Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
    daveluo
    @daveluo
    yea, no rush, and that prioritization makes sense to me
    daveluo
    @daveluo
    for daily cases and effective reproduction rate by state
    daveluo
    @daveluo
    @sunwsusan forgot to point you to this earlier but we have developed a PPE needs estimation model here: https://github.com/covidcaremap/covid19-healthsystemcapacity/blob/master/covidcaremap/ppe.py
    which can be tied to forecasting, example: https://github.com/covidcaremap/covid19-healthsystemcapacity/blob/master/notebooks/processing/PPE_Burn_Rate_from_IHME.ipynb
    2 replies
    we haven't created any frontend for it yet because wasn't sure what would be most useful: i.e., a PPE calculator with adjustable parameters, a map viz of forecasted needs, more demo notebooks? if this is lined up with what you're working on, let us know if you have any thoughts/Qs!
    12 replies
    daveluo
    @daveluo
    FYI, finished a cataloging of all US state dashboards and what data is being publicly reported here: https://docs.google.com/spreadsheets/d/1XQ4we4sComBn-DUHS6_YBqkr8gG-nLZ4m3U604QtfMY/edit?usp=sharing, let me know if any Qs. The purpose of this is to index what's being publicly shared by states and compare it to recommendations about what metrics/data should be collected on the "roadmap to recovery" by the National Governors Assoc and others
    Some observations/notes:
    • i'm impressed by a number of states that have done a pretty great job of publicly reporting a lot of breadth and depth in relation to metrics on Cases/Deaths, Syndromic Surveillance, Testing Capacity, Health System Capacity, Public Health Capacity categories. Especially those states that have hospital capacity and/or PPE data. For example: AK, IL, IN, MA, MI, MN, NC, OK, OR, PA, TX, WI
    • a lot of reporting on outbreaks and deaths in congregate living facilities, more than i expected but this is a key high-risk population so makes sense
    • almost no official data shared about public health capacity, i.e. contact tracing. Only North Carolina has talked about it in any way
    • almost nothing on symptom reporting and tracking in the community (before encountering medical system) per Jasmin's suggestion. Only state that has something like it is Oklahoma
      data sources/collection systems for each metric category runs the gamut from local --> national systems per state
    • so. many. dashboards. arcgis and tableau being the most common vendors, then power bi. Lots of PDFs too which actually work pretty well IMO as a shareable data visualization/comm format.
    • not nearly enough open data hub access or ability to download the source data for aggregation and further analysis
    • hospital capacity and PPE data is particularly inconsistent and spotty in reporting, whether by what categories of beds/equipment/supplies or the spatial aggregations used
    Rob Emanuele
    @lossyrob
    This is really awesome work! How should this be shared?
    Susan Sun
    @sunwsusan
    agree with @lossyrob . @daveluo if this is open to be shared publicly, i would like to share the collecton of dashboards with the provider assessment team at Project N95. i think this will help guide our conversations with state govs seeking PPE
    daveluo
    @daveluo
    Yup! Can be shared, will be continuing to work on it to fill in or update data so continuous WIP but it's complete as a first pass of all states. I've shared w Dave Yoon directly but happy if u share w provider team as u see fit
    Susan Sun
    @sunwsusan
    Sgtm!
    Grace Doherty
    @gracedoherty
    The dashboard catalog is going to be so useful for researchers! Glad to see it realized

    On mapathon continuation: Would Friday 4pm ET (same time as last event) work for folks here? Any argument for moving it to a different day of the week?

    I'll budget some time this week to go over the workflow and make it less complex for volunteers. Curating a list of source links for them beforehand being the most important change.

    daveluo
    @daveluo
    @gracedoherty , that time (Fri 4pm ET) may be tough for me - solo-parenting on friday - but don't let me be the blocker if it's the best timing for you and others. I should still be able to be present but not fully active. We should also sync up re: the workflow and potential other tasks that would be really helpful
    Grace Doherty
    @gracedoherty

    @daveluo yes please, maybe tomorrow afternoon?

