Mapping Patient Journeys in the COVID-19 Era
About This Webinar:
Routine chronic disease screenings fell off by nearly 70% in March and April; the number of healthcare providers now offering care via telemedicine is six times higher than it was last year; nationwide hospital outpatient volume was down 35% through June of this year. Against this backdrop of fragmented care and widespread disruption to the status quo, re-assessing patient journey analyses has become more important – and more difficult – than ever before.
In this 30-minute webinar, we discuss how Komodo Health’s payer-complete map of individual patient journeys addresses capturing real-time data and real-world touch points with the healthcare system to understand the full spectrum of patient experience, as it’s happening.
Key topics covered will include:
- Real-world insights drawn from Komodo’s Healthcare Map™
- Different ways of conducting a patient journey analysis and the benefits & downsides
- Key business questions and use cases that patient journey analyses can support
- Komodo’s POV on leveraging real-time datasets to inform patient journeys
00:00 Jesse Leung: Hello everyone. I hope you are well wherever you are joining with us on this webinar today around, “Mapping the Patient Journey.” My name is Jesse Leung, I am a Product Marketing Manager here at Komodo Health. I am joined here today with Shawn Keeler. So Shawn is our lead in the Advanced Analytics group here at Komodo Health. He has experience working on, both the customer side of things, starting his career off on the Pfizer Management Science Group, and has also had experience working in the analytical consulting space when working in roles, ranging from Salesforce sizing and structuring, to guiding sales and building marketing marketing mix models.
00:44 JL: So what we are looking to do today is to do five things, hopefully, in the short amount of time that we have together. The first, really, is to just sort of give you a sense and feel of who we are as a company and what we do. The second thing is to walk you through a little bit of our approach to the patient journey and how it might differ with some of the other vendors that are out there. And the third piece is blend taking together our brand identity, as well as our credo, and moving the needle with software solutions – and thinking about how to build software solutions to move and build a patient journey as a service. The fourth thing we're going to do, depending on time, we’ll fold in a few case studies and some output examples of some of the patient journey work that we've done as a company. And the fifth thing we will like to do is to open up the panel as an open forum to sort of answer any questions that you might have around our patient journey offering or some of our software solutions. And with that, I am going to pass it over to Shawn, who will walk you through some of the content for today.
01:54 Shawn Keeler: Awesome. Thank you, Jesse. So I'd like to start today's webinar by just sharing Komodo’s mission. As a company, we hold dearly the goal of reducing the burden of disease. Our work is centered on improving outcomes and doing so in partnership with amazing teams across pharma, medical devices, and patient advocacy organizations, leveraging the strengths of our Healthcare Map which captures the ground truth on U.S. healthcare.
02:23 SK: Founded in 2014 Komodo is now over 250 professionals. Made up of engineers, data scientists, clinicians, and former Pharma and consulting executives who share enthusiasm for bringing meaning to an ever growing ecosystem of healthcare data. Komodo is heavily invested in data science and technology. We are a technology company by DNA, dedicated to building software products from the ground up to provide innovative solutions to help solve some of the most difficult challenges and healthcare. We currently partner with more than 65 healthcare organizations with thousands of users across those companies spanning life sciences, government agencies, patient advocacy, payers, labs, and also research groups.
03:12 SK: Komodo’s Healthcare Map™ provides a view into the phenotype of disease. And we use this map to understand where latent diseases or complications may exist – and also to pinpoint opportunities for earlier intervention and to close gaps in care. We've undertaken the first effort at scale with a goal of mapping every single interaction with the healthcare system. And this foundation allows us to identify the specialists who are treating relevant patients, to develop a deeper understanding of disease, to uncover the right interventions that healthcare stakeholders can take to close gaps in care, and to promote the standard of care and get patients therapies they need when they need them.
03:57 SK: So at Komodo, we look at the Healthcare Map™ across three core dimensions. First, its breadth – and that means de-identifying and linking data for 320 million patients in the U.S. We look at across depth. Through a large reach, we can see more encounters per patient and provide an understanding of the full level of services behind each encounter. And thirdly, continual enhancement. We are updating our map on a daily basis to get closer to a real-time view of the market.
04:33 SK: When building a patient journey, I can't overstate the power of having a Healthcare Map™ with the highest depth and breadth you can attain. This example I'm showing here is taken from a recent project where Komodo evaluated the journey of patients who have been diagnosed with multiple sclerosis. Well, this particular illustration focuses on a fairly narrow area. MS patients in Texas from a single provider. It shows how a more limited data set could have the potential deployment analyst in a very different direction depending on the biases of the sample.
