文件列表:

fig/ (0, 2023-09-29)
fig/cv_age_case.png (16795, 2023-09-29)
fig/cv_age_case2.png (9535, 2023-09-29)
fig/cv_age_case3.png (22970, 2023-09-29)
fig/cv_age_case4.png (9614, 2023-09-29)
fig/cv_age_delay.png (7320, 2023-09-29)
fig/cv_age_hosp.png (20997, 2023-09-29)
fig/cv_age_hosp2.png (19478, 2023-09-29)
fig/cv_age_hosp2_full.png (36270, 2023-09-29)
fig/cv_age_hosp_covid.png (348, 2023-09-29)
fig/cv_age_hosp_loess.png (19288, 2023-09-29)
fig/cv_cos_60.png (25750, 2023-09-29)
fig/cv_glob_60.png (15278, 2023-09-29)
fig/cv_kb_ibc.png (25584, 2023-09-29)
fig/cv_kb_ibc_log.png (30257, 2023-09-29)
fig/cv_lcps.png (14712, 2023-09-29)
fig/cv_lcps_60.png (7602, 2023-09-29)
fig/cv_lcps_log10_60.png (8647, 2023-09-29)
fig/cv_n_kb.png (30624, 2023-09-29)
fig/cv_n_kib.png (20269, 2023-09-29)
fig/cv_n_kib_log.png (25528, 2023-09-29)
fig/cv_n_knoc.png (30361, 2023-09-29)
fig/cv_n_nd.png (28047, 2023-09-29)
fig/cv_n_nd_log.png (32778, 2023-09-29)
fig/cv_n_noc.png (34237, 2023-09-29)
fig/cv_n_noc_log.png (35762, 2023-09-29)
fig/cv_nd_per_beds.png (7096, 2023-09-29)
fig/cv_nearby.png (10782, 2023-09-29)
fig/cv_nearby_60.png (9365, 2023-09-29)
fig/cv_nl_05_60.png (19328, 2023-09-29)
fig/cv_nl_deaths.png (5443, 2023-09-29)
fig/cv_nl_deaths_60.png (3112, 2023-09-29)
fig/cv_nl_nat.png (271, 2023-09-29)
fig/cv_nl_nat_60.png (7344, 2023-09-29)
fig/cv_nl_nat_log10.png (6979, 2023-09-29)
fig/cv_nl_prov_60.png (10412, 2023-09-29)
fig/cv_nl_top_60.png (17805, 2023-09-29)
fig/cv_nl_v_usa.png (5681, 2023-09-29)
fig/cv_nl_v_usa_deaths.png (7515, 2023-09-29)
... ...

