Canvas: A Bet On New EMRs

TL:DR 

Canvas is an EMR for companies that want to engineer custom workflows. They’re completely redesigning the interface and backend tooling to create an EMR that lets clinicians and developers work side-by-side, and works with providers that are in any type of payment model (direct-to-consumer, paid by insurance, etc.). The future of the company will depend on whether health tech startups choose to build their own systems from scratch or choose to build on top of Canvas, and whether legacy providers will switch their EMR.

Read more below about the pain point they’re targeting, a deep dive into the product, and some of my thoughts on cool parts and expected challenges for Canvas.

This is a sponsored post - you can read more about my rules/thoughts on sponsored posts here. If you’re interested in having a sponsored post done, email nikhil@outofpocket.health.

Company Name - Canvas

Canvas is building a new EMR designed to be programmable, modular, allow you to build your own custom workflows on top of it, and be scalable from the beginning. 

The company named itself Canvas because it’s one letter away from Canva and that company is worth $40 billion dollars. Some accidental google searches are bound to work in their favor.

The company was founded by Andrew HInes and has raised a total of $20 million from Inspired Capital, IA Ventures, Upfront Ventures, and others.

What pain points do they solve?

It’s hard to imagine what pain points Canvas solves when everyone so universally loves their EMR.

I mean the reality is that EMRs work for the purpose which they’re bought and built for: billing in a fee-for-service payment arrangement. They’re great at getting a receipt for a list of services they want to get paid for and have a fun byproduct of user cortisol spikes.

There are lots of problems with this, but a few big ones:

• The UI just...sucks lol. A million different alerts, 90s looking interfaces, copy-paste text fields, switching between side scrolling and vertical scrolling, etc. Across the board just a generally atrocious user experience mostly because they’re trying to have features for a million different workflows within one tool.
• Making custom workflow changes for an EMR sucks both from a technical perspective (these things are old!) and from a “needing to play by the EMRs rules” perspective. You frequently will need to use a CDS hook which goes to a remote service, increasing the latency of when information shows up in the EMR.

• For companies in alternative payment models, the current fee-for-service EMRs are not designed well. If you have any telemedicine workflows, or need to make sure specific screenings are done to hit quality metrics, or need to shift tasks to different members of a patient’s care team it’s pretty difficult to create totally new workflows that accommodate this with current EMRs.

Canvas thinks it might have a solution to help with this by completely changing the EMR’s form (for the users), function (for the developers), and friendship (for all of us). 

What does the company do?

Canvas is building a new EMR meant for engineers and clinicians at tech-forward care delivery companies aka. the ones that don’t have any departments named “Health IT”. There are a few key components to understand.

The first is its “narrative charting” user experience. It’s sort of like the Gmail auto-complete feature on steroids and actually useful because it understands when to end phrases with exclamation points and when not to in order to avoid sounding passive-aggressive but not too excited.

Narrative charting is designed to auto-complete likely parts of the visit and guide physicians to appropriate next steps. Think of it like starting with a blank canvas and then the software starts structuring and...filling in the gaps...holy shit I just realized why it’s called Canvas lol.

The auto-charting aims to make charting faster and easier. The goal is to surface everything within the workflow instead of needing to pop into other tabs, find the right drop down menu, scroll around, etc. If you’ve ever watched any physician try and order labs, you’ll understand how insane it is that they need to traverse 3 separate windows to do this. Lots of neck veins start popping out and the mouse clicks get faster and angrier.  Here’s a video on how narrative charting actually looks in the product, and watch how the drop-downs and checkboxes fill in as you type + the sidebar keeps a running tab of the important patient information.

Two other byproducts of auto-charting is that the data comes out more structured and it guides physicians to do stuff that might help avoid a claim denial down the road from the payer. The company claims its narrative charting system is 3x faster and 80% fewer clicks when tested against other menu-based EMRs like Allscripts.

