Impact of Nonpharmacological Public Health Interventions on Epidemiological Parameters of COVID-19 Pandemic in India

Background Public health interventions are epidemiologically sound and cost-effective methods to control disease burden. Non-pharmacological public health interventions are the only mode to control diseases in the absence of medication. Objective To find the impact of public health interventions on the epidemiological indicators of disease progression. Methods This is a secondary data analysis done on COVID-19 data. The median doubling time and R0 were calculated for a rolling period of seven days. Interventions were scored from zero to three with an increasing level of stringency. Multivariate linear regression was performed to find the role of individual interventions on R0 and the median doubling time. Results The highest intervention score was reported in the lockdown phase, which gradually decreased to the lowest level of 22. The R0 values settled to a level of 1.25, and the median doubling time increased to 20 days at the end of the study. Public awareness and public health laws were found to be related to both R0 and the median doubling time in the pre-lockdown phase only. Conclusion The implementation of interventions at the ground level is one of the key factors in the success of public health interventions. Post implementation, poor effectiveness of many interventions is evident from the study. Further, studies related to the sequence of interventions are required to further analyze the poor effect of the interventions.

The recent COVID-19 pandemic has activated the public health system of every country. Every country has started intervening the problem in the most innovative public health approaches, apart from clinical approaches. In the absence of appropriate treatment and vaccine for COVID-19, PHIs remain the only measure to tackle this pandemic. PHIs like wearing masks, physical distancing, and hand washing signi cantly controlled the pandemic in the past.
Since its rst noti cation in Wuhan province in China, the pandemic has spread to almost 214 countries and severely impacted human lives. (World Health organization, Situation Report, COVID 19) Many countries like New Zealand, Singapore, Sri Lanka, and Germany could contain the pandemic to a large extent with the help of PHIs. However, countries like the United States, China, Italy, and the United Kingdom could not contain the pandemic using the same interventions. In India, the COVID-19 pandemic began in late January 2020, approximately one month after the start of the pandemic in the world. (Govt. of India, 2020) The Government of India (GOI) accordingly took many steps to stop the progress of the pandemic in India.
The main objective of the study is to assess the nature and type of non-pharmacological PHIs in different phases of the COVID-19 pandemic in India and their impact on the various epidemiological indicators related to disease progression.

Study type:
This study is a secondary data analysis conducted on data collected from different sources reporting COVID-19 caseload and intervention as mentioned below.

Study duration:
The study was conducted from January 2020 to June 2020.

Methodology proper:
The required data regarding the COVID-19 pandemic was collected from the ministry of health and family welfare, govt. of India, press information bureau India. Interventions related to COVID-19 were collected from the ministry of home affairs, govt. of India website, ministry of health and family welfare, govt. of India, the press information bureau of India. Classi cation of time periods: The whole period of the epidemic was classi ed into seven phases according to the interventions and caseload. The detailed method of classi cation was mentioned in table-1.
Classi cation of interventions: The non-pharmacological PHI measures taken by the central government related to COVID-19 can be categorized to domains like those related to restriction in the o ce places, restriction in industry, agriculture and construction, restriction in the local transport, restriction in the interstate air transport, restriction in the out-migration, restriction in immigration, social distancing, closure of educational institutes, closure of hospitality sector, restriction of public gathering, health system preparedness, public awareness, and public health laws.
Intervention Scoring: Each intervention is scored from zero to three based on the strictness with which it was prescribed. The higher score indicates stricter intervention, whereas the lower score indicates lower intervention. The highest intervention score was thus 42 and the lowest is zero.
Analysis Proper: Various parameters like median doubling time, death rate, recovery rate, and R0 (basic reproduction number) related to disease were calculated. The median intervention score, R0, median doubling time, death rate, and recovery rate were calculated in each period. Multivariate linear regression was performed to nd the impact of each intervention on the summary measures of the COVID-19 pandemic, i.e. median doubling time and R0. For this purpose, we divided the whole study period into two categories, i.e. pre-lockdown period and lockdown & post lockdown period.
From the number of cases, the median doubling time and R0 were calculated using the following formula.
Median doubling time = ln (2)/(C1 /C2) Where C1 is the number of cases in a given day and C2 is the number of cases in the previous day. R0 = 1+(I * ln (2))/Td)where "I" is the infectious period and "Td" is the doubling time.
R0 and median doubling time was calculated using a sliding time period of seven days as mentioned by Li Q et al, 2020. Thus, and R0 was calculated from the caseload in the previous seven days.

Study de nitions:
1. COVID-19 Cases: The laboratory diagnosed cases as mentioned in the ministry of health and family welfare, govt. of India.
2. COVID-19 related Deaths: A death due to COVID-19 is de ned for surveillance purposes as a death resulting from a clinically compatible illness, in a probable or con rmed COVID-19 case, unless there is a clear alternative cause of death that cannot be related to COVID disease (e.g. trauma). There should be no period of complete recovery from COVID-19 between illness and death. This de nition was adopted from the Govt. of India guidelines.

