Cross-cultural adaptation and psychometric validation of the Portuguese Maternal Health Literacy Inventory in Pregnancy (MHELIP)

Technical Summary: Cross-cultural Adaptation and Psychometric Validation of the Portuguese Maternal Health Literacy Inventory in Pregnancy (MHELIP)

Problem Statement
Maternal health literacy (MHL) is a critical determinant of informed decision-making, self-care, and pregnancy outcomes, yet it remains an underexplored dimension of maternal health, particularly regarding interactive and critical literacy skills. While general health literacy instruments exist, they often fail to capture the specific information-processing and behavioral demands inherent to pregnancy. Although the Maternal Health Literacy Inventory in Pregnancy (MHELIP) was originally developed and validated in Iran, and subsequently adapted for Turkish populations, no culturally adapted or psychometrically tested version existed for Portuguese-speaking populations. This gap limits the ability to assess MHL vulnerability among Portuguese pregnant women and hinders the development of targeted, evidence-based interventions.

Methodology
This methodological study followed a rigorous cross-cultural adaptation and psychometric validation protocol for the Portuguese version of the MHELIP.

• Adaptation Process: The study adhered to the five-stage process recommended by Beaton et al. (2000), involving forward translation by two bilingual translators, synthesis of versions, back-translation by certified English speakers, review by an expert committee (including nursing researchers and clinicians), and pretesting with a convenience sample of 50 pregnant women. Cognitive interviews were conducted during pretesting to ensure semantic accuracy and cultural relevance.
• Participants: The final validation sample consisted of 886 pregnant women recruited via a non-probabilistic, multimodal convenience strategy from public hospitals, primary healthcare units, and private clinics in the Viseu district of Portugal. Data collection occurred between October 2023 and May 2024. Participants were required to be aged 18 or older, at least 10 weeks pregnant, and able to read Portuguese.
• Statistical Analysis: Psychometric evaluation utilized IBM SPSS (v30.0), jamovi (v2.7.34.0), and AMOS (v29.0). The analysis included:
  • Descriptive statistics and assessment of normality (skewness and kurtosis).
  • Internal consistency assessment using Cronbach's alpha and McDonald's omega ($\omega$).
  • Confirmatory Factor Analysis (CFA) using maximum likelihood estimation to evaluate the four-factor structure (Maternal Health Knowledge, Search for Maternal Health Information, Assessment of Maternal Health Information, and Maternal Health Decision-Making and Behavior).
  • Model fit evaluation using $\chi^2/df$, GFI, CFI, RMSEA, RMR, and SRMR.
  • Evaluation of convergent validity (Average Variance Extracted - AVE) and discriminant validity (Heterotrait–Monotrait ratio of correlations - HTMT).
  • A second-order hierarchical model was specified to test the multidimensional nature of the construct.

Key Contributions and Results
The study successfully validated a shortened, culturally adapted version of the MHELIP for the Portuguese context.

• Item Refinement and Structure: Through iterative CFA and item refinement based on factor loadings and reliability coefficients, 25 items were removed from the original 48-item scale. The final instrument comprises 23 items while retaining the original four-factor theoretical structure.
• Model Fit: The final second-order hierarchical model demonstrated acceptable fit indices ($\chi^2/df = 4.41$; GFI = 0.91; CFI = 0.90; RMSEA = 0.062; SRMR = 0.047). This supports the conceptualization of maternal health literacy as a multidimensional construct composed of four interrelated domains.
• Reliability: The total scale exhibited excellent internal consistency ($\omega = 0.92$). Subscale reliability was acceptable to good, with $\omega$ values ranging from 0.75 to 0.82.
• Validity:
  • Convergent Validity: Three of the four domains (Knowledge, Search, and Assessment) met the recommended AVE threshold (>0.50). The "Maternal Health Decision-Making and Behavior" domain showed an AVE of 0.389, indicating a need for cautious interpretation regarding this specific domain.
  • Discriminant Validity: Generally satisfactory, with most HTMT values below the 0.90 threshold. However, a high conceptual overlap was observed between the "Search for Maternal Health Information" and "Assessment of Maternal Health Information" domains (HTMT = 0.912), suggesting these dimensions are closely related.
• Factor Loadings: All standardized factor loadings in the final model ranged from 0.62 to 0.86, indicating moderate to strong associations between items and their respective latent factors.

Significance and Claims
The authors claim that the Portuguese MHELIP offers a valid, reliable, and theoretically coherent measure of maternal health literacy that is shorter and more feasible for use in antenatal care settings and research than the original 48-item version. The study emphasizes that the reduction in items, driven by psychometric criteria and cultural adaptation, does not compromise the theoretical framework.

The paper positions the instrument as a tool to support evidence-based interventions and public health strategies aimed at improving informed decision-making during the perinatal period. The authors note that the brevity of the scale may facilitate its integration into routine clinical practice where time is limited. However, the study modestly acknowledges limitations, including the cross-sectional design (precluding temporal stability assessment), the recruitment from a single geographic region (limiting generalizability), and the need for further investigation into the predictive validity and performance in more diverse Portuguese populations. Future studies are recommended to examine test-retest reliability and the instrument's ability to inform targeted interventions.