Medicare at Scale: Embracing the Modern Web in Healthcare


Thank you, Phil. I really appreciate that. So, as you said, my name is Ryan McKenna. I'm a Staff Software Engineer at RVO Health. In addition to this, I co-chair our company's disabilities employee impact group, which is called "Encourage." This talk is "Medicare at Scale: Embracing the Modern Web in Healthcare." To get started, we have to talk about who we are as a company. First, RVO Health is a name you might not have ever heard of, but there's a pretty decent chance that you're familiar with one of our brands. The most recognizable one is It was considered the number one health information website in 2020, attracting an average of 68 million unique monthly visitors. This means, on average, we reach approximately 26 people every second.

The RVO Health portfolio assists nearly 100 million people every month by providing health information, helping them find doctors, saving money, and managing their health and wellbeing effectively. Our family of industry-leading websites, products, and services guide consumers along the path to wellness. We strive to make healthcare easier to navigate, more accessible, and more affordable for everyone. This context is crucial because it underscores the scale of our operations. I am just one of many exceptionally talented individuals who help create this ecosystem of health and wellness products. I consider myself very fortunate to be here today, sharing their work with all of you.

Despite being a relatively young company, we have been in this business for a while. I'm not really going to go into too much detail here. I'm pretty confident that you guys can find our "About Us" page. I'm going to introduce you to 2 products that our team worked on this year where Netlify played a crucial role in our success. These products exist in a very specific market of the American healthcare system: Medicare.

Yeah, this is that unfortunate part of the talk where we actually have to talk about the American healthcare industry, just a tiny bit. I don't imagine that this is what you thought you would learn about today, but it's part of the story, so we've got to go through it. It's not exactly a secret that the healthcare industry can be slow to adopt new technologies. Shout out to Joe earlier for calling us out. In America, we can't really talk about the healthcare industry without talking about the health insurance industry. That means managing risk is a big part of the game. The healthcare industry is highly regulated with high stakes. Operating in this space requires careful consideration. Moving too quickly can lead to problems. We're dealing with an industry that impacts the lives of millions of Americans in very tangible ways.

To truly grasp the scale of the Medicare and Medicaid service in America, let's look at some of these numbers. The Centers for Medicare and Medicaid Services, also known as CMS, is the largest single health payer in the United States. CMS is responsible for administering care through Medicare, Medicaid, the Children's Health Insurance Program (also known as CHIP), and the federal marketplace, Over 150 million Americans depend on CMS programs for high-quality health coverage. The president's fiscal year budget for 2023 estimated $1.4 trillion in mandatory and discretionary outlays for CMS. This is about a net increase of $53 billion from the previous year. Medicare Services alone provide health insurance coverage to 65 million people in the United States. This includes 57 million older adults and nearly 8 million younger adults with disabilities.

The products we built operate in the privatized Medicare Advantage portion of the market. In 2023, 30.8 million people are enrolled in a Medicare Advantage plan. This accounts for more than half, about 51%, of the eligible Medicare population and $454 billion of total federal Medicare spending net of premiums. The average Medicare beneficiary in 2023 has access to 43 Medicare Advantage plans, the largest number of options ever. Just want to comment that I've never seen a note delineating that the following tables are in millions of dollars; that's something.

So, there's something very unique about these programs that we have to talk about, which is the open enrollment period. Medicare health and drug plans can make annual changes to aspects such as cost, coverage, and the providers and services in their networks. The period from October 15 to December 7 is when all people with Medicare can alter their health plans and prescription drug coverage for the following year to better suit their needs. During this time, millions of Americans review their current plan coverage, and if their plans are changing, they evaluate whether the new plan will meet their needs for the upcoming year. Some attentive listeners might realize that, as I'm giving this presentation, the open enrollment period is actively underway. This presents a very unique challenge. There are eight weeks of the year when traffic to Medicare-related properties and web searches peak. All marketing is concentrated around this crucial time period. Consequently, for the rest of the year, we receive very little traffic, if any. During most times of the day, however, the average session lengths of those users are upwards of 20 minutes each.

