https://orcid.org/0000-0003-2881-9723
Figures
Abstract
Schools were closed all over Pakistan on November 26, 2020 to reduce community transmission of COVID-19 and reopened between January 18 and February 1, 2021. However, these closures were associated with significant economic and social costs, prompting a review of effectiveness of school closures to reduce the spread of COVID-19 infections in a developing country like Pakistan. A single-group interrupted time series analysis (ITSA) was used to measure the impact of school closures, as well as reopening schools, on daily new COVID-19 cases in 6 major cities across Pakistan: Lahore, Karachi, Islamabad, Quetta, Peshawar, and Muzaffarabad. However, any benefits were contingent on continued closure of schools, as cases bounced back once schools reopened. School closures are associated with a clear and statistically significant reduction in COVID-19 cases by 0.07 to 0.63 cases per 100,000 population, while reopening schools is associated with a statistically significant increase. Lahore is an exception to the effect of school closures, but it too saw an increase in COVID-19 cases after schools reopened in early 2021. We show that closing schools was a viable policy option, especially before vaccines became available. However, its social and economic costs must also be considered.
Citation: Mueed A, Ahmad T, Abdullah M, Sultan F, Khan AA (2022) Impact of school closures and reopening on COVID-19 caseload in 6 cities of Pakistan: An Interrupted Time Series Analysis. PLOS Glob Public Health 2(9): e0000648. https://doi.org/10.1371/journal.pgph.0000648
Editor: Xerxes Tesoro Seposo, Nagasaki University: Nagasaki Daigaku, JAPAN
Received: May 24, 2022; Accepted: August 24, 2022; Published: September 19, 2022
Copyright: © 2022 Mueed et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The data was provided to the Akhter Hameed Khan Foundation team for this study as part of its work with Pakistan's Federal Ministry of National Health Services, Regulations & Coordination (MoNHSR&C) and the National Command & Operation Centre (NCOC) in Islamabad, which lead Pakistan's response to the COVID-19 pandemic. The AHKRC team has provided analytical support to the above entities, and such created knowledge that has directly informed pandemic policy-making in Pakistan. COVID-19 data is compiled and shared in daily National Situation Reports, or Sitreps, by the National Emergency Operation Centre (NEOC). Each day's Sitrep is compiled as a PDF file. The data used for this study was manually compiled from these PDF files and then used in STATA. The parentage of this data is with the NCOC and the MoNHSR&C. The AHKRC team received this data with the express understanding that it would be kept confidential. However, the data can be obtained independently from the NEOC, through a data request procedure, which is subject to approval from the MoNHSR&C. The data request form as well as the standard operating procedures for a data request are provided as individual documents in the attachments for this submission. The data request itself is to be addressed to: Dr. Shahzad Baig, National Coordinator, National Emergency & Operation Center, D Block, EPI Building, Chak Shahzad, Park Road, Islamabad. Email: eocpakistan@gmail.com Phone: +92-51-8730879.
Funding: This work was supported, in whole or in part, by the Bill & Melinda Gates Foundation [grant number: INV-025171]. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Since the beginning of the global spread of COVID-19 and before effective vaccines became available, non-pharmaceutical interventions (NPIs), either barriers or means to limit contact between individuals, were the mainstay to control the spread of COVID-19. Perhaps the most widely debated among these NPIs was the closure of schools, which drew criticism for the significant social, learning, economic [1, 2], and physical and mental health costs [3–6] associated with them. Notably, these costs are disproportionately borne by already disadvantaged families [1, 7], thereby exacerbating social and economic inequalities [8].
Prior studies suggest that children infected with COVID-19 are often asymptomatic or have mild symptoms identical to other common respiratory infections [9, 10], and yet they can transmit the infection even when they feel well. Children have also been key spreaders in other respiratory infections such as influenza, because of prolonged contact in close proximities with other children at schools [11].
Early evidence on the effect of school closures on epidemic transmission of COVID-19 seemed mixed. Initial, and often modeling-based studies, suggested that closing schools may not help reduce COVID-19 transmission in communities [12–15]. However, more recent, and more empirically based studies have tended to show a role for school closure in reducing cases in the community [16–23]. In low- and middle-income countries such as Pakistan where learning is already inadequate and remote learning solutions are all too often unavailable for most students [24], it is paramount that such a social policy be used only if absolutely supported by evidence of a benefit in limiting COVID-19 transmission and then too, only as a means of last resort. We explore the changes in daily cases on COVID-19 pre and post school closures in Pakistan using a single-group Interrupted Time Series Analysis (ITSA).
