FORESIGHT COVID-19

FORESIGHT COVID-19: OUTBREAK AFTER WEAKENING OF QUARANTINE MEASURES

Date of publication 21.06.2020

Contents

1. Peculiar features of the COVID-19 pandemic spread in Ukraine after weakening of the quarantine measures

2. Impact of quarantine measures weakening on territorial distribution of the COVID-19 pandemic in Ukraine

2.1. General tendencies

2.2. Detailed analysis by Ukraine’s regions

3. Foresight modeling of the COVID-19 pandemic spread in Ukraine up to the end of June, 2020

Outcomes

References

Project taskforce

1. Peculiar features of the COVID-19 pandemic spread in Ukraine after weakening of the quarantine measures

The fourth month of the COVID-19 pandemic spread comes to an end in Ukraine. Measures taken by the national government (Decree of the Cabinet of Ministers of March 11, 2020, No. 211 and others) in the time period between the beginning of March and the end of April 2020, in principal, helped to considerably decelerate the process of the disease rapid spread and moderate its heavy consequences. Under the quarantine restrictions the evolution of the epidemiologic situation in Ukraine was under control, it was characterized with the comparably low number of cases and flattening of the morbidity curve (350 – 550 reported cases daily, except some outbreaks that didn’t influence the general trend), which was stretched out in time. In the last third of May, 2020, a slight downtrend of the decease spread emerged (see for more details).

However, the four-stage weakening of the quarantine measures and transition to adaptive quarantine resulted in the considerable relaxation of social discipline of citizens, their carelessness, increased mobility and other factors that caused the fast growth in the decease rate and the exponential pandemic spread in the first half of June, 2020. Dramatic growth in the number of reported sick persons in the first and second thirds of June, 2020, comes from the fact that the number of medical tests went up one-and-a–half times during this period of time (Table 1-2, Fig. 1-3).

Table 1. Change in the COVID-19 pandemic spread in Ukraine in the process of weakening of quarantine measures

	Cases	Total tests	New tests	Event
12.03.20-11.05.20				Strict quarantine
01.05.2020		118545	6686
02.05.2020	550	122752	4207
03.05.2020	502	129723	6971
04.05.2020	418	134592	4869
05.05.2020	366	139759	5167
06.05.2020	487	144283	4524
07.05.2020	507	151569	7286
08.05.2020	504	159155	7586
09.05.2020	515	167107	7952
10.05.2020	522	176403	9296
11.05.2020	416	181552	5149	The first stage of the quarantine weakening – work was permitted for dental departments, beauty and hairdressing salons, consumer service enterprises; it was allowed to go to parklands, outdoor children’s playgrounds and athletic fields; public catering enterprises were permitted for carry-out service; public catering enterprises were permitted to use outdoor grounds in the course of their business activity; nonfood goods trading was allowed in shops (particularly, in shops that work in shopping and entertainment centers) etc.
12.05.2020	375	187307	5755
13.05.2020	402	192247	4940
14.05.2020	422	202495	10248
15.05.2020	483	211614	9119
16.05.2020	528	220638	9024
17.05.2020	433	227801	7163
18.05.2020	325	232899	5098
19.05.2020	260	239961	7062
20.05.2020	354	248529	8568
21.05.2020	476
22.05.2020	442	267185		At the second stage of the quarantine weakening the following activities were permitted: operation of industrial markets, big shops, overland city transport; running sport crowd-free events with the number of participants not exceeding 50, religious exercises under 1 person per 10 sq. meters restriction, resumption of urban and suburban transportation in the ordinary course, resumption of hotel activity (except hostels).
23.05.2020	432	277712	10527	Resumption of urban traffic in most of Ukrainian cities.
24.05.2020	406	285626	7914
25.05.2020	259	291868	6242	The Metro was allowed to operate without special permits, preschool centers were allowed to work.
26.05.2020	339	301736	9868
27.05.2020	321	312532	10796
28.05.2020	477	322746	10214
29.05.2020	429	337318	14572	The indicators did not allow to ease the quarantine in 4 regions only. They include: Chernivtsy, Lviv, Rivne and Donetsk regions
30.05.2020	393	348279	10961
31.05.2020	468	356565	8286
01.06.2020	340	363187	6622	The third stage of the quarantine weakening – partial resumption of interregional and suburban railway and motor-vehicle transportation. Operation was resumed in fitness-centers, sporting halls, swimming pools, educational institutions (to provide EIT and attestations). The obligatory self-isolation was cancelled for those 60+ persons, who were representatives of clubs and coaching staff.
02.06.2020	328	371668	8481
03.06.2020	483	381552	9884
04.06.2020	588	392316	10764
05.06.2020	553	403551	11235	Resumption of work of restaurants and cafes in the ordinary course, resumption of divine services near and inside churches under 1 person per 5 sq. meters restriction, cancellation of obligatory self-isolation for 60-year persons, remote accreditation of educational programs. 7 regions (Lviv, Chernivtsy, Luhansk, Donetsk, Zhytomyr, Rivne and Volyn), as well as the city of Kyiv appeared unready for the weakening of restrictions
06.06.2020	550	414542	10991
07.06.2020	485	424046	9504
08.06.2020	463	431085	7039
09.06.2020	394	437444	6359
10.06.2020	525	445940	8496	The fourth stage of the quarantine weakening – it was allowed to open cultural institutions and hold cultural events. Functioning of cinemas and fan-zones remains under the ban at this time. Operation of healthcare and recreation facilities was resumed (except children camps).
11.06.2020	689	456509	10569
12.06.2020	683	468172	11663
13.06.2020	753	479111	10939
14.06.2020	648	489334	10223
15.06.2020	656	497284	7950	International air communication was partially resumed, but entry into EU and Schengen countries is possible only for the purpose of work, education and medical treatment.
16.06.2020	666	507251	9967
17.06.2020	758	517995	10744
18.06.2020	829	530442	12447
19.06.2020	921	542247	11805
20.06.2020	841	553737	11490

