Materials and methods
We conducted a single-center cross-sectional study among the attendees of an outpatient COVID-19 clinic at a tertiary care hospital in New Delhi, India. This clinic is run for providing healthcare services to the workers who are under institutional health schemes. The institute ethics committee approved the study protocol, and appropriate consent was taken from the participants before their enrolment.
Between June 17 to July 1, 2020, 1,066 patients were tested for SARS-CoV-2 by reverse transcription-polymerase chain reaction (RT-PCR) in the screening clinic. Three hundred and eighty-four patients were included in the study based on successful telephonic communication and informed consent availability. Five patients were excluded from the study due to the non-availability of their test results due to pre-analytical issues. We analyzed socio-demographic and infection prevention practices related to data collected from 379 patients. Each patient was telephonically interviewed in the local language by a single investigator, and the answers were documented in a preformed questionnaire proforma. To minimize bias, the telephonic communications were completed before their test results were generated.
Study participants were tested for COVID-19 after a thorough clinical examination. Real-time RT-PCR for SARS-CoV-2 was performed for the diagnosis of COVID-19 in all patients included in our study. Testing indications were based on the national regulatory authority's advisory, the Indian Council of Medical Research (ICMR) [4]. One nasal and one throat swab were collected from each patient. Two swabs were put into a single vial of the viral transport medium (VTM) and were sent to the testing laboratory maintaining the cold chain. The test reports were accessed from the hospital information system.
A questionnaire was developed to collect socio-economic, demographic, behavioral, and infection prevention practices information from the enrolled participants (Table 1). Socio-economic status was documented as per the modified Kuppuswamy index [5]. More than three individuals residing in one room were considered as overcrowding [6]. In India, ICMR, the national regulatory authority, has recommended prophylaxis with hydroxychloroquine for workers who are at occupational risk of acquiring this infection [7]. Therefore, the healthcare workers were enquired about the use of hydroxychloroquine for prophylaxis against COVID-19. The prophylactic use of other medications (alternative medicines) was also recorded from the study population.
Table 1  Risk factors and preventive practices enquired from the participants
PPE - personal protective equipment
Parameters evaluated  Questions asked
Commute to workplace  Do you commute to your workplace alone or do you use public transport?
     Yes
     No
Hydroxychloroquine prophylaxis  Are you taking hydroxychloroquine tablets for prophylaxis against COVID-19?
     Yes
     No
Social distancing  How frequently do you maintain social distancing which means staying 6 feet away from each other?
     Almost always (more than 90%)  
     Mostly (about 75%)  
     Commonly (about 50%)  
     Sometimes (about 25%)  
     Rarely (less than 10%)
Hand hygiene  How frequently do you wash your hands with soap water or alcohol-based hand rubs?
     Once every hour
     Every 2-3 hourly
     Every 3-4 hourly
     Less than that  
Do you ensure that you spend 40 seconds every time washing your hands?
     Yes
     No
Use of PPE by healthcare workers  What PPE do you use while working in the healthcare setting?
     Mask  
     Gown  
     Gloves  
     Cap  
     Shoe covers  
     Goggles/ Eyeshield   Statistical analysis
Data was recorded on a pre-designed proforma and managed in an excel spreadsheet. All the entries were checked for any possible keyboard errors. Quantitative variables were assessed for approximate normality and summarized as mean ± SD or median (Q1, Q3). In step 1, logistic regression analysis was used with the outcome and each of the independent variables separately. In step 2, multiple logistic regression analysis was performed with all the variables simultaneously with the outcome variable separately. Stata 15.0 statistical software (StataCorp LLC, College Station, USA) was used for data analysis. In this study, a p-value <0.05 was considered as statistically significant.