Kirstin Oliveira Roster
@kroster.bsky.social
100 followers 100 following 12 posts
Infectious disease modeling and forecasting, global health, surveillance of drug resistance. Postdoc @ Harvard School of Public Health
Posts Media Videos Starter Packs
kroster.bsky.social
Thanks to @YHGrad and Peter White for the supervision and collaboration!
kroster.bsky.social
A critical point: CDC data defined the trends and lay the foundation to guide and optimize interventions. We need continued and expanded public health efforts, not cuts, else we risk losing these gains in health.
kroster.bsky.social
We outline next steps in research, data collection, and surveillance to address each of these possible factors.
kroster.bsky.social
We review possible drivers and their expected impact on gonorrhea, chlamydia, and syphilis:
🧪 Changes in screening, reporting, & diagnostic accuracy
👥 Impact of COVID-19 & mpox on sexual behavior
💉 Protection through vaccination with Bexsero
💊 Doxy-PEP
📋 Changes in antibiotic treatment guidelines
Reposted by Kirstin Oliveira Roster
franciscorodrigues.bsky.social
In our new paper in PLOS Neglected Tropical Diseases, we quantified the gap between expected and observed dengue cases in Brazil. We used an interrupted time series approach, incorporating predictions generated by machine learning models. With @kroster.bsky.social
journals.plos.org/plosntds/art...
Reposted by Kirstin Oliveira Roster
kroster.bsky.social
Thank you to all our colleagues at MDPH, CDC, Stanford, and HSPH for the support on this project! yhgrad.bsky.social mintturonn.bsky.social jsalomon.bsky.social as well as Heather Elder, Kathy Hsu, Kathleen Roosevelt (MDPH), and Tom Gift (CDC), not (yet) on Bluesky.
kroster.bsky.social
Scaling up surveillance is costly – our model helps improve interpretation of existing data by estimating the infection landscapes that could underlie what we observe through surveillance.
kroster.bsky.social
If we stop detecting new cases, how long must we wait before we can be confident that there is no ongoing transmission? The likelihood of strain elimination rose from only 34% when the strain was first detected to 98% after 6 months without detected cases.
kroster.bsky.social
When a strain is first identified, we estimated that an additional 5.4 cases would be undetected; but the estimate dropped to 2.5 undetected cases with no new infections reported over the next 180 days.
kroster.bsky.social
In scenarios with two initial detected cases, our simulations varied from quick elimination to persistence, highlighting the challenges facing public health agencies as they weigh how to respond.
kroster.bsky.social
We worked to address this question using a stochastic model, in collaboration with the Massachusetts Department of Public Health, as the first strains in the US with reduced susceptibility to all approved antibiotics were detected here in MA in 2022. www.thelancet.com/journals/lan...