Particularly, studies of the human mobility show that it considerably exceeds initial indicators as of the beginning of the current year. Thus, at the beginning of June, 2020, the average driving mobility in Ukraine reached 150% since the beginning of the year, and walking mobility exceeded 100%; since the 11th of May Ukraine practically became on a par with Sweden as to the mobility (Fig. 1). It is important to emphasize that in Poland and Romania, where the decease spread was similar to that in Ukraine during May, 2020, a substantial increase in the number of new infected persons was not observed in June, 2020; and in Italy this process monotonically decreases (Fig. 4). Meanwhile, during the first and second thirds of June, 2020, the trends of the COVID-19 morbidity in Ukraine were taking on the signs of similitude to respective tendencies in Sweden, though the latter did not introduce substantial quarantine restrictions (Fig. 2).

Figure 1. Dynamics of human mobility in Ukraine and other countries of Europe during the quarantine period

Figure 2. Correlation between COVID-19 cases and general human mobility in Ukraine during 4-stage weakening of the quarantine measures

Figure 3. Correlation between the number of reported COVID-19 cases and the number of performed tests

Figure 4. COVID-19 pandemic spread pattern in Ukraine and some European countries

In order to escalate anti-epidemic measures the Government specifies three key parameters that reflect the virus transmission control, level of capacity of the healthcare network, epidemiologic centers and healthcare system as a whole to resist the spread of the COVID-19 acute respiratory disease, which is caused by the SARS-CoV-2 coronavirus and, consequently, enable the epidemic situation control at the level as of the current date:

incidence (number of new COVID-19 cases during the recent 7 days per 100 000 persons) shall be less than 12 per 100 000 persons (the seven-day morbidity level per 100 000 persons in Ukraine is + 50% during the period of conventional pandemic peak);
occupancy of beds in health care facilities intended for the patients with confirmed COVID-19 cases is less than 50% (it is the sufficient number of beds to ensure in-patient care, given the specified morbidity rate and the needed reserve of beds in case of morbidity rate increase);
testing coverage at the level which is not lower than the incidence level – that is, the average number of tests (using the method of polymerase chain reaction and enzyme-linked immunosorbent assay) during recent 7 days shall be over 12 per 100 000 persons (Table 2).

2. Impact of the quarantine measures weakening on the territorial distribution of the COVID-19 pandemic in Ukraine

2.1. General tendencies. The analysis of the number of persons in Ukraine infected with COVID-19 during the first and second ten-day periods of June 2020, shows an increase in territorial irregularity of their distribution dynamics in different regions and settlements of the country (Tables 2, 3; Fig. 5). These disproportions are caused by specific features of population communication in different regions of Ukraine, non-uniformity of migration flows, regional features of counteraction and control of the epidemic. In particular, due to weakening of quarantine measures and intensification of communications with Suchava parish (Romania), Belgorodskaya oblast (Russian Federation), Byelorussia, where the worse epidemic situation is observed with regard to incidence indicators, in the last weeks significant increase in coronavirus cases is observed in some border areas of Ukraine. As a result, epidemic situation in adjacent regions is gradually worsening due to renewal of transportation between the regions of Ukraine.

Figure 5. Level of incidence in different regions of Ukraine

Table 2. Analysis of epidemic situation as of the end of the second ten-day period of June, 2020, in the regions of Ukraine

Region	Indicator 1 (incidence)	Indicator 2 (occupancy of beds)	Indicator 3 (test coverage) 22.05.2020	Indicator 3 (test coverage) 03.06.2020	Indicator 3 (test coverage) 18.06.2020	Readiness for the stage ІІ
Targeted indicator	< 12 per 100 000	< 50%	> 12 per 100 000	> 12 per 100 000	> 12 per 100 000	✔ or ✘
City of Kyiv	13,92	38,65	43,49	63,89	98,86	✘
Vinnytsia region	21,18	40,94	34,65	40,22	88,57	✘
Volyn region	33,36	40,41	14,50	26,79	70,10	✘
Dnipropetrovsk region	1,26	1,26	12,15	9,94	17,96	✔
Donetsk region	4,20	1,99	8,32	8,85	22,66	✔
Zhytomyr region	13,42	3,08	21,05	12,58	27,24	✘
Transcarpathian region	32,78	49,70	26,48	35,54	41,44	✘
Zaporizhia region	1,84	6,33	18,92	27,95	30,21	✔
Ivano-Frankivsk region	17,84	18,14	18,34	21,63	30,77	✘
Kyiv region	11,22	25,21	17,70	34,84	44,74	✔
Kirovohrad region	4,40	21,96	22,45	43,18	54,40	✔
Luhansk region	2,09	1,82	7,26	13,30	17,68	✔
Lviv region	39,19	29,48	12,12	22,83	47,42	✘
Mykolaiv region	1,07	1,84	13,33	19,61	22,63	✔
Odesa region	6,02	3,77	16,94	25,01	20,75	✔
Poltava region	0,79	1,31	11,81	16,91	37,64	✔
Rivne region	34,44	21,60	38,01	48,18	72,41	✘
Sumy region	3,09	3,47	17,78	79,93	33,99	✔
Ternopil region	18,59	9,85	31,94	34,08	45,12	✘
Kharkiv region	12,53	7,95	13,65	11,61	28,11	✘
Kherson region	0,49	1,40	27,10	47,13	49,95	✔
Khmelnytskyi region	5,10	8,95	13,59	17,51	31,65	✔
Cherkasy region	5,21	7,29	25,13	91,94	65,28	✔
Chernivtsi region	37,95	52,37	42,15	47,46	55,64	✘
Chernihiv region	10,81	11,71	15,65	28,58	24,47	✔