# COVID-19 in the Netherlands _Author: Jay Lee_ Last update: Fri, Sep 29, 2023 7:34:35 PM Table of Contents ================= header-includes: # Links and notes * This document will be updated every few days or so. - [**HTML**](https://bit.ly/covid_nl_html) (download and open in your own browser , easier to navigate since there's no pagination); - [**PDF**](https://bit.ly/covid_nl_pdf) (view on Dropbox or download from there); - [**GitHub**](https://github.com/JaySLee/covid/blob/main/README.md) (but in-text color not allowed on GitHub) * [121021:] [Report is now on GitHub](https://github.com/JaySLee/covid/blob/main/README.md) - But the inline math is bottom-aligned with the surrounding text, so looks a bit clumsy. - Also, colored text isn't rendered on GitHub. * Some dates are written U.S. style, so, e.g., 17.11.21 could appear as [111721:] * "Cases" = "infections". ## Updates * 19.10.22: [Possible new behavior or waning vax efficacy](#interesting) * 16.01.22: [Omicron seems to result in fewer hospitalizations and deaths](#omicron) * 29.12.21: Sources now include direct links to (some) csv data files. * 28.12.21: 1. [Comparison to Spanish flu (U.S.)](#comparison-to-spanish-flu-in-the-us) 2. [Lockdown prediction/expectation](#lockdown-prediction) * 20.11.21 - [Hospitalization risk by age](#hospitalization-risk-by-age) * 17.11.21 - [Covid by age](#covid-by-age) # Netherlands ## National trend * [042223:] Latest new cases = 203 (RIVMc), 203 (RIVMn) → +21 (since yesterday or last RIVM data update) * Updates: - [101922:] - Added 60 day new deaths plot, given recent up-tick. - Clipped non-log plot at 100K new cases (still visible in log10 plot). - [08??22:] RIVM updating only twice a week. - [041522:] LCPS (NL hospital data) no longer updated over the weekend. - [040922:] RIVM no longer updated over the weekend. ~~LCPS is however.~~ Also, PCR testing not as common from this point forward, so less reliable of an indicator of infections. - [011422:] It looks pretty clear -- given the lowering hospitalizations despite massive increase in infections -- that omicron is indeed less severe. + Hence, one might want to focus on hospitalization numbers (to determine how much one should be alarmed or not on a given day :wink:); see [analysis below](#hospital-occupancy-trends). - RIVMc is computed from the cumulative file while RIVMn is specifically new cases data. Not sure why there's a difference. - [121421:] Dropping numbers may be deceiving. From Dutchnews.nl: > However, changes in the testing regime may have had an impact on the number of official cases. People with mild symptoms are now being told to take a self test first and only report to the regional health board testing centres if that test is positive. It is not yet clear if everyone who has a positive self test is following the new guidelines and requesting a confirmation check. - [122921:] For occasional sudden drops then sudden increase on consecutive days (NLTimes): > ... it was about xxxx below average due to an IT error. - Just remember: there are about 10x as many infectious ~~zombies~~ people wandering about compared to the latest reported daily new cases. + [111821:] 23591*10 = ~236K = 1.35% of NL population + [112121:] Looking at the infectious estimates produced by RIVM (that they kept updated until July, 2021), it seems this multiplier is more like 15x -- 20x. * [ca.120121:] We're ~~25%~~ ~~50% higher~~ nearly double than the last highest peak (Christmas/New Years 2020-21). * Legend: - The figures below show raw (and log10) new case counts and deaths. - Orange line is 7-day moving average. - [022022:] GGD catch-up peak (400K) clipped in some of the plots. - Horizontal lines: The dashed gray horizontal line indicates last count of new cases. - Vaccinations: - [010722:] Thick **green** line indicates % of total population ~~double~~ vaxxed (where 100% is level with the highest new cases point on the graph). Numbers are weekly. - Dashed green line indicates % with only single dose. - [010722:] Thick **blue** line indicates % of 18+ population boostered. RIVM updates numbers twice a week, so some flat areas will appear in the plot. I started collecting this only recently and RIVM does not publish historical numbers. - [020222:] It seems double vaxxed numbers drops in the past month, ~~probably taking into account waning effectiveness of vax~~ because booster numbers seem to account for those double vaxxed, so double vaxxed numbers exclude boostered individuals. So, now green line depict _any_ vax. - Lockdowns/measures: - Colored _vertical lines_ indicate when restrictions are enacted (solid red vertical line) or lifted (dashed green vertical line). - Thickness of line indicates severity of lockdown or extent of the loosening of restrictions. - Based on observations farther below (and emphasis by the Prime Minister on hospital beds), it would seem that restrictions are imposed or lifted according to hospital occupancy and not number of infections. [See below](#lockdown-prediction).

* Source: [RIVM](https://data.rivm.nl/covid-19/COVID-19_aantallen_gemeente_per_dag.csv) _(releases updated data daily at 15:15 CET, clicking on link acquires the new cases/city csv)_. ## Hospital occupancy trends * [121021:] A few days ago, LCPS split ICU numbers by NL and "International" (i.e. in beds in Germany). Those are combined here. - The latter includes "the number of COVID patients moved abroad from the Netherlands. This currently concerns COVID IC patients who have gone from the Netherlands to a hospital in Germany." * [112721:] Hospitalizations are 😟. * [111621:] Hospitalizations are not as dire as earlier this year and last year, but almost there :-/.

