Translation and psychometric validation of the Vietnamese versions of the Drinker Inventory of Consequences and Protective Behavioral Strategies Scale

This study was part of a larger doctoral research project conducted in Thai Nguyen, a northeastern province of Vietnam. The broader project included a randomized controlled trial (RCT) evaluating the effectiveness of a promoting protective behavioral strategies program to reduce alcohol use, alcohol-related consequences, and intention to drink among adult men with HAU.

The current study focused on the translation, cultural adaptation, and psychometric validation of two instruments used in the RCT: the DrInC¹⁴ and PBSS-20.¹⁵ Psychometric properties were assessed using data from a pilot study to examine inter-rater reliability and baseline data from the RCT to evaluate internal consistency.

The study was conducted in a rural area of Thai Nguyen Province using stratified random sampling. One rural district was randomly selected from 6 in the province, followed by the random selection of one commune within that district. Eligible individuals were screened and randomly recruited using a computerized selection process.

Ethical approval was granted by the Ethical Review Board for Biomedical Research at Hanoi University of Public Health, Vietnam (No. 454/2024/YTCC-HD3). All participants provided written informed consent. Data were anonymized to ensure confidentiality. This study adheres to the Strengthening the reporting of observational studies in epidemiology (STROBE) guidelines for cross-sectional studies, as detailed in the completed STROBE checklist (Appendix 1).²⁴

The target population consisted of men aged 35–44 years residing in Thai Nguyen Province who met criteria for HAU.

All data were analyzed using IBM SPSS Statistics version 28 (IBM Corp., Armonk, NY, USA). Descriptive statistics were used to summarize participant characteristics. CVI, Cronbach’s alpha coefficients, and ICC values were computed to evaluate the instruments’ psychometric properties.

Table 1 summarizes the demographic characteristics of participants in both the pilot and main studies. The mean age was comparable across samples (39.23 ± 2.57 years in the pilot and 39.53 ± 2.85 years in the main study). Most participants were married, had completed secondary education, and were primarily employed as farmers. Income levels varied, with nearly half of the main study participants falling into the high-income group. Notably, a substantial proportion reported early alcohol initiation (before age 18 years), and nearly half reported a family history of alcohol use.

Table 2 presents scores on the DrInC and PBSS-20 among participants in the main study. The mean total DrInC score was 25.17 (SD = 7.83), indicating a low to moderate level of alcohol-related consequences. The highest mean scores were observed in the physical, interpersonal, and intrapersonal domains, while lower scores were reported in impulse control and social responsibility.

The mean PBSS-20 score was 78.43 (SD = 14.15), indicating moderate use of protective behavioral strategies. Among subscales, SHR strategies were used most frequently (M = 36.10), followed by MOD (M = 21.71) and stopping/limiting drinking (M = 20.63).

Expert review confirmed strong linguistic and conceptual equivalence between the original English and Vietnamese versions of both instruments. Mean ratings for language comparability and interpretability were below 2 on a 7-point scale, indicating excellent equivalence, (Table 3).

Table 4 presents the psychometric outcomes for both instruments. The I-CVI ranged from 0.80 to 1.00 for both tools, exceeding the recommended threshold of 0.78. The S-CVI was 0.97 for the DrInC and 0.95 for the PBSS-20, indicating strong content validity.

Inter-rater reliability was excellent for the DrInC, with ICCs of 0.95 (single measures) and 0.97 (average measures). The PBSS-20 demonstrated good inter-rater reliability, with ICCs of 0.75 and 0.86, respectively. All ICCs were statistically significant (P < 0.001).

Internal consistency, assessed via Cronbach’s alpha, was acceptable for both instruments: 0.75 for the DrInC and 0.79 for the PBSS-20.

For item-total correlations, the values of DrInC ranged from −0.08 to 0.58. Several items (such as items 10, 18, 23, 42, 43) showed strong alignment with the overall scale (>0.50), most items were acceptable, and a few items were weaker (such as items 8, 12, 17, 20, 22, 25, 28) or negative (items 24, 37, 50). Cronbach’s alpha if item deleted ranged only from 0.74 to 0.76, indicating that removing these items would not meaningfully improve reliability.