From there, the group with negative epidemiologic situation stands out among the regions. It includes Chernivtsi, Transcarpathian, Ivano-Frankivsk, Lviv, Volyn, Rivne, Zhytomyr, Vinnytsia, Kyiv, Kharkiv regions and the city of Kyiv (Table 2, 3; Fig. 5). Maximal number of incidences in Chernivtsi and Lviv regions amounts to about 40 cases per 100 000 population. The threshold of 12 cases per 100 000 population is exceeded in these regions. The existing epidemiologic situation is characterized by existence of many regions with low incidence of the decease in the South, East and Center of the country. The character of the decease spread in these regions drastically differs from the same in the West regions.

Table 3. Escalation / Weakening of quarantine measures in the regions of Ukraine

Region	Weakening	Escalation
Ukraine
City of Kyiv	Quarantine was not eased up (public catering enterprises)
Vinnytsia region		12.06 Cultural institutions do not work
Volyn region		04.06 Kindergartens and sporting clubs do not work
Dnipropetrovsk region
Donetsk region
Zhytomyr region		06.06 catering enterprises, cultural institutions, accommodation services, divine services do not work
Transcarpathian region		18.06 escalation. Kindergartens and sporting clubs do not work, interregional transportation is banned
Zaporizhia region
Ivano-Frankivsk region		15.06 fitness-centers and cultural institutions do not work, religious exercises are restricted
Kyiv region
Kirovohrad region
Lugansk region
Lviv region	Kindergartens, sporting clubs, special vehicles do not work till 19.06
Mykolaiv region
Odesa region
Poltava region
Rivne region	Kindergartens and sporting clubs do not work, intercity transport does not function
Sumy region
Ternopil region		17.06 ban on public events
Kharkiv region
Kherson region
Khmelnytskyi region
Cherkasy region
Chernivtsi region	Public catering enterprises do not work
Chernihiv region

The greatest occupancy of beds is observed in Chernivtsi region where it exceeds 50% threshold of occupancy, as well as in Transcarpathian, Vinnytsia, Volyn regions and the city of Kyiv. Lviv, Ivano-Frankivsk, Rivne, Kyiv and Kirovohrad regions take the high risk to exceed the occupancy limit (Fig. 6).

Figure 6. Level of occupancy of beds in the regions of Ukraine

The number of tests that are performed on an everyday basis, grows permanently over a period of May – June. As a result of comparison, the average number of tests for 7 days per 100 000 population went up 2-3 times for the regions with complicated epidemiologic situation as compared to the figures of May, 22. As of today, all the regions have the rates that exceed 12 tests per 100 000 persons. It also has an impact on the growth of statistical fixation of new reported cases. Respectively, according to the data as of the end of the second ten-day period of June, 2020, the number of tests performed in Vinnytsia, Volyn, Rivne regions and the city of Kyiv is the most (Fig. 7).

Figure 7. Regional testing coverage in Ukraine

The beginning of increase in number of incidences coincides with implementation of the 3rd stage of quarantine weakening, which means that the situation began to worsen during the 2nd stage when commercial markets were opened and the work of intercity and municipal transport restarted (Fig. 8).

Figure 8. Comparison of quarantine weakening stages and daily dynamics of persons that sickened, recovered, and their difference

The situation began to worsen during the first two weeks of June in the most problematic regions. The worst COVID-19 morbidity dynamics is observed in Lviv region. The situation is traditionally complicated in Chernivtsi region and in the city of Kyiv (daily increase equals 1,5%). The situation is worsening at a swift rate in Rivne, Kyiv, Transcarpathian and Ivano-Frankivsk regions. As to eastern regions, the complicated situation is observed in Kharkiv region and, in particular, in the city of Kharkiv. At the same time the number of cases is practically unchanged in a number of regions (Fig. 9). Here, the maximum daily increase in a number of COVID-19 sick persons equals 3,5% - 4% of the total number of cases.

Figure 9. COVID-19 morbidity dynamics in the regions of Ukraine

2.2. Detailed analysis by Ukraine’s regions. In this section the continuation of the first two regional studies is given, which are focused on the analysis of particular features of COVID-19 pandemic spread in Ukraine’s regions:

Foresight COVID-19: regional aspect;
Foresight COVID-19: transfer to the phase of coronavirus pandemic decrement.

