Comments on “Psychometric Properties,Stability,and Predictive Validity of the Hindi Version of the Prolonged Grief Disorder Scale (PG-13-R-H) Among Hindi-speaking Adults in the United States”

Discussion

The scale adaptations and/or translations are recommended to follow established guidelines and frameworks, such as those provided by the International Test Commission guidelines, ² to ensure methodological rigor, cultural equivalence, objectivity, and accuracy. However, it remains unclear whether Mewani et al. ¹ adhered to such frameworks, especially given that they used only a single translator, whose selection criteria were not documented, despite recommendations for utilizing between two and four bilingual translators. ³ Ideally, translators should be familiar with the construct, have relevant clinical/subject knowledge, and possess prior translation experience. ³ To compound this, the back translation was conducted by a co-investigator, whereas best practices recommend that this person be blinded to the original instrument, ³ which may have introduced bias and affected the conceptual and linguistic equivalence between the translated and original versions.

Mewani et al. ¹ employed Cronbach’s alpha as the sole index of internal consistency, despite its reliance on tau-equivalence.^4,5 Psychometric studies encourage an array of reliability indicators, including McDonald’s omega and composite reliability.^4,5 Although it is positive to see item–total correlations reported, most items yielded low correlations (<0.50), indicating potential concerns regarding discriminant ability. ⁶ Noteworthy inter-item correlation would have facilitated the examination of whether each item over- or underfitted the construct. ⁶

Mewani et al. ¹ used the Patient Health Questionnaire (PHQ-9) and the Warwick–Edinburgh Mental Well-being Scale (WEMWBS) to assess convergent/divergent validity, with only a weak correlation (r = 0.23) observed between the PG-13-R-H and PHQ-9. Thus, the authors’ claim of convergent validity is limited, with strong correlations with related constructs preferred (r > 0.50). ⁷ Furthermore, most PG-13-R-H items showed negligible/non-significant correlations with the PHQ-9, raising validity-related concerns for the Hindi version given the conceptual overlap between prolonged grief disorder (PGD) and depression, including sadness, social isolation, worthlessness, and suicidal ideation. ⁸ Additionally, the suitability of WEMWBS as a measure for demonstrating the discriminant validity of PG-13-R-H may be questioned, given the established association of PGD with poor well-being and related functional impairments, including poor health indicators (e.g., high blood pressure), higher rates of suicidal ideation and death by suicide, low life satisfaction, and increased use of medical services.^8,9 Thus, the findings reported in Mewani et al. ¹ could indicate a constraint in the Hindi version that fails to adequately capture the broader psychological impact of PGD on well-being; ultimately raising concerns about its validity.

Though it is positive to see that no assumption violations regarding normality, homoscedasticity, and multicollinearity were reported, the statistical tests used to evaluate these assumptions (the associated values) were not clearly described. The principal component analysis accounted for 32.12% of the variance, followed by 9.38%, suggesting that the PG-13-R-H is two-dimensional rather than unidimensional. ¹ However, greater detail on the factor structure, including factor loadings, rotation methods, and interpretations, is required to fully interpret these findings. Furthermore, a major limitation is the absence of a confirmatory factor analysis (CFA) to assess construct validity, which could facilitate the evaluation of factor loadings, composite reliability, convergent validity via average variance extracted, and discriminant validity via heterotrait–monotrait ratio of correlations.^4,5 The absence of which is not adequately acknowledged.

Though sampling methods were acknowledged within the article as potentially impacting the data (i.e., USA-based Hindi-speaking individuals potentially being may proficient in English, and thus having lower comprehension of the Hindi items), in light of the above methodological concerns, the reported PGD prevalence of 15.6% in Hindi-speaking adults, relative to the 13.4% observed in a diverse USA college sample, ¹ should be interpreted with further caution. Thus, it is indeed important to replicate this study within a sample recruited from India.

Overall, Mewani et al. ¹ Hindi version of the PG-13-R is timely and valuable given what we know about how symptoms of psychiatric conditions can vary across cultures, thus requiring a culturally sensitive diagnostic criterion and tools. ¹⁰

Conclusions

We recommend future studies to establish the strong psychometric properties of the PG-13-R-H as a culturally sensitive tool using robust methodological and statistical approaches, such as following standard guidelines in translation and adaptation, conducting CFA, measurement invariance across gender and culture, test-retest reliability, using a large, diverse sample size, including gold standard measures to assess criterion validity, such as the Diagnostic Interview Schedule.

Supplemental Material