Page 68 - GHES-2-4

Global Health Economics and
Sustainability
Assessing Vietnam’s pandemic lockdown

pandemic. Most research indicates that Vietnam’s dummy variable indicating the pre-intervention period
government successfully managed COVID-19 throughout (coded as 0) and the post-intervention period (coded as 1);
the first three pandemic waves by responding swiftly to and Z represents the number of days since the intervention
the virus. Tran et al. (2020) synthesized and compared was implemented.
Vietnam’s results with those of nine other Southeast Asian For example, the dataset prepared for the ITS model of
nations using the Oxford COVID-19 Government Response Lang Son province, with one intervention on May 29, 2021,
Tracker dataset (Hale et al., 2021). Le et al. (2021) evaluated is described in Table 2.
Vietnam’s policy responses to the COVID-19 pandemic by
synthesizing and assessing 959 pertinent policy documents In Equation I, β represents the baseline level at t = 0;
0
across various categories, finding that Vietnam’s policy β and β represent the change in the level of the outcome
2
1
system responded quickly, proactively, and successfully at for each day in the pre- and post-intervention periods,
multiple levels of authority. Nguyen et al. (2021) utilized a respectively, whereas β is a primary parameter of interest,
3
case study approach to highlight the importance of early representing the difference in the slope of the post-
technical preparedness and solid political commitment in intervention period.
ensuring an effective COVID-19 response in Vietnam. The β coefficient is considered to be positive, as the
0
However, none of the previous Vietnam-based research number of cases typically grows before the intervention
applied quantitative methods, and thus, no rigorous occurs. Since local governments generally issue interventions
analysis has been performed. Therefore, to address this when infection cases are rising, the β coefficient is also
2
gap, the present study employs the ITS analysis approach to expected to be positive. Based on previous studies (Islam
examine the effectiveness of lockdown policy intervention et al., 2020; Silva et al., 2020; Thayer et al., 2021; Tobías, 2020),
in Vietnam. Furthermore, unlike previous research, which a negative trend in new infection cases is expected following
focused on the first three waves, this study examines the the policy intervention, so β is assumed to be negative.
3
fourth wave of the pandemic (from April 27, 2021, to During the study period, local administrations typically
October 1, 2021), during which the number of COVID-19 implemented multiple interventions rather than just one.
cases grew exponentially compared to earlier periods. Therefore, to estimate the effects of each intervention
The remainder of this article is organized as follows: simultaneously, this study expands Equation I to employ
Section 2 explains the analytical framework, data, and the following regression model:
models used in this study. Section 3 discusses the empirical Y = β +β T+(β X +β Z )+ (β X +β Z )+...+(β X +β Z )
results. Section 4 concludes the study and provides policy t 0 1 x1 1 z1 1 x2 2 z2 2 xk k zk k (II)
implications.
where k is the number of interventions; X is the dummy
J
2. Methods variable, taking the value 0 for the pre-intervention period
and 1 for the post-intervention period of intervention
2.1. ITS models j. The Z variable takes either a value of 0 before the
J
ITS models are increasingly employed to evaluate public implementation of intervention j or the number of
health interventions, particularly well-suited to those
implemented at a population level over a clearly defined Table 2. Example of a dataset for interrupted time series
period (Bernal et al., 2017; Wagner et al., 2002). In addition, model of Lang Son province with a study period from April
some studies consider ITS design to be the most reliable 27, 2021, to August 10, 2021
quasi-experimental approach for evaluating the effectiveness t Y T X Z
of interventions in time series data (Cook et al., 2002). The 2021-04-27 0 1 0 0
main advantage of the ITS approach is its intuitive and
graphical illustration, allowing for a clear comparison of the 2021-04-28 0 2 0 0
effect of an intervention by visualizing the distribution of the 2021-04-29 0 3 0 0
outcome variable before and after the intervention. 2021-05-28 10 33 0 0
An ITS regression model with a single intervention can 2021-05-29 a 7 34 1 1
be explicitly expressed by the following equation: 2021-08-08 2 104 1 71
2021-08-09 0 105 1 72
Y= β + β T+β X+β Z (I) 2021-08-10 0 106 1 73
2
t
t
0
1
3
where Y represents the number of new infection cases Notes: Y represents the number of new infection cases which is
t
at time t; T represents the number of days since the start collected from the Ministry of Health’s data gateway. X and Z values
a
of the study period (i.e., April 27, 2021); X is a binary are derived from the intervention date of May 29, 2021
t
Volume 2 Issue 4 (2024) 4 https://doi.org/10.36922/ghes.3423