The study was carried out with due regard to the change of the process character after quarantine measures weakening, its considerable territorial irregularity, different communication variances of the population, uneven migration flows, regional peculiarities of resistive action against and control of the decease, etc.

As in preceding investigations, in order to reveal long-term trends, the methods of technical analysis of time series based on the base indicators, in particular, «zigzag» and «supertrend», that are useful for monitoring the main trends and revealing of “trading signals” at stock markets, were used [1-3]. The «supertrend» indicator is selected due to the fact that it is an effective tool of technical analysis for revealing a trend on highly volatile data. When the output curve shows ascending tendency, the values of «supertrend» indicator are below the curve; respectively, in case of descending trend the values of this indicator are over the output data chart. Crossing of the indicator and data curves may demonstrate completion or violation of the preceding tendency. Multiple crossing of these curves shows that the data does not represent an evident tendency. Along with a majority of technical indicators, the «supertrend» indicator shows a data trend change with certain delay, however, in the opinion of the project taskforce, this feature of the indicator is substantial for answering the question “Whether the incidence peak remained in the past?” and may be appraised as “tentativeness” of the respective opinion.

«Zigzag» indicator joins the most substantial local extremums on a data chart and is not sensitive to small fluctuations. This indicator is easy to use for analyzing preceding data fluctuations. Beginning from the second stage of the analysis of coronavirus pandemic in Ukraine, «ivar» technical indicator (trend strength indicator) and «АТР» indicator (volatility indicator) were involved. The first makes it possible to analyze the existing trend strength or confirm its absence in the data. Impact of this indicator is based on fractal characteristics of time series. The rules of its use are very simple: if the indicator’s value exceeds 0.5, it means that a trend is absent; and what is more, the closer this value to 1, the more reliable this statement; and otherwise, if the indicator’s value is less than 0.5, it means that a trend (regardless what its nature is) exists, and, the closer this value to 0, the stronger the respective tendency.

«АТР» indicator makes it possible to measure the volatility that shows a level of data volatility in time. One of the principles of «АТР» use lies in the following: the higher this indicator’s value, the higher the possibility for the existing tendency to change.

1. Ukraine in its entirety (except temporarily occupied territories):

Figure 10. Analysis of new reported cases in Ukraine

The recent peak of the number of new reported sick persons – June 10, 2020. The «Ivar» trend strength indicator is below 0.5, which means the beginning of a new tendency of the process development. Values of the volatility indicator (existence of risks) in June grew up to that level when the first peak values were observed in Ukraine – April 22, 2020. However, high volatility exists for the output chart of number of new reported cases per day because this data fluctuated in the range of 350 to 550 during the recent month.

The official data as of the end of May testified that the descending tendency of the pandemic development may possibly begin, as indicated by the findings of the preceding study made by the project taskforce on 01.05.20 and 30.05.2020. It was confirmed by the chart of difference between the number of new reported cases and number of recovered persons, per day (Fig. 1a). The lower the value of blue curve, the more confidence in the pandemic decrease. However, the volatility of this curve remained high enough, and the volatility indicator demonstrated it. Consequently, the number of sick persons considerably grew after the quarantine weakening. It should be noted that during the quarantine the number of sick persons per day did not exceed 550 persons per day, but after the quarantine weakening their number reached, approximately, 700-900. The volatility during the first and second ten-day periods began to grow, consequently, risks of lack of control over epidemic situation also grow. Values of trend strength indicator decrease, which is to say that they go out of stable range of 350-550 sick persons in the direction of increasing.

As to the correlation of the number of new sick persons and new recovered persons per day - the tendency is also not optimistic. In the course of time the number of persons who were taken ill within 24 hours exceeds the number of those who came through the illness. Volatility of the blue chart is down, and it may indicate that this tendency will continue in the nearest future.

Figure 10a. Analysis of correlation dynamics of the number of new reported sick and recovered persons in Ukraine as of the middle of June

2. Detailed analysis by regions of Ukraine and the city of Kyiv

2.1. the city of Kyiv:

Figure 11. Analysis of dynamics of new reported cases in the city of Kyiv

Recent peak of the number of new reported cases per day – April 16, 2020. Within June the lateral motion of the chart with descending trend is observed. Volatility decreased. As to correlation of the number of sick and recovered persons, some stabilization of the pandemic process may be expected with cautious optimism.

Figure 11a. Analysis of correlation dynamics of new reported sick and recovered persons in the city of Kyiv

According to the data given by the Ministry of Health, the incidence factor (the number of sick persons per 100 000 population) is 13.82 as on June 19, which is confirmed by the findings of the study.

2.2. Vinnytsia region

Figure 12. Analysis of dynamics of new reported cases in the Vinnytsia region

Since May 27, 2020, the increasing trend of the number of new reported cases per day took place. The recent peak: June 10. High volatility (volatility of data from day to day) is observed; it continuously grows since June 9. The curve of correlation of the number of persons deceased and recovered since April 26, fluctuates near zero, and the volatility of the process decreases since May 13.

Figure 12a. Analysis of correlation dynamics of new reported sick and recovered persons in the Vinnytsia region

According to the data from the Ministry of Health 18.07 sick persons per 100 000 population are observed in the Vinnytsia region as on June 19.

