FORECAST, 10.09.2020

Short-term forecasts for the development of the epidemic in Poland after September 10, 2020

Since September 4, 2020 Poland reports mostly symptomatic cases, which impacted daily number of reported cases.

Changes in testing after 03/09/2020

On September 3, 2020, Health Ministry announced changes in registering infections due to adjustments to testing. They shared this announcement on twitter, that they will test only symptomatic cases. Those adjustments impacted detected cases reported after September 3, 2020, what is clearly visible below:
Fig.1. Number of new daily cases as well as 7 and 14-day rolling average of new cases from March 2, 2020 to September 10, 2020.
Forecasts of cumulative and incident COVID-19 deaths and cases in Poland in a standardized format made by MOCOS model is presented below. Forecasts were shared for the German and Polish COVID-19 Forecast Hub.
Fig.2. Forecast of cumulative deaths in 4 weeks from September 7, 2020. Upper-left corner presents point predictions of various models on October 3, 2020. MOCOS model indicates around 2444 cumulative deaths on October 3, 2020. An inner bundle represents quantiles 25%-75%, while an outer bundle represents quantiles 2.5%-97.5% from MOCOS model.
Fig.3. Forecast of weekly deaths in 4 weeks from September 7, 2020. Upper-left corner presents point predictions of various models on October 3, 2020. MOCOS model indicates around 85 weekly deaths for the week ending on October 3, 2020. An inner bundle represents quantiles 25%-75%, while an outer bundle represents quantiles 2.5%-97.5% from MOCOS model.
Fig.4. Forecast of cumulative incidents in 4 weeks from September 7, 2020. Upper-left corner presents point predictions of various models on October 3, 2020. MOCOS model indicates around 84 302 cumulative incidents on October 3, 2020. An inner bundle represents quantiles 25%-75%, while an outer bundle represents quantiles 2.5%-97.5% from MOCOS model.
Fig.5. Forecast of weekly incidents in 4 weeks from September 7, 2020. Upper-left corner presents point predictions of various models on October 3, 2020. MOCOS model indicates around 3694 weekly incidents for the week ending on October 3, 2020. An inner bundle represents quantiles 25%-75%, while an outer bundle represents quantiles 2.5%-97.5% from MOCOS model.

ANALYSIS RESULTS, August 30, 2020

Short-term forecasts for the development of the epidemic in Poland after August 30, 2020

Until September 29, 2020, the 7-day average of daily new cases in Poland may increase to around 870 from the current level of around 730.

An upward trend

Fig.1. Number of new daily cases as well as 7 and 14-day rolling average of new cases from April 12, 2020 to August 30, 2020.

As of July 31, the 7-day mean daily increment in new cases increased from about 520 to about 730 on August 30.

If the current rate of growth does not change, then in the next month the 7-day average of daily increases in new cases will most likely increase to about 870 new cases per day (90% confidence interval for the 7-day mean is about (660, 1280)).

Fig.2. Probable trajectories of the daily number of newly diagnosed people until September 29, 2020 based on the microsimulation model. The trajectories were adjusted to the data from July 31, 2020 to August 30, 2020. The analysis assumed that the same restrictions and the same level of testing would be maintained in the future. The daily increases in new cases in Poland in the last month are marked with black dots. The black solid line marks daily increases in new cases in Poland that are additionally smoothed using a 7-day moving average. Each of the blue lines corresponds to the result of one model simulation that approximated the situation in Poland over the last month. The black dashed lines show the 5%, 50%, and 95% forecast quantiles from bottom to top. The red dashed lines show the minimum and maximum of the forecasts generated by the model, they are additionally smoothed using a 7-day moving average.

Recommended Strategy

In the current phase, the epidemic is characterized by a constant number of new infections in a fixed unit of time (e.g. during the week) - this is the so-called endemic state. The goal of all should be to bring the epidemic to an end as soon as possible by reducing the opportunities for new infections. This is not only possible, but also much more cost-effective economically and socially than prolonging the duration of the threat.

Detect

The authorities should carry out extensive testing for the presence of the virus, with the aim of increasing the detectability of infections in persons with asymptomatic or mild symptoms.

Backtrack

Epidemiological services should quickly and effectively monitor the contacts of infected persons. Citizens should use the contact tracing application, as at least half of the population needs to use it to be an effective tool to reduce the epidemic.

Reduce

To extinguish the epidemic, restrictions on social contacts must be maintained.

We can consider the epidemic to be over if we do not record new infections for 14 days. Remember that even an increase of 5 cases per day is a threat - every country that has experienced a rapid increase in infections was previously only a few cases.

ARCHIVE RESULTS, August 24, 2020

Short-term forecasts for the development of the epidemic in Poland after August 23, 2020

Until September 22, 2020, the 7-day average of daily new cases in Poland may increase to around 800 from the current level of 725.