This paper is a continuation of our earlier work, which examined the effects of school closures on the daily cases of COVID-19 in Islamabad vs. Peshawar, during the same period as in this study [25]. However, this study attempts to examine the effect of school closures with a different methodology, and also with a larger sample of cities.
Methods
In this paper we conduct a pre- and post-school closures and reopening analysis of changes in the daily incidence of COVID-19 cases (per 100,000 population) in 6 cities of Pakistan: Lahore, Karachi, Islamabad, Quetta, Peshawar, and Muzaffarabad using a single-group ITSA. These cities are provincial capitals except Islamabad which is the federal capital of Pakistan. We choose these cities for the analysis as these are the largest cities of relevant provinces in Pakistan and much of the COVID-19 caseload was concentrated in these cities during the analysis period. To estimate treatment effects of school closures and reopening, we use a single-group ITSA because it is a quasi-experimental tool that is particularly useful when data cannot be fully randomized, there is no comparison group, and there is a need to consider the effect of only one intervention.
This suits our study as, in Pakistan, all non-school NPIs were enacted in groups–except for the closure of schools. For example, marriage hall restrictions and ban on large scale gatherings were notified at the same time, as were mask-wearing, broader “smart” lockdowns (lockdowns in parts of cities), and reduced market timings. Mask-wearing and social distancing were constant across time with similar compliance across the country. Two major non-school NPIs that were implemented in these cities during our period of study were marriage hall restrictions and smart lockdowns (see S1 Table). These NPIs targeted only a small proportion of the population and hence their impact was assumed to be limited. School closures, on the other hand, were universally enforced and applied to all schools–public or private, day or boarding–and to students of all grades across Pakistan [26].
Data for this analysis was sourced from the daily National Situation Reports (Sitreps) published by the National Emergency Operations Centre (NEOC) in Islamabad, Pakistan. It is the only Ministry of Health’s official COVID-19 data that was being analyzed and used for Pakistan’s pandemic response, nationally and internationally. This data is anonymized and aggregated by city, with no disaggregation by age, gender, ethnicity, or any other potentially identifying characteristic. We use this data for an inferential analysis of the change in daily COVID-19 incidence in the overall populations of the 6 aforementioned cities, regardless of demographic characteristics, due to the change in one particular NPI. It is because this NPI is the only policy intervention that could be isolated in our chosen time period of observation. Populations of cities were derived from Population Census 2017 of Pakistan, which is publicly available [27].
We estimated 2 sets of ordinary least square (OLS) regressions for each city using a 10- or 20-day delay since COVID-19 incidence changes from school-related NPIs take effect 10 [28] or more days [17, 29, 30] after closures or reopening. Daily new COVID-19 cases (per 100,000 population) were taken for equally spaced time frames with 10- and 20-days delay after the actual school closures and reopening dates. In order to analyze the effect of school closures and reopening, we took a total of 60 days for pre- and post-intervention periods. Table 1 shows the dates of school closures and reopening. For 10-days delay models of school closures and reopening, we took the following dates from November 06, 2020 to January 04, 2021 and January 12, 2021 to March 12, 2021, respectively. For 20-days delay models of school closures and reopening, we took the time periods from November 16, 2020 to January 14, 2021 and January 22, 2021 to March 22, 2021 respectively. We chose a 60-days period because a sample size of at least 50 time points is recommended by Box and Jenkens [31], as larger sample size increases power in the case of segmented time series [32]. We did not take a much longer time period since schools had started to reopen after one-month post-closures period in our 20-days delay model [33].
Model specification
Our model specification is adapted from Linden and detailed below [34]:
Where
- Yti, our outcome variable, is the daily number of new COVID-19 cases (per 100,000 population) in city i;
- β0, the constant term, is the starting level of the daily new COVID-19 cases (per 100,000 population) in city i;
- Tti is the time period since the beginning of this study, and the coefficient β1 shows the slope of daily new COVID-19 cases (per 100,000 population) until the start of the intervention for city i;
- Xti is a dummy variable indicating the intervention period (post intervention = 1, and 0 otherwise) for city i;
- β2 explains the change in daily COVID-19 cases (per 100,000 population) that occurs in the time period immediately followed by the school closure/reopening (our interventions) in city i;
- XtiTti is the interaction term between the intervention period and the time since the start of the study; and,
- β3 represents the difference between the pre- and post-intervention slopes for daily new COVID-19 cases (per 100,000 population) in city i.
