nurse team, and other staff members. So how many patients did we see today? Did we run out of chairs? If so, for how long? What was the average wait time? Were we fully staffed today? Did my nurses got to lunch? How was the workload of my pharmacy? Did we stay pass close, and for how long?
By knowing the answers to these questions, you can really begin to assess how your center is performing on a day to day basis. Possibly, even more importantly when looking at your past timestamps, whether that be a month over month or a quarter over quarter, you should be able to find trends and patterns that could possibly lead to helpful operational changes that could improve your day to day flow. So when we look at how a day unfolded, it is important to focus on two different aspects. How each individual patient traveled through your infusion center, and how their journey stepped up with other patients’ journeys in order to view the full picture of your day.
By delving into these patients’ specific timestamps, one can gain a more clear vision of how and why the day played out the way it did. So here is a typical patient journey. The patient started off with an oncologist visit. Then the patient checks into the infusion center and sits in that waiting area while the nurse preps for the patient. The patient then gets taken back and is seated in a chair. The nurse takes vitals, start the treatment, checks on the patient a few times while the treatment is going on. Take down the last drug, and then the patient checks out. These are all key timestamps to collect and measure in order to gain the full scope of visibility.
However, if you do not have all of these timestamps, it is still OK. Just know, the more data points there are, the more accurate you will be in mapping out each patient’s journey. Whatever time points you can gather from your end will help build a set of metrics that can provide important insights about your center. When starting with a new center, our team shares this list of timestamps that exist within almost every EHR. This list really drove down into almost every metric needed to drive operational change and to make sure the infusion center is on track.
These timestamps can be named a variety of things within different EHR’s. What we care more is what they’re actually measuring, depending on our center’s workflow. These timestamps can be broken into two large categories. The first category, which is on the left-hand side here, is everything related to the infusion visit itself, the schedule details of the appointment, as well as the timestamps that are collected when the patient is physically in the infusion center and interacting with our infusion staff.
The second one is the clinic visit data, the schedule details and a few timestamps collected during the clinic visit. You might ask why we need a clinic visit data? This is certainly not required. But we’d like to, at least, get the clinic appointment time if it is the same day visit prior to the infusion appointment. Because it is actually very helpful when trying to understand the drivers behind some of the demand patterns for infusion appointments at certain times and/or days. For example, a center might notice they have a lot more appointments on a third Wednesday of the month compared to the rest of the Wednesdays.
With the clinic data, we would be able to understand it is because five providers work on a third Wednesday as opposed to the normal three or all other Wednesdays. If recommending level loading, your provider schedule throughout the week will not take effect soon enough due to all the patients already booked on a clinic’s schedule you can at least adjust your staffing accordingly to better serve the demand. We can’t even get down to individual provider level analysis, such as how the pattern coming from specific clinics affect infusion operation.
Among all the patients who have visited clinic providers before their infusion appointment, if we observe that Dr.. John Smith’s patients tend to show up late, we can then leave a slightly larger window while booking infusion appointments for these particular patients. Just so we don’t waste a chair time waiting for a patient who will unlikely to show up on time. At the same time, we can also communicate this delay observed from the infusion side back to Dr. John Smith’s clinics. So they can help look into why and potentially improve on their end as well.
Either way, it’s a win-win situation for a patients and staff on both the infusion team and clinic team. So as I said before, your. EHR is great at storing these timestamps that enables you to recreate any patient’s journey. The next step is to take this further and gain insights into your operations. Let’s start from the minimal information that you will have for sure, which is the appointment time. With appointment time, you can turn these times into actionable insights. So for those of you who are not familiar, this is a chair utilization shape that tells you the numbers of patients in a chair at any given time based on your schedule.
For example, here you’ll see that your chair utilization is over your capacity at 1:30 PM on this day. You noticed ahead of time, so you are able to possibly bring back some patients earlier, or even reschedule them. Now, let’s take one step further. If you have check-in time, this is the moment that the patient checks into your infusion center, you can then analyze the arrival pattern. Are my patients always on time? Are they consistently showing up late in the morning?
Once you find a pattern, you can then factor that into the drivers behind. For example, in this chart, you can see that the patients are mostly on time with a higher portion of patients arriving late in the morning– from this 07:30 AM up to around 10:30 AM. And a slightly higher portion of patients arriving early during late afternoon, late morning, and lunchtime. If some patients are consistently showing up late in the morning, is it because they’re getting a delay from the same day’s previous clinic appointments? And also to scheduled appointment time for infusion treatment becomes unreasonable for them to make it?
