Skip to content
  • Home Home Home, current page.
  • Moments Moments Moments, current page.

Saved searches

  • Remove
  • In this conversation
    Verified accountProtected Tweets @
Suggested users
  • Verified accountProtected Tweets @
  • Verified accountProtected Tweets @
  • Language: English
    • Bahasa Indonesia
    • Bahasa Melayu
    • Català
    • Čeština
    • Dansk
    • Deutsch
    • English UK
    • Español
    • Filipino
    • Français
    • Hrvatski
    • Italiano
    • Magyar
    • Nederlands
    • Norsk
    • Polski
    • Português
    • Română
    • Slovenčina
    • Suomi
    • Svenska
    • Tiếng Việt
    • Türkçe
    • Ελληνικά
    • Български език
    • Русский
    • Српски
    • Українська мова
    • עִבְרִית
    • العربية
    • فارسی
    • मराठी
    • हिन्दी
    • বাংলা
    • ગુજરાતી
    • தமிழ்
    • ಕನ್ನಡ
    • ภาษาไทย
    • 한국어
    • 日本語
    • 简体中文
    • 繁體中文
  • Have an account? Log in
    Have an account?
    · Forgot password?

    New to Twitter?
    Sign up
jameshay218's profile
James Hay
James Hay
James Hay
@jameshay218

Tweets

James Hay

@jameshay218

Postdoc at Harvard Chan School of Public Health using maths and stats to understand infectious disease dynamics. Mostly viral kinetics and serology.

London, England
github.com/jameshay218
Joined December 2015

Tweets

  • © 2022 Twitter
  • About
  • Help Center
  • Terms
  • Privacy policy
  • Cookies
  • Ads info
Dismiss
Previous
Next

Go to a person's profile

Saved searches

  • Remove
  • In this conversation
    Verified accountProtected Tweets @
Suggested users
  • Verified accountProtected Tweets @
  • Verified accountProtected Tweets @

Promote this Tweet

Block

  • Tweet with a location

    You can add location information to your Tweets, such as your city or precise location, from the web and via third-party applications. You always have the option to delete your Tweet location history. Learn more

    Your lists

    Create a new list


    Under 100 characters, optional

    Privacy

    Copy link to Tweet

    Embed this Tweet

    Embed this Video

    Add this Tweet to your website by copying the code below. Learn more

    Add this video to your website by copying the code below. Learn more

    Hmm, there was a problem reaching the server.

    By embedding Twitter content in your website or app, you are agreeing to the Twitter Developer Agreement and Developer Policy.

    Preview

    Why you're seeing this ad

    Log in to Twitter

    · Forgot password?
    Don't have an account? Sign up »

    Sign up for Twitter

    Not on Twitter? Sign up, tune into the things you care about, and get updates as they happen.

    Sign up
    Have an account? Log in »

    Two-way (sending and receiving) short codes:

    Country Code For customers of
    United States 40404 (any)
    Canada 21212 (any)
    United Kingdom 86444 Vodafone, Orange, 3, O2
    Brazil 40404 Nextel, TIM
    Haiti 40404 Digicel, Voila
    Ireland 51210 Vodafone, O2
    India 53000 Bharti Airtel, Videocon, Reliance
    Indonesia 89887 AXIS, 3, Telkomsel, Indosat, XL Axiata
    Italy 4880804 Wind
    3424486444 Vodafone
    » See SMS short codes for other countries

    Confirmation

     

    Welcome home!

    This timeline is where you’ll spend most of your time, getting instant updates about what matters to you.

    Tweets not working for you?

    Hover over the profile pic and click the Following button to unfollow any account.

    Say a lot with a little

    When you see a Tweet you love, tap the heart — it lets the person who wrote it know you shared the love.

    Spread the word

    The fastest way to share someone else’s Tweet with your followers is with a Retweet. Tap the icon to send it instantly.

    Join the conversation

    Add your thoughts about any Tweet with a Reply. Find a topic you’re passionate about, and jump right in.

    Learn the latest

    Get instant insight into what people are talking about now.

    Get more of what you love

    Follow more accounts to get instant updates about topics you care about.

    Find what's happening

    See the latest conversations about any topic instantly.

    Never miss a Moment

    Catch up instantly on the best stories happening as they unfold.

