Abstract
Background:
The Coronavirus Anxiety Scale (CAS) was developed in April 2020 as an instrument to gauge the symptoms of stress and anxiety occurring secondary to the COVID-19 pandemic. It was constructed using the English language due to which its application to a multi-linguistic country like India became a hindrance.
Aim:
To establish psychometric validity and reliability of the Bengali translation of CAS.
Materials and Methods:
Experts well-versed in the English and Bengali languages translated CAS from English to Bengali. This was followed by further correction by forward and backward translation processes until a version with high accuracy and low redundancy was procured. The original scale was applied to 91 volunteers, and the translated version was applied to the same group after 14 days.
Results:
The mean scores of the original version and the translated version showed a significant correlation. The Bengali version of CAS had high internal consistency, significant concurrent validity, and acceptable split-half reliability. It was hence deemed easily understandable and capable of measuring anxiety due to COVID-19 similar to the original scale.
Conclusion:
The Bengali translation of the CAS can be accurately used to assess symptoms of anxiety and stress with high reliability and validity in those who are able to read the Bengali language.
Keywords: Anxiety, Bengali, coronavirus, COVID-19, lockdown, pandemic, stress
The World Health Organization (WHO) declared the SARS-CoV-2 outbreak as a pandemic on March 11, 2020, which was within a few months after its first case on November 17, 2019, in Wuhan, China. This viral strain of the coronavirus family took only a few days to catapult the entire world into a hotbed of a novel and unpredictable flu, and a breeding ground of chaos, confusion, and crises. Glimpses of historic plagues and pandemics were observed in the modern 21st century as the virus shattered our apparent veneer of invincibility to lay bare the fragile and vulnerable nature of mortal existence.[1,2,3] The COVID-19 pandemic redefined the daily lifestyle and habits of the average individual, as it progressed through its undulating phases.[4,5,6] The Indian state of West Bengal recorded its first confirmed case on March 17, 2020, after which the case numbers only kept surging. Although the number of cases in West Bengal were lower than the national average (0.618 vs. 0.843), the number of deaths were higher than the national average (1.777 vs. 1.443). Kolkata was the most affected, followed by North 24 Parganas and Howrah, all of which together comprised 80% of the cases in West Bengal. Following the first wave in India, the percentage of positive cases was 8.86%, the mortality rate was 2.08%, and the recovery rate was 74.48%; all of which were positive indicators.[7] The public health recommendations and various government measures meted out in an attempt to curb the rampant transmission had imminent corollaries. Labor markets, consumption patterns, and economic activities, lifestyle patterns were grossly disrupted. It upended much of society in unprecedented ways and the entire world came into the grasp of a mental health crisis. The potential for adverse effects of COVID-19 on mental health in India was further heightened due to the fragmented social structure and insufficient healthcare systems prevalent in our country.[8] Although adequate attention has been meted out for measures to identify people with the coronavirus infection, the same cannot be said about measures and steps taken to identify the mental health issues of people affected by this pandemic. Therefore, the Corona Anxiety Scale (CAS) was developed to help gauge mental health disruptions in response to the coronavirus pandemic by reliably identifying individuals suffering from symptoms of anxiety as a consequence of the pandemic.[9] However, the scale is not available in many Indian languages.[10] India has been one of the most severely hit countries by the novel coronavirus and its subsequent mutations. Due to CAS being available only in the English language, it provided a language barrier to Indians. As per Census 2011, Bengali language is the second most common first language of Indians, right after the Hindi language. Thus, a Bengali translation of the scale was undertaken to cater to this majority. The present study was undertaken to establish the reliability and psychometric validity of the Bengali version of the CAS.
MATERIALS AND METHODS
After gaining approval from the institutional ethics committee, this cross-sectional analytical study was conducted over 6 months on 91 consenting adults who were well-versed in both the English and Bengali languages. The snowball technique was used for data collection. Those with any comorbid psychiatric disorders preventing them from partaking in the study were excluded.
Preparation of the Bengali translation of the CAS
Two independent Bengali translations of the CAS were arrived at by two linguistic experts proficient in English and Bengali, giving us versions 1 and 2 of the scale. Following that, the subject experts intervened and arrived at version 3 of the scale. Version 3 was back-translated into the English language by two other linguistic experts, giving us version 4 of the scale. Discrepancies between the original scale and version 4 were brought out by subject experts and were addressed. This was once again translated to Bengali, and version 5 was hence arrived at. Version 5 was applied to 30 individuals, who were requested to provide feedback in terms of difficulties encountered in understanding any of the items: all of which were addressed to arrive at the final version 6 of the scale, which was used in the present study.
Tools
Sociodemographic proforma
This self-constructed proforma was utilized to assess sociodemographic variables.