For the PBSS-20, corrected item-total correlations ranged from −0.12 to 0.60. Several items (such as items 2, 5, 8, 14, 15, 18) demonstrated strong correlations with the total scale (>0.50), most items were acceptable, and a few items were weaker (3, 7, 10, 11, 20) or negative (item 4). Cronbach’s alpha if item deleted ranged from 0.76 to 0.81, indicating that all items contributed meaningfully to overall reliability.

Split-half reliability provided additional evidence of internal consistency. The Spearman–Brown coefficients were about 0.67 for both DrInC and PBSS-20, consistent with the overall Cronbach’s alpha.

The expert evaluation process demonstrated that the translated instruments retained both linguistic accuracy and conceptual equivalence with the original versions. The language comparability and interpretability similarity scores indicated that the translations preserved the meaning of the original items while ensuring clarity for Vietnamese population. This result aligns with previous research suggesting that systematic translation procedures, such as Brislin’s (1970) method, contribute to maintaining conceptual consistency across different languages.²⁶

Content validity was confirmed through high I-CVI: 0.8–1.0; S-CVI >0.9, exceeding standard benchmarks.²⁷ These values indicate a high degree of expert agreement on the relevance and clarity of each item. An I-CVI of 0.78 or higher is typically considered acceptable when using 5 or more raters, and the S-CVI values in this study meet the threshold for excellent content validity.²⁷ The expert panel’s diverse backgrounds including 2 nursing educators and 3 researchers with experience in addiction health and behavioral science, further enhance the credibility of these evaluations.^33,34 These results also affirm the success of the rigorous translation and adaptation process, which prioritized semantic accuracy, cultural alignment, and linguistic clarity.

Reliability analysis supported the overall suitability of both scales for use in the Vietnamese context. The DrInC demonstrated excellent inter-rater reliability and acceptable internal consistency (Cronbach’s α = 0.75), while the PBSS-20 showed good internal consistency (α = 0.79). These values fall within the established thresholds for research instruments,²⁸ although they are slightly lower than those reported in validation studies conducted in Western or high-income country settings, as discussed below.

For the PBSS-20, internal consistency in this study (α = 0.79) is comparable to that reported in the Spanish version (subscale α = 0.71–0.77) among university students,³⁵ supporting its cross-cultural reliability. However, higher internal consistency was reported in the original validation, with subscale alphas ranging from 0.81 to 0.88.¹⁵ These differences may be due to demographic and contextual variation, such as the use of a rural, older adult male population in this study compared to younger, university-based samples in earlier validations. Cultural norms around alcohol use and behavioral self-regulation may also influence how consistently individuals respond to protective strategy items, particularly when harm-reduction concepts are less familiar or socially reinforced.³⁶ This is consistent with behavioral health theories such as the Theory of Planned Behavior (TPB), which emphasizes the role of subjective norms and perceived behavioral control in shaping protective health behaviors.³⁷

For the DrInC, the internal consistency in this study was 0.75, which is lower than that reported for the English version (α = 0.937)³⁸ and the Swahili version (α = 0.96).¹⁸ These discrepancies may stem from differences in health literacy, cultural framing of alcohol-related harm, and the salience of certain consequences across contexts. For instance, items assessing moral or spiritual harm may be interpreted differently in collectivist versus individualist cultures. Nevertheless, the Vietnamese DrInC maintained acceptable reliability and item performance, suggesting that the core structure of the instrument remains robust even across culturally distinct populations. Furthermore, the high inter-rater reliability of the DrInC suggests strong agreement in item ratings across expert reviewers, which further supports its usability in clinical and research settings requiring consistent scoring.

Additionally, split-half reliability analyzes for both instruments yielded coefficients slightly below the conventional 0.70 threshold,³⁰ but closely aligned with overall Cronbach’s alpha values. This indicates moderate internal consistency and supports the reliability of using total or subscale scores for research and clinical purposes, while cautioning against overreliance on single items.