2.3. Volyn region

Figure 13. Analysis of dynamics of new reported cases in the Volyn region

Ascending trend of the number of new reported cases per day is observed from May 4, moreover, this tendency considerably increased from June 6. Increased volatility of the data is also observed from June 6. The recent peak took place on June 9. «Supertrend» indicator line does not show a suspension of incidence level increasing

Figure 13a. Analysis of correlation dynamics of new reported sick and recovered persons in the Volyn region

Correlation dynamics of new sick and recovered persons demonstrates the worrying ascending trend. Within the recent month the number of recovered persons exceeded the number of sick persons only on June 12, 2020. There were 34.91 sick persons per 100 000 population in Volyn region as of June 19.

2.4. Dnipropetrovsk region

Figure 14. Analysis of dynamics of new reported cases in the Dnipropetrovsk region

From May 5 (the date of the last peak of new reported cases) the descending trend for the number of new reported cases per day remains. The volatility of the process is low.

Figure 14a. Analysis of correlation dynamics of new sick and recovered persons in the Dnipropetrovsk region

Correlation dynamics of new sick and recovered persons shows a stabilization of the incidence process in the Dnipropetrovsk region. 1.20 sick persons per 100 000 population are observed as on June 19.

2.5. Donetsk region

Figure 15. Analysis of dynamics of new reported cases in the Donetsk region

The recent peak of the number of new identified sick persons: June 15 (42 cases), then the descending trend of the number of new reported cases per day is observed. The volatility (in the data) grows from June 2. The chart of correlation of new reported cases and the number of recovered persons shows instability of the situation. From June 3, the number of sick persons exceeds the number of recovered persons (except June 16).

Figure 15a. Analysis of correlation dynamics of new reported sick and recovered persons in the Donetsk region

According to the data from the Ministry of Health the number of sick persons per 100 000 population is 3.73 as on June 19.

2.6 . Zhytomyr region

Figure 16. Analysis of dynamics of new reported cases in the Zhytomyr region

The indicators show that the tendency of increasing new reported cases still persists; the volatility (of the process data) grows from May 30 and now it is high.

Analysis of the recovery process dynamics from May 28 shows an instability and uneasiness of the situation. Within the last month the number of recovered persons exceeded the number of those who took sick on June 11 only.

Figure 16a. Analysis of correlation dynamics of new reported sick and recovered persons in the Zhytomyr region

According to the data from the Ministry of Health 15.57 sick persons per 100 000 population are observed as on June 19.

2.7. Transcarpathian region

Figure 17. Analysis of dynamics of new reported cases in the Transcarpathian region

The tendency of increase in new cases is observed from June 8. The last peak took place on June 11. The volatility (in the new decease incidence data) is high. Correlation dynamics of decease incidence and recovery processes is unstable. Within the month of June this indicator was highly volatile (the volatility indicator), and the number of cured persons never exceeded the number of new incidence cases; it shows the danger of the further process.

Figure 17a. Analysis of correlation dynamics of new reported sick and recovered persons in the Transcarpathian region

According to the data from the Ministry of Health the number of sick persons per 100 000 population is 36.21 as on June 19, the occupancy of beds in healthcare facilities is equal to 52,71%.

2.8. Zaporizhia region

Figure 18. Analysis of dynamics of new reported cases in the Zaporizhia region

Decrease in the number of new reported cases per day is observed from May 15. The volatility indicator (volatility of daily morbidity process) is not high; it points to the controlled morbidity situation in the region. The last peak was observed on April 29, 2020. The curve of correlation of new reported sick and recovered persons is stable from May 3. The volatility of this process is also low.

Figure 18a. Analysis of correlation dynamics of new reported sick and recovered persons in the Zaporizhia region

According to the data from the Ministry of Health, as on June 19, the incidence is less than 12 sick persons, the occupancy of beds in healthcare facilities intended for hospitalization of patients with confirmed COVID-19 cases is less than 50%, PLR-testing coverage is wide enough.

2.19. Ivano-Frankivsk region

Figure 19. Analysis of dynamics of new reported cases in the Ivano-Frankivsk region

The volatility is high. The last peak of the number of new identified sick persons per day was observed on April 8. Slight ascending trend is observed from June 10, but on June 17, the «supertrend» line crossed the chart of the number of new reported cases, which hold a promise of the trend change. Analysis of correlation dynamics of decease incidence and recovery processes points to uneasiness of the situation, because from June 7 the number of new reported cases permanently exceeds the number of recovered persons.

Figure 19a. Analysis of correlation dynamics of new reported sick and recovered persons in the Ivano-Frankivsk region

According to the data from the Ministry of Health, as on June 19, the number of sick persons per 100 000 population is 19.92.

2.10. Kyiv region

Figure 20. Analysis of dynamics of new reported cases in the Kyiv region

The last peak of the pandemic took place on April 11. The volatility (volatility of daily morbidity) began to decrease from May 4, and now it is not high. Descending trend is observed from June 12. Nevertheless, from June 1 and now the number of new reported cases exceeds the number of recovered persons.

Figure 20a. Analysis of correlation dynamics of new reported sick and recovered persons in the Kyiv region

According to the data from the Ministry of Health, as on June 19, the number of sick persons per 100 000 population in the region is 11.84.

