The participants included in the current study were mothers of children participating in a double-blinded clinical trial (ClinicalTrials.gov Identifier: NCT02272842; Universal Trial Number: U1111-1161-5187) entitled “The effect of Vitamin B12 supplementation in Nepali Infants on Growth and Neurodevelopment” [24]. A total of 600 children of 6–11 months enrolled in the trial were followed every 12 months after a one-year supplementation period. From 9 September 2020 to 10 February 2021, we interviewed 493 of the mothers on COVID-19-related questions. The age of the children of the participating mothers were 54–71 months.

A total of 67 children from the original study of 600 children were lost to follow up (due to migration and refusal) when we initiated the COVID-19 exposure interviews with the mothers. Of the remaining 533 participants, we were not able to complete the COVID-19 exposure questionnaire with 40 mothers and thus the final sample consisted of 493 mothers. The flow of participants is depicted in Fig 1.

At enrollment, mothers answered questions on demographics and household characteristics. Information on the family members included mother’s age, age category of family members, caste/ethnicity of family, level of education, occupation group, and nutritional status of the mother. Family members were categorized in age groups for which we considered children <16 and adults >60 more vulnerable to consequences of a COVID-19 infection. Caste/ethnicity of the families was categorized as Newar, Brahmin, Chhetri, Tamang and others. The educational level was assessed as illiterate, primary school, secondary school, School Leaving Certificate/intermediate school, bachelor’s degree, and above. For our analysis, we dichotomized education up to 10^th years (i.e., completed secondary school) and above. For occupation, we categorized the occupation groups of both mothers and fathers as no formal work, agriculture, carpet worker, daily wage earner, self-employed, working in the service, and working abroad. Trained staffs measured height and weight in order to calculate body mass index (BMI) for the nutritional status of the mothers using a stadiometer for height (Prestige, HardikMedi Tech, India) and an electronic scale for weight (Salter/HoMedics Group, UK and Seca, Germany).

Socioeconomic status was measured by a multidimensional measure, the WAMI-index with the components; water and sanitation, household assets, maternal education, and income [25]. In contrast to previous studies, our data did not have income information and hence the index was calculated using the first three dimensions [26] (see S1 Table). For the current study, we set the cutoff for a low WAMI score at the 33^rd percentile of our study sample.

In follow-ups of the original study when the children were 30–35 months old, the mothers answered the Self-Reporting Questionnaire-20 (SRQ-20). The scores from SRQ-20 were from 2 years prior to the current study and covered previous maternal symptoms of depression.

The COVID-19 follow-up was completed during a visit to the study clinic when the children were in between 54 and 71 months old. Trained study staff administered the questionnaires with the mothers in a private room at the study clinic. Women who did not attend this follow-up visit for the present study (approximately 50%) were interviewed through phone calls due to restrictions of movement (lockdown).

We assessed COVID-19 exposure through a questionnaire particularly developed for this study (See S1 File). The presence of symptoms during the last two weeks that had been described for COVID-19 was recorded by the major symptom categories (fever, cold/cough/difficulty breathing, sore throat and diarrhea). To identify whether the mother and her family members had a chronic illness or diseases that constituted a risk for severe COVID-19 illness, a checklist with the following illnesses was included; chronic respiratory illness, renal disease, high blood pressure, diabetes, other chronic diseases. If the mother or another family member had such medical condition, we further asked the concerns about the shortage of essential medicines related to the illnesses. Furthermore, we recorded whether the mother had been exposed to someone who was COVID-19 positive, if she or anyone from her family had received COVID-19 positive test results, and how many days she previously had been quarantined or isolated. Prior to the study, our study team had thorough discussion of the items in which physicians and psychologists from Nepal and clinical psychologists from Norway were involved.

We also asked if the mothers believed that the life of someone close to her or her own life was in danger (response categories “not at all,” “Some/Just a little,” and “a great deal”). These items of the questionnaire are similar to previous traumatic exposure questionnaires [27, 28]. The impact of the pandemic on different aspects of life (i.e., economic, food security, employment, health related issues, daily life) was assessed on a five-point scale: “Not at all” (1); “a little” (2); “some” (3); “quite a lot” (4); and “a great deal” (5). This is a similar approach to what was used in the same region in relation to the impact of earthquakes [28].

