Coronavirus Thread: Worldwide Pandemic
- Thread starter 88m3
- Start date
edit i had wrong info ignore
ADevilYouKhow
Rhyme Reason
DrBanneker
Space is the Place
From a ProMed email:
Date: 26 January 2020
From: David Fisman, MD MPH [edited]
I wish to offer the following observations on the epidemiology of 2019-nCoV in Hubei Province, China, over the past few weeks. I hope that the thoughts below will be helpful to others trying to organize and interpret the flood of information that has emerged about this new pathogen. Information from a variety of sources suggests that this novel virus is a recombinant beta-coronavirus of animal origin that emerged in November or December 2019, likely at the Wuhan Seafood Market. Epidemiological analysis was initiated after recognition of a market-linked pneumonia cluster in late December. Notwithstanding the name of the "Wuhan Seafood Market", the market sells large numbers of live animals, including wild animals, which are kept in close proximity to one another, perhaps facilitating viral recombination. Similar disrupted ecology contributed to the emergence of SARS.
The emergence of many cases of a novel, animal-derived pathogen in a live animal market, over a short time period was suggestive of a point source outbreak with animal-to-human spread, and I'll assume that the initial cluster of approximately 40 cases was largely a result of such transmission, with little human-to-human transmission.
However, on January 23, 2020, the WHO released the report of its IHR Emergency Committee for nCoV; the report noted that "fourth generation transmission" was occurring, and cited internal analyses placing the basic reproduction number (R0) at between 1.4 and 2.5; this report noted that 557 cases (which I'll round up to 600 cases) had been confirmed as of January 22, 2020 (ref.1).
Several estimates of R0 appeared from independent groups around the same time; these estimates were remarkable in their consistency, ranging from 1.4 to 3.8 (refs.2-7). Such consistency despite limited data availability and disparate methods employed for estimation provides a degree of face validity to these estimates.
I note that these estimates are likely skewed upwards by the greater recognition of larger case clusters and super-spreader events (there has been at least one 14-case cluster in a hospital), and also by the possibility that later cases are being recognized more completely than earlier cases, all of which would have the tendency of biasing R0 estimates upwards. I'll assume that the lower bound R0 (around 2) is probably about right, and also note that this is consistent with estimates from SARS coronavirus, which shares substantial genetic similarity with nCoV.
What would be the implications of a disease with R0 ~ 2, with four generations of transmission over a period of around a month? This timeline would be consistent with growth in the number of recognized cases from 40 to 600 during that time interval. If nCoV has a generation time of approximately 10 days (similar to that described for SARS), we would have expected the initial 40 cases from late December to cause 80 secondary cases in early January (120 total cases); these 80 cases would create 160 incident cases around mid-January (280 total cases), which would in turn create another 320 cases around January 22 (600 total cases).
These numbers fit very nicely to case data available as of January 22, 2020, but unfortunately, they are wrong. The abrupt surge in confirmed case counts (to 1423 cases as of January 26, 2020) is not compatible with the growth process described above, certainly not with a SARS-compatible generation time of 6-10 days.
Indeed, the authors of the MRC model (3) noted in one of their earlier reports that the volume of observed exported cases in countries outside China suggested a much larger underlying epidemic than had been reported at that time, and this epidemic may have begun a month prior to the recognition of the market-associated outbreak, consistent with the reported timing of viral emergence based on phylogenetic analyses (5).
The authors of several analyses cited above have incorporated the MRC estimates of under-reporting in order to fit their models (2, 3, 5).
A second line of evidence suggesting undercounting of cases relates to the older age of cases (median 59 years in early reports), and the even older age of fatal cases (averaging around 75 years in the first 17 deaths) as contrasted with a median age of 37 or 38 years in China.
Increased age in cases as compared to the population as a whole suggests that younger (likely milder) cases have been under-reported. As such, it would seem likely that at least part of the sudden apparent growth in case counts does not reflect changes in transmission, but rather increasing ascertainment of previously undercounted cases.
Why is R0 so important? As R0 is proportional to duration of infectivity, reducing the infective period of cases would reduce the effective reproduction number. If the effective duration of infectivity is reduced by over 50% for a pathogen with R0 ~ 2, the average reproduction number would be reduced to less than 1, which should control viral spread over time.
It is encouraging that one of the reports cited above suggests that the mean time from symptom onset to isolation has decreased from more than 6 days to less than 1 day as control measures have been implemented (6). Social distancing measures (like suspension of public gatherings and transportation) and reduced transmission per contact (e.g., through the use of personal protective items by healthcare workers) would also result in proportionate reductions in the reproduction number. Precise predictions in the face of substantial uncertainty are not appropriate, but given the large size of the epidemic as of the time of writing, some simple back-of-the-envelope math can demonstrate that large numbers of incident cases should be expected in the coming weeks, even in the face of effective control efforts.