Coronavirus Anxiety Scale (CAS)
The CAS is a screening tool to identify dysfunctional anxiety as well as the degree of severity in individuals, with 90% sensitivity and 85% specificity.[9]
Bengali translation (version 6) of the CAS
This CAS translation was made for the present study.
Methodology
The study was conducted in two parts. The subjects included in the study were 91 residents of the city of Kolkata in West Bengal and Pune in Maharashtra who were conversant in English and Bengali. After obtaining written informed consent and recording the sociodemographic details, the original CAS was applied to the participants After a hiatus of 14 days, the Bengali translation of the CAS was applied to the same participants. It was scored along the lines of the original version and was tabulated alongside the original version. The collected data were scored and tabulated. The scores were compared and evaluated.
Statistical analysis
The validity and reliability of the Bengali version of CAS were assessed with the help of Classical Test Theory (CTT). The scale was tested for validity and reliability as a whole, as well as for each item. The entire scale was assessed using methods of internal consistency. Individual items were assessed using item–item correlation and item–total correlation. A confirmatory factor analysis (CFA) was also performed. Statistical analysis was performed using SPSS 20.0 (IBM, Atlanta, USA). Calculations of McDonald's omega, the greatest lower bound, and the CFA were performed using the JASP software (University of Amsterdam, Netherlands).
RESULTS
The sample comprised 50 females and 41 males. The mean (±standard deviation [SD]) age of the subjects was 30.68 (±10.87) years. The range of age was from 20 to 61 years. The majority of the sample had completed their graduation. Apart from six students and two retired persons, all subjects were employed [Table 1]. The mean (±SD) score of the original version was 3.451 ± 3.357 and that of the Bengali translation was 3.396 ± 3.678. The difference between the scores was not significant. Skewness and kurtosis for items on the Bengali version of the CAS scale were in the range of −1.031 to 2.052 and −0.374 to 3.715, respectively, which is acceptable [Table 2]. The scores on the English and Bengali versions were significantly correlated. Kendall's tau b was 0.796; P < 0.001 (2-tailed). All items of the Bengali CAS have corrected item-total correlation values of more than 0.4, which indicates very good discrimination. Further, apart from item 1, all the item-total correlations were greater than 0.5 [Table 3]. Inter-item correlation using the Pearson correlation revealed that except for item 1, all the inter-item correlation coefficients were higher than 0.3 [Table 4], indicating an acceptable validity for the instrument.[11]
Table 1.
Demographic characteristics of the sample
| Background | Category | n | % |
|---|---|---|---|
| Age distribution | 25 years and below | 29 | 31.87% |
| More than 25 to 50 years old | 52 | 57.14% | |
| More than 50 years old | 10 | 10.99% | |
| Gender | Male | 41 | 45.06% |
| Female | 50 | 54.95% | |
| Education level | High school | 3 | 3.30% |
| Diploma | 3 | 3.30% | |
| Bachelor degree | 54 | 59.34% | |
| Master degree | 29 | 31.87% | |
| Doctoral degree | 2 | 2.20% | |
| Occupation | Student | 6 | 6.59% |
| Lawyer | 6 | 6.59% | |
| IT professionals | 11 | 12.09% | |
| Doctor | 24 | 26.37% | |
| Self-employed | 16 | 17.58% | |
| Service | 26 | 28.57% | |
| Retired | 2 | 2.20% |
Table 2.
Scores on the Bengali and English versions of the Coronavirus Anxiety Scale
| Variables | BTOT | ETOT |
|---|---|---|
| Mean | 3.3956 | 3.4505 |
| Median | 2.0000 | 2.0000 |
| Std. deviation | 3.67840 | 3.35746 |
| Variance | 13.531 | 11.273 |
| Skewness | 1.514 | 1.779 |
| Kurtosis | 3.497 | 5.664 |
| Minimum | 0.00 | 0.00 |
| Maximum | 20.00 | 20.00 |
Table 3.
Bengali version of CAS: Item-total statistics
| Items | Mean | Std. deviation | Scale mean if the item deleted | Scale variance if item deleted | Corrected item-total correlation | Squared multiple correlation | Cronbach’s alpha if the item deleted |
|---|---|---|---|---|---|---|---|
| B1 | 0.4066 | 0.68277 | 2.9890 | 11.078 | 0.437 | 0.259 | 0.918 |
| B2 | 0.8132 | 0.96508 | 2.5824 | 8.046 | 0.832 | 0.733 | 0.838 |
| B3 | 0.9890 | 0.90055 | 2.4066 | 8.866 | 0.719 | 0.571 | 0.866 |
| B4 | 0.5714 | 0.94449 | 2.8242 | 8.147 | 0.833 | 0.827 | 0.838 |
| B5 | 0.6154 | 0.90393 | 2.7802 | 8.373 | 0.830 | 0.847 | 0.840 |
Item total correlation values of 0.4 and above indicate very good discrimination.