We run the model specified above for each city i separately. Since we want to estimate city-level effects as compared to a national or combined effect, we do not opt for a panel data model. To get a singular, direct estimate of the effect of closing/reopening schools, we used the lincom estimate which is the sum of β1 and β3 [35].
This generates a separate variable that sums the values of β1 and β3.
To adjust for autocorrelation and possible heteroskedasticity, we used Newey-West standard errors in our regression models [36]. Cumby-Huizinga test for autocorrelation was performed on each regression model to identify correct lag structure. Linktest was applied to check if the models were correctly specified. Stata 16 software package was used for the analysis.
Results
Our descriptive results showing means and standard deviations of daily COVID-19 cases (per 100,000 population) are presented in Table 2. These were calculated for school closures and reopening periods of each city, separately for pre- and post-intervention periods. For 10-day delay, Islamabad showed highest cases (per 100,000 population) (18.80, SD: 4.411), while Quetta showed the lowest (0.722, SD: 0.294) in the pre-intervention period of school closures; same trend followed in school reopening pre-intervention period. Islamabad and Quetta also had the highest and the lowest cases (per 100,000 population) in pre-intervention periods of 20-day delay school closures and reopening.
After adding a 10-day delay after the actual date of school closures (Table 3), the rate of change in daily COVID-19 cases declined following closure of schools in Karachi, Islamabad, Quetta, and Peshawar; the reductions per 100,000 population were by -0.16 cases (95% CI: -0.23, -0.13) in Karachi, -0.41 cases (95% CI: -0.53, -0.30) in Islamabad, -0.01 cases (95% CI: -0.01, -0.00) in Quetta, and -0.06 cases (95% CI: -0.08, -0.03) in Peshawar. In Lahore, daily COVID-19 cases continued to rise at a rate of 0.03 cases (95% CI: 0.01, 0.05) per 100,000 population after the closure of schools. For Muzaffarabad, the rate of change of COVID-19 cases was declining both before and after the school closure, at -0.11 cases (95% CI: -0.16, -0.06) and -0.05 cases (95% CI: -0.06, -0.03) per 100,000 population, respectively.
The opposite trend was seen following schools reopening in early 2021. Before schools reopened, accounting for a 10-day delay from the actual date of reopening, the rate of change of daily COVID-19 cases per 100,000 population was declining in every city by: -0.03 cases (95% CI: -0.05, -0.02) in Lahore, -0.20 cases (95% CI: -0.24, -0.16) in Karachi, -0.08 cases (95% CI: 0.12, -0.04) in Islamabad, -0.01 cases (95% CI: -0.01, -0.00) in Quetta, and by -0.04 cases (95% CI: -0.06, -0.04) in Peshawar. Muzaffarabad’s pre-reopening trend is not statistically significant at 95% confidence level.
After schools reopened, the rate of change of daily COVID-19 cases became positive in every city–except in Karachi, where the rate remained negative at -0.03 cases (95% CI: -0.04, -0.02) per 100,000 population. For the remaining cities, daily new COVID-19 cases began to increase at a rate of 0.09 cases (95% CI: 0.06, 0.11) per 100,000 population in Lahore, 0.34 cases (95% CI: 0.18, 0.50) in Islamabad, 0.01 cases (95% CI: 0.00, 0.01) in Quetta, 0.02 cases (95% CI: 0.01, 0.04) in Peshawar, and 0.09 cases (95% CI: 0.04, 0.16) in Muzaffarabad.
These effects were similar but more modest when allowing for a 20-day delay after the actual date of school closures. The post-closure trend remained statistically significant in Lahore, where the rate of change in cases continued to rise at 0.01 cases (95% CI: 0.00, 0.02) per 100,000 population, while Islamabad’s post-closure trend showed a decline at a rate of -0.12 (95% CI: -0.21, -0.03). Post-reopening, in Islamabad, the rate of daily new COVID-19 cases changed from 0.05 cases (95% CI: 0.00, 0.09) per 100,000 to 0.98 cases (95% CI: 0.69, 1.26) per 100,000 population while in Muzaffarabad, rate of change went from 0.01 cases (95% CI: 0.00, 0.01) before schools were reopened to 0.09 cases (95% CI: 0.04, 0.16) once schools opened. These results are presented graphically in Figs 1–4 below.
