-
Categories
-
Pharmaceutical Intermediates
-
Active Pharmaceutical Ingredients
-
Food Additives
- Industrial Coatings
- Agrochemicals
- Dyes and Pigments
- Surfactant
- Flavors and Fragrances
- Chemical Reagents
- Catalyst and Auxiliary
- Natural Products
- Inorganic Chemistry
-
Organic Chemistry
-
Biochemical Engineering
- Analytical Chemistry
-
Cosmetic Ingredient
- Water Treatment Chemical
-
Pharmaceutical Intermediates
Promotion
ECHEMI Mall
Wholesale
Weekly Price
Exhibition
News
-
Trade Service
Recently, Beijing, Guangdong, Zhejiang, Sichuan, Henan and other places released the latest research judgment, saying that the peak of the first wave of infection has passed, and the number of fever outpatient visits continues to decline
.
So far, nationwide data on the outbreak is still missing
.
On December 25, 2022, after 1,079 days of continuous updates, the official website of the National Health Commission announced that it would stop publishing daily epidemic information
.
Subsequently, the relevant data was transferred to the Chinese Center for Disease Control and Prevention to release to the public, but due to the inertia of the channels of information acquisition, most of the public did not know that the epidemic daily was continuing
.
In fact, since the "Twenty Articles" and "New Ten Articles" were successively introduced at the end of 2022, which led to the cancellation of large-scale nucleic acid testing, the epidemic data we can see has lacked reference value
.
In the past three years, the daily epidemic data accompanied by 1.
4 billion Chinese people is a complete set of "active" epidemic surveillance system, that is, with a variety of "active" means such as social nucleic acid screening, nucleic acid testing in medical institutions, and entry-exit monitoring, to capture new crown cases
of various types and stages from the source at the first time.
From the private sector to the official, the demand for epidemic data from all walks of life is still there, and it is more severe than in the past, especially at the critical moment
when new crown infections in various places have peaked or exceeded the peak, and the peak of severe disease has arrived or is approaching.
The information vacuum needs to be filled
.
In the face of the flood of new crown infection, official, non-governmental and academic groups are working hard to rebuild the reference system from various angles, using epidemiological models, questionnaires, search index predictions and other methods to try to study and judge the epidemic, in order to have a more accurate grasp
of the trend of the epidemic.
According to disease control sources, the relevant departments have issued a new version of the new crown surveillance plan, requiring the existing infectious disease direct reporting system and influenza surveillance system to be applied to the new crown infection surveillance
.
Some epidemiologists also stressed that while using the existing passive and normalized surveillance system, it is also necessary to mobilize all feasible active monitoring methods, including community surveillance and sampling of key populations
.
The National Disease Control Administration has also recently asked provinces to actively investigate local epidemic data, emphasizing that "methods are not limited"
.
When traditional epidemiological models fail
When traditional epidemiological models failFor two weeks after the promulgation of the "20 Articles" on November 11 last year, Professor Huang Senzhong and his team from the Institute of Public Health and Health of Nankai University have been painstakingly searching for new "references" to provide a basis
for their modeling.
But like all epidemiologists in China, Huang Senzhong faced the same problem: after the liberalization of epidemic prevention policies, the SEIR model of infectious disease dynamics, widely used in academia, suddenly
failed.
In this classic model, the population is divided into four categories: S is susceptible, E is exposed, I is infected, R is recover, and individuals are transferred
between categories.
To measure the change in the number of people in different groups, you can calculate
it from a number of indicators that measure the transmissibility of the virus.
Among them, the effective regeneration number Rt value over time is critical
.
The Rt value is usually understood as the real-time transmission index of the outbreak, that is, the number of
infected people who can transmit the pathogen to other susceptible people during the infection period.
When the policy was liberalized and human interventions were greatly reduced, the Rt value of the domestic new crown fluctuated sharply, "Many people came to ask me how much the Rt value was equal to now, and I joked with them that it was equal to 1,000
.
" This value is currently difficult to estimate
.
”
Huang Senzhong told Eight Points Jianwen
.
When the underlying variables are no longer exact, the model is heavily biased
.
As a result, the road to continue purely following the SEIR model is blocked
.
After spending two weeks searching for a "reference" similar to large-scale nucleic acid results, Huang Senzhong decided to abandon this classic path and start
from scratch.
In the past three years, Huang Senzhong's team has conducted epidemic research and trend prediction
in many cities.
During the epidemic in Shanghai at the turn of spring and summer last year, in order to predict whether there would be a run on medical resources, the team had predicted the "pressure value of inpatient beds"
.
Huang Senzhong believes that when a new epidemic occurs, the first thing to feel pressure on is medical resources
.
Being able to predict the peak number of beds admitted to critically ill and general patients, and the timing of the peak, would better assist public health departments in decision-making
.
This value brought Huang Senzhong new inspiration
.
He revisited existing thinking and found that it was imperative to gauge how much health resources would be hit everywhere and what policymakers needed to prepare
.