Successful control of this outbreak would be expected to take many months (again, as was the case with SARS). While average estimates of R0 are helpful, it is also important to note that other beta-coronaviruses of public health importance (SARS, MERS) have been notable for the "overdispersion" of their reproductive numbers. Without getting too technical, this means that the average R0 is quite different from the variability in the R0.
We actually have a distribution of R0 with a long "tail", which is a mathematical shorthand for superspreader events, where a case infects a large number of individuals. For example, there was at least one SARS superspreader who generated 76 downstream cases. However, with an overdispersed R0 many cases are "dead ends" and will not transmit.
The three key insights here for the contour of this epidemic are:
1. It is the average R0 that determines whether, and how, the disease can be controlled. By analogy with SARS and MERS, with which nCoV seems to share many characteristics, the spread of this virus should be controllable.
2. Superspreader events are likely (and have already occurred) and are important to outbreak control efforts: they are demoralizing and dangerous to response personnel. They often occur in hospitals during aerosol-generating procedures like intubation. These events make it feel like the battle is being lost. They should be anticipated, and it is important to emphasize that their occurrence will represent a temporary setback which is likely to be overcome.
3. While superspreader events are unwelcome, their occurrence may, in fact, be a salutary sign for the control of this outbreak. An average R0 of, say, 2, with an overdispersed R0 means that many cases are also likely to be "dead ends" epidemiologically. Inasmuch as superspreader events may be more likely to be recognized due to their dramatic nature, an outbreak driven by superspreaders may be more likely to attract the control measures needed to disrupt transmission.
By contrast, a more homogeneous outbreak, where each and every case has the potential to create a downstream cascade of cases in the absence of recognition, may be much more difficult to control.
As I note above, this outbreak is in its early weeks and understanding and knowledge will doubtless change. However, analysis of the cases counts, rate of growth of the epidemic, reproduction number estimates, and estimates of likely undercounting that have emerged over the past two weeks can provide a coherent view of the likely early dynamics of this outbreak, and also suggest what the contours of the outbreak may look like if control efforts are successful.
David N. Fisman, MD MPH FRCP(C) Professor, Epidemiology Dalla Lana School of Public Health University of Toronto
References
1. World Health Organization. Statement on the meeting of the International Health Regulations (2005) Emergency Committee regarding the outbreak of novel coronavirus (2019-nCoV). Available via the Internet at Statement on the meeting of the International Health Regulations (2005) Emergency Committee regarding the outbreak of novel coronavirus 2019 (n-CoV) on 23 January 2020. Last accessed January 25, 2020. Geneva, Switzerland; 2020.
2. Riou J, Althus C. Pattern of early human-to-human transmission of Wuhan 2019-nCoV. Preprint. Available via the Internet at jriou/wcov. Last accessed January 25, 2019. 2020.
3. Imai N, Cori A, Dorigatti I, Baguelin M, Donnelly C, Riley S, et al. Report 3: Transmissibility of 2019-nCoV. Available via the Internet at News / Wuhan Coronavirus. Last accessed January 25, 2020.; 2020.
4. Majumdar M, Mandl K. Early transmissibility assessment of a novel coronavirus in Wuhan, China. Preprint. Available via the Internet at Early Transmissibility Assessment of a Novel Coronavirus in Wuhan, China by Maimuna Majumder, Kenneth D. Mandl :: SSRN. Last accessed January 25, 2020. 2020.
5. Bedford T, Neher R, Hadfield J, Hodcroft E, Ilcisin M, Müller N. Genomic analysis of nCoV spread. Situation report 2020-01-23. Available via the Internet at auspice. Last accessed January 25, 2020.; 2020.
6. Liu T, Hu J, Lin L, Zhong H, Xiao J, He G, et al. Transmission dynamics of 2019 novel coronavirus (2019-nCoV). Available via the Internet at . Last accessed January 26, 2020. bioRxiv 2020.01.25.919787; doi: . 2020.
7. Read J, Bridgen J, Cummings D, Ho A, Jewell C. Novel coronavirus 2019-nCoV: early estimation of epidemiological parameters and epidemic predictions. Note that the authors have stated that the R0 estimates here are overestimates and will be revised downward. Available via the Internet at Novel coronavirus 2019-nCoV: early estimation of epidemiological parameters and epidemic predictions. Last accessed January 26, 2020. medRxiv. 2020; 2020.01.23.20018549.