### Lockdown prediction * [122821:] _The findings here pertain to the above hospital plots._ * As described above for infections/cases, _non-gray vertical lines in the above plots indicate when lockdowns are enacted_ (solid red vertical line) or lifted (dashed green vertical line). - Thickness of line indicates severity of lockdown or extent of the loosening of restrictions. * Lockdown observations and prediction: - Interestingly (and probably verifiable through LCPS), (hard) lockdowns appear to occur when non-ICU hospital bed occupancy is ~1750 and/or ICU bed occupancy is about 500. - Significant ease of lockdowns seems to occur when non-ICU bed occupancy decreases to 500. - [122821:] **Based on this pattern and assuming the occupancy trends continue, we can expect ease of current lockdown (at the earliest) near the beginning of February 2022** ~~the end of January~~. + Note that the announcement of a lockdown (or its removal) occurs a few days prior to the actual lockdown. Still, the observed intersections is likely not be coincidental since one can assume policy is partly based on predictive models. - [011822:] Light loosening of restrictions occurred on 011522. The hospitalization numbers interestingly correspond with the 1st loosening back ca. middle of May, 2020 and June, 2021. * Source: [LCPS](https://lcps.nu/wp-content/uploads/covid-19-datafeed.csv) _(releases data daily between 1pm-2pm CET)_ and [Wikipedia](https://en.wikipedia.org/wiki/COVID-19_pandemic_in_the_Netherlands) for lockdown moments ### Interesting * [101822:] I just noticed a spike in recent deaths (relative to number of ICU beds occupancy); maybe it's just noise or due to delayed reporting of deaths. * But if it isn't, then this could be indicative of new variant behavior or waning efficacy of vax (for a subpopulation). * Legend: - The colors correspond to the ratio between new deaths to ICU bed occupancy (red) and to regular bed occupancy (green) - Lower levels of red could indicate higher survival rate (of people in the ICU).

## Omicron * The colors indicate the date: - Dark blue points are the earliest days of the pandemic (Mar 2020) - _Hospital admissions data was unavailable until Oct 2020, while occupancy data was available as early as Mar 2020._ - Light green/yellow points are the middle of the pandemic. - Dark red/brown points are the most recent dates. * Faint arrows indicated the temporal trajectory. * The points are sized by new daily deaths. * The grey, dashed diagonal bisects the plot. * Y-axis: - Left plot: New daily hospitalizations (regular+ICU) - Right plot: Hospital bed occupancy (regular+ICU) * [021722:] Visual space clipped to not show 400K new infections point (GGD catching up on delayed data).

* Observations: - **[020122:] Sudden uptick of admissions (ICU+non-ICU).** - Either this is an artifact of data (collection) numbers or something more troubling, like an omicron variant or its penetrating a vulnerable subpopulation. Hospital by age data (published this Wednesday/020222) may tell us more. - The dark blue arm -- at 12 o'clock -- represents the 1st wave when seemingly fewer infections incurred high hospitalization occupancy as well as deaths. + However, covid testing was at its infancy in this period. - The light blue arm represents the 2nd wave of the pandemic when fewer infections caused more hospitalizations and deaths. - The red arm represents vaccination era, when infections led to fewer hospitalizations and deaths, than in the earlier days. - **The dark red arm -- at three o'clock -- represents the omicron era, in which hospitalizations and deaths have seemed invariant to (even high) number of infections.** + Interesting: there appears to be a similar horizontal invariance around the middle of the pandemic: orange points whose trajectory is rectangular/counter-clockwise. + The daily ICU admissions (as _y_-axis, but plot not shown) exhibit a similar pattern. The plots below log transform the _x_-axis (# of daily new infections), to stretch out those numbers.

### ICU v. non-ICU Here, non-ICU hospital bed occupancy (_x_-axis) plotted against ICU bed occupancy (_y_-axis). - Points left of the diagonal could indicate more severe covid. - The unidimensional ratio can be seen in the purple line of plots of [2.2](#hospital-occupancy-trends). - Semi-circular trajectories might be explained by time gap for those moved from hospital to ICU bed.

### Infections vs. new deaths

## Within NL cities/towns/provinces * More new cases per capita in the smaller towns than national level. * Last 60 days only (otherwise, the plot becomes messy). * Both plots show 7 day moving average. * Cities ordered from worst to best (within those displayed). * Left plot: worst cities; right plot: Largest cities in NL (top 5%) * [NEW 121021:] Lower plot: provinces

* Source: RIVM # Trends in nearby countries * As for countries near us: Belgium, Germany, Austria, and Denmark are also experiencing their own massive peaks, record-breaking for Germany and Austria. France is experiencing a minor surge.