In the process of estimating the response capacity of medical resources, if you can understand the number of new crown patients in the fever clinic and the new crown attendance rate, or the number of severe cases and severe disease rate and other medical data, supplemented by SEIR model prediction, you can reverse the size of the number of
infections in a region.
"For example, around December 13, many people said that four to five million people in Beijing had been infected, and even 60 percent of Beijingers had already been infected
.
" Huang Senzhong explained, "But based on the severe cases and deaths of the new crown in Beijing that we had at hand, and at the same time referred to the critical illness and mortality rate in hospitals during the epidemic in Shanghai, Jilin and Hong Kong, and then divided by the 10% attendance rate, we ran out of the model and speculated that the number of infected people in Beijing at that time was only 720,000
at most.
" ”
In the estimation process, the most difficult is to estimate the rate of visits, because it fluctuates the most and is subject to many
interference factors.
Senzhong Huang and his team's long-standing work model has been to combine mathematical model prediction with "information analysis," which aims to analyze the living habits and behaviors of residents everywhere
.
Under the cognitive framework of "information analysis", the rate of
visits is constantly fluctuating.
"After the initial policy was relaxed, everyone was very nervous, rushed to the hospital, and then found that they could heal themselves at home, so they stopped crowding the hospital, and many hospital wards were more than half empty, which we defined as 'residents' self-limited sexual behavior
'.
But then there was an increase in the number of critically ill patients, and hospital visits surged
again.
Huang Senzhong said
.
According to the actual situation observed, Huang Senzhong used it as a framework for "information analysis" and dynamically adjusted the value of the medical attendance rate accordingly
.
Because of the country's vast data volume, his goal is to estimate the order of
magnitude of quasi-infections as much as possible.
In mid-December, Huang Senzhong and his team made a prediction: 2022.
12.
20~2023.
1.
15 will usher in the peak period of hospitalization, which will be the first round of impact
after release.
After that, from the beginning of 2023 to the first half of 2023, the country will usher in two rounds of smaller shocks
.
Based on this, he figuratively described this predicted trend as "a twist and turn"
.
Behind the "fever" search index,
Behind the "fever" search index,Behavioral insights for infected people
Behavioral insights for infected peopleDuring this information vacuum period when official data lost its reference value, some non-governmental epidemic prediction models have also entered the public eye
.
Among them, a WeChat mini program called "Datagroup+" has been widely circulated on social media, with more than 10 million
views in just two weeks of its launch.
The Mini Program can predict the epidemic progress index of major cities, specific to the start and end dates of the first round of
peaking.
For example, it predicts that the first wave of the outbreak in Beijing will begin on November 26, 2022, peak on December 17, 2022, and end
on January 12, 2023.
△ Screenshot of "Data Blob+" Mini Program
Dr.
Chen Qin, who used to teach at the School of Economics of Fudan University and is currently the chief economist of Pulse Technology, co-developed this mini program
with his team.
They used Baidu Search and ByteDance to predict the progress of the epidemic in each region by using the changes in the search index of
"fever" keywords in each region on Baidu Search and ByteDance.
Chen Qin told Eight Points that since early November, he has been thinking about the impact of
the policy shift.
"Is this the best time to let go? How much will China's medical resources be affected, and will there be a large number of excess deaths? How is the current impact different from the past? These questions haunted Chen's mind, and he and his team wrote many discussion articles
.
Chen Qin found that predicting and monitoring the disease status of the region through the "fever" search index is a feasible means
.
In 2008, Google used this method to predict flu conditions in various states in the United States in advance, with greater sensitivity than the CDC
.
After fitting the data, Chen Qin found that the search index during the new crown epidemic is still highly correlated
with the increase in cases in various countries and regions around the world.
For example, the previous epidemics in Singapore, Japan, Hong Kong Special Administrative Region and Taiwan have kept pace
with the "fever" search index in time and proportion.
So he decided to try searching for exponential forecasting
.
During the epidemic, the search index of "fever" keywords will be higher than the benchmark level during the non-epidemic period, which Chen Qin defines as "oversearch multiple"
.
Looking back at the previous epidemics in Shanghai and Jilin, Chen Qin found that the "over-search multiples" of the two places were 1.
57 times and 1.
74 times
.
Since the end of September 2022, Urumqi's fever index has changed for the first time, which is 2.
6 times
the benchmark level.
Following Urumqi, the search index of Shijiazhuang, Baoding, Beijing and other places also rose
rapidly at the end of November.
When analyzing the Google index and the changes in the epidemic situation in various regions, Chen Qin found that when the "cumulative excess multiple" reaches a value related to the logarithm of the total local population, the local epidemic will reach its peak; After the other threshold is exceeded, the local outbreak tends to end
.
Therefore, Chen Qin used the epidemic search thresholds in different cities as the simulation threshold for the inflection point and end of the epidemic, monitoring the current progress of the epidemic in each region, the number of new cases per day in the future, and the beginning and end time of the epidemic peak
.