The first product that we're entering this year is called "Hello Medicare." We've been working on this for nearly a year, and we're thrilled with how it's turned out. Hello Medicare provides resources, comparison tools, and guides to help you find the right Medicare coverage plan for your health, budget, and lifestyle. With Hello Medicare's comparison engine, you can search, compare, and purchase Medicare Advantage (known as Medicare Part C) or Medicare supplement insurance (Medigap) plans from major providers in the United States. Basically, it's an all-in-one stop to help you identify the type of coverage you need, discover plans available in your area, and compare plans side by side. We're excited about the impact of this product and truly believe that it will create substantial value for those affected by annual changes in Medicare plans. Our hope is that we can help millions of Americans make more informed decisions about the plans available to them, leading to greater satisfaction with their healthcare coverage.

The second product differs slightly yet shares many of the same objectives. Precision Medicare is a platform powered by machine learning and artificial intelligence. The aim is to provide comprehensive recommendations for Medicare plan subscribers. Once you import your health records and enrich them with your healthcare goals and preferences, we forecast your health needs for the upcoming year. This information is transformed into a prediction of potential healthcare treatments and services that you might need. This is achieved by analyzing all your diagnoses, treatment codes, and lab results. We review the benefits of thousands of plans to help find the coverage you need for your specific conditions. In addition to this, we also consider network restrictions, approval ease, usage caps, necessary prior authorizations, referral guidelines, benefit limits, and total out-of-pocket costs. It's a mouthful for the American healthcare system. The result is a detailed breakdown of plans available in your area. These plans do not only include your preferred doctors but also offer the best benefits for you and feature the lowest out-of-pocket costs. This product is genuinely exciting as it harnesses the power of artificial intelligence and machine learning for the benefit of millions of people across this country.

Now we've got to talk about challenges in the story, right? One of the big ones is that this is our first open enrollment period. This is the first time we've ever released market products in the Medicare marketplace. This is not an exhaustive list of the challenges we face by any means, but these were some of the big ones. Time's definitely not our friend. We must be prepared for the next open enrollment period. There really isn't an option to launch at any other time of the year. It's all or nothing. So, we need to have a plan that allows us to move swiftly in design and development, utilizing the resources that we had at our disposal. It's essential to plan for scale and be able to iterate on our solutions quickly, taking into account feedback. Velocity is the name of the game.

And then the elephant in the room, there's HIPAA, also known as the Health Insurance Portability Act of 1996. This mandated that the Department of Health and Human Services adopt national standards for electronic healthcare transactions, code sets, security, etc. HIPAA mandated the adoption of federal privacy protections for individually identifiable health information. Broadly, the high-level description may seem wordy and verbose, and that's kind of because it is. While HIPAA compliance isn't the focus of this conversation, the crucial point to grasp is that we had to implement very specific, non-negotiable technical requirements for these products to ensure compliance with federal law. If you've ever been involved in such a process, you'd know it's not the same as launching a regular product. The government takes this matter very seriously, and we have the CMS approval of content. This is significant and had a profound impact on our development and design processes. We needed approval from the Centers for Medicare and Medicaid Services to launch our products without exception. It's like building a regular product, but then you add in another stakeholder that actually determines whether or not anything happens.

Then we have the scale and performance. As I mentioned earlier, these products experience a high volume of consistent traffic within a relatively small time frame. It was crucial for us to architect a solution to scale easily. This optimization was aimed at balancing costs and performance during both high and low traffic periods. So, this is going all in on serverless. It's a little bit more like going all in on the more composable approach. In the health industry, it’s our easiest in. We had recently standardized a technology stack for building out our client experiences by migrating our existing code bases to Next.js. Which was fortunate because when we considered the spikes in traffic that we would experience, employing autoscaling solutions by taking a serverless approach became the clear direction for the architecture of the project. There were several advantages to this, but a few of the big ones were that the pay-for-what-you-use model meant that our cost scaled as our usage did. No more forecasting and provisioning resources in advance. So, we could build at a low cost and really optimize around the open enrollment period. Additionally, by leaning into a composable architecture, we could leverage Edge networks to drive substantial performance gains. There are millions of Americans who rely on the Medicare and Medicaid services who do not have access to robust high-speed internet. Ensuring that our products were performant and accessible to those individuals was critical for the success.