* Source: [CSSE](https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_confirmed_global.csv) _(data is one day behind the RIVM and LCPS data, clicking on link acquires the global confirmed cases csv)_ # Covid by age ## Cases by age group * Just the age groups of our students, at 3 levels of measurement. - National (NL) - Province (Zuid-Holland) - Local (GGD Rotterdam) * RIVM groups ages by decade. * Smooth curves fit the data points (loess). * Upper plots show 7-day moving average. * Lower plots no moving average (but dates are more restricted for zooming). * Left plots shows **relative** proportion across all age groups. * Right plots shows absolute percentages within population of the age group. - Also, in the right plots: **_the last few rightmost points are incomplete (data), thus the sharp drop._** - On Nov 15, 2021, ~0.15% of all NL 20-29 year-olds were newly infected. + They also constitute ~18% of the infected on the same date. - Note: The denominator for the right plots is all youths in each age group across NL, for all three sources. Thus, the Province and Local percentages are less informative. * Sometimes, GGD Rotterdam throws up some extreme (high or low) values for last date. * Observations: - A recent relative uptick of cases in the two relevant age groups. - The percentages within age groups are more striking.

Source: RIVM ([case/age csv](https://data.rivm.nl/covid-19/COVID-19_casus_landelijk.csv)) ### Delay of age data * Dates of case/age data are offset by ~~+3~~ +1 day in the plot below. * So two kinds of delays: - RIVM country totals data (of all new cases) is behind the age data by ~~3~~ 1 days. - But comprehensive age data for the past week is incomplete/delayed significantly. * Mid-Feb 2022 spike seems back-distributed. ## Hospitalization by age group * _New hospitalization data comes out once a week, Wednesdays_. * [112421:] _There was a bug in my code that made these graphs inaccurate; fixed now._ * Age groups of just EUR students and lecturers shown here. * LCPS groups age in 5-year bins. * LCPS releases these on a weekly basis. * First two plots show 7-day moving average. * Upper left plot shows relative proportion across all age groups. - Young age groups are merged due to misalignment between hospital and case data. - [112521:] _Relative proportion generally going down largely because of (relatively) more hospitizalizations in the older (not shown) age groups._ * Upper right plot shows ratio of hospitalizations to number of cases. - These proportions is an upper bound, as there are far more contagious cases thanindicated by a daily case number and also considering not all cases are registered by RIVM. - [112521:] - _Under that assumption, there is currently < 1% chance of hospitalization after infection._ - _The diminishing curves for the older age groups may be indicative of the vaccine's effectiveness._ * The lower two plot shows _percentage_ within population of each age group. - The right plot shows a LOESS regression fit. - Because the recent week's LCPS age data (i.e. most recent date) are incomplete, points are omitted in these two plots. - [112521:] _Hospitalizations (within age group) are going up for all age groups._

Full time line (no moving average)

Source: RIVM ([hospital/age csv](https://data.rivm.nl/covid-19/COVID-19_ziekenhuis_ic_opnames_per_leeftijdsgroep.csv)) # Other countries ## Global and other countries * Left plot is top 25 countries (in descending order), over the last 60 days. * Right plot contains various countries that came to my attention (e.g., being in the news) or of personal interest. - Ordered by when the country came to my attention, and not by new cases. * The orange denotes the latest NL moving average (7 day) and not the latest daily new cases. * Both plots use 7-day moving average ("7day").

## USA v. NL * Top plot: - New infections per 100K - 7 day moving average, last 60 days, top 25 states - The red, , blue horizontal line is the US national average. * Lower left plot: - New infections per 1M - 7 day moving average, whole pandemic (for which there is data) * Lower right plot - Same as previous but with deaths data.

# Risk calculations ## Booster protection 021722: - Based on information presented in this [RIVM page](https://www.rivm.nl/en/news/booster-vaccination-prevents-many-hospital-and-ICU-admissions) ... &space;=&space;.0032\%) &space;=&space;.014\%) &space;=&space;.033\%) * Being boostered offers 10x more protection from hospitalization than being unvaxxed. * Being just vaxxed offers ~2.5x more protection. ## Hospitalization risk **tl;dr:** (based on Nov 13, 2021 numbers) * Risk of getting covid, while being double vaxxed = 0.27% * Risk of hospitalization, giv ... ...

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：