Particle Health And Pulling Patient Data

One API, one dream

Looking to hire the best talent in healthcare? Check out the OOP Talent Collective - where vetted candidates are looking for their next gig. Learn more here or check it out yourself.

Hire from the Out-Of-Pocket talent collective

EHR Data 101

Hands on, practical introduction to working with data from electronic health record (EHR) systems. Learn practical strategies to extract useful insights and knowledge from messy and fragmented EHR data.
Learn more

Featured Jobs

Finance Associate - Spark Advisors

  • Spark Advisors helps seniors enroll in Medicare and understand their benefits by monitoring coverage, figuring out the right benefits, and deal with insurance issues. They're hiring a finance associate.

Data Engineer - firsthand

  • firsthand is building technology and services to dramatically change the lives of those with serious mental illness who have fallen through the gaps in the safety net. They are hiring a data engineer to build first of its kind infrastructure to empower their peer-led care team.

Data Scientist - J2 Health

  • J2 Health brings together best in class data and purpose built software to enable healthcare organizations to optimize provider network performance. They're hiring a data scientist.

Looking for a job in health tech? Check out the other awesome healthcare jobs on the job board + give your preferences to get alerted to new postings.

Check Out The Job Board


Particle Health provides an API that allows companies to query patient records using demographics like name, date of birth, address, across a bunch of different data sources (EMRs, prescription data, etc.) riding the new interoperability rules. Particle Health gives you a single API to plug into and gives the data output in a structured FHIR format.

The company drops the cost and speed of collecting patient data. It is going to be dependent on dealing with the limitations of FHIR and getting access to more data sources to maintain defensibility.

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

Company Name

Particle Health provides a single API for companies to build on and retrieve patient medical records across a ton of different data sources. The company got its name by combining two areas where no one understands how the math actually works out and the numbers are scary: physics and healthcare.

The company was started in 2018 by Troy Bannister and Dan Horbatt. The company has raised $15M from Menlo Ventures, Collaborative Fund, and some angels like Nat/Zach from Flatiron, Elliot Cohen from Pillpack and Vivek from Clover Health.

What pain points do they solve?

If you’ve been in healthcare for 5 seconds or more, you’ve heard someone talk about data silos and the lack of interoperability. Honestly, people in this industry don’t have much else to talk about.

Our data is spread out all over the place and our information never follows us. I’ve confusingly repeated my medical history more than I’ve repeated the same opening line on dating apps (which is coincidentally my medical history). When a physician wants to get my past medical history it’s from some combination of me trying to recollect my medical history and the physician pinging previous places I’ve gotten care to get full records faxed to them (which they then have to parse out the relevant information).

What if I told you…that things don’t have to be that way. Thanks to the new interoperability/anti-information blocking rules, your data can be much more easily accessed.

With these rules, we have:

  • A decision on which healthcare data classes and elements need to be accessible via API + the format they need to be in (USCDI)
  • An agreed upon standard for how the data should be transported (FHIR)
  • A timeline for when this is going to happen. Things actually kicked off on April 5th, 2021 and the next 18 months have different milestones for when different levels of accessibility need to be met.
  • Real fines and penalties for non-compliance (up to $1M per violation!)

But the problem is not going to be solved suddenly. One of the big issues is still going to be the fact that the data lives in a million different places. If you’re someone building a new healthcare company that requires getting that data, you’d have to hook into all of the different health information exchanges, EMRs, ADT feeds, prescription data networks, etc. to actually pull that data and contract with each of them separately. Honestly that’s pretty annoying + you’ll discover that many of the APIs are maintained really poorly.

Another issue is that we don’t have unique patient identifiers in health records, so there’s actually some work required in making sure it’s the same patient across these different records.

Wouldn’t it be nice if someone else handled those disparate network integrations, contracts, etc. and could actually perform a query to get all the data about a patient across all of the networks?

What does the company do?

Particle Health is essentially a network of health data networks. The company connects into all of these different data providers themselves (EMRs, health information exchanges, etc.), standardizes the data output, and allows developers to deal with one API and vendor.

This generally means you need less engineering resources dedicated to setting up pipelines to the wide array of data vendors that might have the relevant data. Each contract + implementation can take anywhere from 3-24 months depending on how much leverage your company has, the vendor, the complexity of the integration, etc. Using Particle, you just have to interface with Particle and according to Particle the implementation time is about 7-8 weeks.

Particle plugs into these different data sources so you can retrieve patient data quickly across data sources. If you’re a company, you sign a Business Associates Agreement (BAA) with Particle which allows you to use patient data for the specific use cases outlined in the BAA. Particle has their own BAAs with different vendors.