In this case, not only the Komodo find more than two and a half times the number of MS patients, but we actually demonstrated that each patient had five times the number of encounters versus what was observed in the data to which our customer previously could access. And this is a tremendous amount of signal that would have been missed in a patient journey or any predictive model that followed.
05:26 SK: So we found that the concept of a patient journey has different meanings for different people. At Komodo, we've defined a patient journey as a means of understanding the encounters and interactions that patients have with the healthcare system – from that first presentation disease through diagnosis and treatment. And through this lens, we can use the patient journey framework as an anchor for future research and decisions. It enables, for example, the segmentation of patients and treatment writers perhaps for different communication strategies. It enables outcomes research and cost modeling across various treatment paths. And it opens up the opportunity for better predictive modeling by providing a guide for engineers during that feature engineering phase of mobility.
06:18 SK: So, a diverse set of data assets helps tremendously in illuminating a robust patient journey. Different datasets have different strengths and limitations. For example:
– Primary market research offers a pretty quick turnaround from a variety of healthcare stakeholders, yet typically offers a small number of data points and is also pretty challenging to refresh as the market changes.
– Medical claims provide, typically, the vast amount of coverage but depending on the source that can vary in terms of the depth and the information available.
– Lab data offers an important view into the health profile patients yet doesn't always illuminate the events that are proceeding, or following the test
– HR data can help provide deeper insight but has a common drawback of being restricted from linking and identifying the treating HCPs
– And finally, health system data provides a really important view for in patient encounters, for you have all those are often bundled in claims data, however visibility into post discharge trends is often missing when looking at health system data
07:33 SK: Many of these data sets can be combined, and should be combined, to enhance a patient journey. And with industry to standard tokenization and strong coverage across the U.S., Komodo can easily join to other data sets, and other assets with high overlap, to enrich our targets – looking backward and forward years from the key events of interest.
07:58 SK: So a patient journey can be customized to meet the needs of different stakeholders throughout an organization. Commercial teams often seek to identify additional points of provider engagement or inform messaging strategies. Patient journeys can facilitate the identification of the right provider and the right time for communication. Medical teams at life science companies may leverage patient journeys to guide further research or spark an interest in a more formalized study. And HEOR teams can use patient journey to uncover how the cost of care or the burden of disease varies across treatment cohorts and also across pair dimensions.
08:43 SK: So Komodo’s patient journey analyses are clinically focused and have four primary components.
First we develop an understanding of the patients in the study and the differences across those key cohorts. Next, we uncover the common patterns and healthcare encounters prior to diagnosis. Third, we follow the past patient take across providers. And fourth, we identify how patients are treated and can highlight the differences among us treatment protocols. For example, lines of therapy for complex cancers or outcomes and healthcare costs of those different treatment protocols.
An important outcome of each of these steps is the relevant information to identify points in the journey where earlier intervention could benefit a patient, or to discover where, and with which providers, these educational opportunities arise. So next, I'd like to walk through each of these steps in a little bit more detail.
09:43 SK: A convenient starting point for a patient journey is a milestone event. A milestone event could be a first true diagnosis or it can be the start of treatment. One primary output from this stage is the identification of encounters during the 24 to perhaps 36 months prior to the event of interest or that anchor event. Specifically, you know, what were the earlier recorded diagnoses? treatments? procedures? and even tests? And we can visualize these in that chart that highlights when the signals appeared in patients' collective history.
10:19 SK: Common business questions that could be answered here are for common symptoms – How far back are they reported? How does the prevalence of the change over time? And also we can get into insights, such as the order of events or the sequence of events.
10:39 SK: We can also focus on signals that are truly differentiated to that positively diagnosed group with a simple odds ratio report. And odds ratio essentially compares the rate of events in a cohort of interest against another cohort, or perhaps a non diagnosed control group. We can use a simple report such as this to recommend rules for alerts, to guide the feature engineering of the predictive model, or to build more informed patient cohorts downline. These reports right here showing essentially a single procedure, but the power of this analysis increases when we dig a bit deeper and perhaps show code pairs observed together.
11:26 SK: A patient journey should also examine the pathways that patients traveled on the way to eventual diagnosis or treatment. This can answer your questions such as – What are the referral patterns? For example, did a patient start with a primary care physician and then visit a series of specialists? Did a patient start at a community hospital and eventually get diagnosed or treated in an academic setting? And when is the diagnosing physician different than the treating physician?
12:02 SK: With claims data we can sometimes see these referrals explicitly but we also have the ability to use the temporal data, or network information, to determine referrals on a more implicit basis.