Discussion
We show that school closures are associated with fewer daily new COVID-19 cases compared to pre-closures by 5 to 62 actual daily cases in individual cities. Correspondingly, reopening schools appear to increase them by 1 to 35 daily cases. The association is the strongest in Karachi, Lahore, Peshawar, and Islamabad [37], which are larger, denser cities, and had the most cases; while the association was modest for sparse and smaller cities of Quetta and Muzaffarabad, that also had fewer overall cases.
Pakistan saw a much lower reduction in COVID-19 cases with school closures than was seen in many other countries. For example, our reduction of 5 to 62 daily cases is considerably smaller than the reduction of 424 cases per 100,000 population seen in the USA [17]. However, this is consistent with the fact that cities with fewer cases had the least reductions in cases when schools were closed. Pakistan has also seen much fewer cases, hospitalizations, and deaths from COVID-19 than in Europe or North America, perhaps relating to its sparser social networks leading to fewer contacts among individuals within the community [38], or perhaps higher levels of nonspecific immunity from prior infections with disparate viruses [39]. However, both of these possible explanations are speculative at the moment.
Reductions in cases with school closure is better understood in the context of the education set up in Pakistan. Most school-going children in cities go to low-cost-private schools where they sit in small classrooms with little space for social distancing [40]. Additionally, most children commute to and from schools in small vehicles–up to 15–17 children in the back of a minivan. Children then come into contact with adults at school (teachers and custodian staff), then at home (parents and elder family members), and in so doing become a conduit for COVID-19 spread outside of schools. As 31% of total population of Pakistan falls in the school going age [41], schools in Pakistan then essentially function as “super spreader” locations for COVID-19.
Our findings are consistent with the global evidence, as well as the results of our own previous work [25], that school closures are associated with reduction in COVID-19 transmission in communities [17, 30]. In the US, school closures were associated with reduced COVID-19 caseloads [15], deaths [19, 42, 43], and hospitalizations by as much as half [19]. Similarly, the timely closure of schools and high education institutes were found to lower COVID-19 transmission rates in the European Union and other developed countries [44–47]. Earlier in the epidemic, a number of modeling studies had predicted more modest effects of such closures [19, 21, 24, 44–46]. However, more recent studies using empirical community transmission data have generally shown a more robust association between school closures and reductions in community cases of COVID-19 [21].
Limitations
There are limitations of this analysis. The daily COVID-19 data are aggregated nationally, regionally, and by certain major cities, with no disaggregation by age or gender. Additionally, we acknowledge that a pre- and post-intervention analysis itself has limitations. For example, it may not effectively discern the effects of an intervention from that of a long-term trend on an outcome variable. This is referred to as the “maturation” threat to the internal validity of a pre- and post-intervention analysis. However, this has negligible impact on our analysis, as we consider a total of 60 days for each ITSA regression, in each of the 6 cities, when examining the effects of the intervention after accounting for sufficient delays–at 10 and 20 days–to be sure of the effects of the interventions.
It is difficult to explain why Lahore did not show any reduction in cases after school closures. It is possible that the epidemic affected cities at different points in time and that it was at a relatively lower level in Lahore during the study period. We also acknowledge that there could be potential cross-contamination of COVID-19 cases between Islamabad and Peshawar, which are separated by a 2-hours commute by road, and between Islamabad and Lahore, which are 4-hours apart by road. However, there are no data on the magnitude of any potential contamination due to bilateral intra-city travel. Nevertheless, were there significant contamination between the cities, one would have expected to see convergence in COVID-19 caseloads between them, and there is no evidence that this occurred. Finally, measurement of the serious social, economic, and educational attainment costs from school closures was beyond the scope of our study.
Conclusions
School closures may be associated with lower transmission of COVID-19 in communities and such closures are an important policy tool to stop the spread of COVID-19. However, their social and economic costs are high, perhaps more so in a developing country. The balance of these costs and benefits must inform this effective NPI specially when other measures, including vaccines, are being planned.
Supporting information
S1 Table. Non-school closures non-pharmaceutical interventions (NPIs) during study period (November 6, 2020 to March 22, 2021).
https://doi.org/10.1371/journal.pgph.0000648.s001
(PDF)
Acknowledgments
We thank Testing, Tracing and Quarantining (TTQ) team at the National Command and Operation Centre (NCOC) and the National Emergency Operations Centre (NEOC) for facilitating our work.
