Getting Started with Project Alvearie
By David Ranum | Published March 8, 2021
No matter who you are or what you do, everyone has an interest in healthcare. Whether you are a provider, a payer, a patient, a researcher, or some other important part of the complex healthcare web, the way that healthcare is managed and delivered matters to you. Simply put, you care about the way healthcare happens.
In this day and age where technology is pervasive across almost every aspect of daily life, one of the most important drivers for healthcare offerings is the availability and usability of data. In particular, data that is complete, accurate, consistent across sources, and can be used to make healthcare better.
In a previous post, Project Alvearie is described as a new open-source project that “seeks to help healthcare technology professionals unlock and effectively use healthcare data to solve some of the biggest challenges in healthcare.” If you are browsing through the Project Alvearie site, looking at the resources, visiting this blog and reading the posts, then it is likely that you have the interest and passion that we are looking for in users and collaborators.
The only thing missing perhaps is a hint or idea about how to get started. Even though Project Alvearie is new, it already has a number of resources for you to explore and opportunities for you to get involved. So, what are some things that you can do right now? That might depend a little on who you are and what your interests are, so let’s try to describe a few possibilities and jump-start your involvement.
One of the best ways to get started is to do what you are doing now-read and explore the resources that Project Alvearie has to offer. Start with the main Alvearie page and read through the introductory descriptions and rationale. Is the project well-defined? Are the purpose and description clear? Are you able to understand the basic objectives? Move on to explore the architecture and components of the project. Read the documentation related to each. Is it clear? Does it answer your questions?
The goal of these pages is to provide an overview of the project and provide a launching point to learn more about the overall focus and design as well as to detail the individual parts of Alvearie. Often, we think that contributing to an open-source project is only for the original developers; the coders. But that is simply not true. There is plenty of creative work to present the project in a way that is inviting and useful for those who are coming to explore and use the resources.
As you are getting deeper and deeper into the various pages and repositories, if you find documentation that is not complete or seems a bit inconsistent or lacking, propose an issue that describes the problem and how it could be improved. Or, if you are just getting started and find a mistake such as a typo or some other simple error, how about contributing a fix?
No matter who you are, contributing to the project at any level starts here. Request access and join our Slack channels, follow discussions and issue boards, and in general become part of the community. These are things you can do right now!
If you have an interest in healthcare-related data, perhaps as a data producer or consumer, data scientist, or related to some
other data integration tasks, then using tools and techniques that allow access, transformation, enhancement, and storage of
data are fundamental to your success. One of the best places to start exploring is to experiment with the
Alvearie Clinical Ingestion Pattern.
You will find descriptions and details of this pattern in a previous blog post, as well as in source code and deployment repositories.
What is the Clinical Ingestion Pattern? It is an assemblage of open-source tools and technologies that allows you to take healthcare data, likely in different formats and from different sources, and normalize, process, enhance, and eventually store the data in a consistent manner, making it accessible for further analysis and use (such as advanced data analytics or machine learning). The storage destination in this case is a FHIR Server, an industry standard healthcare data storage solution adhering to the HL7 FHIR specification, making it extremely robust and broadly applicable to all types of data. The pattern is referred to as a reference implementation since it shows one possible configuration for doing realistic work and producing a usable end-product that can be modified or extended as needed.
So, what can you do right now? How about trying out the ingestion pattern? Installing the project into your Kubernetes cluster is straight-forward using the Helm charts provided. Just follow the step-by-step instructions you found in the repository and the previous blog post.
Once you have installed the pattern, what can you do? Since the pattern is dependent on a logical flow running in a tool known as NiFi, the best way to get going is to spend a few minutes exploring the open-source NiFi project. Then, jump in by injecting some data into the flow and watch it proceed through the steps on the “canvas” until it finally ends up safely resting in the FHIR server.
