Comments on psychological impact of COVID-19 lockdown: An online survey from India

Sir,

We thank Singh[1] and Shukla[2] for their comments on our article.[3] Both the authors have raised issues related to the translation of questionnaire to multiple languages. We agree that carrying out the psychometric evaluation of the translated versions in all languages could have added to the whole effort. However, the research was done at the beginning of the COVID-19 pandemic, and carrying out psychometric evaluation of the scales, in the absence of in-person contact, was not feasible, and this itself would have taken lot of time. The other option was to conduct the study by using the available English and the Hindi version of the scales. The Hindi version of the scales has some psychometric data available.[4,5,6,7]

However, to have a wider perspective, we translated the scales into various languages to avoid limiting the study population to those who are able to read only English/Hindi. However, at the end of the survey, we realized that there were few takers for those with local languages. Majority of the responses were received for the English (n = 1072; 63.62%), and this was followed by the Hindi (with English version as part of bilingual) version (n = 213; 12.64%). Other languages contributed to the remaining 400 (23.73%) responses.

We reduced the time frame for the assessment on perceived stress scale, keeping the timing of the survey. The suggestion of labeling it as a modified version of perceived stress scale is well taken. However, the same was not considered, because, except for the time frame, we did not make other adaptations in the scale. Regarding the Warwick Edinburgh scale and the age norms, the survey actually included subjects aged 18 and above. As already mentioned in the article, we actually intended to exclude all the responses of people aged less than 18 years. Unfortunately, we missed out on 4 responses, 2 from those aged 14 years, 1 aged 15 years, and one aged 16–year-old participant. Accordingly, considering the total number of subjects aged less than 18 years to be 0.25% of the total study sample, this should not be considered as a major issue, which could have influenced the findings. Singh[1] pointed out a difference in the “frequencies in the variable occupation and profession as retired mentioned are differently documented.” We are aware of the same, and the same can be understood from the perspective that many professionals who had retired formally from their government jobs, marked themselves as retired, but at other place while answering about their profession, marked themselves as per their professional degrees. Hence, we did not alter the responses to align these and presented the responses as such. Unfortunately, the survey questionnaire had no such provision to record the specialty of the healthcare professionals, which was raised. Considering that this was an online survey and we were not sure about the response rate, we had formally not estimated the sample size. Singh[1] raised the issue that the “the article did not discuss the effects of lockdown on severe depression and suicide.” We did present the information about the prevalence of severe depression as per the PHQ-9. This study did not aim to evaluate the incidence of completed suicide; hence, this issue is off-placed.

Shukla[2] has brought out the issue of snowball sampling. It is a pertinent issue, but it needs to be understood that, at the time when the survey was done, the use of other sampling methods could have led to the loss of opportunity to carry out such a survey. In terms of issues related to the face validity of the traditional assessment tools (e.g., PHQ-9 and GAD-7), it is important to understand that the same has been used during the current pandemic across the globe to evaluate the psychological impact on general population and healthcare workers. Hence, this limitation applies to all the surveys carried out during the ongoing pandemic. Shukla[2] has pointed about the failure to elaborate on the geographical location of the participants of this enormous survey involving several languages. We could not present the breakup of the geographical location of the participants as this variable was not part of the survey. In terms of time to response, the fact that the finding of the current survey of median time to respond to the survey was the 2^nd day of the survey itself is understandable, considering the fact that, in most of the surveys, majority of the responses are received during the initial days of the survey. For the variable “level of working,” we have rechecked the data and the total figures do add up to 100%. We did not provide the description of the variable “other,” for which 147 (8.7%) of the responses were given, as the numbers pertaining to each combinations were small, and this was not the major outcome variable.

Further, Shukla[2] pointed out the issue of SurveyMonkey site not technically supporting one device one response. In response, we would like to say that we verified the responses for the IP addresses to the best of our understanding and no effort was made to purchase the responses. We understand that SurveyMonkey platform has the option of purchasing the responses. If we wanted to utilize all these strategies, we could have used the same to have much more number of responses. As pointed out in the comments of Singh,[1] we have tried to be very brief in drawing any conclusions. At the end, Shukla[2] again has misunderstood when we say that the response rate is low. In a snow ball sampling technique, it is not possible to know the denominator; hence, it is not possible to calculate the response rate. When we said that the response rate was low, we basically wanted to put forth the issue that we expected a wider response, from various parts of the country, and we did not get the desirable level of response. We do not claim that the survey reflects the exact picture at the ground level in every part of the country, as there are many limitations of this survey, as documented in the paper. It is important to understand that every research endeavor will have its limitations and the findings of the same must be interpreted in light of the same.[8]

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.