An upward trend

Since the first loosening of the restrictions announced at the end of April, the mobility of Poles has consistently increased from the level of -60% compared to the reference value, i.e. the level of mobility observed from the beginning of the year to mid-March, and has remained at the level of + 70-80% of the value since mid-July[1]. At the same time, from mid-July to approx. August 10, we observed a strong increase in the daily increase in new cases, which significantly slowed down over the past two weeks.

Fig.1. Number of new daily cases as well as 7 and 14-day rolling average of new cases from April 12, 2020 to August 24, 2020.

As of July 24, the 7-day mean daily increment in new cases increased from 375.86 to 725.43 on August 23.

If the current rate of growth does not change, then in the next month the 7-day average of daily increases in new cases will most likely increase to about 800 new cases per day (90% confidence interval for the 7-day mean is about (600, 1200)).

Fig.2. Probable trajectories of the daily number of newly diagnosed people until September 22, 2020 based on the microsimulation model. The trajectories were adjusted to the data from July 24, 2020 to August 23, 2020. The analysis assumed that the same restrictions and the same level of testing would be maintained in the future. The daily increases in new cases in Poland in the last month are marked with black dots. Each of the blue lines corresponds to the result of one model simulation that approximated the situation in Poland over the last month. The black dashed lines show the 5%, 50%, and 95% forecast quantiles from bottom to top. The red dashed lines show the minimum and maximum of the forecasts generated by the model, they are additionally smoothed using a 7-day moving average.

TECHNICAL REPORT, 04.06.2020

COVID-19 Pandemic: Analysis of the current situation and short-term forecasts

The MOCOS Group has developed an advanced microsimulation model that simulates the behaviour of individual agents (individuals). It allows to assess the current level of virus reproduction and the stage of development of the epidemic, the scale of reduction of human contacts achieved, the level of detection of virus-infected people with mild symptoms and the impact of recommended actions on the possibility of reducing the development of the epidemic.

REPORT 🇵🇱
PRESS RELEASE 🇵🇱

Summary

The epidemic in Poland is growing again, and outbreaks of infections appear like mushrooms after rain. A month ago we reported that the epidemic in Poland has gone to extinction, this trend has been reversed. Now the country is again at the stage of steady growth of new infections. The largest share in the number of newly diagnosed cases is in Śląskie, Mazowieckie and Łódzkie Voivodeships.

In Lower Silesia, the growth of new infections is decreasing with each day, but we are far from a lack of new infections.

The epidemic in Wrocław is also slowly dying out, but we have only 10% chance to finish it before the new school year. For Poland and Lower Silesia the forecasts do not give even such small chances for the end of the epidemic before September.

Home and workplace contacts are the main sources of infections, and the proportion of workplace infections is likely to increase in the future.

Car mobility among the population - both in Poland and Lower Silesia or Wrocław itself - has reached the state before the epidemic. We move as often as we do, and we come into contact with other people, which is a natural mechanism of virus reproduction.

If the record numbers of new infections reported in recent days are not limited to the households of mine workers and other workplaces that are outbreaks of infections, the more pessimistic variants of our forecast for the development of the epidemic must be taken seriously.


Poland

Map of probable parameter configurations for the COVID-19 epidemic state in Poland on 03/06/2020. The shades of blue indicate the probable location of the parameter set.

State of the Epidemic

Modelling as of 03.06.2020 indicates that the epidemic in Poland lies on the critical line. The location of the epidemic on the critical curve resembles a location on the edge, from where it can easily lean towards both the extinction and growth phases.

Probable trajectories of the daily number of diagnosed cases in Poland up to 03/09/2020. The brown line is a 7-day average of the number of diagnosed cases in Poland from 04/03/2020. Each of the blue lines corresponds to the result of one of the model simulations that best described the situation in Poland in the last few weeks from 1/04/2020. The bundles of curves were generated on the basis of a microsimulation model with a contact tracking module.

Future Scenarios

The average forecast indicates a slow decline of the epidemic in Poland (below 300 diagnosed daily from the second half of July). The probable trajectories of the daily number diagnosed in Poland reflect this effect of being on the critical edge.

Lower Silesia


Map of probable parameter configurations for the COVID-19 epidemic state in Lower Silesia on 04/06/2020. Shades of blue indicate probable location of the parameter set.

State of the Epidemic

The scope of the epidemiological situation in Lower Silesia is very close to the critical line on the extinction side of the epidemic. The proximity of the critical line translates into a long period of extinction.

Probable trajectories of daily number of diagnosed cases in Lower Silesia to 03/09/2020. The brown line marks the 7-day average of the number of diagnosed cases in Lower Silesia from 16/04/2020. Each line corresponds to the result of one of the model simulations that best approximated the situation in the region in the last few weeks. The bundles of curves were generated on the basis of a microsimulation model with a contact tracking module.