In order to experiment with the pattern, find some data, either in FHIR format or in HL7 format, and provide it to one of the two entry points either via Kafka topic or HTTP push with curl or Postman. You can find example data as well as an example command here. As noted in the instructions, if you don’t have your own data, it is also possible to generate synthetic data using the Synthea. Either way, just follow the directions and you’ll be processing data in no time.
Currently, this pattern allows you to ingest and store patient data, validate that the data adheres to the FHIR standards, de-identify a patient by removing identifying information using the De-Identification project, or translate codes to make them consistent with some desired standard.
As you are working through the instructions and reading about the pattern itself, remember, if you find things that are not clear or if you see a mistake or a need to add an additional step, create an issue. Or, better yet, create a fix and propose contributing it back to the project. Feeling more advanced, perhaps you can think of an additional data enhancement you would like to make, or a change you would like to add to the Clinical Ingestion experience. Don’t hesitate to dig into the documentation, learn a bit more about NiFi, and explore the possibilities.
Finally, maybe you have cloud or Kubernetes experience, and perhaps you see additional steps that could be taken to make the install smoother and more robust. Or, you have experience with other cloud platforms and can document steps for installation, where they are similar and where they diverge from those that are provided. If so, add those ideas and instructions to the list. One of the goals of Project Alvearie is to provide resources that can run on many platforms and provide immediate value to the user. This requires a robust implementation and a cloud/platform-agnostic mindset.
These are things you can do right now!
If you are a coder or a tester, beginner or advanced, then Project Alvearie has a number of opportunities that might align with your talents as a software developer. As described above, reading the overview documentation and installing resources such as the Clinical Ingestion Pattern will provide you with a necessary foundation to propose extensions, new features, and new functionality while at the same time allowing you to contribute back to make the documentation better.
But, what can you do right now? How about bugs? All software applications likely have some bugs that need to be identified and fixed. For Alvearie, these bugs are tracked as issues in the repositories. One of the best ways to get involved, if you are so inclined, is to grab one of these existing issues (or perhaps one that you just created), investigate it, and propose and contribute a fix. Remember, not all bugs are easy (you might want to look for the “good first issue” tag to help identify issues that might be appropriate for getting started). In order to understand the issue, you’ll need to be able to reproduce it which will likely mean installing the resource, getting it to run, and then designing test cases that show the issue has been resolved. In the end, you will issue a “pull request” that contains your proposed solution for consideration by the project community.
Not all issues are bugs. Some describe larger features that involve design and discussion with project maintainers. Again, if after using the resource, you have an idea for a new feature, add it as an issue describing what it is, why it would be a good addition, what problem it would solve, and ideas related to the implementation and testing. For example, as a seasoned developer, you might consider the following ideas related to the Clinical Ingestion Pattern:
- Data currently arrives in either HL7 or FHIR format. Perhaps you have a new type of data and you would like to propose a NiFi processor group that would accept that data and convert it into a FHIR format for storage in the back-end server.
- As a follow-on to that idea, you could learn about and propose a custom built NiFi processor capable of carrying out the data transform.
- As part of the data enrichment piece of the pipeline, we might want to read “unstructured text” fields and code them according to some standard. Perhaps an existing open-source natural language processing project could be injected into the flow as a module capable of doing the enrichment by producing FHIR-formatted data corresponding to the data hidden in the unstructured text.
None of these ideas is easy. But they represent the types of advanced features that one might find in a typical open-source project. Features that are challenging to understand, rewarding to develop and deliver, and worthwhile to the community of users. In the end, whether you update documentation, fix a bug, or add a significant feature, you are contributing to the overall success and usefulness of the project.
Using an open-source project can be a great way to get access to tools and applications that support the work of building solutions for problems you are trying to solve. On top of that, contributing to an open-source project like Project Alvearie can be very rewarding, and fun. The activities mentioned above are all things that you can do right now! They range from basic to complex and can make a real difference. Feel free to do any or all of these or come up with your own and add them to our growing list of opportunities.