Date: 26 January 2020
From: David Fisman, MD MPH [edited]
I wish to offer the following observations on the epidemiology of 2019-nCoV in Hubei Province, China, over the past few weeks. I hope that the thoughts below will be helpful to others trying to organize and interpret the flood of information that has emerged about this new pathogen. Information from a variety of sources suggests that this novel virus is a recombinant beta-coronavirus of animal origin that emerged in November or December 2019, likely at the Wuhan Seafood Market. Epidemiological analysis was initiated after recognition of a market-linked pneumonia cluster in late December. Notwithstanding the name of the "Wuhan Seafood Market", the market sells large numbers of live animals, including wild animals, which are kept in close proximity to one another, perhaps facilitating viral recombination. Similar disrupted ecology contributed to the emergence of SARS.
The emergence of many cases of a novel, animal-derived pathogen in a live animal market, over a short time period was suggestive of a point source outbreak with animal-to-human spread, and I'll assume that the initial cluster of approximately 40 cases was largely a result of such transmission, with little human-to-human transmission.
However, on January 23, 2020, the WHO released the report of its IHR Emergency Committee for nCoV; the report noted that "fourth generation transmission" was occurring, and cited internal analyses placing the basic reproduction number (R0) at between 1.4 and 2.5; this report noted that 557 cases (which I'll round up to 600 cases) had been confirmed as of January 22, 2020 (ref.1).
Several estimates of R0 appeared from independent groups around the same time; these estimates were remarkable in their consistency, ranging from 1.4 to 3.8 (refs.2-7). Such consistency despite limited data availability and disparate methods employed for estimation provides a degree of face validity to these estimates.
I note that these estimates are likely skewed upwards by the greater recognition of larger case clusters and super-spreader events (there has been at least one 14-case cluster in a hospital), and also by the possibility that later cases are being recognized more completely than earlier cases, all of which would have the tendency of biasing R0 estimates upwards. I'll assume that the lower bound R0 (around 2) is probably about right, and also note that this is consistent with estimates from SARS coronavirus, which shares substantial genetic similarity with nCoV.
What would be the implications of a disease with R0 ~ 2, with four generations of transmission over a period of around a month? This timeline would be consistent with growth in the number of recognized cases from 40 to 600 during that time interval. If nCoV has a generation time of approximately 10 days (similar to that described for SARS), we would have expected the initial 40 cases from late December to cause 80 secondary cases in early January (120 total cases); these 80 cases would create 160 incident cases around mid-January (280 total cases), which would in turn create another 320 cases around January 22 (600 total cases).
These numbers fit very nicely to case data available as of January 22, 2020, but unfortunately, they are wrong. The abrupt surge in confirmed case counts (to 1423 cases as of January 26, 2020) is not compatible with the growth process described above, certainly not with a SARS-compatible generation time of 6-10 days.
Indeed, the authors of the MRC model (3) noted in one of their earlier reports that the volume of observed exported cases in countries outside China suggested a much larger underlying epidemic than had been reported at that time, and this epidemic may have begun a month prior to the recognition of the market-associated outbreak, consistent with the reported timing of viral emergence based on phylogenetic analyses (5).
The authors of several analyses cited above have incorporated the MRC estimates of under-reporting in order to fit their models (2, 3, 5).
A second line of evidence suggesting undercounting of cases relates to the older age of cases (median 59 years in early reports), and the even older age of fatal cases (averaging around 75 years in the first 17 deaths) as contrasted with a median age of 37 or 38 years in China.
Increased age in cases as compared to the population as a whole suggests that younger (likely milder) cases have been under-reported. As such, it would seem likely that at least part of the sudden apparent growth in case counts does not reflect changes in transmission, but rather increasing ascertainment of previously undercounted cases.
Why is R0 so important? As R0 is proportional to duration of infectivity, reducing the infective period of cases would reduce the effective reproduction number. If the effective duration of infectivity is reduced by over 50% for a pathogen with R0 ~ 2, the average reproduction number would be reduced to less than 1, which should control viral spread over time.
It is encouraging that one of the reports cited above suggests that the mean time from symptom onset to isolation has decreased from more than 6 days to less than 1 day as control measures have been implemented (6). Social distancing measures (like suspension of public gatherings and transportation) and reduced transmission per contact (e.g., through the use of personal protective items by healthcare workers) would also result in proportionate reductions in the reproduction number. Precise predictions in the face of substantial uncertainty are not appropriate, but given the large size of the epidemic as of the time of writing, some simple back-of-the-envelope math can demonstrate that large numbers of incident cases should be expected in the coming weeks, even in the face of effective control efforts.
Successful control of this outbreak would be expected to take many months (again, as was the case with SARS). While average estimates of R0 are helpful, it is also important to note that other beta-coronaviruses of public health importance (SARS, MERS) have been notable for the "overdispersion" of their reproductive numbers. Without getting too technical, this means that the average R0 is quite different from the variability in the R0.
