OHDSI Home | Forums | Wiki | Github

New WG - Book of OHDSI

New Working Group: The Book of OHDSI

Organizers: Martijn Schuemie and David Madigan

The “Book of OHDSI” will be a community-created open resource that will serve as a knowledge repository for all things OHDSI. We envision that the book will be especially useful as an entry point into OHDSI but also to capture the rapidly growing global OHDSI knowledge base. The book will cover topics such as:

  • The OHDSI Community and History
  • The Common Data Model
  • OHDSI Analytics Tools
  • Cohort Construction
  • Characterization
  • Population-level Estimation
  • Patient-level Prediction
  • LEGEND

We will use GitHub to manage and distribute the book and make use of bookdown. We may consider a paper version of a snapshot of the book at some point in the future.

Please join us! The first weekly meeting will take place (URL, etc. to follow later):

Tuesday November 27th

11am Eastern Time

8am Pacific Coast Time

5pm Central European Time

5 Likes

So many researchers in Korea keep asking me why there is no book for OHDSI. Now I can tell them that we’re making it!

Thank you for leading this great work! @David_Madigan @schuemie

1 Like

This is a great idea. Count me in.

Good idea! But I think we should aim at publishing it as a textbook, and marketing it to the graduate schools, so that to make OHDSI OMOP CDM de facto standard.

Agree that this is a great idea, and also agree with the textbook idea. Yes, textbooks are obsolete on the day of publication, but, as I think I might have brought up in the past, a book that you can find on Amazon (or at a bookstore) provides a level of credibility to something, as in “See, this real!”. Plus, would it not look great to see a copy of “Book of OHDSI, 1st Edition” on your bookshelf?

CWRU and the Cleveland Institute for Computational Biology (CICB) will contribute to these efforts, and we will discuss at our next team meeting.

Yes, sounds cool. Would be interested in contributing to that effort as well. I would add a few topics, such as OMOP Standardized Vocabs as well as Network Studies

I am also very interested in this topic, I think it could be very beneficial to anyone who is new to OHDSI and not sure where to start.

I’m super interested but will be at AWS Reinvent next week. Hope to join in early Dec!

Also very interested in the topic, and learning more on how to effectively communicate the dynamic work OHDSI!

That’s Cool! Count me in, if the book needs something relate to my study :slight_smile:

Great idea! Will certainly contribute a section on probabilistic phenotyping with Aphrodite to it.

I would welcome the opportunity to contribute, I’m looking forward to our first call.
Thank you for the leadership on this topic!

What an excellent project! Thank you for spearheading this endeavor. I’m happy to help and contribute however I can!

Thanks everyone for your interest! I’ve started drafting an outline here (everyone can comment. If you want to edit I think you need to request that).

Please let us know any topics you think should be added, either in that outline or here in the forums.

@SCYou: just so you know, the idea for this workgroup was borne out of the discussion you, me, and @rwpark had in that small restaurant in Taipei :wink:

Examples of other community-developed books are:

2 Likes

We discussed this earlier but for the record count me in for prediction and i can contribute to chapter on international collaborations for OHDSI Europe and EHDEN.

1 Like

@schuemie These chapters represent natural divisions of OHDSI resources. That makes sense and has a lot to recommend it. It allows the related materials already developed for each resource to be easily brought to bear on its chapter. I think that is an important advantage.

Another way to organize the book could be to reflect the process of going from initial interest in OHDSI to becoming a fully participating member that derives great value from participation. Chapters might cover the initial work of getting a resource stood up, learning about the community, starting to evaluate and use the resource, … developing collaborations, leading a study… publishing results.

In that structure different aspects of each OHDSI resource would be highlighted in the chapter devoted to that stage of the process. The community, for example, plays a role at each stage of that journey, as does data characterization.

A journey structure might more easily use examples that carry through the book. Examples could be more than a study. There could be a set of examples that reflect prototypical person + institution + motivation scenarios. A journey structure also might more naturally demonstrate the functional relationships among OHDSI components and makes their purpose and value clear.

The same materials for each resources would still be provided, but in a context that is focused on the needs and activities at a different stage of involvement. That might be especially helpful to newer or prospective community members and might be less helpful to more experienced members who just want to look things up. I doubt the drawbacks for the latter group would be significant.

A more significant potential drawback to a journey structure might be a greater need for new material development: writing up how people in different roles at different institutions become involved, get resources stood up, identify how OHDSI helps meet their needs, use OHDSI to meet those needs, etc. Its possible that could be very valuable new material, but it would almost certainly slow the pace of writing the book.

Some of that additional work would be required of any illustration of an example that persists throughout the book. But structuring the book around that journey would almost demand more new material about people in different roles at different types of institutions with different reasons for participation.

Even if there are benefits like the ones I ascribe to the journey structure, it might not be better than the structure you proposed or might not be enough of an improvement to justify the additional work. So I’m not advocating for it, just proposing it as food for thought.

I will follow Andrew’s approach to suggest changes through this forum post.

@schuemie - great work with creating an initial outline. I actually like that you bundled OMOP Standardized Vocabularies, ETL, Themis, and other data/CDM related activities into one chapter.

I would propose the following addition - a separate chapter on “OHDSI Network Research” to describe the goals, challenges and process. I think it is important to emphasize that OHDSI is not only about developing tools and methods to conduct analytics on a single site but also enabling a large scale research across a network of participating organizations.

I’d like a better understanding of the purpose of this book. Will it be for academics or builders?

I ask because I taught college courses in Relational Database Design for 12 years. The first couple of years, I taught from college texts, but I noticed that my students were struggling. After that, I switched to teaching from trade press books (Database Design for Mere Mortals by Michael Hernandez in particular), and my students became more successful. One reason is that my students were taking my class to learn how to build databases, and not to advance the study of Relational Theory.

So my questions are: Who is the intended audience? Architects or builders? Will this book will be at textbook or a trade press book? Will it be focused on understanding the theory, or implementing OMOP?

Don’t get me wrong, both have their places. I just don’t think a single book can address them both. For one thing, textbooks and trade press have different tones and writing styles. Textbooks tend to start with theoretical and proceed to concrete examples. Trade press books usually skimp on the theory and quickly move to the implementation.

1 Like

Martijn - count me in if you need content for NLP and China working group activities. thanks

Hua

I would like to contribute.
Possible topics are data quality, network studies, community/history, vocabulary, data model (advancing measurements, procedures)

1 Like
t