Profile picture
, 3 tweets, 1 min read
My Authors
Read all threads
I always thought this was a longshot to get published in NEJM but the past week has given me hope about our chances... ImageImageImage
Disclaimer: This is a preprint. We preformatted and typesetted it for NEJM to save them some time when we finally submit it.
Some of you are reporting you can’t find it on arXiv or medRxiv. That’s because we didn’t submit it there.

We submitted it to twittRxiv.
Missing some Tweet in this thread? You can try to force a refresh.

Enjoying this thread?

Keep Current with Karandeep Singh

Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

Twitter may remove this content at anytime, convert it as a PDF, save and print for later use!

Try unrolling a thread yourself!

how to unroll video

1) Follow Thread Reader App on Twitter so you can easily mention us!

2) Go to a Twitter thread (series of Tweets by the same owner) and mention us with a keyword "unroll" @threadreaderapp unroll

You can practice here first or read more on our help page!

More from @kdpsinghlab see all

Profile picture
Karandeep Singh
@kdpsinghlab
I’ll be giving a talk on implementing predictive models at @HDAA_Official on Oct 23 in Ann Arbor. Here’s the Twitter version.

Model developers have been taught to carefully think thru development/validation/calibration. This talk is not about that. It’s about what comes after...
But before we move onto implementation, let’s think thru what model discrimination and calibration are:

- discrimination: how well can you distinguish higher from lower risk people?

- calibrations: how close are the predicted probabilities to reality?

... with that in mind ...
Which of the following statements is true?

A. It’s possible to have good discrimination but poor calibration.

B. It’s possible to have good calibration but poor discrimination.
Read 26 tweets
Profile picture
Karandeep Singh
@kdpsinghlab
The DeepMind team (now “Google Health”) developed a model to “continuously predict” AKI within a 48-hr window with an AUC of 92% in a VA population, published in @nature.

Did DeepMind do the impossible? What can we learn from this? A step-by-step guide.

nature.com/articles/s4158…
To understand this work’s contribution, it’s first useful to know what was previously the state of art. I would point to two of @jaykoyner’s papers.

cjasn.asnjournals.org/content/11/11/… and
insights.ovid.com/crossref?an=00…

The 2016 @CJASN paper used logistic regression and 2018 paper used GBMs.
The 2016 CJASN paper is particularly relevant because it was also modeled on a national VA population. Altho the two papers used different modeling approaches, one key similarity is in how the data are prepared: using a discrete time survival method.

What the heck is that?
Read 35 tweets
Profile picture
Karandeep Singh
@kdpsinghlab
I recently had a manuscript accepted that was outlined, written, and revised entirely in Google Docs. Am writing this thread to address Qs I had during the process in case they are of use to others.

#AcademicTwitter
The elephant in the room: how to handle references?

This one is easy: use the PaperPile add-on for Google Docs (not the Chrome extension). It’s free and you can handle citation styles for pretty much every journal.

1. Search paper
2. Click “Cite”
3. Click “Update bib” to format
Do you have a collaborator who likes to make edits in Word?

1. Save the GDoc as a .docx file and share
2. Reimport revised Word doc with tracked changes into GDocs (as new file)
3. Accept/reject suggestions in GDocs

Citations still work! (bc PaperPile refs are roundtrip-safe)
Read 9 tweets

Embed code for your website

×
Follow Us on Twitter!

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just three indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3.00/month or $30.00/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!