Now let’s say a patient comes to your company. The patient gives your app authorization in advance + a few identifiers (name, date of birth, etc.). Particle will find any instances of patient data across the different data sources it’s plugged into and bring it into your application. If the patient is already established as receiving care, you can streamline this process and query data into their record without needing to get express authorization from every patient every time if you’re operating under clear Treatment use cases as articulated under the HIPAA Treatment, Payment, and Healthcare Operations guidelines. That’s the nice part about everyone having a BAA!

Particle claims to have an 87% success rate of finding patient records that are queried across its network and finds 105 records per query. They have built their own “record locator service” that uses the patient's home location and performs a radius search in conjunction with other heuristics and patient registries they’re hooked into in order to figure out where your records are. Once your record is found, the output is standardized entirely in a FHIR format. You can play around in their sandbox with synthetic patient data if you want to actually see how it works.

Synthetic data above

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

Particle Health is targeting tech-first healthcare service providers. As Troy, the CEO, said to me:

“Our North Star is powering the two entrepreneurs in the garage that are trying to build Venmo for healthcare.”

As someone who lives in New York I didn’t understand what a “garage” was, so he rephrased it to me as “two entrepreneurs working in a coffee shop that buy one coffee each and stay basically the entire day while randomly ordering pastries when they’re feeling peckish...building Venmo for healthcare”.

Some use cases Particle has highlighted:

Digital pharmacies

  • Pull health history including current + past medications, allergies, diagnoses, labs, etc.
  • Flag contradictions - any other medications or health issues which might cause an issue with any new prescription written
  • Trigger interventions - If you want to do something when a patient gets a new prescription written or refilled (e.g. send an SMS or trigger a call when that happens)
  • Monitor outcomes - Get a feed of vitals, functional status, labs, and encounters/admits for patients that are on meds.

Primary Care, Virtual Health, and Chronic Disease Management Companies in Value-Based Care Contracts

  • Pull health history including current + past medications, allergies, diagnoses, labs, etc.
  • Generate simple summaries for clinical team on patient’s health history
  • Monitor patients - Get a feed of vitals, functional status, labs, and encounters/admits for patients that you’re at-risk for.
  • Report population level and individual level diagnoses, treatment plans, etc. to whoever is paying you to keep the population healthy.

Home Care and Remote Monitoring

  • Pull health history, are you noticing a pattern here?
  • Speed up onboarding for new users by pre-filling questionnaires and populating health history where relevant
  • Set up triggers on when to intervene with patients if different health screening values fall out of certain parameters to make sure patients are escalated to urgent care, ER, etc. appropriately.

Particle charges companies $1 per query, and has enterprising pricing for larger volumes. The company expects to grow alongside the companies it sells to.

Job Openings

Particle is hiring for several areas, you can check it out below.

Out-Of-Pocket Take

At a high level, I think companies like Particle will be very useful for any new, tech-first company that is resource constrained and wants to bring patient data into their application. It feels with these new rules coming into effect we’re really seeing the first innings of true interoperability in healthcare.

Particle has likened themselves to “Plaid for healthcare”. I can assure you, I definitely have not heard this pitch 1000 times in the last 2 months for every company that has some semblance of an identity layer from customer and nice looking API docs. Definitely. Have. Not.

Jokes aside though, Particle does have Plaid-like qualities. For the uninitiated, Plaid creates a link between any personal finance app and your bank. Each bank basically has its own janky APIs, ways they authenticate users, and ways they receive and pass back data. Before Plaid, these apps would have to create integrations with every bank and maintain those integrations (and they’d break all the time). Plaid did the integrations for them and created a standardized API. Now, if you use one of these personal finance apps you can connect with Plaid and Plaid maintains those data pipes between the app and your bank. This is one of the reasons why there’s been an explosion of personal finance apps recently.

Particle does something similar by creating and maintaining those integrations for you so that you don’t have to. I’m a big believer that in general we just need more small-medium sized businesses in healthcare, and a core part of that is reducing the costs to start companies. Particle reduces the traditional integration costs and associated resources it would require, which lets companies start and experiment much quicker.

I think every company should focus as much of their efforts on whatever they believe their core competency is. The fact that there are companies who pitch their access to data via integrations as part of their defensibility should be something the healthcare industry moves away from, and instead actually competes on the services rendered instead. Maybe we’ll see a similar explosion in health tech as we’ve seen in fintech thanks to this infrastructure.

Particle does some specific nifty things.