References
- 1. Hanushek EA, Woessmann L. The economic impacts of learning losses. 2020. https://doi.org/10.1787/21908d74-en.
- 2. Sadique MZ, Adams EJ, Edmunds WJ. Estimating the costs of school closure for mitigating an influenza pandemic. BMC Public Health. 2008;8:135. Epub 2008/04/26. pmid:18435855; PubMed Central PMCID: PMC2377259.
- 3. Kinsey EW, Hecht AA, Dunn CG, Levi R, Read MA, Smith C, et al. School Closures During COVID-19: Opportunities for Innovation in Meal Service. Am J Public Health. 2020;110(11):1635–43. Epub 20200917. pmid:32941069; PubMed Central PMCID: PMC7542295.
- 4. Kishida K, Tsuda M, Waite P, Creswell C, Ishikawa SI. Relationships between local school closures due to the COVID-19 and mental health problems of children, adolescents, and parents in Japan. Psychiatry Res. 2021;306:114276. Epub 20211108. pmid:34798486; PubMed Central PMCID: PMC8585496.
- 5. Amran MS, Jamaludin KA. The Impact of Unplanned School Closures on Adolescent Behavioral Health During the Covid-19 Pandemic in Malaysia. Front Public Health. 2021;9:639041. Epub 20210607. pmid:34164364; PubMed Central PMCID: PMC8215535.
- 6. Tang S, Xiang M, Cheung T, Xiang YT. Mental health and its correlates among children and adolescents during COVID-19 school closure: The importance of parent-child discussion. J Affect Disord. 2021;279:353–60. Epub 2020/10/26. pmid:33099049; PubMed Central PMCID: PMC7550131.
- 7. Andrew A, Cattan S, Costa Dias M, Farquharson C, Kraftman L, Krutikova S, et al. Inequalities in Children’s Experiences of Home Learning during the COVID-19 Lockdown in England*. Fiscal Studies. 2020;41(3):653–83. pmid:33362314
- 8. Armitage R, Nellums LB. Considering inequalities in the school closure response to COVID-19. Lancet Glob Health. 2020;8(5):e644. Epub 2020/03/31. pmid:32222161; PubMed Central PMCID: PMC7195275.
- 9. Wong JJM, Abbas Q, Chuah SL, Malisie RF, Pon KM, Katsuta T, et al. Comparative Analysis of Pediatric COVID-19 Infection in Southeast Asia, South Asia, Japan, and China. Am J Trop Med Hyg. 2021;105(2):413–20. Epub 20210615. pmid:34129517; PubMed Central PMCID: PMC8437183.
- 10. Musa OAH, Chivese T, Bansal D, Abdulmajeed J, Ameen O, Islam N, et al. Prevalence and determinants of symptomatic COVID-19 infection among children and adolescents in Qatar: a cross-sectional analysis of 11 445 individuals. Epidemiol Infect. 2021;149:e203. Epub 20210914. pmid:34517936; PubMed Central PMCID: PMC8446592.
- 11. Goldstein E, Nguyen HH, Liu P, Viboud C, Steiner CA, Worby CJ, et al. On the Relative Role of Different Age Groups During Epidemics Associated With Respiratory Syncytial Virus. The Journal of infectious diseases. 2018;217(2):238–44. pmid:29112722; PubMed Central PMCID: PMC5853559.
- 12. Fukumoto K, McClean CT, Nakagawa K. No causal effect of school closures in Japan on the spread of COVID-19 in spring 2020. Nat Med. 2021;27(12):2111–9. Epub 20211027. pmid:34707318; PubMed Central PMCID: PMC8674136.
- 13. Isphording IE, Lipfert M, Pestel N. Does re-opening schools contribute to the spread of SARS-CoV-2? Evidence from staggered summer breaks in Germany. Journal of Public Economics. 2021;198:104426. https://doi.org/10.1016/j.jpubeco.2021.104426.
- 14. Iwata K, Doi A, Miyakoshi C. Was school closure effective in mitigating coronavirus disease 2019 (COVID-19)? Time series analysis using Bayesian inference. Int J Infect Dis. 2020;99:57–61. Epub 2020/08/04. pmid:32745628; PubMed Central PMCID: PMC7836901.
- 15. Davies NG, Kucharski AJ, Eggo RM, Gimma A, Edmunds WJ, Centre for the Mathematical Modelling of Infectious Diseases C-wg. Effects of non-pharmaceutical interventions on COVID-19 cases, deaths, and demand for hospital services in the UK: a modelling study. Lancet Public Health. 2020;5(7):e375–e85. Epub 20200602. pmid:32502389; PubMed Central PMCID: PMC7266572.
