A panel analysis on preventing and controlling efficiency of COVID-19

the numbers of COVID-2019 total confirmed cases (TC), total dead cases (TD), dead cases (ND), newly confirmed cases (NC), newly healed cases (NH), total healed cases (TH), and newly recovered cases (NR). Pooled, dynamic, and event study analyses conducted to reflect the associations among them. RESULTS: Descriptive analyses showed that NH/NC ratio in China are bigger than NR/NC in USA and the world. Pooled analysis showed that various roles of regions in NH and NR in a specific sample. Dynamic analysis showed significant roles of lags and NC in NH and NR. Panel event study showed that key events influence ND and NH in China significantly and NR in the world rather than NR in USA. CONCLUSION: The findings in this study indirectly confirmed the relationship between spreading growth, treatment efficiency, and death increase. China’s control strategies of COVID-19 pandemic are worth of learning by the globe.


Introduction
Now, health agencies worldwide are facing the challenges of the coronavirus disease 2019 (COVID-19) endemic. Saliva droplets and direct and indirect contact via surfaces identified as the possible risk factors for interhuman transmission of COVID-19 [1]. Since late January, massive public health interventions have been implemented by national bodies to contain the spread of the virus and infection.
Without validated drugs [2], successful strategies for COVID-19 from China such as wartime control measures [3], active surveillance, contact tracing, quarantine, and social distancing [4], enhanced traffic control bundling [5], and outbreak city shutdown [6]result in dramatic reductions in number of daily newly confirmed cases and deaths. Under the condition of inter-human transmission and lockdown measures, the temporal dynamics and characteristics of the COVID-19 epidemic in a high-risk city were reported in China [7]. In order to reduce morbidity and mortality, several medical decision-making tools have the potential to be a major breakthrough in efforts to control the current pandemic [8][9][10].
Despite travel restrictions and border control measures [11] and travel limitations [12], transmission law of COVID-19 has rapidly spread across the globe. Several theoretical studies depicted the COVID-19 increase of transmission between persons [13]. Based on the publicly available epidemiological data, a simulation forecast a trend of the COVID-19 spreading in Hubei, China [14]. With estimated 2.68 reproductive number for COVID-19, a study inferred that COVID-19 epidemics are already growing exponentially in multiple major cities of China with a lag time behind the Wuhan outbreak of about 1-2 weeks [15]. A simulation study indicated that around 4,000 and 18,300 people will die of COVID-19 in China and across the world, respectively [16]. The global spread of COVID-19 pandemic accelerates the speed and volume of clinical trials launched to investigate potential therapies for COVID-19 [17].
From December 12, 2019 till now, a series of daily policies and regulations released by the Chinese government, global organizations, and western countries which could be documented in China Data Lab [18]. A growing number of countries impose a nationwide containment policy in the fight against epidemic COVID-19 viral infection. A series of extreme measures in mitigating against further development of contagion were released for the national public health interest. Thus, in order to indirectly assess national struggling efforts against COVID-19 pandemic, this study with daily cases aims to analyze the associations of ND, TD, NH, TH, and NR with NC, TC, ND, and TD.

Data
China COVID-19 daily cases included the numbers of COVID-2019 TC, TD,   ND, NC, NH, and TH in China province-level units and city-level units from January   22, 2020 to March 18, 2020 [19]. United States of America (USA) COVID-19 daily cases included the numbers of COVID-2019 ND, NC, and newly recovered cases (NR) with state and county-level data available from March 17, 2020 to April 12, 2020 [20].
The World (outside Antarctica, China, and USA) COVID-19 daily cases included the numbers of COVID-2019 ND, NC, and NR from January 22, 2020 to April 4, 2020 [21]. China dataset contains information on 31 provinces and 181 cities. The USA dataset contains information on 51 states and 3143 counties. World dataset contains information on 246 countries and regions.

Pooled regression analysis
Here, OLS linear regression is employed to estimate:

Dynamic estimation
Here, a dynamic panel corresponding to a series of lag coefficients and covariates is employed to analyze the associations of ND, TD, NH, TH, and NR with their lags and covariates. Arellano and Bond estimator (AB estimator) and Arellano and Bover/Blundell and Bond system estimator (ABS estimator) are employed to explore the associations.