The real magic though happens in the backend. Canvas has a Software Developer Kit that makes it easy to create different rules, protocols, workflows, etc. that can then get surfaced in the narrative charting interface when they come up and aren’t a pain in your backend (nice). The key benefits are increasing the speed in which information is surfaced, and collapsing processes into faster commands. 

In terms of speed, by making the EMR itself run on custom code that executes in real-time it can:

• Check itself for inconsistencies in a patient record without a manual review (e.g. this patient is on X med but doesn’t seem to have a diagnosis for Y condition in their record)
• Surface measures around cost-effectiveness based on the patients plan (e.g. a patient needs a lower limb MRI - these are data fields that are missing to get a quick prior authorization and here is the closest/cheapest in-network facility)
• Highlight information a physician would end up needing to send to a payer later (e.g. screenings they need to do to hit quality bonuses)
  

By bringing this stuff up in the actual interface during the visit, the physician can have the conversation with the patient in the room and prioritize the things that need to get done at that moment when you have a fixed amount of time in the visit. Otherwise the physician might need to chase patients down after a visit for missing information or hire staff to manually make sure these things happen/are highlighted before the patient comes in next.

The other way Canvas speeds things up is by creating programmable commands that fit into a phrase. For example, you could program a phrase like “AWV” for a patient’s annual wellness visit and it would check to see which tests need to be run, automatically set up the order forms, create the questionnaires that need answers, and automatically generate a care plan that gets sent to the patient. If any of you know “dot phrases” in existing EMRs or TextExpander, you can use those to turn phrases into much longer text templates. Canvas commands are like those on steroids and can execute functions, not just create copyable text.

‍

Canvas is an EMR that’s also purpose-built for developers by developers. Clean documentation, easy to use developer kits, modular functionalities that let you pick what you want to add, and more. The other nice part is that the custom workflows can plug-in to the existing tech stack a company uses - if you want to build your own patient-facing software and only want Canvas to run in the back, you can connect the two simply with custom workflows.

Canvas actually just launched something pretty cool - any developer can spin up an instance of Canvas, get the API keys, and just start messing around with it using synthetic patient data. 

What is the business model and who is the end user?

The business model is pretty straightforward, providers pay Canvas for their tool. It’s $0.80 per active patient + 2.8% of payments received. Payments include:

• Patient fee for service payments
• Patient subscription payment
• Patient cost share payments
• Insurance claim remit payments
• Insurance capitation payments
• Employer direct payments
• Duffel bags of cash (jokes)

But exclude shared savings payments or quality bonus payments. You can see more pricing info here.

The users are both clinical and development teams at care delivery companies. Canvas is largely targeting companies that have engineers that can use their APIs and SDKs to build the workflows they need. These custom workflows are particularly helpful for companies that are either in value-based care arrangements or want to move to some at-risk financial agreement that requires relatively complex rules around patient care.  But any company that is trying to optimize their patient workflows can make use of the custom logic you build behind the scenes.

Canvas is targeting three main customer types:

Specialty Telehealth: Now that people are realizing modern specialty groups will be a combination of in-person and virtual care delivery, there are lots of new specialty telehealth startups popping up. These groups can build all of the patient-facing pieces that they consider a differentiator while outsourcing the provider-facing and back-office software to Canvas (scheduling, patient payment, medical record management, etc.) Some examples might be weight loss companies creating the sequence of medications they plan to recommend based on cost/efficacy/patient risk factors or a D2C pharmacy trying to assess which parts of a patient’s questionnaire should yield further follow up questions, a video consult, or a specific medication.

Modern care delivery organizations with complex patients- A lot of Medicare and Medicaid patients have complex health needs that require some combination of in-person care, home care, virtual care, and lots of staff coordination between people that work at the clinics. Canvas acts as the EMR chassis but allows for advanced task management, custom risk stratification, etc. so that clinical teams can do outreach when necessary to make sure patients are coming in for follow-up visits or get their meds, alert clinicians to what’s needed during a visit to hit quality care metrics depending on the patient’s health insurance, pull in relevant data feeds like Admission-Discharge-Transfer (ADT) in case they need to be alerted that a patient showed in an emergency room, and more. 