If some patients are consistently showing up early, is it because the infusion appointment scheduled is almost a couple of hours after the clinic appointment, hence leaving a too big of a gap in between. These can further lead you to think about what would be a reasonable window to leave between same-day clinic appointment and the infusion appointment to provide the best experience for the patients as well as not flooding the infusion center at some particular time. So now we have an appointment time and a check-in time point. If you also have the indication on the moment the patient sits in the chair, or some capture time point on when the first drug is hung.
You now have everything that you need to measure our wait time. With in-chair time, you can measure how long the patient waits before he gets seated into the chair. If this part of the wait time is too long, is it because all of the nurses are busy and nobody’s available to take this patient? Or maybe the order is not right. Or the order is not signed, and a nurse, a MA, or another staff member has to go chase someone else in order to get everything ready for this patient? With an indication when the first drug has been hung, you can further measure how long it takes to get the treatment really started. If there is a delay here, you might look into, is it because my pharmacy has a backlog and it’s causing a delay in delivering the drugs?
If so, any areas that we can potentially change to improve this? You can also look at patients broken down by wait time categories as it gives you a really strong sense on the proportion of patients waited for how long. So now we have covered all the time points indicating the appointment start. If you have any time point indicated on when the appointment actually ended, you can then measure how long each appointment actually lasted. For example, this chart tells you how your appointments actually lasted comparing with planned duration. The lighter blue you see here means the appointment ran shorter than its planned duration. This medium blue means the appointment ran in range. Lastly, this darker blue means the appointment ran longer than expected.
For this particular example, you can see that for most of the appointment groups, such as the first, the third, and the fourth appointment groups, it is roughly about 30% of the appointments with each appointment group runs long. Where it is almost 50% of the appointments for this second appointment group. What action does this entail? You might want to look into these appointments run long for this particular appointment groups, and see whether it is due to reactions. Or it is time to revisit the planned duration tag for this appointment ER, EHR. To make sure the planned duration really reflects the true estimate of this treatment type. With this chart, it keeps track of how your actual duration compares with the planned duration.
Having an accurate planned duration is really critical in terms of estimating how your day will play out and plan resources around that. So above, are the different time points for both the starting and ending point of the treatment. It would be great for us to have all of them. However, like I mentioned earlier, if you’re not able to get all of them, we can work with what we get. We have the flexibility to design your own priority and offset to estimate a rough appointment start time and end time. For example, when that chair-in time is really missing for some particular appointments, and we make assumptions such that typically patients sit in the chair about 20 minutes after being checked in or 10 minutes before the vitals gets taken.
With this information, we can then estimate check-in time plus 20 minutes, as a rough estimated start time, or vitals time minus 10 minutes. This chart tells you all of the derived appointments start an appointment end, what is the proportion of each field getting used? This is also a great way to examine the overall timestamp quality. For this example, we can see that most of the appointments start are using chair time, which means chair time [INAUDIBLE] data is sufficiently populated. As for appointment end timestamps, the majority of the appointments are using a drug downtime from the very last drug. With a small portion of the appointment using checkout time with some offset applied.
Only very, very few appointments– you can see those little purple ones– do not have any clear indication of when this appointment actually ended. In this case, we estimate the appointment time by assuming the appointment runs to planned duration. The last piece here is a little bit of a separate topic from the patient’s journey. But it provides really important visibility into the future to help with their planning. So in one of the earlier slides, we have talked about how the planned chair utilization can help you gain insights based on the scheduled appointments. In addition to that, we can know a lot more from your scheduling pattern.
So let’s start from this table. How to read this table. The first row is just the numbers of appointments that are already on the book. The second row is our volume projections that tells you the numbers of patients that you will likely end up seeing for that particular day. The last item here is our flags, which are any specific aspects that you should watch out for, try to adjust, or manage to. With appointment make date, the date this appointment was originally booked, not the last modified date, we can gain insights to your specialty patters. From your EHR, you can already know the numbers of appointments that are already on the book for any days in the future, which essentially is this first row that we get from your data.
We can predict the numbers of patients you will end up seeing for any particular day in the future by considering the three following aspects. What is their lead time? What are the patterns observed from your scheduling changes that we should incorporate? What are the patterns observed from your day of operations that we should factor in, such as numbers of same-day arrivals and same-day cancels you tend to get. Numbers of no-shows you typically end up with on any particular day of the week. For a day with higher predictive volumes, you may want to call for extra staffing support if it is not possible to reschedule some of the patients to a slightly open day.