Panel event study
Here, a panel event study corresponding to a difference-in-difference style model with a series of lag and lead coefficients and confidence intervals with the method implemented by the program EVENTDD in Stata [22] is employed to analyze how the key events influence ND, TD, NH, TH, and NR, respectively. In the context, there three key events were adopted as treatments in China province-level units and  [ Table 2]

Dynamic analysis
The dynamic panel regression results between confirmed cases and dead cases in China province-level units are presented in Table 3. In AB estimator, lags ( [ [ Table 5] Panel event study Figure 1 shows that ND increase and then decline steadily over the twenty days following the key event. Figure 2 shows that TD decline steadily over the twenty days and increase rapidly following the key event. Figure 3 shows that TD decline steadily over the twenty days and increase rapidly following the key event. Figure 4 shows that TD decline steadily over the twenty days and increase rapidly following the key event. Figure 5 shows that the effect of the key event on NR in USA States is a relatively precisely estimated zero, reflecting both the lack of correlation between NR and key event and the relatively small number of NRs. Figure 6 shows that the effect of the key event on NR in USA Counties is a relatively precisely estimated zero, reflecting both the lack of correlation between NR and key event and the relatively small number of NRs. Figure 7 shows that NR decline steadily over the twenty days and increase rapidly following the key event.

Discussion
This study employed public available daily data in China, USA, and the world (outside Antarctica, China, and USA) and obtained the nexus between NC and ND, the nexus between NC and NR, and the nexus between NC, ND, and NR. Pooled regression showed the association between confirmed and dead cases, and the association between ND and NR, the association between NC and NR. Dynamic regression showed that high orders of lag terms had significant associations with ND, TD, NH, TH, and NR. In panel event study, curve lines showed key events influence ND, TD, NH, and TH significantly, while straight line showed key events had no significant influence on NR.
This study was in agreement with several studies of COVID-19 speed of geographical expansion. Another theoretical study mathematically demonstrated that the relatively high per-capita rate of transmission and the low rate of changes in behavior had caused a large-scale transmission of COVID-19 spatially [23]. Due to transmission between persons, the COVID-19 speed of geographical expansion overwhelmed the increase supply system of public health service in the national regions. This study was in line with another study which revealed that the effect of NC on the ND was heterogeneous across Provinces/States in China [24]. But, the finding in this study is not consistent with their results that an increase in NC by 1% increases ND by 0.10%-1.71% (95% CI). This was the same in USA and the world.
This study is in line with several theoretical studies. Traditional public health measures are effective in reducing peak incidence and global deaths of COVID-19 [25]. Therefore, combining emergency response with regular prevention and control measures could compete the tough and long battle against the COVID-19 epidemic.
This study is in line with several empirical studies. For example, an exploratory data analysis with visualizations has been made to understand the number of different cases reported (confirmed, death, and recovered) in different provinces of China and outside of China [26]. Regarding the results in the dynamic regression, this study is congruent with several theoretical studies. In particular, the lags of time series study could be explained by the estimated R0 was between 2.7 and 4.2 [27]. Higher orders of lags indicated that countries of the world should take a series of continuous and austere measures to contain the spread of COVID-19.
The explanations of ineffective policy invention could be over freedom of lifestyle. Facing covid-19 pandemic, a growing number of countries announce at war against further development of contagion. In-house isolation, quarantine, and promoting general awareness about transmission routes are the practical strategies used to tackle the spread of COVID-19 [28]. But, the common persons do not heeded government advice about the importance of individual behavioral attitudes to counter disease propagation. Consequently, the situation rapidly deteriorated with increasing number of cases that started to overwhelm health services [29], especially in western countries. The other causes to ineefective prevention and controlling COVID-19 were ill-prepared countries [30], facemask shortage [31], and poor traveller screening [32].

Strengths and weaknesses of the study
Regarding data sources, this study employed three data. The current study has a large sample size which increased the precision of the study. Additionally, the 60-day period could provide reliable results regarding epidemic control and daily changes in the prevalence of COVID-19 conditions. Regarding statistical methods, this study adopted three advanced panel regression methods. Especially, the event study with difference in difference is used to analyze the role of key events.
There are several limitations. First, several variables included demographics, financial support, and international aids were not taken into account. Statistically, a study in South Korea found that sex, region, and infection reasons affected on both the reported numbers of recovered and deceased cases [33]. Uncontrollable COVID-19 spread in India may be due to inadequately equipped and dedicated health facilities (e.g., space, infection control, waste disposal, safety of healthcare workers, partners to be involved in design and plan) to manage sick patients while protecting healthcare workers and the environment [34]. Second, the spread and transmission of COVID-19 were not considered mathematically and globally. Changes in case definitions affected inferences on the transmission dynamics of COVID-19 allow detection of more cases as knowledge increased in China [35].

Conclusions
Using panel analysis and data collected in China province-level units and