2.11. Kirovohrad region

Figure 21. Analysis of dynamics of new reported cases in the Kirovohrad region

The peak of morbidity cases took place on April 14. The process is characterized by stable dynamics from May 6, and it continues so far. The volatility (volatility in the data of new reported cases) is low. Within the period from May 3 to May 23 the number of recovered persons permanently exceeded the number of sick persons. Now the chart fluctuates near zero.

Figure 21a. Analysis of correlation dynamics of new reported sick and recovered persons in the Kirovohrad region

According to the data from the Ministry of Health, as on June 19, the incidence is less than 12 sick persons, the occupancy of beds in healthcare facilities intended for hospitalization of patients with confirmed COVID-19 case is less than 50%, testing coverage is wide enough.

2.12. Luhansk region

Figure 22. Analysis of dynamics of new reported cases in the Luhansk region

Slight ascending tendency was observed from June 13. Unfortunately, researches had not enough data to analyze the situation in the region in details. The chart of correlation of new morbidity cases and recovered persons permanently fluctuates near zero.

Figure 22a. Analysis of correlation dynamics of new reported sick and recovered persons in the Luhansk region

According to the data from the Ministry of Health, as on June 19, the incidence is less than 12 sick persons, the occupancy of beds in healthcare facilities intended for hospitalization of patients with confirmed COVID-19 case is less than 50%, the testing coverage is wide enough.

2.13. Lviv region

Figure 23. Analysis of dynamics of new reported sick persons in the Lviv region

The number of new reported cases per day is steadily increasing; the last observed peak – June 10. The volatility (volatility in the morbidity data) increases from the end of May, and now it is high. The number of new cases per day permanently exceeds the number of recovered persons; it confirms unsafety of the situation.

Figure 23a. Analysis of correlation dynamics of new reported sick and recovered persons in the Lviv region

According to the data from the Ministry of Health, as on June 19, the number of sick persons per 100 000 population is 40.38, and it substantially exceeds the population morbidity threshold set in the region.

2.14. Mykolaiv region

Figure 24. Analysis of dynamics of new reported cases in the Mykolaiv region

The morbidity peak was observed on April 20. Then, there was a decrease of the number of new reported cases per day, followed by volatile changes of the chart in the range from 20 to 0 new reported cases per day. The following chart (Fig. 24a) confirms the stable situation of the morbidity progression.

Figure 24a. Analysis of correlation dynamics of new reported sick and recovered persons in the Mykolaiv region

2.15. Odesa region

Figure 25. Analysis of dynamics of new reported cases in the Odesa region

The volatility (volatility of the data related to new identified incidences) decreases from May 7 to June 12, then a slight volatility increase is observed. Now the lateral motion of the chart is observed. However, from June 6, the number of new reported cases exceeds the number of recovered persons.

Figure 25a. Analysis of correlation dynamics of new reported sick and recovered persons in the Odesa region

According to the data from the Ministry of Health, as on June 19, the incidence was less than 12 sick persons, the occupancy of beds in healthcare facilities intended for hospitalization of patients with confirmed COVID-19 cases was less than 50%, the testing coverage was wide enough.

2.16. Poltava region

Figure 26. Analysis of dynamics of new reported cases in the Poltava region

The pandemic peak was observed on April 22. Steady descending trend of the number of new reported cases per day takes place, as well as descending volatility of new reported cases per day. Now the number of new morbidity cases is close to zero. From May 3, the number of new reported recovered persons generally exceeds the number of sick persons.

Figure 26a. Analysis of correlation dynamics of new reported sick and recovered persons in the Poltava region

2.17. Rivne region

Figure 27. Analysis of dynamics of new reported cases in the Rivne region

The last peak in the number of new reported cases per day took place on June 11. It is not certain that a stable ascending or descending tendency exists for the number of new morbidity cases in the region. Quite high volatility (volatility of the pandemic progression) takes place, that points to high risks of unfavorable pandemic progression. High volatility is demonstrated also by the correlation dynamics of new reported sick and recovered persons.

Figure 27a. Analysis of correlation dynamics of new reported sick and recovered persons in the Rivne region

According to the data from the Ministry of Health, as on June 19, in the Rivne region the threshold number of sick persons per 100 000 population is 40.08.

2.18. Sumy region

Figure 28. Analysis of dynamics of new reported cases in the Sumy region

The last defined peak – June 13. On June 14, the «supertrend» line crossed the chart of the number of new reported morbidity cases, and it holds a promise of easing the situation. From the end of May, the volatility (volatility of daily morbidity data) considerably increased, that points to the loss of situation stability. Correlation dynamics of new reported sick and recovered persons is also unstable.

Figure 28a. Analysis of correlation dynamics of new reported sick and recovered persons in the Sumy region

2.19. Ternopil region

Figure 29. Analysis of dynamics of new reported cases in the Ternopil region

The pandemic peak took place on April 29. Then there was a steady decrease in the number of new reported cases per day. The ascending trend was observed from June 1. The correlation of the number of new reported morbidity cases and recovered persons has the worrying ascending tendency.

Figure 29a. Analysis of correlation dynamics of new reported sick and recovered persons in the Ternopil region

According to the data from the Ministry of Health, as on June 19, the number of sick persons per 100 000 population is 22.44.