We've seen a variation of this slide a couple of times today. I didn't know that was going to happen, but anyone who's ever engaged in this debate knows that the answer is always some form of, "It depends." I've spoken to many engineers who are truly passionate about the process of creating amazing things, the true purists of software engineering. There are times when I am this engineer, and it can often be difficult to pause and say, "Is this the right thing to do?" Sometimes we tend to write code first and measure the impacts later. In San Francisco, we might be able to relate to this one. As someone who truly enjoys building things, I find it difficult to admit the following: sometimes, not always, but sometimes, companies with a large pool of engineering talent at their disposal overestimate their ability to create internal tools and systems. These often just do not compare to the value of a product that might already be on the market. Just because we have the ability to do something doesn't really mean we should.

So at this point, we want to talk about change. We're always having a moment where we're having new technical requirements coming our way. We know the product is going to change. Oftentimes engineering is treated as over the hedge. We're reactive to problems. In this moment, we can take composable and choose to be a bit more proactive in engaging with these problems as they're coming our way. So, some of these questions are important. They seem easy to answer, but some of these frequently overlooked aspects, especially about this first one, if we solve this problem, what does it really cost us? Who will be responsible for ensuring this solution remains viable? Everybody wants to be the one who solves the problem; very few people want to be the person who maintains it afterwards. If we solve this problem, why is it actually better? Cost is a significant factor to consider, but it's not the only one. Can we replicate the most valuable features? Do we have the right talents? Do we have the time? Do our engineers want to work like this?

So, this is really the heart of the matter. Not every problem is an engineering problem, although we can sometimes be hasty in resorting to that tool in our workbench. Something about, if every problem is a nail, you become pretty good at using a hammer, right? The point being is that this was really a business problem, and I want to frame it for you in that way. We had numerous engineering concerns and requirements for these products to be successful, and we were certainly capable of devising solutions. I would argue that from a business perspective, the most valuable code is the code that you do not have to write. So, when we evaluated Netlify, many things just kind of made sense for us. Sure, we were capable of handling these challenges on our own, but that really wasn't the point. Partnering with Netlify allowed us to concentrate on our strengths, to focus on building the products that only we could make. We could utilize the existing infrastructure for their Edge deployments.

Since this was our inaugural open enrollment period in the market, Netlify gave us confidence that we could handle the pressure that these applications would face. They assisted us in identifying stress test scenarios and provided excellent support for our team throughout our launch, which was a significant relief for us. Then we had features like enterprise-grade security, analytics, log drains. All of these were built in; these were many challenges we just didn't have to deal with because Netlify was there for us on it.

So, this is one of my favorite ones: "The process of putting things on the internet should be easy." A prevalent issue in large companies is the creation of bottlenecks around specific roles and specialties. Frontend teams handling only frontend things, infrastructure teams dealing with all networking infrastructure-related issues. Each team has its specialty and every problem with a preordained solution. While this approach has its advantages, it can sometimes be an impediment to progress and innovation. We need to empower our generalists, not suffocate them. A simple mantra can be applied here: If I can accomplish a task on my own with less effort than it would take with my company's support, then my company is not supporting me properly.

Netlify significantly simplified the deployment of our applications. Instead of building for extended periods and then conducting large releases for quality assurance, we can now identify and resolve issues promptly. In the past 8 months alone, we've completed over 700 deploy previews. That might not seem like a big number when you think of all the other numbers we've seen today, but for us that's a big deal. This change has also greatly enhanced our engineers’ experience. When people feel in control of their work, they tend to be much happier. The days when we depended on other teams to fully service and maintain our applications are gone. Engineers now feel like owners and this sense of ownership is evident in the work.

This is a good one. “You can take branch deployments from my cold dead hands.” This came from one of our product managers, that's the key aspect here. We were working quickly leading to several instances where our teams were performing quality assurance while development was happening in parallel. It wasn't just our engineering team that saw advantages of having Netlify as a partner. Our designers, product managers, and stakeholders were also able to quickly identify issues as they arose with password protected deployments and subdomains. We could easily share our work with partners and external stakeholders for review without any additional effort. While we could have figured this out on our own, I'm really glad that we did not have to.

To bring us to a close here, around this time last year we initiated our pilot with Netlify intending to start with these products and evaluate from there. Our partnership with Netlify instilled confidence in our architecture which allowed us to concentrate on creating excellent products. As other teams began to recognize the benefits that we were gaining, adoption quickly spread in our organization. Standardizing our client experiences to be deployed through Netlify particularly as we transition them to composable modern tech stacks has saved us time and energy while empowering our developers every step of the way. Thank you for your time today.