Table 4.
Item correlation matrix Bengali version (Kendall's tau_b)
| Item | Correlation | B1 | B2 | B3 | B4 |
|---|---|---|---|---|---|
| B2 | Correlation coefficient | 0.294** | |||
| Sig. (2-tailed) | 0.002 | . | |||
| B3 | Correlation coefficient | 0.310** | 0.697** | ||
| Sig. (2-tailed) | 0.001 | 0.000 | . | . | |
| B4 | Correlation coefficient | 0.226* | 0.664** | 0.498** | |
| Sig. (2-tailed) | 0.021 | 0.000 | 0.000 | . | |
| B5 | Correlation coefficient | 0.155 | 0.694** | 0.530** | 0.844** |
| Sig. (2-tailed) | 0.113 | 0.000 | 0.000 | 0.000 |
An exploratory factor analysis (principal component analysis with varimax rotation) of the Bengali version of CAS was undertaken to examine its factor structure. approximate χ2 was 329.798, df = 10, P < 0.001 in Bartlett's test of sphericity. The results identified one factor accounting for 69.243% of the total variance explained. The component matrix ranged from 0.562 to 0.909, commonality coefficients ranged from 0.316 to 0.823, and cross-loadings ranged from 0.23 to 0.29 [Table 5]. A CFA was performed to confirm the unidimensional factor structure suggested by the original authors. The single-factor model (Chi-square (df 5) =7.259; P < 0.202) in the CFA yielded a good model fit for all indices (GFI = 0.998, MFI = 0.988, CFI = 0.999, TLI = 0.999, RMSEA = 0.071, SRMR = 0.053, and χ2/df ratio = 1.45) The results of all psychometric measures were acceptable and confirmed the validity and reliability of the Bengali version of the CAS [Table 6 and Figure 1].
Table 5.
Factor analysis of Bengali version of CAS: Total variance explained
| Component | Initial eigenvalues |
Extraction sums of squared loadings |
||||
|---|---|---|---|---|---|---|
| Total | % of Variance | Cumulative % | Total | % of Variance | Cumulative % | |
| 1 | 3.462 | 69.243 | 69.243 | 3.462 | 69.243 | 69.243 |
| 2 | 0.777 | 15.531 | 84.774 | |||
| 3 | 0.467 | 9.335 | 94.109 | |||
| 4 | 0.205 | 4.095 | 98.204 | |||
| 5 | 0.090 | 1.796 | 100.000 | |||
Extraction method: principal component analysis
Table 6.
Psychometric properties for the Bengali version of the COVID Anxiety scale at the scale level (n=91)
| Psychometric measure | Result | Suggested cut-off |
|---|---|---|
| Cronbach’s alpha | 0.888 | >0.7 |
| McDonald’s omega | 0.902 | >0.7 |
| Greatest lower bound | 0.943 | >0.7 |
| Average variance extracted (AVE) | 0.821 | >0.5 |
| Composite reliability | 0.916 | >0.7 |
| Standard error of measurement | 0.385 | <1.839 |
| Model fits of confirmatory factor analysis | ||
| X2 (df, P) | 7.259 (5, 0.202) | ns |
| X2/df | 1.4518 | <5 |
| Goodness of Fit Index (GFI) | 0.998 | >0.95 |
| Comparative Fit Index (CFI) | 0.999 | >0.95 |
| Tucker–Lewis Index (TLI) | 0.999 | >0.95 |
| Root mean square error of approximation (RMSEA) | 0.071 | <0.08 |
| Standardized root mean square residual (SRMR) | 0.053 | <0.08 |
| Bollen’s Incremental Fit Index (IFI) | 0.999 | >0.90 |
| Bentler–Bonett Normed Fit Index (NFI) | 0.998 | >0.90 |
Figure 1.