2.20. Kharkiv region

Figure 30. Analysis of dynamics of new reported cases in the Kharkiv region

The ascending trend of the number of new diurnal morbidity cases takes place given quite high volatility. Negative correlation dynamics of new reported sick and recovered persons is observed from June 1.

Figure 30a. Analysis of correlation dynamics of new reported sick and recovered persons in the Kharkiv region

According to the data from the Ministry of Health, as on June 19, the number of morbidity cases is 11.97, the occupancy of beds in healthcare facilities intended for hospitalization of patients with confirmed COVID-19 cases is less than 50%, the testing coverage is wide enough.

2.21. Kherson region

Figure 31. Analysis of dynamics of new reported cases in the Kherson region

The pandemic peak took place on April 28. Now the number of new reported cases per day is close to zero. Figure 31a shows the positive dynamics of the process.

Figure 31а. Analysis of correlation dynamics of new reported sick and recovered persons in the Kherson region

2.22. Khmelnytskyi region

Figure 32. Analysis of dynamics of new reported cases in the Khmelnytskyi region

The last peak took place on June 2. A lateral motion of the chart is observed. The chart of correlation of new reported sick and recovered persons shows a positive tendency.

Figure 32а. Analysis of correlation dynamics of new reported sick and recovered persons in the Khmelnytskyi region

2.23. Cherkasy region

Figure 33. Analysis of dynamics of new reported cases in the Cherkasy region

The pandemic peak was observed on March 28. After sustained decrease, from May 26 the ascending trend of the number of new reported cases per day takes place. During the last month the number of new recovered persons exceeded the number of new morbidity cases only on June 17 (Fig. 33a).

Figure 33а. Analysis of correlation dynamics of new reported sick and recovered persons in the Cherkasy region

2.24. Chernivtsi region

Fig. 34. Analysis of dynamics of new reported cases in the Chernivtsi region

The morbidity peak took place on April 22-24, then a decrease in the number of new reported sick persons per day discontinued, the volatility (volatility of data of the pandemic process progression) is permanently low within June. A lateral motion of the chart is observed. During the month of June the number of new cases permanently exceeds the number of recovered persons. Moreover, volatility indicators of this process decrease from May 25, and it evidences that such negative tendency may last further.

Figure 34а. Analysis of correlation dynamics of new reported sick and recovered persons in the Chernivtsi region

According to the data from the Ministry of Health, the number of new morbidity cases per 100 000 population in the Chernivtsi region is 42.17, the occupancy of beds in healthcare facilities is 49.53%, which is close to the threshold value of 50%.

2.25. Chernihiv region

Figure 35. Analysis of dynamics of new reported cases in the Chernihiv region

The volatility was quite high within June. The last reported peak – June 3. Currently a lateral motion of the chart is observed. Figure 35a shows that from the end of May the number of reported cases exceeds the number of persons who recovered. There are risks of negative development of the pandemic process.

Figure 35а. Analysis of correlation dynamics of new reported sick and recovered persons in the Chernigiv region

3. Foresight modeling of the COVID-19 pandemic spread in Ukraine up to the end of June, 2020

Given essential uncertainties of the COVID-19 pandemic spread in Ukraine, caused, in the first turn, by unpredictable social behavior of individuals after weakening of quarantine measures, the “Foresight COVID-19” team gives very conservative foresight data as to the lower and upper bounds of the confidence interval of the disease spread up to the end of June, 2020. The confidence interval for the forecast of new identified sick persons in the second half of June was “robustly” determined using a combination of the following methods:

SARIMA model or ATIMA season model [4,5], which is based on the analysis of non-stationary time series with regard to seasons (in our case it is associated with specific registration of suspects and new identified sick persons, human factor, specifics of the medical statistician work, etc.). Using Box-Jenkins method, the time series stationary was estimated. If a series proved to be non-stationary, single roots and integration order was found for it. If the integration order was more than zero, the series was transformed into a (weakly) “stationary” one [6,7].
Linear regression models, in order to determine interrelation between different data sets using linear functions where unknown parameters of the model were estimated using the least-square method [8].
The method of Gradient Boosting Decision Tree on whose leaves (“branches”) the attributes are written on which the target function depends. The values of the target function are written on the “leaves” of the tree, while the attributes that characterize the difference between the cases are written in the nodes. The classification of new cases was made by “descending” the tree to the leaf and determining the respective value [9]. The XGBoost open library was used for training and structure optimization, where, in particular, the methods of clever penalization of trees, proportional reduction of leaf nodes, Newton boosting, are implemented, and the additional randomization parameter is used [10].

The foresight data with respect to the lower and upper bounds of the confidence interval of the disease spread till the end of June, 2020, which were received using the aforesaid methods, are shown in Fig. 36 and Table 4. The statistical error of calculations equals 0,03%.

Figure 36. Confidence intervals for the foresight new identified cases till July 1, 2020.

