A year+ ago, a team from @ResearchKi โ @Yonatan_Bilu , Guy Amit, @IrenaFrenklach, and myself โ participated in, and won ๐, the #SNOMED Entity Linking Challenge!
We're excited to share that the paper describing the challenge and the winning submissions has just been published!
These findings highlight the complex and multisystemic nature of endometriosis. They also emphasize the need for a multidisciplinary approach to both diagnosis and management, one that takes into account a broader range of comorbidities than currently considered.
Just published in npj Womenโs Health: Unveiling endometriosis hidden comorbidities using a data-driven approach: a retrospective matched cohort study.
https://t.co/44op9u48VW
@TZelovich
More details ๐
Our analysis confirmed known comorbidities (e.g., migraines and allergic disorders) and identified new ones (e.g., acute laryngitis, sinusitis, viral infections, and sciatica)
Our study, published in @npjDigitalMed, presents a method may become an important accelerator of robust model deployment. With the synergetic team Jenna Reps and @ChenYanover .
https://t.co/bGiWSxpTH2
An awesome gift for the Holidays:
The first @OHDSI network study that we, at the KI Research Institute (@ResearchKi), coordinated in close collaboration with Rambam Health Care Campus gastroenterologists, has been just published
https://t.co/LTOakyYaWm
It characterizes the disease trajectory of over a million patients with inflammatory bowel disease (#IBD), using routinely collected healthcare data from sixteen databases in seven countries (USA ๐บ๐ธ, UK ๐ฌ๐ง, France ๐ซ๐ท, Germany ๐ฉ๐ช, Japan ๐ฏ๐ต, Korea ๐ฐ๐ท, and Australia ๐ฆ๐บ)
I'm excited to share my recent study with Chen Yanover and Vered Klaitman with perfect timing for Endometriosis Awareness Month.
In this work, we explore endometriosis and comorbidities, providing additional evidence that endometriosis is a multi-system disease.
Just published: an @OHDSI network study that investigates the effect of feature selection on modelsโ external performance, across 4 countries: ๐บ๐ธUS, ๐ฌ๐งUK, ๐ซ๐ฎFinland, ๐ฐ๐ทKorea
Led by @naderalvojoud, @tmboussard; w/ @TalElHay, @JGT_bio_eng, Byungjin Choi, Thomas Falconer, et al.
Our latest study showcases the power of international collaboration, developing predictive models for post-surgery outcomes with unparalleled generalizability using theย OHDSI network https://t.co/7B0xau17sF @ccurtinprs
In https://t.co/UoHviwWmkC we developed a method to do that. Now we benchmark it in real-world #clinical settings using data from five US datasets and prediction models for various outcomes, and demonstrate its accuracy. More to come ...
@TalElHay, Jenna Reps, @OHDSI
An exciting project in the making:
Say you trained a classifier on your data, obtaining good performance. Next, you want to deploy it to a different environment but all you know โ or can obtain โ are limited statistical characteristics of the external data.
Estimating model performance on external data sources from their summary statistics: a real-world benchmark #OHDSISocialShowcase
Lead: Tal El-Hay
Co-Authors: Jenna M Reps, Chen Yanover
Learn more: https://t.co/3zmEYwFnOQ
#JoinTheJourney
Data-driven assessment of mental health among children and adolescents with food allergy #OHDSISocialShowcase
Lead: Natalie Flaks-Manov
Team: Inbal Goldshtein, Chen Yanover
Learn more: https://t.co/GuxOftt9S6
#JoinTheJourney
... ranitidine (which treats gastroesophageal reflux and peptic ulcer) does not increase the risk of cancer compared with use of other H2ย receptor antagonists. Thanks @SengChanYou1 for leading this effort.