The item-level analyzes of the Vietnamese DrInC offer insights into how these instruments function in the Vietnamese cultural context. While several items demonstrated strong alignment with the overall scale, others showed weaker or even negative correlations. These weaker items often reflect consequences that are either very common but non-discriminating like hangover,³⁹ rare in this population such as accidents, absenteeism from work,⁴⁰ or socially sensitive and prone to underreporting as drunk driving.⁴¹ The limited variability or under-endorsement of these items likely reduced their correlations with the overall scale. Importantly, Cronbach’s alpha remained stable even when these items were excluded, suggesting that the instrument as a whole maintains acceptable reliability.

At the subscale level, the findings further highlight the role of cultural and contextual influences in shaping item responses. For example, behaviors in the impulse control domain such as drunk driving, physical fights, and other reckless actions may be interpreted less as personal failures of restraint and more as normalized outcomes of heavy group drinking.^2,3 Similarly, legal-related items, such as arrests for drinking and driving, may be underreported due to shame, and social desirability, because of Vietnam’s strict zero-tolerance laws.⁴² Items within the social responsibility domain, including missed work or financial problems, implicate family reputation and collective honor. In Vietnam’s collectivist culture, acknowledging such failures can bring shame not only to the individual but also to their family,⁴³ contributing to underreporting despite their prevalence. These cultural nuances underscore the importance of interpreting DrInC subscales through the lenses of collectivism and social obligation within the Vietnamese sociocultural environment.

For the PBSS-20, items with stronger correlations included a mix of self-regulation strategies (such as not exceeding a set number of drinks, avoiding pregaming, avoiding keeping up with others) and safety-oriented harm reduction strategies (e.g. refusing to ride with a drunk driver, avoiding drinking games, monitoring one’s drink).¹⁵ These findings suggest that both domains of protective strategies resonate meaningfully with Vietnamese participants. In contrast, several items showed weaker or negative correlations, particularly strategies that may be less common or less culturally relevant in this setting. For example, alternating alcoholic and non-alcoholic drinks, drinking water, or adding extra ice¹⁵ are not typical practices in Vietnamese drinking culture, when one has had enough may conflict with prevailing peer norms that encourage drinking.² Such patterns suggest that while many protective strategies are broadly applicable, others may require cultural adaptation or additional explanation to ensure relevance and adoption in Vietnam.

Although exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) are critical methods for examining construct validity, their reliability depends heavily on adequate sample size.⁴⁴ The current study’s sample of 104 participants which falls below commonly recommended thresholds for factor analysis. For example, guidelines suggest at least 5–10 participants per item or a minimum of 200–300 participants to ensure stable factor solutions, particularly for scales with complex, multidimensional structures such as the 50-item DrInC.^44,45 Similarly, CFA typically requires samples of at least 200 participants to achieve stable estimates and reliable model fit indices.^44,45 Conducting factor analysis with the current sample risked unstable or uninterpretable results, which could undermine the validation process.⁴⁴ Therefore, this study prioritized rigorous translation, cultural adaptation, content validation, inter-rater reliability, and internal consistency as foundational steps in the validation process. Future studies with larger, more diverse samples should use EFA, CFA, and invariance testing to confirm construct validity, factorial stability, and generalizability across Vietnamese populations.

Finally, robustness checks such as placebo tests and variable substitutions were not conducted in this study. These analyzes widely used in fields that rely on causal inference or regression modeling, where the goal is to verify that findings are not biased covariate selection.⁴⁶ The present study, however, focused exclusively on the psychometric validation of translated measurement instruments. For this reason, robustness analyzes were not applicable. The absence of such analyzes may limit the extent to which we can comment on the robustness of these measures under different analytic specifications.⁴⁶ Future studies that apply the Vietnamese DrInC or PBSS-20 in intervention trials, longitudinal cohorts, or predictive modeling could incorporate robustness analyzes.

Participant characteristics aligned with prior epidemiological profiles of hazardous drinkers in Vietnam,^3,47 including early initiation of alcohol use and family history of drinking, both of which are known risk factors of HAU.^20,48 Notably, the rate of early initiation in this study was higher than in some Western countries,⁴⁹ suggesting cultural influences on alcohol exposure during adolescence.