From this example, you can see a few days are getting flat. This means that some specific areas of this day needs our attention in addition to the projected volume. Whether your schedule has pushed you to run out of chairs. Or there are too many appointments getting booked within any particular hour that will likely drive up the nurse workload for that particular [INAUDIBLE]. As you can see, timestamps are very important for mapping out a patient’s journey, for retrospective analysis, and eventually, arm you to make strong and data-driven operational decisions. What would make these timestamps even more valuable would be to increase the frequency at which we are able to receive them.
For all that, like your customers on the phone, our application receives the daily data feed in our [INAUDIBLE] location once per day. That includes the completed appointment for the past one or two days, and the scheduled appointments for the future two months. Since we are only getting this feed once per day, the schedule data is not constantly updated as changes are made to the schedule. However, our application is fully equipped to handle more frequent data, or as we call it, real-time or near real-time data. Having more frequent data updates will help support scheduling needs and your team’s day of operations. It’s just a whole lot of potential that can be unblocked with real-time data.
So let’s first look at how real-time data can help with your scheduling. We have to say schedulers have a really difficult job that they have to keep in mind about multiple stakeholders, the patients, the infusion staff, the pharmacy staff, et cetera. And they have to find a well-balance between all of them, pick a slot that is convenient to every patient, and also try to make sure the schedule ended up with will not cause a flooding to the infusion center staff. With real-time data, at the point of scheduling, iQueue can be a great scheduling aide to help schedulers make the best possible decisions on where to book patients, especially when a schedule is already filling up. When the schedule is relatively open, it matters slightly less what slots to book into.
However, you can still get a sense of how you’re building out the schedule, basically building out your day. This takes us to the second point here, getting immediate visual feedback on each scheduling decision made. So let’s go back to the very first visual that we saw here. The numbers of patients in-chair at any point of time. Imagine with the real-time data support with every single scheduling changes, new appointments getting booked, existing appointments getting canceled, any reschedules. You can literally see in this visual how your day is scheduled to run. With this visual, it is a lot more clear to you what are the more ideal slots to book into to keep your day still flat.
So let’s go back to our slide here. Now, if we stretched out the timeline just a little bit from the point of scheduling, another thing that some of you might already be doing, and we also highly recommend doing, is future scheduling grooming and management. So what is future schedule grooming? It is essentially the schedule cleanup once a schedule is already booked. What are the cases that you should be doing for future scheduling grooming, there are mainly two scenarios. The first one, for a day in the future, if somehow your day ends up with a peaky or sub-optimal shape, then this is the chance for you to still improve the schedule for that day by moving some patients around.
Maybe some patients were putting down on a schedule relatively earlier, are eligible, and also open to take another time on the same day to free up some space at some particular busier hours. Maybe it is possible to spread some of the demand to different days. Maybe some patients can be seen at a different center that is relatively less busy on the same day if you have multiple centers that the patients can be seen interchangeably. The second scenario is if the scheduler needs help figuring out where to best squeeze in the patient to an already really, really full day. You’re basically looking for a place in the same visual that can allow you to squeeze in this patient without running out of chairs at any other point of the day, and also not making your nurse team run around like crazy.
Basically, for a single stand-alone center, try to move the patients to a different time of the day if possible. If not, spread the demand to different days. If spreading the demand is not an option for you, you are just really, really full, and there are no other places to move that demand to, at the very minimum, you can try adjust the staff schedule, if permitted to at least try to prep ahead of time. If you segment infusion operations, the previous slide is really talking about the first part, which is what schedule did you start with at the beginning of your day? Sometimes, even if you start your day with a relatively good schedule, you still cannot rest assure your mind because you know there will always be variability or changes that you will for sure need to accommodate.
This is the second part, which is also really critical. So how can real-time data help on your day of operations? On the day of, I’m sure there will be some decisions that need to be made fast, such as the following. Where to steer add-ons when a patient is calling on a phone who needs to be seen today? When can we take the patients back early? Do we foresee any bottlenecks or congestion? If we also have the flow time point updated in our system in real-time, such as the check-in time, the chair time, the first [INAUDIBLE] start, we can then update the same visual with the flow time point as your day goes. It’s almost like creating a video on your day, as it is still playing out. You will be able to see exactly whether a patient is checking in late, has started late, whether a patient stays in chair longer than expected, maybe due to a reaction, et cetera.