Payers + independent PCPs - Canvas is appealing to payers working with independent primary care providers even if they don’t have their own developers on staff. Payers want independent primary care practices - they’re typically cheaper and are more likely to avoid referring to expensive downstream care vs. their hospital owned/affiliated colleagues. However, it takes a lot of work and frustration to submit data to a million different payers especially if you’re trying to be in some sort of value-based care arrangement. 

Canvas has a software developer kit that payers can use to make it easy for data to be sent between the PCP and payer. This makes it easier for payers to flag things in the EMR for the provider to make sure they get paid, automates data flows so the providers don’t need to hire third-party data abstractors and chart reviewers to review and structure the data, etc. Payers and independent providers SHOULD theoretically be on better terms but the onerous reporting requirements and squeezed margins have made it difficult to create a mutually beneficial relationship. Lowering the admin workload through real-time data sharing might be the first step to a more productive relationship.

{{sub-form}}

Job Openings

Canvas is hiring for:

• VP of Engineering
• VP of Product
• Senior Software Engineers
• Product Manager

___

Out-Of-Pocket Take

Remember when I wrote a whole thing about why building new EMRs was hard? Canvas is basically saying “no, you’re just soft.”

More seriously, I actually think it’s potentially a good time to start this business because many new tech-centric care delivery companies have been started that are trying to customize existing EMRs or build their own from scratch. It’s hard to teach an old EMR new tricks (like... moving data between the EMR and other front-end applications). As more of these companies hire engineers with certain expectations of the software they work with, platforms like Canvas become more appealing.

Beyond that, there are more companies today whose business models today rely on improved customer experience (e.g. companies targeting cash pay patients) or managing a complex set of rules in the back-end (e.g. providers in value-based care arrangements). An EMR that can help providers prioritize the right things to talk about during + between visits is not only better for the patient experience but also for payers that need to make sure certain quality metrics are captured. In the future as US healthcare explores an increasingly diverse array of payment models, providers have to be more flexible in how they deliver care. An adaptable EMR that can reflect the payment model changes in the care delivery itself is critical for this.

Some specific things I like about Canvas:

Radical redesign of the interface - From just a personal taste, aesthetic, and user experience perspective, Canvas is really refreshing to see in action. But some pretty typeface and auto-complete isn’t the powerful part - it’s having control of that experience on both sides. If you build nice tools for developers but don’t have control over changing the interface presented to the clinician, then you’re severely hampered. By giving a better user experience to BOTH developers and clinicians, changes in the back end actually get presented usefully or change the workflow itself to the front end that clinicians use. 

Self-serve tooling - I think in general self-serve tooling is a missed opportunity in healthcare. Right now most products are so feature bloated because they need to check off all the boxes their end buyer needs, so the products need entire service arms and consulting teams to get you set up. If you need to get accredited to demonstrate proficiency in the software, then it means your software is too complicated and you have a bunch of users who enjoy posting certificates on LinkedIn. Canvas gives you the tools to build things yourself and clear documentation that lets you experiment without needing to hit them up every time. Doing this will hopefully not only result in better products but potentially lower service costs too. 

I’m excited about their open sandbox experiment for developers to take Canvas for a free spin. This tactic has been pretty popular in non-healthcare enterprise software land, where individual developers mess around with these tools to see what they can do and then bring the tool with them to enterprises when they go work there. I’m very curious to see if this tactic works in healthcare software as well.

Collaborative software and task-shifting - A key hallmark of most value-based care models is that it incentivizes both the use of a team-based care model instead of clinicians billing however much they can as an individual AND incentivizes having more touchpoints with a patient outside of the four walls of a clinic. Canvas does a good job of letting providers create different care teams, assign different tasks to different members of the care team, and doing as much as possible outside of the clinic. By doing this, every person on the care team is working at the top of their license and patient-facing time within the clinic isn’t spent on things that probably could have been done over the phone. Yes, you can do a lot of this task managing stuff in EMRs today but they’re usually quite clunky and most of the task routing ends up happening through ad-hoc messaging, emails, etc. 