Worry of contracting the corona virus was measured by two items, of which one was worry about contracting the corona virus yourself, and one was the worry that someone in the family would contract the corona virus. The worry items are based on a previous questionnaire on worry for infections [29]. Three items were on sleep related problems covering if the mothers slept worse than before due to worry about health-related, social and economic consequences due to the corona virus. Scoring for both worry and sleep problems was recorded on a five-point Likert scale (strongly disagree to strongly agree). The score ranges from 2 to 10 and 3 to 15 for worry and sleep respectively. The internal consistency of the worry and sleep score were excellent, with a Chronbach’s Alpha for itemized data of 0.93 and 0.95 respectively.

The Self-Reporting Questionnaire-20 (SRQ-20) [30] is a brief screening questionnaire for mental health problems consisting of 20 items on symptoms of depression, anxiety and psychosomatic complaints. It can be administrated through interview or the paper-pencil method. The response from each item was recorded as Yes or No depending on whether the symptoms were present or not during the last 30 days. The scores are summed up to a total score that ranges from 0 to 20. It takes 5–10 minutes to administer. In the current study, we used 10 as a cutoff for depression [31]. The feasibility of the questionnaire has already been confirmed in a previous study within the same population [32].

The instruments were translated into Nepali by a psychologist and medical doctors and then back-translated by a qualified person independent from the study who was fluent in both Nepali and English language, following the standard guidelines for translation processes [33]. Discrepancies were discussed and adjusted.

Ethical approvals were obtained from Nepal Health and Research Council (NHRC; #73/2017, #820/2020) and from the Regional Committee for Medical and Health Research Ethics (REC; No.2014/1528) in Norway. Mothers were asked to visit the central clinic and we obtained informed written consent for the extended follow ups.

All the continuous variables are presented in means and standard deviations, and categorical variables in numbers and percentages. Exposure variables were categorized into a three-point scale: “Not at all” (1 from original questionnaire), “Some” (merged 2 and 3 from original questionnaire) and “A great deal” (merged 4 and 5 from original questionnaire). We compared the mean worry and sleep scores between groups using the student’s t-test. The high risk groups were dichotomized as follows: education of mother and father (up to 10^th years vs. above 10^th years), symptoms similar to COVID-19 (absence vs. presence of at least one symptom), chronic disease of mother (absence vs. presence of at least one chronic disease), chronic disease in family member (absence vs. presence of at least one chronic disease), previous symptoms of depression (scores above 10 vs. 10 and under), and the WAMI-index score (≤33^rd percentile vs. >33^rd percentile of the WAMI score). The missing values were handled by listwise deletion. Data was analyzed using the STATA 16.0 software.

The mothers reported that the economic and employment status were greatly affected by the pandemic. At baseline for the original study, many reported to have no formal work, being in agricultural work, carpet work and being a daily wage earner and about half of their husbands had the same occupations signifying that the type of work in Bhaktapur is mainly based on daily wage earnings. These were occupations that were almost impossible to keep functioning during lock down or other restrictions. Other aspects of the daily life, like food insecurity, health related issues, shortage of essential medicines were also marginally prevalent. These results are in line with a previous review of published articles related to the psychosocial impact of COVID-19 in Nepal showing substantial impact on economy, agriculture, employment and the health sector [4]. We found a similar pattern of responses in the same study sample during the massive earthquakes in Nepal in 2015, where the most affected aspect of daily life was employment, and less on health-related issues. However, in comparison to the present results, the perceived impact on food insecurity was higher while economic impact was almost not affected [28].

Although many mothers reported that they or someone in their family had a chronic disease, the mothers were not so much concerned about supplies of essential medicines. This could be related to the fact that the study setting is close to the capital city of Nepal where the government made more efforts to prevent the pandemic compared to other regions of the country. In another study from Eastern less urban areas of Nepal, people perceived that medical supply was the most affected aspect of the health care system in Nepal during the pandemic [4].

Around 10% of the participants believed that their own life or the lives of someone close to them were in danger. Approximately 40% of the participants reported being worried that they themselves or their family members would contract the corona virus. A high rate of worry is expected during a pandemic and in line with previous studies that demonstrated an increase in public anxiety during epidemics or pandemics such as the 2003 SARS pandemic, the 2009/2010 H1N1 pandemic and the 2014/2016 Ebola pandemic [34–37]. Worry about the COVID-19 pandemic can also induce negative effects on mental health in the general public [38].