    Happy to run the mapathon by myself no problem, but the volunteers really benefit from hearing someone from this team's perspectives on the state of data for public health response. Most of the volunteers were students who are in final exams, so their schedules are pretty unpredictable. We could easily change the day. Sometime during east coast evening is probably still best.

    daveluo
    @daveluo
    sounds good! tmrw afternoon to sync up works for me, will DM u about timing
    daveluo
    @daveluo

    WA has a new risk assessment dashboard out, i think to go with their phased reopening plan:

    Ticks almost all the boxes (except about healthcare worker exposure), could use more granularity of detail, better data reporting formats than PDFs, more unified viz but hey, it's a promising start

    daveluo
    @daveluo
    been looking a bit into the number of healthcare workers (HCWs) who've been infected, or who's even reporting this number...and it's shockingly sparse data. Only came across 6 states that's reporting it so far (AL, GA, ID, MN, NH, WI). That plus cobbling together news articles from various states, i get ~18K confirmed cases among HCWs. Which is already way more than the 9K that the CDC reported, with caveats that it's known underreporting (https://www.cdc.gov/mmwr/volumes/69/wr/mm6915e6.htm?s_cid=mm6915e6_w)
    every news source cites that ~11% of cases are among healthcare workers which lines up with what i've found. Given there are now 1.2M cases in the US, that's an estimated ~130K cases among HCW. Of which we only know ~1/8th about (18K/130K)
    how can we protect our HCWs if we don't even know how many are getting sick?
    daveluo
    @daveluo
    anyone come across any resources collecting data on HCW cases and outcomes? I've mainly seen collections of stories and names:

    https://khn.org/news/true-toll-of-covid-19-on-u-s-health-care-workers-unknown/

    The number of health care workers who have tested positive for the coronavirus is likely far higher than the reported tally of 9,200, and U.S. officials say they have no comprehensive way to count those who lose their lives trying to save others.

    daveluo
    @daveluo

    https://www.ecdc.europa.eu/sites/default/files/documents/covid-19-rapid-risk-assessment-coronavirus-disease-2019-ninth-update-23-april-2020.pdf:

    Of the confirmed cases in China, 3.8% (1 716/44 672) were healthcare workers. Of those, 14.8% were severely or critically ill and five of the severe cases died [74]. Latest figures reported from Italy show that 10% of COVID-19 cases are healthcare workers [69], with the Lombardy region reporting up to 20% of cases in healthcare workers [75]. In Spain, the latest COVID-19 situation overview from the Ministry of Health reports that 20% of COVID-19 cases are in healthcare workers [66]. In the US, overall, only 3% (9 282/315 531) of reported cases were among healthcare workers; however, among states with more complete reporting, healthcare workers accounted for 11% of reported cases [76]. In a Dutch study, healthcare workers were tested voluntarily for COVID-19 and 6% tested
    positive [77]. In a report on 30 cases in healthcare workers in China, all cases had a history of direct contact (distance within 1 metre) with COVID-19 patients, with an average number of 12 contacts (7 ̶16), and the average cumulative contact time being two hours (1.5 ̶2.7) [78]. In the Dutch study, only 3% of healthcare workers reported being exposed to hospital patients with COVID-19 prior to onset of symptoms and 63% had worked while asymptomatic [77]. In the US, from 1 423 healthcare workers, 55% reported a known contact with a laboratoryconfirmed COVID-19 patient in the 14 days before illness onset [76].

    daveluo
    @daveluo

    "What's Behind South Korea's COVID-19 Exceptionalism?" https://www.theatlantic.com/ideas/archive/2020/05/whats-south-koreas-secret/611215/
    Fast Testing:

    In late January, just one week after the country’s first case was diagnosed, government officials urged medical companies to develop coronavirus test kits and told manufacturers to prepare for mass production. By mid-February—while the U.K. was talking about “herd immunity” and President Donald Trump was predicting that the virus would “miraculously” disappear in weeks—South Korea was churning out thousands of test kits every day. By March 5, South Korea had tested 145,000 people—more than the U.S., the U.K., France, Italy, and Japan combined.

    High Tech Expansive Tracing:

    The national mapping of citizen activity yields fast results. On a Saturday in April, a 58-year-old man was diagnosed with the coronavirus. Surveillance data showed that he had voted in the election and visited several restaurants in previous days. Within 48 hours, South Korea had identified—and, in some cases, interviewed—more than 1,000 people who had potentially come into contact with him. All of them were instructed to self-isolate, thus cordoning off the virus’s spread. By the end of the month, no new clusters appeared in the Korean infection data.