- 16. Matzinger P, Skinner J. Strong impact of closing schools, closing bars and wearing masks during the Covid-19 pandemic: results from a simple and revealing analysis. medRxiv. 2020. Epub 20200928. pmid:33024976; PubMed Central PMCID: PMC7536875.
- 17. Auger KA, Shah SS, Richardson T, Hartley D, Hall M, Warniment A, et al. Association Between Statewide School Closure and COVID-19 Incidence and Mortality in the US. JAMA. 2020;324(9):859–70. Epub 2020/08/04. pmid:32745200; PubMed Central PMCID: PMC7391181.
- 18. Banholzer N, van Weenen E, Lison A, Cenedese A, Seeliger A, Kratzwald B, et al. Estimating the effects of non-pharmaceutical interventions on the number of new infections with COVID-19 during the first epidemic wave. PloS one. 2021;16(6):e0252827. Epub 20210602. pmid:34077448; PubMed Central PMCID: PMC8171941.
- 19. Brauner JM, Mindermann S, Sharma M, Johnston D, Salvatier J, Gavenciak T, et al. Inferring the effectiveness of government interventions against COVID-19. Science. 2021;371(6531). Epub 20201215. pmid:33323424; PubMed Central PMCID: PMC7877495.
- 20. Li Y, Campbell H, Kulkarni D, Harpur A, Nundy M, Wang X, et al. The temporal association of introducing and lifting non-pharmaceutical interventions with the time-varying reproduction number (R) of SARS-CoV-2: a modelling study across 131 countries. The Lancet Infectious diseases. 2021;21(2):193–202. Epub 20201022. pmid:33729915; PubMed Central PMCID: PMC7581351.
- 21. Stage HB, Shingleton J, Ghosh S, Scarabel F, Pellis L, Finnie T. Shut and re-open: the role of schools in the spread of COVID-19 in Europe. Philosophical transactions of the Royal Society of London Series B, Biological sciences. 2021;376(1829):20200277. Epub 20210531. pmid:34053270; PubMed Central PMCID: PMC8165592.
- 22. Davies NG, Klepac P, Liu Y, Prem K, Jit M, group CC-w, et al. Age-dependent effects in the transmission and control of COVID-19 epidemics. Nat Med. 2020;26(8):1205–11. Epub 20200616. pmid:32546824.
- 23. Liu Y, Morgenstern C, Kelly J, Lowe R, Group CC-W, Jit M. The impact of non-pharmaceutical interventions on SARS-CoV-2 transmission across 130 countries and territories. BMC Med. 2021;19(1):40. Epub 20210205. pmid:33541353; PubMed Central PMCID: PMC7861967.
- 24. Alban Conto C, Akseer S, Dreesen T, Kamei A, Mizunoya S, Rigole A. Potential effects of COVID-19 school closures on foundational skills and Country responses for mitigating learning loss. Int J Educ Dev. 2021;87:102434. Epub 20211011. pmid:34658500; PubMed Central PMCID: PMC8504478.
- 25. Mueed A, Aliani R, Abdullah M, Kazmi T, Sultan F, Khan A. School closures help reduce the spread of COVID-19: A pre- and post-intervention analysis in Pakistan. PLOS Global Public Health. 2022;2(4):e0000266.
- 26.
All education institutions to close down from Nov 26 as Covid-19 positivity rises. Dawn. 2020.
- 27.
PBS. Final Results of Census-2017: PBS; [cited 2022]. Available from: https://www.pbs.gov.pk/content/final-results-census-2017-0.
- 28. Nader IW, Zeilinger EL, Jomar D, Zauchner C. Onset of effects of non-pharmaceutical interventions on COVID-19 infection rates in 176 countries. BMC public health. 2021;21(1):1472. pmid:34320982
- 29. Ingelbeen B, Peckeu L, Laga M, Hendrix I, Neven I, van der Sande MA, et al. Reducing contacts to stop SARS-CoV-2 transmission during the second pandemic wave in Brussels, Belgium, August to November 2020. Euro surveillance: bulletin Europeen sur les maladies transmissibles = European communicable disease bulletin. 2021;26(7). pmid:33602386; PubMed Central PMCID: PMC7897911.