The relatively moderate mean DrInC scores (M = 25.17) are consistent with the AUDIT classification used in this study, which focuses on hazardous, rather than harmful or dependent drinking. Importantly, the highest DrInC subscale scores were observed in the physical, interpersonal, and intrapersonal domains, indicating that even non-dependent drinking levels were associated with early disruptions in health, relationships, and emotional well-being.¹²

Similarly, PBSS-20 scores revealed moderate use of protective strategies, with a strong preference for SHR behaviors (e.g. using a designated driver, eating before drinking). In contrast, strategies aimed at reducing the quantity or pace of drinking were less common, pointing to a reliance on externally observable harm mitigation rather than internal behavior control.¹⁵ This pattern may reflect the sociocultural context in Vietnam, where alcohol use is deeply linked to rituals of masculinity, celebration, and hospitality.² In such contexts, reducing quantity or abstaining may be socially discouraged, whereas minimizing visible harm may be more acceptable.

These findings point to the need for culturally grounded interventions that not only promote protective strategies but also challenge drinking norms embedded in social identity, family roles, and occupational environments, particularly in rural settings.

A notable strength of this study lies in the detailed and culturally sensitive approach to translation and adaptation. While Brislin’s forward–backward translation model ensured semantic accuracy,²³ the study went further by conducting systematic item-level refinements guided by iterative expert review, cultural relevance, and field context.²¹

Several items from the PBSS-20 required careful localization. For instance, culturally specific concepts such as “pregaming” or “designated driver” lack direct Vietnamese equivalents and were therefore reworded to convey meaning through locally relevant behaviors and contexts. Likewise, DrInC items involving morality or spirituality, such as Item 36 with the term “my spiritual or moral life” were adapted using broader, more inclusive language to ensure clarity across varying belief systems and educational backgrounds.

These thoughtful adaptations were driven by iterative expert reviews and guided by both linguistic precision and cultural humility. In doing so, the study ensured that the translated instruments were not only psychometrically sound but also emotionally and socially appropriate for Vietnamese men in rural settings, who may be less familiar with formal or Westernized phrasing.²¹

This approach enhances the tools’ validity, usability, and acceptability, contributing to greater participant engagement and more accurate data collection in future research and clinical applications. A detailed summary of key item adaptations, including original phrases, final Vietnamese versions, and adaptation rationales, is presented in Appendix 2. By documenting this process, the study provides a valuable reference for researchers engaged in cross-cultural validation of psychological and behavioral measures, particularly in low- and middle-income countries where local adaptations are crucial for meaningful implementation.

A key limitation of this study is the relatively small sample size, which restricted psychometric evaluation to internal consistency and inter-rater reliability, without the ability to assess construct validity through factor analysis. As such, the factorial structure and measurement invariance of the Vietnamese DrInC and PBSS-20 remain to be empirically tested. Future validation efforts should employ factor analytic techniques with larger and more diverse populations to confirm structural validity and enhance the robustness and generalizability of these culturally adapted tools.

Second, the study sample comprised rural men aged 35-44 years with HAU, which limits the generalizability to women, adolescents, urban populations, and individuals with alcohol dependence. Future validation studies should include these groups to extend applicability and ensure cultural relevance across the broader Vietnamese population.

Third, test-retest reliability was not assessed. Establishing temporal stability is critical for confirming measurement consistency over time, particularly in the context of intervention studies. Future validation studies should include longitudinal testing and cognitive interviews to further evaluate score stability and enhance the clarity and cultural relevance of scale items.

The validated Vietnamese DrInC and PBSS-20 instruments provide frontline healthcare workers, particularly nurses and primary care providers, with standardized, culturally appropriate tools for assessing alcohol-related problems and evaluating harm-reduction strategies. Their ease of use makes them ideal for integration into community health outreach, rural clinics, and public health screening programs, where formal diagnostic tools may be scarce.

In the nursing context, these tools can support early identification of risky drinking behaviors, facilitate patientcentered care planning, and provide metrics to monitor intervention outcomes over time. Their availability also opens new possibilities for multisite studies, cross-cultural comparisons, and international collaborations focused on alcohol harm reduction in Southeast Asia.