12:22 SK: So a patient journey doesn't stop with a diagnosis. We can examine claims data to track therapy patterns across time.
For example, a line of therapy analysis can show how different providers progress patients across medications and complex therapies or complex oncologies. Without the notes, and the detailed data provided with electronic medical record, there's a bit more upfront work to create the business rules that can infer a change in line. For example, days between claims or look back periods to ensure that we're seeing that initial diagnosis.
That said, once those rules are defined the benefits are significant. Line of therapy analyses with claims data allow us to connect the insights directly to the physicians and to the health care providers. And we also benefit from a broader universe including, not just the community physicians, but also the academic centers in our analysis as well.
Merging claims for treatment with service line data can illuminate how disease progression may change across treatment cohorts, whether a medication is being used prophylactically or to treat an acute event such as a flare, or even the factors that are associated with higher persistency on medication as well.
13:57 SK: For each of the final cohorts, we can report out on a number of factors. We can look at patient demographics, such as age, gender, or geography. We can provide insight into the payer mix to highlight how different insurance types may influence care decisions. And we can look at provider insights, perhaps breaking down by specialty or affiliations. These attributions allow the earlier output to be cut across many different dimensions and illustrate how treatment or disease identification can change across these various segments.
14:41 SK: I'd like to move on and share a brief case study highlighting how we executed a line of therapy study in the past. Project background, in this case, our customer was in the pre-launch preparation phase for an oncology drug. They had planned on entering into a category where there were many established chemotherapy agents and targeted therapies often used in combination. A competing product had recently launched potentially disrupting existing patterns.
15:!4 SK: And we followed essentially a two-tiered approach to this. First we conducted a retrospective analysis to understand that baseline product utilization prior to the competing therapy launch. And it’s involved, as I alluded to earlier, the customizing of business rules, right. So how many days following a claim do we wait prior to calling the next drug claim a combination therapy or perhaps a change in line? How many days following a claim, do we assume that therapy may have ceased? And there are a number of other business rules that can be added in there to fine tune the analysis.
15:51 SK: Once that baseline is established, Komodo applied those same business rules on a monthly basis to track changes over time and identify which providers changed their behavior. And we cut this across different dimensions to provide additional insights across payer type and geography.
In the end we shared how therapy usage changed over time at a national level, but also at a very granular HCO level. We're able to illustrate how pattern shifted with a competitor launch and then with our customers launch. And our customer was able to set a targeting strategy including messaging differentiation based on these insights.
As an added bonus, since this customer is also a subscriber to our Aperture platform, this allowed us to tag physicians of interest within the software and help inform other workflow decisions that they were making.
16:49 SK: So a deep dive consultant lead project is not always needed to answer the key business questions that you may have. On the next slide, I'll highlight how we've started to automate some key components the patient journey into our technology platforms directly
17:08 SK: Within our Aperture and our Iris platforms, we've rolled up patient data to the provider level. And along with its base functionality to more deeply understand the HCPs and the HCOs that are treating in a category, the software highlights inbound and outbound referral patterns at a granular level.
Iris for commercial teams has an added feature we can illustrate how patients may progress across complex treatment protocols with visualizations for the line of therapy I described earlier.
Pulse, our clinical alerting software, has been designed to facilitate engagement at key points along the patient journey.
Prism is a powerful cohort creator and allows, and offers an optional feature to export a list of events. The diagnosis, so the therapies that the procedures that may have proceeded an anchor event that you select.
And finally, Sentinel, a data platform where customers can upload and link external data as Komodo’s Healthcare Map™ enables customer driven journey analyses that can also be bundled with an automated patient journey tool.
18:24 SK: So I'd like to leave you with just a few key takeaways. First, again our definition of patient journey is that overview of encounters and interactions between patients and the healthcare system from the first presentation of disease through diagnosis and treatment. Second, a strong patient journey requires robust data sources with high granularity – ideally for multiple years prior to and following the events of interest. And third, patient journeys will shift over time. A consultant lead journey can provide a strong foundation of knowledge that explores multiple facets of encounters; however, there are also software solutions to offer insight into key components of that journey, in a more streamlined manner. These can be used to refresh on a much more frequent cadence as well.
So at Komodo, we're continually taking strides to improve our Healthcare Map™ and build software that enables access to these insights in near real-time as the healthcare system continues to evolve.