- 30. Stein-Zamir C, Abramson N, Shoob H, Libal E, Bitan M, Cardash T, et al. A large COVID-19 outbreak in a high school 10 days after schools’ reopening, Israel, May 2020. Euro surveillance: bulletin Europeen sur les maladies transmissibles = European communicable disease bulletin. 2020;25(29). pmid:32720636; PubMed Central PMCID: PMC7384285.
- 31.
Box GEP, Jenkins GM, Reinsel GC. Time Series Analysis: Forecasting and Control. 4 ed. Hoboken, NJ: John Wiley & Sons, Inc.; 2008.
- 32. Zhang F, Wagner AK, Ross-Degnan D. Simulation-based power calculation for designing interrupted time series analyses of health policy interventions. J Clin Epidemiol. 2011;64(11):1252–61. pmid:21640554.
- 33.
Dawn. Classes 9–12 to reopen from Jan 18 as planned, says Shafqat Mahmood [Internet]. Dawn; 2021 [cited 2022 July 15]. Available from: https://www.dawn.com/news/1601645.
- 34. Linden A. Conducting Interrupted Time-series Analysis for Single- and Multiple-group Comparisons. The Stata Journal. 2015;15(2):480–500.
- 35. Linden A. A Comprehensive set of Postestimation Measures to Enrich Interrupted Time-series Analysis. The Stata Journal. 2017;17(1):73–88.
- 36. Newey WK, West KD. A Simple, Positive Semi-Definite, Heteroskedasticity and Autocorrelation Consistent Covariance Matrix. Econometrica. 1987;55(3):703–8.
- 37.
Statistics PBo. 2017 Census—Area, Population by sex, sex ratio, population density, urban proportion, household size and annual growth rate (Sindh). Islamabad, Pakistan: Pakistan Bureau of Statistics, 2021.
- 38.
North DC, Wallis JJ, Wieingast BR. Violence and Social Orders: A Conceptual Framework for Interpreting Recorded Human History. Cambridge: Cambridge University Press; 2009. 346 p.
- 39. Abela IA, Pasin C, Schwarzmüller M, Epp S, Sickmann ME, Schanz MM, et al. Multifactorial seroprofiling dissects the contribution of pre-existing human coronaviruses responses to SARS-CoV-2 immunity. Nature Communications. 2021;12(1):6703. pmid:34795285
- 40.
Shah D, Amin N, Kakli MB, Piracha ZF, Zia MA. Pakistan Education Statistics 2016–17. National Education Management Information System (NEMIS). Academy of Educational Planning and Management (AEPAM), 2018.
- 41.
Table 4—Population by single year age, sex and rural/urban. Islamabad, Pakistan: PBS; 2018.
- 42. Rauscher E. Lower State COVID-19 Deaths and Cases with Earlier School Closure in the U.S. medRxiv. 2020:2020.05.09.20096594.
- 43. Yehya N, Venkataramani A, Harhay MO. Statewide Interventions and Coronavirus Disease 2019 Mortality in the United States: An Observational Study. Clinical infectious diseases: an official publication of the Infectious Diseases Society of America. 2021;73(7):e1863–e9. pmid:32634828; PubMed Central PMCID: PMC7454446.
- 44. Klimek-Tulwin M, Tulwin T. Early school closures can reduce the first-wave of the COVID-19 pandemic development. Z Gesundh Wiss. 2020:1–7. Epub 20201015. pmid:33078090; PubMed Central PMCID: PMC7557316.
- 45. Piovani D, Christodoulou MN, Hadjidemetriou A, Pantavou K, Zaza P, Bagos PG, et al. Effect of early application of social distancing interventions on COVID-19 mortality over the first pandemic wave: An analysis of longitudinal data from 37 countries. J Infect. 2021;82(1):133–42. Epub 20201201. pmid:33275956; PubMed Central PMCID: PMC7706420.
- 46. Garchitorena A, Gruson H, Cazelles B, Karki T, Sudre B, Roche B. Integrated packages of non-pharmaceutical interventions increased public health response efficiency against COVID-19 during the first European wave: evidence from 32 European countries. medRxiv. 2020:2020.08.17.20174821.
- 47. Chang SL, Harding N, Zachreson C, Cliff OM, Prokopenko M. Modelling transmission and control of the COVID-19 pandemic in Australia. Nat Commun. 2020;11(1):5710. Epub 20201111. pmid:33177507; PubMed Central PMCID: PMC7659014.