Results
The pandemic was started in India approximately one month after the Wuhan outbreak. The number of cases remained stagnant for one month, i.e. in February 2020, before it began to rise in the rst week of March 2020. The number of COVID-19 cases along with the number of deaths rose steadily.
Epidemiological features of the pandemic: Overall, the pandemic recovery rate increased over time from 10.0% in the early week of March 2020 to 59.1% at the end of June. However, a sudden rise in the recovered number was observed in the last week of May 2020. The death rate, on the other hand, remained on the lower side, i.e. around 3.0%. The median doubling time of the pandemic was found to be increasing throughout the study period from 1.0 days at the start of the pandemic to almost 20 days at the end of June 2020. However, the R0 remained static at around 1.2 at the end of the study period after an initial rise in R0 observed in February and March 2020. The phase-wise representation of these summary epidemiological variables is provided in the Table-2. Interventions & Intervention Scores: India started its preparedness to combat the pandemic before the detection of the 1st case in the country. Initial preparation was related to the advisory related to the out migration from India followed by health system preparedness as surveillance. Thus, in initial phases of intervention started from controlling the source of infection followed by interventions related to the prevention of new infection and blocking the transmission. From 25th March 2020, India announced a complete lockdown followed by stepwise unlocking of the services and continue. The detailed interventions are provided in Figure-1.
Intervention scores were found to be increasing gradually till 25th March, to achieve the highest score of 42. After the aforementioned, there was a gradual decrease in the score in the later phases of lockdown to reach a score of 22 at the end of the study period. The detailed variation of the intervention score is given in Figure-1. Phase wise variation in the median intervention score is provided in table 2.
Relation Between Intervention Score and Epidemiological Features: Multivariate logistic regression, it was found that public awareness (p-value: 0.003) and public health laws (p-value: 0.000) were signi cantly associated with the median doubling time in the pre-lock down phase. Similarly, R0 was negatively associated with the public health laws (p-value: 0.000) and positively associated with public awareness (p-value: 0.000) in the pre-lock down phase. However, the immigration restrictions, out-migration restrictions, restrictions in a public gathering, and social distancing were found statistically not related to the R0 and median doubling time. (Table-3) However, in the lockdown & post lockdown era the restriction in industry, agriculture, etc. (p-value: 0.000), restrictions in local transport (pvalue: 0.000), closure of places of worship (p-value: 0.000), were found to be negatively associated with the median doubling time, and the same were found to be positively associated with the R0. Similarly, closure of the hospitality sector was found to be negatively associated (p-value: 0.000) with the R0. The rest all the factors were found to be not signi cantly associated with the median doubling time and R0. (Table-3)

Discussion
The study found that proactive intervention started in the early phase of the pandemic in the country. Initially, the interventions are linked to controlling the source of infection and gradually it shifted to blocking the transmission and prevention of new infections. The median doubling time gradually increased from 1.4 days in phase one to 18.6 days in phase seven. Similarly, the recovery rate increased gradually to 64.57% in phase seven. However, R0 decreased gradually after initial uctuations to a lower level of around 1.2. Death rate however remained at a lower level around 3.0% throughout the study period. Intervention scores gradually increased till the end of the march, however after this decreasing trend was observed.
The initial intervention was found to be limited to entering the infection in the country which was logically correct as the pandemic was started in the countries outside India. This was followed by the prevention of new infections and followed by blocking the transmission. The interventions that undertook by the govt. of India reached maximum stringency level in the latter part of March 2020, i.e. in the third phase of the study. Despite the interventions, the number of cases began to rise in phase four and later in the study period. This may be related to the poor stringency of the interventions or may be due to poor screening of the cases that may lead to continued transmission in the background. In the pre-lock down phase, the relation between public awareness and public health laws were found to be signi cantly associated with the median doubling time. Public awareness was found to be positively associated with R0. This may be due to the fact that the implementation of public awareness activity may not have transformed into action immediately. However, implementation of the public health laws was found to be negatively associated with the R0, i.e. with strict implementation of public health laws, the R0 value decreased. No relation was observed between the factors like social distancing, immigration, out migration found not statistically signi cant with the R0 and MDT. This may be attributed to the fact that the short period of action of these interventions in the pre lockdown era. In the lockdown and post-lockdown period, the interventions like the restriction of industry/ agriculture and construction, restriction of local transport were found to be negatively associated with MDT and positively associated with R0. A study by Pan et al.

Conclusions
Although the strict interventions were planned in the country, but the proper execution of the interventions may be a problem across the nation, that might have affected the spread of the COVID-19 pandemic in the country. Various interventions that have shown to be helpful in controlling the different parts of the world were found to be ineffective in India. Thus, further research is required pertaining to the intervention implementation strategies in epidemic situations, human behavioural pattern related to the pandemic situation may be required to further analyse the poor effectiveness of interventions as depicted by the current study.

Declarations
Compliance with Ethical Standards:   Figure 1 Non-Pharmacological Public Health Interventions with interventions cores and its impact on the R0, Median Doubling Time