As soon as the Mini Program was launched, it became a hit, and the attention it received far exceeded the expectations
of Chen Qin and his team members.
Of course, some doubts have also come, and some scholars believe that this year is a year with a high incidence of influenza, and it is impossible to distinguish the specific reasons for users' search for "fever"; Some netizens believe that obtaining data through search engines has natural drawbacks, such as excluding groups that do not use the Internet
.
An epidemiologist commented to Eight Points that Google predicted influenza through more than 40 indicators, but the results were still
out of track.
Working with big data requires special care, as there can be "overfitting", where too many parameters and too little data can lead to so-called dimensional collapse
.
Chen Qin encountered another problem
in the later process of collecting data.
On December 19, he found that the "fever" index of both Baidu and Big Engine had dropped
sharply.
Chen Qin suspected that the data pollution was caused by human interference, and had to temporarily change the search terms
.
On December 27, Chen Qin wrote on social platforms: Up to now, the Mini Program predicts that the epidemic situation in the country has passed its peak
, except for a few areas with a small population and inaccurate search index calculations.
From Xingtai and Tianmen in Hubei on December 13 to Fuzhou, Xiamen and other places in Fujian on December 26, in just two weeks, the stage of the largest number of new infections in the country has passed
.
Chen Qin predicted that before the 2023 Spring Festival, the new epidemic situation in most areas will fall
sharply.
Officials are rebuilding their coronavirus surveillance system
Officials are rebuilding their coronavirus surveillance systemIn addition to the epidemic predictions of scholars and civil organizations, the disease control departments of Hainan, Zhejiang, Sichuan and other places have also released questionnaires on new crown infection in recent days and announced the results of the survey
.
Many public health experts have expressed caution against the online questionnaire method, believing that it has similar shortcomings
to the search index prediction method.
An epidemiologist told Eight Points that most questionnaires would ask participants whether they had symptoms such as fever, cough, muscle aches, or overlapping with flu or other respiratory infections, resulting in a decrease
in accuracy.
The nucleic acid and antigen results obtained from the questionnaire may not be accurate
.
Moreover, the questionnaire cannot reach residents who do not use the Internet, such as the elderly population, who may be the "silent majority"
.
Eight Points Health News also learned that nationwide, the existing two sets of information monitoring systems of the national disease control department have also been launched to monitor and predict the trend
of the new crown epidemic.
The two systems, one is the national infectious disease information direct reporting system established after SARS in 2003; The first is the Chinese influenza surveillance network, officially established in 2004, which covers 408 network laboratories and 554 sentinel hospitals
across the country.
Some public health experts said the two monitoring systems also had a number of flaws
.
For example, influenza data reporting is not strictly implemented, which leads to low
data quality.
What's more, most hospital outpatient clinics no longer require nucleic acid or antigen results from patients, so they cannot capture and report infection data
at the terminal.
Wei Sheng, professor of epidemiology at the School of Public Health, Tongji Medical College of Huazhong University of Science and Technology, told Eight Points that epidemiological surveillance is a long-term, dynamic and continuous collection of disease status data in the population, so as to understand the current situation and trend of the epidemic of diseases in the population, and determine the effect of prevention and control measures, which is one of the characteristics of
modern public health.
Monitoring can be divided into active
monitoring and passive monitoring.
Active surveillance refers to the disease control department taking the initiative to investigate and collect data
on the disease status in the population.
Passive surveillance, such as the current online direct reporting system, is detected by medical institutions and reported after diagnosis
.
Wei Sheng suggested that when the passive monitoring data is insufficient to judge the epidemic situation, the form of active monitoring can be used to understand the epidemic situation in time, such as the current method of influenza monitoring, by collecting samples from patients in medical institutions for new crown virus testing, to understand the epidemic trend
of the new crown epidemic.
It is also possible to sample some residents in the community for nucleic acid testing to understand the level of
new coronavirus infection among community residents.
The Zhejiang Zhoushan CDC conducted a comprehensive analysis of the city's nucleic acid positive detection rate, community sentinel monitoring and sampling survey of key populations, and estimated that as of December 29 last year, the proportion of infected people in the city reached 30%-40%, and the total number of infected people was close to 400,000
.
An epidemiological expert told Eight Points that the National Disease Control Administration requires provinces to investigate local epidemic data "without limitation", "Provinces and cities with sufficient funds can engage in antibody monitoring and community sampling, and insufficient funds can only do online questionnaire surveys
.
" ”
The expert also revealed that a conference call organized by the National Health Commission on December 21 reported that from December 1 to 20, the cumulative infection rate of the national population reached 17.
56%.
This national estimate is derived from the sum
of the estimates reported by the provinces.
In the new year, many megacities have moved from the peak of infection to the peak of severe disease, and the number of infections in small and medium-sized cities and rural areas continues to surge
.
With the Spring Festival just around the corner, the national epidemic situation will be more complicated, the new crown surveillance system still needs to be improved, and the public expects higher quality public information
on the epidemic.