---

I really like the product, but I do think distribution is still going to be tricky. Here are some questions I’d have about a company like Canvas.

EMR switching - It’s just...really hard to sell an EMR to providers that already have an EMR in place. The switching costs are no joke - lost patients from scheduling downtime, carving out time for training, etc. will already make practices wary of switching over. The irony is that for many practices the upfront cost of switching EMRs might be lower than long term time cost + physician frustration of sticking with the existing EMR. But when many physicians have sunk a bunch of hours learning the specific (probably terrible) workflows of their existing EMR - they’re going to be reluctant to change to a new one that they need to relearn.

This is one of the core reasons Canvas seems to work well with newer providers that are choosing which EMR to use for the first time instead. Canvas claims it takes 2-12 weeks to get up and running including data migration, training, testing, and go live. The open question is whether companies choose to build EMRs in-house (including the certifications, etc. needed) or build on top of a flexible purpose-built EMR like Canvas.

The market size - Canvas is most useful for care delivery organizations that have developers on staff as well. Despite the amount of VC funding that has flooded the space, that’s still a pretty small universe of companies and it’s a bet that it will grow. It’s possible they can start with this small-medium size segment and slowly move into other old school provider groups, but the product will probably have to adapt and an engineering services arm will need to help them get set up. Starting with small-medium businesses and moving to larger enterprises is common in the enterprise SaaS world, with companies like Slack, etc. doing this successfully.

Product complexity - There’s an open question around whether clinicians will want pre-built systems they can just plug into that are simple, standardized, and don’t require developers. Canvas’s superpower is the ability to be customizable, programmable, and modular which might not be for everyone (and could potentially continue to get even more complex as it aims to satisfy for more and more use cases). Right now, third-party applications don’t build on top of Canvas for you to load them up. You would need a developer to connect the API of one product to the Canvas API - which is great for people that need to customize but tough if you want something that works out of the box. You can think of this as a Shopify vs. Stripe battle, and it’s possible the market is big enough to satisfy both depending on your needs. 

Even beyond that, creating products that appeal to small-medium size businesses and larger enterprises is hard because of how feature hungry the latter is. As the tech-enabled care delivery startups get larger and Canvas also starts targeting legacy players, will the current smoothness of the product persist? Or will it become a four letter curse word (the most offensive being JIRA). 

---

In healthcare it’s rare for good, usable products to win in the enterprise space. However, I’m rooting for Canvas because they’ve clearly been thinking really deeply about the functionality of the product and it shows. If we want providers to be able to choose the care models and business models they participate in, they need a flexible EMR that can be modified and adapted by the provider themselves. Anecdotally, I’m already seeing so many companies deciding whether it’s worthwhile to build their own EMR internally and assessing how much of it they want to build. As developers become a bigger part of the healthcare ecosystem and new tech-first care models spring up, this feels like a growing opportunity for companies like Canvas. 

Dr. Saharsh Patel, their VP of Strategy/Ops said this to me and I felt like it was a good encapsulating quote of Canvas: 

“Our end goal is for clinicians to have hope that someone is building for them; operators to see our tool as a cost controller; and for devs to know they can build better/faster/cheaper on top of us with industry leading documentation (that reminds them of non-healthcare dev experiences). 

We have too much at stake and this is our moment. Together on Canvas, clinicians, developers, and believers will create industry leading patient experiences and outcomes.”

I think that’s a good goal to strive for. 

Thinkboi out,

Nikhil aka. “Van Gogh cause I'm about the Canvas”

Twitter: @nikillinit
IG: @outofpockethealth

---

If you’re enjoying the newsletter, do me a solid and shoot this over to a friend or healthcare slack channel and tell them to sign up. The line between unemployment and founder of a startup is traction and whether your parents believe you have a job.