Single factor confirmatory factor analysis of the Bengali version of the Coronavirus Anxiety Scale
DISCUSSION
The atmosphere was emotionally charged with feelings of anxiety, fear, worry, and various other psychological disturbances to the extent that they were being identified as a secondary pandemic running concurrently with the COVID-19 pandemic. Imminently, there was a need to assess symptoms of anxiety and worry arising due to the pandemic as they became both a potential confounding factor in the outcome of COVID-19 cases and an independent entity mandating requirements on its own. The traditional tools to measure these symptoms of anxiety were either under-diagnosing or over-diagnosing due to the uniqueness of the background against which it was arising. The Hamilton Anxiety Scale, Beck's Anxiety Inventory, and Generalized Anxiety Disorder-7 were inevitably lapsing in measuring anxiety symptoms as they were not constructed to measure symptoms of anxiety in the background of a pandemic. Hence, new scales were constructed that not only had adequate reliability and validity to measure symptoms but also remained pertinent to the pandemic. The CAS was one of the first of its kind; it comprised 5 items and had a 90% sensitivity and an 85% specificity.[9] Statistical analysis revealed that the items of the Bengali translation had good item–item as well as item–total correlation, that is, all items were unique and neither redundant nor repetitive,[12] and the items were independently also measuring the same attribute, respectively.[13] In the present study, apart from item 1, all the item–total correlation was greater than 0.5 [Table 3]. Inter-item correlation using the Pearson correlation revealed that except for item 1, all inter-item correlation coefficients were higher than 0.3 [Table 4]. This concurs with recommendations that in an empirical approach, as a rule of thumb, if the score of the item-to-total correlations is more than 0.50 and the inter-item correlations exceed 0.30, the construct validity is satisfied.[11]
The translated scale demonstrated good internal consistency reliability and composite reliability. The Kaiser-Meyer-Olkin of our study demonstrated a middling degree of common variance among the variables, and the Bartlett test of sphericity was significant, indicating that the correlation matrix was not an identity matrix such that each variable correlated perfectly with itself, but not with other variables. Overall scores found the Bengali translation to be an adequate tool to assess anxiety in the context of the COVID-19 pandemic.
Limitations
The sample population of the study comprised well-educated persons residing in urban areas and knowing English and Bengali. The sample size was modest due to the limitation of time and the pandemic.
CONCLUSION
The Bengali version of CAS is a valid and reliable instrument for assessment among the literate Bengali-speaking population.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
REFERENCES
- 1.Srivastava K, Chaudhry S, Sowmya AV, Prakash J. Mental health aspects of pandemics with special reference to COVID19. Ind Psychiatry J. 2020;29:1–8. doi: 10.4103/ipj.ipj_64_20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Dhamija S, Samudra M, Davis S, Gupta N, Chaudhury S. COVID19 lockdown – Blessing or disaster? Ind Psychiatry J. 2021;30(Suppl S1):294–6. doi: 10.4103/0972-6748.328834. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Chag J, Chaudhury S, Saldanha D. Economic and psychological impact of COVID 19 lockdown: Strategies to combat the crisis. Ind Psychiatry J. 2020;29:362–8. doi: 10.4103/ipj.ipj_120_20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Davis S, Samudra M, Dhamija S, Chaudhury S, Saldanha D. Stigma associated with COVID19. Ind Psychiatry J. 2021;30(Suppl 1):S270–2. doi: 10.4103/0972-6748.328827. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Davis S, Samudra M, Dhamija S, Chaudhury S, Saldanha D. Quarantine: Psychological aspects. Ind Psychiatry J. 2021;30(Suppl 1):S277–81. doi: 10.4103/0972-6748.328829. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Chopra S, Ranjan P, Singh V, Kumar S, Arora M, Hasan MS, et al. Impact of COVID-19 on lifestyle-related behaviours-a cross-sectional audit of responses from nine hundred and ninety-five participants from India. Diabetes Metab Syndr. 2020;14:2021–30. doi: 10.1016/j.dsx.2020.09.034. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Mondal B.K., Sahoo S., Paria P., Chakraborti S, Alamri AM. Multi-sectoral impact assessment during the 1st wave of COVID-19 pandemic in West Bengal (India) for sustainable planning and management. Arab J Geosci. 2021;14:2448. [Google Scholar]
- 8.Kene P. Mental health implications of the COVID-19 pandemic in India. Psychol Trauma. 2020;12:585–7. doi: 10.1037/tra0000750. [DOI] [PubMed] [Google Scholar]
- 9.Lee SA. Coronavirus Anxiety Scale: A brief mental health screener for COVID-19 related anxiety. Death Stud. 2020;44:393–401. doi: 10.1080/07481187.2020.1748481. [DOI] [PubMed] [Google Scholar]
- 10.Deshmukh S, Chaudhury S, Diwan C, Kumar S, Prakash J. Reliability and psychometric validity of the Marathi adaptation of coronavirus anxiety scale. Ind Psychiatry J. 2023;32:417–22. doi: 10.4103/ipj.ipj_45_22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Cohen J. Statistical power analysis. Curr Dir Psychol Sci. 1992;1:98–101. [Google Scholar]
- 12.Piedmont RL. Inter-Item Correlations. In: Michalos AC, editor. Encyclopedia of Quality of Life and Well-Being Research. Dordrecht: Springer; 2014. pp. 3303–4. [Google Scholar]
- 13.Zijlmans EA, Tijmstra J, van der Ark LA, Sijtsma K. Item-score reliability in empirical-data sets and its relationship with other item indices. Educ Psychol Meas. 2018;78:998–1020. doi: 10.1177/0013164417728358. [DOI] [PMC free article] [PubMed] [Google Scholar]