With all of these visibilities, you know how your day is playing out without having to check every single chair in your center. And you will be armed with so much information to make any decisions. Let’s take one concrete example. With the flow point constantly getting updated, you will get to check each patient journey very closely. You’ll know exactly how long each patient spent in the waiting room before he or she gets taken back. If you notice that patients begin to stay longer and longer in the waiting room, you will be able to accurately communicate this delay to the patients who might be checking in very shortly.
In addition, you can also communicate this information to the clinics. So they can then accommodate from there end, or communicate directly to the patients who are about to head over to the infusion center. So now we have talked about how real-time can unblock a lot of the possibilities. We’d also like to touch a little bit on how to set up your day real-time data feed to make all of these happen– just a little bit on this part. This slide might look a little bit scary, but it is actually a really straightforward process, so stay with me. We have started working on getting real-time data feed with a few of our customers.
There are mainly two approaches. You are probably more familiar with the first one, which is an HL7 Bridge. There are three steps involved in this first approach. Number 1, we will set up a secure VPN between your network and our network. You can think of it as a really secure tunnel that no one else can access. Then from your side, you will turn on HL7 messages for some specific events, such as scheduling appointments, reschedule appointments, cancel appointments, et cetera. This is a one-time configurational change on your EHR.
Once this part is done, your EHR will send an HL7 message to us whenever the list of events has been turned on. For example, if your turn on scheduled appointment event or whenever a new appointment gets scheduled, an HL7 message will travel through the secure tunnel and get to us. The last step, once your HL7 message travels through to us, we will make sure to be ready to accept and decode those messages, and then further load into our product in order to refresh the information.
The second approach is to do a push. Essentially, just another way of pushing your scheduling information to us. In this approach, the high-level overview is that we will share you an endpoint. You can then directly invoke this endpoint with the list of scheduling changes that you’d like to send through. Once we receive the information, we do the same steps of loading the information into our product to refresh what is being shown. It is all about letting us know any updates that you have made on your schedule. This is the last thing that we have for today, and thank you so much for listening.
HOST: Awesome! So if you have any questions, don’t forget to enter them into the Q&A at the bottom of your screen. We do have a couple of questions already teed up for Luyi. So question number one,. Luyi, is why can’t my EHR do this if it has all the timestamps that your team needs to collect?
LUYI ZHANG: So your. EHR is definitely great at capturing and storing all of the information, whether it is clinical, administrative, financial, or any other aspects. However, what we can provide is what actionable insights this whole set of data can provide, and how these actionable insights can help you make data-driven decision that helps the operations and management.
HOST: OK. Another question is, what is the option 2, push for real-time data?
LUYI ZHANG: So the second option is, basically we set up an endpoint– and we will share that endpoint with you. And what you can do is to invoke that endpoint with a list of scheduling events that you would like to send us over. For example, any newly scheduled appointments, any appointments– existing appointments getting scheduled– rescheduled, or any appointments getting canceled. So it’s like calling our REST– calling LeanTasS REST API.
HOST: OK. Another question we have is, do you have any customers that give your team a real-time data feed already?
LUYI ZHANG: So we started a process with a few customers, and we are actually actively working with two of them right at this moment. And both of them are on APICS. With one customer in particular, that we are working on the first approach, which is the HL7. Bridge approach. We were able to get some tests– HL7 messages flowing through to us last week, which is roughly only two weeks after we had the conversation with the team who will be working on this. So it was actually really exciting for us.
HOST: Great! And we have one more question. ”I am a current iQueue customer. Right now we’re sending our data feed once per day. How much of an IT effort will it be for us to get the real-time data?”
LUYI ZHANG: OK. For most of the customers that we talk to, this is actually the first question that we get asked. So based on our past experience, it really depends on whether we have been putting in touch with the correct team, and also where this item is on the priority list for that particular team. So if this item is relatively high in the priority list, and we’re also directly talking to and working with the team that will be working on this, we definitely observe a lot quicker turnaround. So like I mentioned earlier, we were able to tag a couple of test messages flowing through from one customer only about two weeks after we talked to the team who is working on this.
HOST: Great! OK. I’m not seeing any more questions coming through. So our time is up for today. A huge thanks to Luyi and, of course, to all of you for participating. Keep an eye on your inbox for a link to the recording of the session as well as for announcements about future webinars. There are many other webinars about very detailed topics like this at leantass.com-webinars.
Also please complete the survey at the end that will be presented to you as you leave the online session. It should only take you about 15 seconds. Also remember, you can text your email address and questions to us at 6308845493. Or you can send us an email for a demo request or more information at email@example.com. Thanks again.