    1. James Hay‏ @jameshay218 13 Oct 2020

      Ct values can be used to estimate epidemic dynamics! We show that the distribution of viral loads changes during an epidemic and develop a new method to infer growth rates from cross-sectional virological surveys without using reported case counts. 1/25https://www.medrxiv.org/content/10.1101/2020.10.08.20204222v1 …

      32 replies 226 retweets 501 likes
      Show this thread
    2. James Hay‏ @jameshay218 13 Oct 2020

      Highlights: - The distribution of observed viral loads is determined by recent incidence trends - The median and skew of detectable Cts in Massachusetts were correlated with R(t), as predicted - A novel statistical method to infer the epidemic growth rates using Ct values 2/25

      1 reply 9 retweets 45 likes
      Show this thread
    3. James Hay‏ @jameshay218 13 Oct 2020

      Cts are inversely proportional to log viral loads. The relationship depends on the instrument and sampling variation, but low Cts generally indicate high viral loads. Think about expectations & distributions rather than individual measurements https://academic.oup.com/cid/advance-article/doi/10.1093/cid/ciaa619/5841456 … 3/25

      1 reply 8 retweets 37 likes
      Show this thread
    4. James Hay‏ @jameshay218 13 Oct 2020

      James Hay Retweeted A Marm Kilpatrick

      In acute infections, viral loads generally follow a consistent (but noisy) trajectory. We aren't certain of the full trajectory, but we know that i) viral growth is faster than decline and ii) people can persist at low viral loads for a long time. https://twitter.com/DiseaseEcology/status/1312537935892246528?s=20 … 4/25

      James Hay added,

      A Marm Kilpatrick @DiseaseEcology
      PSA: We (STILL) have no data to know the pattern of viral loads over time from infection to recovery. So we don't know how test sensitivity & infectiousness correlate. Tons of stories w/ quotes from top people are not making this clear & it matters. Clarifying thread
      Show this thread
      1 reply 5 retweets 27 likes
      Show this thread
    5. James Hay‏ @jameshay218 13 Oct 2020

      So, a high viral load likely indicates recent infection, whereas a low viral load is *more likely* (note, not guaranteed) to indicate an older infection. 5/25

      1 reply 4 retweets 33 likes
      Show this thread
    6. James Hay‏ @jameshay218 13 Oct 2020

      OK… where am I going with this? Well, it’s been noted a few times that the distribution of Ct values has changed over the course of the pandemic, eg: https://www.medrxiv.org/content/10.1101/2020.07.20.20157792v1.full.pdf … How puzzling! 6/25

      2 replies 3 retweets 27 likes
      Show this thread
    7. James Hay‏ @jameshay218 13 Oct 2020

      One way this can arise is through biased sampling eg. if we mostly sample people soon after symptom onset then we’ll see low Cts, whereas if we sample people after they’ve recovered, we’ll see higher Cts. But if we sample people at random, this shouldn’t matter. 7/25

      2 replies 5 retweets 26 likes
      Show this thread
    8. James Hay‏ @jameshay218 13 Oct 2020

      Then why have Ct values changed? It’s because of *epidemic dynamics*! Imagine we sampled a bunch of infected people today. If we knew when those people got infected, we’d notice that (8/25):

      1 reply 3 retweets 25 likes
      Show this thread
    9. James Hay‏ @jameshay218 13 Oct 2020

      1. If the epidemic is growing, people were typically infected more recently. If the epidemic is declining, people were typically infected further in the past. The distribution of *times since infection* depends on the incidence of infections. 9/25pic.twitter.com/Ab1yhJr063

      2 replies 15 retweets 56 likes
      Show this thread
      James Hay‏ @jameshay218 13 Oct 2020

      2. The time since infection dictates observed viral loads. People who were infected a long time ago typically have lower viral loads, because the decline phase of viral kinetics takes longer that the growth phase (remember – expectations not individuals!). 10/25pic.twitter.com/r4RNcLwKXa

      7:20 PM - 13 Oct 2020
      • 12 Retweets
      • 34 Likes
      • Juan J Gil V noa-witheringly Matt Flor Javichu Wear a mask, wash your hands, get vaccinated Isaac Demme David Niño Beorn Black Belen Martrat
      1 reply 12 retweets 34 likes
        1. New conversation
        2. James Hay‏ @jameshay218 13 Oct 2020

          3. So the distribution of viral loads sampled today depends on the distribution of times since infection, which in turn depends on the epidemic growth rate. Therefore, with appropriate calibration, the distribution of observed Ct values is informative of epidemic dynamics. 11/25pic.twitter.com/5QJuhoM8We

          2 replies 24 retweets 71 likes
          Show this thread
        3. James Hay‏ @jameshay218 13 Oct 2020

          This is a simple cartoon, but it shows how epidemic dynamics -> time since infection distribution -> viral load distribution -> observed Ct values. How these steps link might change depending on eg. the instrument used to measure Ct values, but the general principle holds. 12/25

          1 reply 5 retweets 29 likes
          Show this thread
        4. James Hay‏ @jameshay218 13 Oct 2020

