
Data science, epidemiology @genentech. Complex health data modeling & ML. Formerly @HarvardChanSPH @CCDD_HSPH
San Francisco, CA
Joined January 2012
- Tweets154
- Following61
- Followers125
- Likes29
Sarah McGough retweeted
🎓The Turing-@Roche Partnership is delighted to announce our Community Scholars!
10 PhD students join the partnership community, attending a conference & running a community project on topics around the partnership & #AI in health.
Find out more: bit.ly/3tk2AIk
Sarah McGough retweeted
With over 100 incredible contributors – experts, activists and authors – it highlights many of the different faces of the crises we face, and aims to connect the dots between them.
2/4
Interpretable ML *heavens sing* great work @MauSantillana and team!
Our paper documenting a new interpretable deep learning-based approach for predicting the risk of a patient's no-show is out in @Nature_NPJ Digital Medicine rdcu.be/cLP7n Great collaboration with @DianboLiu WY Shin, E Sprecher, K Conroy, O Santiago, G Wachtel
2
Coming up next week: structured missingness workshop with the #AlanTuringInstitute. We are examining a special (and interesting!) form of heterogeneity in multimodal data. @turinginst
EVENT: In partnership with @Roche we are inviting #Researchers to participate in a 3-part #VirtualWorkshop on #StructuredMissingness A call for papers will be announced with funding available for successful proposals
ow.ly/3F1T50GtJC6
Sarah McGough retweeted
Happy to see our paper w/@EsMagoo,@robtibshirani & all on prediction modeling with high-dimensional left-truncated and right-censored survival data published in Statistics in Medicine. Need to be especially careful using C-index for model evaluation. onlinelibrary.wiley.com/doi/…
Sarah McGough retweeted
Our work on "A dynamic, ensemble learning approach to forecast dengue fever epidemic years in Brazil using weather and population susceptibility cycles" just out at the Journal of the @royalsociety interface royalsocietypublishing.org/d… with @EsMagoo @SKNKsmash Nathan Kutz
Uncovering immortal time bias when fitting prognostic survival models in a large clinico-genomic database. Here, our assessment w/ @DevinIncerti, @robtibshirani & all (pre-print).
medrxiv.org/content/10.1101/…
Tl;dr, misleading, and occasionally conflicting, conclusions about model performance if unaddressed (e.g. extra bad calibration, not-extra-bad discrimination, biased coefficients)
Our GitHub: github.com/phcanalytics/coxn…
Sarah McGough retweeted
Analysis of routine datastreams to make timely decisions is usually complicated by reporting lags. New paper from Sharon @nychealthy applies "nowcasting" methods by @rebeccajk13, @EsMagoo, and @CCDD_HSPH to estimate real-time COVID-19 incidence in NYC: publichealth.jmir.org/2021/1…
Sarah McGough retweeted
There is a new ‘variant’ clearly identified today, P.1. And it is worth saying a little about what we have learned about it, and from the variants we have identified so far virological.org/t/genomic-ch… 1/quite_a_few
Sarah McGough retweeted
We're deploying vaccines like we do annual flu shots rather than the mass vaccinations done in Africa & Asia where millions are vaccinated in weeks
Years ago, Joe Rhatigan & I studied these campaigns. They offer several lessons that can inform what we need in the US now
THREAD
Sarah McGough retweeted
Now online led by Sharon Greene, @EsMagoo, and @rebeccajk13 Nowcasting for Real-Time COVID-19 Tracking in New York City: Evaluation Study Using Reportable Disease Data From th… pubmed.ncbi.nlm.nih.gov/3340…
Sarah McGough retweeted
There is much more discussion in this excellent and thorough blog post virological.org/t/preliminar… (thx @WvSchaik for alerting me) but a few headline comments follow 13/n
Sarah McGough retweeted
I've been asked about this a lot so I wrote a blog post @ccdd_hsph COVID-19 Vaccines and Herd Immunity ccdd.hsph.harvard.edu/2020/1…
Thanks for capturing these learnings- we hope they resonate with many who work with complex health data!
Pan-tumor precision medicine via personalized healthcare insights. Great talk at #NeurIPS2020 #ML4H workshop by @EsMagoo, @robtibshirani. "rigorous models require intimate understanding of your data". This is the future of medicine.
1
A pleasure to share the work we are doing in the machine learning and personalized healthcare space! #ML4H #NeurIPS2020
Replying to @_hylandSL
Sarah McGough (@EsMagoo) from Roche is now talking about her work with collaborators Rob Tibshirani (@robtibshirani) and Balasubramanian Narasimhan on pan-tumor oncology, and challenges with real-world datasets. #ML4H #NeurIPS2020
Sarah McGough retweeted
We explore to what extent two rising public health threats — climate change and antibiotic resistance — are related in our recent paper published in @Eurosurveillanc. Great collaboration with @EsMagoo D MacFadden, MW Hattab, @KareMolbak eurosurveillance.org/content…
Sarah McGough retweeted
Our recent work on monitoring emerging disease outbreaks using novel data sources just became available at @PLOSCompBiol Great collaboration with E Aiken, @EsMagoo @maiamajumder G Wachtel, @AndreNguyen16 and C Viboud doi.org/10.1371/journal.pcbi… @Bos_CHIP
Sarah McGough retweeted
This is the method that @EsMagoo @majohanss (@CDCgov) Nick Menzies (@decisionscience) published recently journals.plos.org/ploscompbi… and is in use in health depts around the world. Now easier to use than ever!