Bidirectional data networks - Health data exchanges are no fun if everyone is only querying data and not putting their own data up to be queried. If your company wants to participate in exchanging data with the national networks that Particle is a part of, it's expected that you share data back. Particle handles that for their customers - for example they enable Oak Street Health to query patient data into their own platform, and other companies can query Oak Street’s clinical data into theirs. Particle has another API that allows for the sharing of data back with all the EMRs and other implementers across the US.  

Creating a “true” FHIR API - To give a little context, right now data that gets exported from an EHR comes in the form of a Consolidated Clinical Document Architecture (CCDA). This document gives a single block of data containing a lot of different elements that are specific to whatever the clinical context is which the document is being created for. FHIR, on the other hand, handles each piece of the encounter as a different “resource” (e.g. a medication, a patient, a type of visit, etc.) and breaks it up piecemeal. The reason this is important is because you have way more flexibility in how you query data and ingest it - you probably wouldn’t need an entire patient document every time you were querying data. With FHIR you can query single pieces of data at a time.

In order to be compliant with the new interoperability rules, a lot of companies are trying to turn their CCDA files into FHIR by basically trying to cram the same data elements from the former into the latter without doing the harder part of actually breaking the elements out, linking them to the right resources in FHIR, etc. As you can imagine, most companies that have to make their data available now are putting the least amount of work to be compliant.

Particle is actually doing the work of converting these CCDA exports into “true FHIR” doing a lot of engineering stuff like semantic normalization + converting XML to JSON + deduplicating the data which goes above my pay grade but results in a cleaner API. I’d be curious how they handle the issue of CCDAs being created for a specific clinical context, but they seem to be putting a lot of resources into actually helping the healthcare data standard world transition into FHIR.

New products thanks to cross-data access - As more disparate data sets become available, linking them together to understand the patient journey at an individual level can create new types of products only feasible if you have access to all of the data and have consents for all of them. For example, you can use the EMR data to see what the provider is prescribing the patient and link it to the pharmacy data to see if/when the patient actually picks up the meds and gets refills.

Here are some questions I’d have for a company like this going forward.

How defensible is this business?

Now that this data is mandated to be interoperable + structured, basically anyone can access it. Particle is trying to differentiate itself in a few ways. The first is investing heavily in engineering resources to make a clean, easy-to-use API for developers. Developer lock-in is powerful, though not perfectly defensible.

The second is by increasing the breadth of data available through their single API that isn’t available on the health information exchanges alone. Particle is working with Cureatr to include pharmacy data, fill data, and medication adherence information for example. This becomes a bit of a two-sided marketplace: applications will go to whichever API has the most patient data from across different sources and data vendors will go to the API that has the most demand to sell their feeds.

Finally, they offer other value-add services to their customers. For example, since we don’t have unique patient identifiers in the US each provider has to figure out ways to use the data they have available to ensure it’s the same patient in the data they’re querying. Particle helps customers with patient matching by figuring out what data they have is valid that would result in higher confidence patient matching.

This space is heating up rapidly, and whether or not Particle wins will be an execution challenge.

How will the company increase the number of patient records it matches?

Currently Particle has ~90% of EMRs accessible through its API including all of the big ones (Epic, Cerner, etc.). The available data covers >250M patient lives.

This is a great start, but what does that look like to getting to 100% EHR adoptions especially when the long-tail is looooooong. What are the best ways to increase the number of covered lives to everyone?

How will the limitations of FHIR and the interoperability rules limit the company?

At the end of the day, Particle is currently limited to what FHIR can do and the rollout/compliance of the new regulation. For example, the law only makes the data available if you send an explicit request to query it, but you can’t push data back into the place where the data is coming from. For example, even if you have access to the data within an EMR, you couldn’t push data into the EMR to schedule appointments.

This also means Particle’s product is specific to the US.

And then there’s the question of timelines. it could be years before some of the more important data like imaging, etc. are mandated to be available via APIs (and therefore Particle). Until then it’s still down to faxes, CD-ROMs, etc. to move this data around.


In general I think there will be lots of different approaches to trying to achieve data interoperability + making it easy as an API call to get all of the patient’s data in one place is a dream that feels closer than ever before. Companies like Particle Health are slowly bringing that dream closer and have the potential to make it easier to start a healthcare business by making data lock-in less of an issue.

Companies like Particle make me feel like we’re so close to a flourishing health data ecosystem, but that could be my naive optimism once again. Just let me have this one.

Thinkboi out,

Nikhil aka. “article health"

Twitter: @nikillinit

IG: @outofpockethealth

Let's Keep In Touch

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
search icon