19:33 JL: All right. Thank you, Shawn. I guess at this point what we'd like to do is open up the panel for any questions that you might have. Whether they're in regards to our patient journey offering or some of our software solutions. I know we didn't really necessarily have the time to sort of dive deep into specific use cases but happy to answer any questions that you might have right now.
One of the questions that has come through is, “How do you anticipate the patient journey changing in light of COVID-19 restrictions?”
20:05 SK: So there are a number of different thought leadership pieces that Komodo has put forward since COVID to monitor you know changes that we've seen. Earlier in the crisis, we've looked at rates, for example, of colorectal cancer diagnoses and colonoscopies. We noticed that there was concern over drops and procedures among the healthcare community but had not yet been validated statistically. We were able to leverage our Healthcare Map™ and partnered with Fight CRC, a patient advocacy group that Komodo partners with over the past several years, to track changes and colorectal cancer. And where, you know, early detection and treatment of this fast moving disease can increase success rates.
20:52 SK: What we found in this case was that the number of colonoscopies and biopsies declined by nearly 90% in mid-April compared to the same period in the prior year. And new cases were down by more than 30% by mid-April. Surgeries fell by, I think, 50% compared to a year ago figures.
21:!2 SK: Another example, this week we published a piece on childhood vaccine rates. The CDC sites a drop of 15% and vaccination rates due to the disruptions in preventative care. And this is concerning especially as children returned to school. We've been able to use a Healthcare Map™ to examine vaccines throughout the most recent year and compare them to year-over-year. Our results in this case also showed that vaccinations began to fall sharply, probably as expected in March, hitting a low point in April. And we're going to continue to follow these over time. I think what we've seen so far is thankfully it seems to be a U-shaped recovery as things begin to normalize. But ultimately I think you know COVID has disrupted healthcare patterns. It's pretty obvious to this point main term has some downstream effects on the patient journey. And you know now that we are more than six months – we have more than six months of data into the pandemic – I'm sure that our customers will start asking to show comparisons of patient journeys, you know, pre-COVID ad post-COVID.
22:30 SK: I see the next question up here, “Are you able to identify that a doctor will be seeing a patient for a specific condition in the next 30 or 60 days?”
22:41 SK: So what we can do in those cases is that we can build predictive models, right. So the patient journey is a phenomenal tool for helping to kind of guide a predictive model and start doing the feature engineering of that. But ultimately what we would do to answer a question like this is, we would look to see, essentially, who are the patients who are most at risk for disease? And, you know, one model could be identifying – once we identify those patients who are most at risk, we can identify who those treatment providers are that they have been seeing. So that's one way that we can get at that and we can identify that this patient is likely suffering from this disease and is being cared for by this particular doctor.
In terms of, you know, whether they will see a doctor for their condition in the next 30-60 days. Yeah, I would suspect that we can build a model and kind of time bound that as well. That certainly is possible on that side. We can look to see, I think we would add in that case, both the predictive model and combine that with a referral pattern as well for the doctors that patient has seen in the past.
24:11 JL: Alright, so I know we only have about two minutes left. And I know that we've likely won't have enough time to address all of the questions here. And so I would like to encourage you to send these questions into us, we can address them directly. Shawn if we could advance to the next slide here. Maybe we can sort of talk through a little bit, some of the some of the next steps here.
24:36 JL: So Shawn cited two articles or position papers that we put out recently. One earlier during the COVID crisis around colorectal cancer biopsies and procedures dropping at an alarming rate. And we, more recently this week, published a paper on childhood vaccinations and how we're seeing a U-shaped recovery, which is great concerning that children are returning back to school. You know, we have both of these papers available to you on our website under the insights tab. So I encourage you to check out those two position papers. In addition to that, we also have additional perspectives on health care disparities that we also have available under that tab.
25:23 JL: As also mentioned, you know, we would love – I know we didn't have a lot of time to sort of address all of your questions here, so please feel free to send us questions and we could address those via email, or better yet, we can address them over the phone or we can just hop on the call and really help you find the right solution, software solution or another offering that we have to help move your strategy forward.
25:46 JL: And as always, we're always thinking about how to move the needle in healthcare. And we recognize that the healthcare context is constantly shifting, especially during COVID-19. And so, we also have a forward thinking mindset and building the right products to sort of keep in line with the pace of change in today's environment.
26:07 JL: So feel free to follow us on Komodo Health via Twitter or on LInkedIn. And we’ll continue to, you know, publish interesting findings. And you can follow us along on our journey as things evolve in healthcare. So with that, we look forward, with the opportunity to partnering with you, to reduce the burden of disease. But until then, we look forward to hearing from you, and I hope you have a good rest of your day.