The rate of worry related sleep problems were low in the present study. Despite the major impact of the pandemic on their daily life, the mothers reported that the social, economic and health-related consequences due to the COVID-19 pandemic did not negatively affect their sleep and less than ten percent of the mothers reported these problems. We do not have any possible explanation to the low prevalence of sleep problems since we lack specific information about the sleep patterns of Nepalese women. However, the low prevalence might be due to taking COVID-19 very lightly since most of the people who were infected during the study period were asymptomatic or had mild symptoms.

The mothers who had a previous history of depression reported more sleep problems due to concerns with the COVID-19 pandemic. This finding is in accordance with previous studies showing that depressive symptoms is associated with poorer or disrupted quality of sleep during the pandemic [17, 39], and that previous history of psychological condition such as anxiety, stress and depression is associated with poor sleep [40].

There was a social gradient in worry, but in the opposite direction than expected from previous studies in other parts of the world. In the present study, there were higher levels of worry for contracting the virus among those who had an educational level higher than secondary. This group with the lowest educational qualifications being least concerned about contracting COVID-19, might be the challenging factor in controlling the pandemic due to less safety measures to protect them from virus infection. Hence, public health practitioners should consider the level of education while addressing the psychological impacts of pandemics since our results suggests that education can be the deciding factor for psychological impact of the individuals. Further, and also in contrast to our expectations and the established social gradient of health that has been evident during the pandemic [41], we found that those with higher SES (i.e. higher WAMI index) had higher worry scores than those with lower SES. This pattern of higher level of worry and higher socioeconomic status have been reported previously in high risk communities of Africa with higher academic qualification associated with greater level of psychological distress and worry [42, 43]. This may be related to more access to information about the pandemic and consequences for those with higher education [44]. Health literacy is an individual’s idea, motivation and capacities to recognize, evaluate and judge health information [45]. Individuals with low SES may be less sensitive toward health-related and other consequences of the pandemic due to the lack of health literacy resulting in these being less likely to seek help when they experienced symptoms [46]. Secondly, they might lack the capability of understanding the symptoms and health consequences of the corona virus limiting their interpretation and decision about own health condition [47].Thus, belonging to a low SES family may lead to less worry scores among mothers in our study. Moreover, all people who were infected from the corona virus during the study period, were asymptomatic or had mild symptoms [48]. This may have led to a misinterpretation of the disease in people with low SES.

One in every three of the mothers reported at least one of the major chronic diseases in herself or in someone in her family. Some of these common conditions such as respiratory diseases, diabetes, and chronic kidney diseases, have been related to increased risk of deaths from COVID-19 in Nepal [49]. Still there were no differences in the worry and sleep related worry scores between the participants who had these conditions or with family members with these conditions. The lack of associations with worry might be due to the younger age group and the predominance of the less hazardous 1^st variant of SARS-Cov-2 in Nepal at the time of the data collection [48].

A major strength of our study is the large sample size in a well-maintained cohort of mothers of young children. A limitation is that due to lockdowns, approximately half of the questionnaires were collected through phone calls which might not be as effective as interviews through direct contact. We also found that the worry score was significantly higher when collecting by an in-person interview compared to through a phone call. The fact that people respond differently if they were interviewed through phone calls rather than through direct contact methods have been demonstrated before [50].The study population was also limited to mothers whose infants were enrolled in a randomized controlled trial from 2015 to 2017 which limits the generalizability of the results since the sample may differ from general populations. Although the population of the Bhaktapur district covers different migrants from the whole areas of Nepal, the results can only be generalizable to the urban settings of Nepal with similar characteristics such as availability of daily supplies and health facilities, and the restrictions to the pandemic casted by local governments. We have only included mothers in the present study, and whether the results can be generalized to fathers is uncertain. However, since women often have responsibilities for health and care and may also be important role models for children, their perceptions and attitudes may also impact the family’s preventive measures and health. Our data may present only the results from a certain period of the pandemic in Nepal and hence cannot be generalized to the pandemic as a whole. The questionnaire was brief due to the situations for the mothers, and other aspects that would have been of interest such as more detailed information on sleep and current mental health could not be measured due to time constraints. The measure of depression is limited to a self-report questionnaire that we collected 2 years before, and not by a clinical interview. The final sample was better off than those who were lost to follow up. The result can be the strength of our study to differentiate the SES indicating that the losses to follow up were mainly due to the difficult financial situation during the lockdown (See S3 Table).