          To demonstrate this, we simulated infections from an SEIR model. For each infected person, we simulated Ct values based on when in their infection course they were sampled. R(t) is clearly correlated with the median and skew of the detectable Ct distribution. 13/25pic.twitter.com/sqvw1YwzDI

          1 reply 6 retweets 17 likes
          Show this thread
        5. James Hay‏ @jameshay218 13 Oct 2020

          What about real data? We compared estimated R(t) (using EpiNow2) to the distribution of measured Ct values from Brigham & Women’s Hospital, MA since April. Despite a bit more noise, we saw the same relationship as in our theoretical predictions! 14/25pic.twitter.com/8D26FlCaPg

          3 replies 9 retweets 27 likes
          Show this thread
        6. James Hay‏ @jameshay218 13 Oct 2020

          Finally, (and this is still WIP), we used this relationship to write down a model to infer the epidemic growth rate using the distribution of Ct values observed on a particular day. 15/25

          1 reply 3 retweets 22 likes
          Show this thread
        7. James Hay‏ @jameshay218 13 Oct 2020

          The model has two parts: 1) a viral kinetics model describing the mean and variance of Ct values on each day post infection and 2) a model describing the relative probability of infection on each day prior to the observation day. 16/25

          1 reply 3 retweets 18 likes
          Show this thread
        8. James Hay‏ @jameshay218 13 Oct 2020

          We found that with a single cross section of detectable Cts, we could correctly identify if the epidemic was growing or declining. They’re noisy estimates, but bear in mind this is showing longitudinal dynamics from a *single cross section* of data. 17/25pic.twitter.com/i8wn2cKz9B

          2 replies 16 retweets 48 likes
          Show this thread
        9. James Hay‏ @jameshay218 13 Oct 2020

          Also, to account for uncertainty in the viral kinetics trajectory, we used a Bayesian approach and placed informative priors on most of the model parameters rather than fixing them. If we knew these parameters with more certainty, these estimates would be more constrained. 18/25

          1 reply 4 retweets 22 likes
          Show this thread
        10. James Hay‏ @jameshay218 13 Oct 2020

          We’ve shown how the distribution of Cts changes during an epidemic, which reflects the trajectory. We propose that incorporating Ct values epidemic dynamics inference can retain more information than simply assessing “how many new PCR positive tests did we report today?” 19/25

          1 reply 7 retweets 37 likes
          Show this thread
        11. James Hay‏ @jameshay218 13 Oct 2020

          We can test whether the epidemic is growing or declining based on a single cross-section of viral load measurements. This could be a valuable tool to discern if rising case counts are due to true rises in incidence or just changing testing capacity, as some have suggested. 20/25

          1 reply 10 retweets 37 likes
          Show this thread
        12. James Hay‏ @jameshay218 13 Oct 2020

          We hypothesize this method would be less susceptible to bias than R(t) estimation using case counts, which are hindered by reporting delays and testing capacity. The information here comes from the *distribution* of Cts rather than the (potentially biased) number of cases. 21/25

          2 replies 7 retweets 33 likes
          Show this thread
        13. James Hay‏ @jameshay218 13 Oct 2020

          For anyone still skeptical, I refer to the large literature on using antibody titers to infer epidemic dynamics. Antibody titers -> time since infection -> incidence. This is exactly the same idea, just using viral loads. https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1007840 … 22/25

          1 reply 10 retweets 51 likes
          Show this thread
        14. James Hay‏ @jameshay218 13 Oct 2020

          There are some exciting extensions that we’re working on, so stay tuned for more data and some tighter posteriors! All code here: https://github.com/jameshay218/ct_dynamics_preprint … ... though we're not able to share the raw BWH data at this time. 23/25

          2 replies 4 retweets 22 likes
          Show this thread
        15. James Hay‏ @jameshay218 13 Oct 2020

          I foolishly put this at the end, but this is a joint endeavor with Lee Kennedy-Shaffer (not on twitter), led by @michaelmina_lab, with invaluable input from @SanjatKanjilal and @mlipsitch 24/25

          3 replies 1 retweet 33 likes
          Show this thread
        16. James Hay‏ @jameshay218 13 Oct 2020

          … and I’ll stop there before someone has to start saying “THANK you, Mr. Vice President” over and over again until I shut up. (Topical reference, not suggesting that I’m important 🙃). 25/25

          17 replies 1 retweet 36 likes
          Show this thread
        17. End of conversation

      Loading seems to be taking a while.

      Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.

        Promoted Tweet

        false

        • © 2022 Twitter
        • About
        • Help Center
        • Terms
        • Privacy policy
        • Cookies
        • Ads info