We actually have a distribution of R0 with a long "tail", which is a mathematical shorthand for superspreader events, where a case infects a large number of individuals. For example, there was at least one SARS superspreader who generated 76 downstream cases. However, with an overdispersed R0 many cases are "dead ends" and will not transmit.
The three key insights here for the contour of this epidemic are:
1. It is the average R0 that determines whether, and how, the disease can be controlled. By analogy with SARS and MERS, with which nCoV seems to share many characteristics, the spread of this virus should be controllable.
2. Superspreader events are likely (and have already occurred) and are important to outbreak control efforts: they are demoralizing and dangerous to response personnel. They often occur in hospitals during aerosol-generating procedures like intubation. These events make it feel like the battle is being lost. They should be anticipated, and it is important to emphasize that their occurrence will represent a temporary setback which is likely to be overcome.
3. While superspreader events are unwelcome, their occurrence may, in fact, be a salutary sign for the control of this outbreak. An average R0 of, say, 2, with an overdispersed R0 means that many cases are also likely to be "dead ends" epidemiologically. Inasmuch as superspreader events may be more likely to be recognized due to their dramatic nature, an outbreak driven by superspreaders may be more likely to attract the control measures needed to disrupt transmission.
By contrast, a more homogeneous outbreak, where each and every case has the potential to create a downstream cascade of cases in the absence of recognition, may be much more difficult to control.
As I note above, this outbreak is in its early weeks and understanding and knowledge will doubtless change. However, analysis of the cases counts, rate of growth of the epidemic, reproduction number estimates, and estimates of likely undercounting that have emerged over the past two weeks can provide a coherent view of the likely early dynamics of this outbreak, and also suggest what the contours of the outbreak may look like if control efforts are successful.
David N. Fisman, MD MPH FRCP(C) Professor, Epidemiology Dalla Lana School of Public Health University of Toronto
References
1. World Health Organization. Statement on the meeting of the International Health Regulations (2005) Emergency Committee regarding the outbreak of novel coronavirus (2019-nCoV). Available via the Internet at Statement on the meeting of the International Health Regulations (2005) Emergency Committee regarding the outbreak of novel coronavirus 2019 (n-CoV) on 23 January 2020. Last accessed January 25, 2020. Geneva, Switzerland; 2020.
2. Riou J, Althus C. Pattern of early human-to-human transmission of Wuhan 2019-nCoV. Preprint. Available via the Internet at jriou/wcov. Last accessed January 25, 2019. 2020.
3. Imai N, Cori A, Dorigatti I, Baguelin M, Donnelly C, Riley S, et al. Report 3: Transmissibility of 2019-nCoV. Available via the Internet at News / Wuhan Coronavirus. Last accessed January 25, 2020.; 2020.
4. Majumdar M, Mandl K. Early transmissibility assessment of a novel coronavirus in Wuhan, China. Preprint. Available via the Internet at Early Transmissibility Assessment of a Novel Coronavirus in Wuhan, China by Maimuna Majumder, Kenneth D. Mandl :: SSRN. Last accessed January 25, 2020. 2020.
5. Bedford T, Neher R, Hadfield J, Hodcroft E, Ilcisin M, Müller N. Genomic analysis of nCoV spread. Situation report 2020-01-23. Available via the Internet at auspice. Last accessed January 25, 2020.; 2020.
6. Liu T, Hu J, Lin L, Zhong H, Xiao J, He G, et al. Transmission dynamics of 2019 novel coronavirus (2019-nCoV). Available via the Internet at . Last accessed January 26, 2020. bioRxiv 2020.01.25.919787; doi: . 2020.
7. Read J, Bridgen J, Cummings D, Ho A, Jewell C. Novel coronavirus 2019-nCoV: early estimation of epidemiological parameters and epidemic predictions. Note that the authors have stated that the R0 estimates here are overestimates and will be revised downward. Available via the Internet at Novel coronavirus 2019-nCoV: early estimation of epidemiological parameters and epidemic predictions. Last accessed January 26, 2020. medRxiv. 2020; 2020.01.23.20018549.
ADevilYouKhow
Rhyme Reason
People are going to have to wake up to the idea they are already dead
does HL have a official cornonavirus thread? if not let's make this one it
Check out this sim the did in October
Event 201, a pandemic exercise to illustrate preparedness efforts
Event 201, a pandemic exercise to illustrate preparedness efforts
88m3
Fast Money & Foreign Objects
Hood Critic
The Power Circle
It's going to get to the point where all travel to mainland China will be suspended. That will have an interesting impact on their economy as well as the global economy.
ill
Superstar
Short Chinese stock
This is probably the only time Asians don’t look funny with face masks on at the airport.
Read that they thought the virus was orginally from that open air meat market but it could of orginated from somewhere else 