Future Scenarios

Assuming that the current restrictions remain unchanged and that there is no increase in testing frequency or contact tracking efficiency: 03/07/2020 the confidence interval of the daily number diagnosed in Lower Silesia falls between 8 and 38 cases; 03/08/2020 the confidence interval of the daily number diagnosed in Lower Silesia falls between 3 to 31 cases; 03/09/2020 the confidence interval of the daily number diagnosed in Lower Silesia falls between 2 to 35 cases.

Wrocław


Map of probable parameter configurations for the COVID-19 epidemic state in Wrocław on 04/06/2020. The shades of blue mean probable parameter sets describing the current epidemiological situation.

State of the Epidemic

The epidemiological situation in Wrocław is currently good - the epidemic is in the phase of extinction. Only for a small assumed effectiveness of contact tracing, the epidemic may be on the critical line where the epidemic remains on a similar level. For Poland on the basis of available data we estimate b in the range 0.4-0.6, so this is the most probable area also for Wroc³aw.

Probable trajectories of the daily number of diagnosed cases in Wrocław until 03/09/2020. The brown line marks the 7-day average of the number of diagnosed cases in Wrocław from 16/04/2020. Each line corresponds to the result of one of the model simulations that best approximated the situation in the city in the last few weeks. The bundles of curves were generated on the basis of a microsimulation model with a contact tracking module. The red line marked the maximum trajectory. The dashed black line marked the part of the trajectory that lasts at least until a given day.

Future Scenarios

Assuming that the current restrictions remain unchanged and that there is no increase in testing frequency or contact tracking efficiency, according to the model, the range of likely daily number of people diagnosed from the beginning of July will be around 0 to 11. According to the model, up to 3/09/2020 about 90% of the simulated trajectories are not coming to an end (this means that for 10% of the trajectories until the beginning of September the epidemic ends).

TECHNICAL REPORT & CASE STUDY

Situation in Illigan City, Phillipines

We present two modelling results concerning the state of the pandemic in the city on 7.04 and our evaluation of the effect of the decision to open Mindanao State University - Iligan Institute of Technology as of 21.04.

Case study on opening MSU-IIT

21.04.2020

The simulations indicate that the opening of MSU-IIT can lead to a super-spreading event, if not combined with strict hygiene and social distancing measures, which reduce the number of secondary infections. It is too expect that the contact rate for students at MSU-IIT is higher than in the rest of the population due to teaching in classrooms and laboratories. Even an augmented immunity does not prevent from a massive outbreak. The more the immunity is built up the less intensive is the increased disease spread from MSU-IIT into the city. We strongly suggest to monitor the progression of COVID-19 during the weakening of the lockdown after April 30, 2020 before opening MSU-IIT. In addition we recommend to take strong social distancing and hygiene measures to milder the disease spread.

READ THE MSU-IIT CASE STUDY 🇬🇧

Report

5.04.2020

In different media there is an ongoing debate about weakening the social distancing measures. In this techreport we show for the example of the city of Iligan, Philippines, that, in order not to exceed the capacity of the health system, it it crucial to reduce the outer household contacts. Moreover a mitigation strategy is not recommendable, since the social distancing can not be tuned fine enough. We also simulate how massive testing combined with household contacts can be used in addition to social distancing measures.

READ THE REPORT 🇬🇧
READ THE REPORT SUMMARY 🇬🇧

TECHNICAL REPORT

Situation in Germany

We present the results of the modelling analysis of the effects of opening schools in Berlin and Rhineland-Palatinate.

Berlin


Report on the opening of schools in Berlin

28.04.2020 Number of infected depending on the level of detection of asymptomatic cases (q) and level of contact restriction.

If the contacts of 15-18 year olds in the school are reduced to 18 percent, about 7800 Berliners will be infected within 300 days, and 156 will die in total. To subtract from this number, there are about 5500 infected and 128 dead, which would also happen without the schools closing. If, however, in this age range, i.e. 8 instead of 10 contacts, contacts are reduced to 80 percent, almost thirty times more Berliners, about 200,000, will be infected and about 3,000 people will die. Even half the contacts of high school students (48 percent) will still lead to about 2700 dead and 185,000 infected people. The opening of schools and public life is in our opinion premature at this stage.

REPORT 🇬🇧

Rhineland-Palatinate


Report on the opening of schools in Rhineland-Palatinate

20.04.2020

Microsimulations for Rhineland-Palatinate on the basis of our model give strong indications that the opening of schools together with relaxed contact reduction can only be safely carried out with very high resistance. The opening of schools for individual classes is possible with strong hygiene and social distance if the rest of the population also follows a very strong contact reduction and a significant immunization is ensured. Opening schools and public life is in our opinion premature at this stage.

PRZECZYTAJ RAPORT 🇩🇪