Table 4. Lower and upper bounds of estimated number of new sick persons till July 1, 2020

Date	Lower bound	Upper bound
2020-06-20	649	712
2020-06-21	655	720
2020-06-22	657	724
2020-06-23	656	727
2020-06-24	658	730
2020-06-25	666	731
2020-06-26	676	729
2020-06-27	676	725
2020-06-28	669	724
2020-06-29	662	727
2020-06-30	654	729
2020-07-01	645	732

The results of the short-term foresight modeling of number of COVID-19 sick persons in Ukraine for the period of 21.06.20 - 25.06.20 (Fig.37-40) are obtained using the “Back Propagation” multilayer neuron network on the basis of the “sliding window” mechanism with 12 data points for neuron network training. We see that after the significant pandemic outbreak during the first and second ten-day periods of June 2020, the slow decrease in number of new infected persons per day will be possibly observed (Fig. 37) prior to reaching the foresight confidence interval and remaining within its limits till July 1, 2020 (Fig. 36). The foresight data related to increase in number of deaths, number of recovered persons and difference between the number of SARS-CoV-2 recovered and infected persons in Ukraine are shown in Fig. 38-40, respectively. Such a character of the process evolution may possibly continue till the end of June 2020.

Figure 37. Short-term foresight of increase in the number of SARS-CoV-2 cases in Ukraine

Figure 38. Short-term foresight of increase in the number of SARS-CoV-2-related deaths in Ukraine

Figure 39. Short-term foresight of the number of SARS-CoV-2 recovered persons in Ukraine

Figure 40. Short-term foresight of difference between the number of SARS-CoV-2 cases and recovered persons in Ukraine

Outcomes

The quarantine measures weakening negatively impacted the epidemiologic situation in Ukraine which resulted in increase in daily reported cases up to 800-900 in the first half of June 2020. At the second and third stages of quarantine weakening the incidences dynamics worthened mainly due to opening the urban and intercity public transport for individuals, commercial centers, and substantial breaches of anti-epidemic precautionary measures at border checkpoints with Ukraine’s western neighbors. Statistical increase in number of new cases correlates with increase in number of testing in regions where these indices increase two- and threefold compared with the beginning of May.
Situation by Ukraine’s regions is diametrically different for the Western part of the country where it essentially worthens or continues to be grave, and the South-Eastern part. Hence, it may be deduced that the number of reported cases in the Western regions of Ukraine is critical and considerable risks of further pandemic spread in other regions of Ukraine exist.

As on mid-June 2020, the pandemic rapid proliferation continues in Volyn, Lviv, Rivne, Ivano-Frankivsk, Chernivtsi regions and in the city of Kyiv. During the first half of June the tendency to incidence increase over the average level was observed in Vinnytsia, Transcarpathian, Zhytomyr, Kharkiv, Ternopil and Chernihiv regions; these regions were in the group of middle and low risk of disease spread as of the end of May.
Among the regions that demonstrated the low risk of worse pandemic scenario as on 30.05.2020 (see more details), the Zaporizhia, Kirovohrad, Poltava, Sumy and Kherson regions shall be mentioned, and they remain in this group. The Dnipropetrovsk region joined them, it had a middle risk of the worse epidemic scenario.
As on the end of the second ten-day period of June 2020 the Donetsk, Luhansk, Mykolaiv, Odesa, Khmelnytskyi and Cherkasy regions demonstrate the moderate risk of worse pandemic scenario due to moderate increase in number of new reported cases. However, the increase in PLR-testing number gives the cautious hope of the favorable progress of the process in these regions.
The number of new reported cases per day may vary within the foresight fluctuation range of 600-750 persons per day up to the end of June, depending on quarantine measures escalation/weakening and intensity of PLR-testing.

References

Peter Navarro, When the Market Moves, Will You Be Ready? McGraw-Hill Education, 2003.
Attilio Meucci, Risk and Asset Allocation. (Springer Finance) 1st ed. 2005. Corr. 3rd printing, 2009.
Marcos Lopez de Prado, Advances in Financial Machine Learning. John Wiley & Sons, Inc, 2018.
Operational monitoring of the COVID-19 situation by the National Health Service of Ukraine (dashboard)
Open data on the incidence of COVID-19 in Ukraine
Kissler, Stephen M., Christine Tedijanto, Edward Goldstein, Yonatan H. Grad, and Marc Lipsitch. "Projecting the transmission dynamics of SARS-CoV-2 through the postpandemic period." Science 368, no. 6493 (2020): 860-868.
Tseng, Fang-Mei, and Gwo-Hshiung Tzeng. "A fuzzy seasonal ARIMA model for forecasting." Fuzzy Sets and Systems 126, no. 3 (2002): 367-376.
Milos Hauskrecht, Linear Regression (Machine Learning). University of Pittsburgh, 2020.
Breiman, Leo; Friedman, J. H., Olshen, R. A., & Stone, C. J. (1984). Classification and regression trees. Monterey, CA: Wadsworth & Brooks/Cole Advanced Books & Software. ISBN978-0-412-04841-8.
XGBoost Library Documentation

Project academic advisor: M.Z. Zgurovsky

Project taskforce: O.S. Boyko, N.V. Gorban, I.M. Dzhygyrey, B.R. Dudka, K.V. Yefremov, Yu.P. Zaychenko, P.O. Kasianov, O.P. Kupenko, M.M. Perestiuk, I.O. Pyshnograiev, V.V. Putrenko

WORLD DATA CENTER

FORESIGHT COVID-19

FORESIGHT COVID-19: OUTBREAK AFTER WEAKENING OF QUARANTINE MEASURES

Need some data?

Who are we?

Contact information