    Zero Tolerance Isolation:

    To separate the sick from the healthy—and the somewhat sick from the very sick—South Korea’s patients are divided into several groups. The elderly and those with serious illnesses go straight to hospitals. Moderately sick people are sent to isolation dorms, where they’re monitored. And the asymptomatic “contacts” of recently diagnosed cases are asked to self-quarantine at home and use separate bathrooms, dishes, and towels from their cohabitants. Health-service officials check in twice daily to monitor their symptoms.
    International arrivals are also subject to isolation rules. South Korea requires that foreign arrivals self-isolate for two weeks. Koreans arriving from overseas are required to download an app that registers their symptoms in the days after their arrival.

    daveluo
    @daveluo
    had a good call yesterday with folks from USDR and www.covidexitstrategy.org about what we can do to advance data preparedness at the state level. Next step is to do some design mockups to elicit feedback on how we could visually communicate insights and offer assistance to the state/local level (i.e. governor's office, national governors assoc, county public health depts) based on what's collected in https://docs.google.com/spreadsheets/d/1XQ4we4sComBn-DUHS6_YBqkr8gG-nLZ4m3U604QtfMY/edit?usp=sharing. @jfrankl if you're interested, would you be free to catch up and discuss this today or over weekend?
    9 replies
    Grace Doherty
    @gracedoherty

    FYI: Next mapathon will be Wednesday the 13th!

    CovidCareMap Mapathon
    Wednesday, May 13
    5:30pm – 7:30pm Eastern Daylight Time
    Please RSVP before the event: https://airtable.com/shrWMZWNnL8VLFD7D
    Video meetup: meet.google.com/tsp-dxsh-fvt
    Connect by phone: ‪+1 484-841-9993‬ PIN: ‪729 825 903#‬‬‬‬‬‬‬‬‬

    daveluo
    @daveluo
    :tada:
    Grace Doherty
    @gracedoherty
    Describing the types of temporary hospitals built by USACE and how they triage. Also includes a facilities map from April 16
    https://www.usatoday.com/in-depth/news/2020/04/16/coronavirus-covid-19-field-hospital-army-corps-of-engineers-samaritans-purse-central-park-tcf-center/5119808002/
    image.png
    Each point of interest lists simply facility name, city/state, and number of beds.
    Steven Pousty
    @thesteve0
    Hey folks - I would love to help. Any chance you want to move the data in PostGIS and maybe make an API on top of it?
    daveluo
    @daveluo
    :wave: hi @thesteve0! sorry for the delayed welcome. we haven't discussed much about postgis or APIs yet so they're unexplored areas. what'd you have in mind?
    Steven Pousty
    @thesteve0
    Hey @daveluo thanks for responding. Well I am devrel lead at Crunchy (postgres people) and I am looking for some volunteer work where I could also use Postgres. I was thinking of making a real relational model for the data and storing it in a db (Postgresql with spatial though PostGIS). I think the drawback would be ease of use for adding or updating data. Benefits more data cleanliness and better scalability
    The API would allow people to read the data and maybe do their own visualizations
    I love that you folks have information on ICU beds and percent occupancy - that's one of the most important metrics most people don't pay attention to
    and it's hard to get
    daveluo
    @daveluo
    got it, thanks for these thoughts! The "drawback would be ease of use for adding or updating data" is probably why we implicitly haven't considered it much yet - lots of structural changes happening still and new data being added. For instance, we're starting to add updated ICU bed/occupancy data for certain states and figuring out some data model changes to tally updated counts on a regular basis. @lossyrob has been thinking through this more than I so he'll probably chime in when he circles around to this
    thanks @gracedoherty for organizing another great mapathon! Lower participation this time around due to end-of-semester/finals for many potential folks but we had a very productive and fun time (with choice tunes: https://bit.ly/covidcaremap_tunes) collecting data about underreported Health Care Worker cases..
    daveluo
    @daveluo

    See what we collected at https://airtable.com/shrId9GL5mXSOHEGg/tblDQlqWVsC6e7hpb

    After going thru and finding everything out there for all 50 states, we now know about 36K COVID cases among HCWs, which is the highest count i've seen so far and 4x the CDC reported number from last month. It's also higher than the 31K reported in this investigative report which just came out 2 days ago: https://www.revealnews.org/article/31000-and-counting/

    What we have is still vastly incomplete with semi-outdated news sources. Only 10 states officially reporting HCW case numbers (via dashboards) + PA with daily press announcements

    ^ FYI @Geoyi @andrewwgithub @sunwsusan
    daveluo
    @daveluo
    validated HCW case counts and added screenshot or PDF attachments for each source of data
    Pavel Donin
    @doninpr_gitlab
    Hi everyone! I am experienced on web-map developing (mapbox, openlayers itc) and js (React/Redux, vanilla js). What can I do for your project? I would be glad to help you.
    Steven Pousty
    @thesteve0
    HCWs - what does that stand for?
    daveluo
    @daveluo
    health care workers
    @doninpr_gitlab :wave: hi and welcome! glad to hear about your interest and skills. back in a bit with thoughts on where we could use help, esp w/ React
    Pavel Donin
    @doninpr_gitlab
    @daveluo look forward to hearing from you
    6 replies