Joy Jin (@joyqjin) header
Joy Jin (@joyqjin) profile picture

Incoming PGY-1 @KPSanFrancisco and Dermatology Resident @nyulangone • MD @UCSFMedicine '24 • AB @Harvard '19 • Judge @USFigureSkating • She/her

San Francisco, CA
Joined December 2019
Photo by Joy Jin (@joyqjin)Photo by Joy Jin (@joyqjin)Photo by Joy Jin (@joyqjin)Photo by Joy Jin (@joyqjin)Photo by Joy Jin (@joyqjin)Photo by Joy Jin (@joyqjin)Photo by Joy Jin (@joyqjin)Photo by Joy Jin (@joyqjin)
Excited to share a recent project! With 10+ systemic biologics for psoriasis, it can be difficult to predict which patients respond well to the first one attempted vs. switch multiple times d/t poor efficacy. How prevalent is biologic failure anyway? bitly.ws/P8cK 1/
2
1
3
1,090
we constructed a multivariable regression model. This predicted F sex, shorter psoriasis duration, earlier year of biologic initiation, prior non-biologic systemic therapy use, history of HL, and Medicaid insurance to be significantly associated with MBF. 4/
1
516
Intriguing results and we have another multi-center study ongoing across the U.S. to analyze these findings further! 🔍 5/5
422
Fascinating that AI improves radiologist performance for relatively certain cases, but not for cases with >uncertainty. I wonder if performance was measured based on critical/MNM diagnoses vs. just overall accuracy tho. Impt implications for inevitable AI-human collaborations
We gave 180 radiologists an expert-level AI for diagnosis. But instead of improving, their performance didn't budge. 😮 Why? Radiologists implicitly discount AI predictions, favoring their own judgment - a bias we call "automation neglect". 🧠 Let's dive into our findings. 🚀
870
Doctors at a major US insurance company deny tens of thousands of claims a month without even looking at patients’ files. “We literally click & submit,” a former Cigna doctor told @ProPublica & @Capitol_Forum. “It takes all of 10 seconds to do 50 at a time.” The inside story 🧵
467
8,225
2,828
14,917
8,398,366
Last year, we noticed a surge in the use of Mendelian randomization (MR), a genetics-based epidemiology method, in the dermatology literature as a way to learn how modifiable health behaviors (e.g., smoking) can impact disease development at a population level. 1/
1
1
8
1,648
Some findings: smoking, obesity, cardiovascular disease, Crohn’s (but not UC!), depression were causally associated with PSO/PSA; PSO was causally associated with several comorbidities. We discuss areas of agreement/controversy with prior literature and suggest additional... 6/
1
3
713
strategies to increase the robustness of future MR studies, with the hope that our findings can improve preventive counseling and precision medicine for patients. 7/7
1
443
100+ page "clinician of the future" report by @elsevier - awesome read on tech-enablement of clinical workforces. remarkable that only 59% of physicians in north america agree that "personalized treatment approaches will be more widely used" vs. 88% in APAC (global avg of 77%).
4
3
1
29
After a long Twitter hiatus, I'm excited to share more about the research I've been working on 👀 The Liao lab uses single-cell techniques for several projects and we thought it would be helpful to review the landscape of these studies in psoriasis... 1/ tandfonline.com/eprint/EZNNA…
1
1
3
Immune cell states are dynamic; subpopulations can aberrantly re-activate/differentiate into inflammatory phenotypes depending on surrounding signaling cues. Furthermore, comparison of circulating immune cells with resident skin/joint cells has uncovered... 5/
1
2
specific T cell clonotypes associated with disease. Finally, machine learning models demonstrate great promise in identifying biomarkers to diagnose clinically ambiguous rashes and PsA at earlier stages. 6/6
2