The aim of this study was to cross-culturally adapt and validate the Brazilian Portuguese version of the (GDMKQ). This study was conducted in a high-risk pregnancy outpatient clinic of two university hospitals in southern Brazil (Hospital Materno Infantil Presidente Vargas, HMIPV and Hospital de Clínicas de Porto Alegre, HCPA) from January to December 2023, in accordance with the Declaration of Helsinki. This study was approved by the Scientific Committee and Research Ethics Committee of HCPA (Brazil) (Certificate of Presentation for Ethical Appreciation 62984522.9.1001.5327) and the Scientific Committee and Research Ethics Committee of HMIPV (Certificate of Presentation for Ethical Appreciation 62984522.9.2001.5329). Pregnant women with a diagnosis of GDM or with diabetes diagnosed during pregnancy were selected from the medical record database of the high-risk antenatal clinics at both hospitals. Those who met the eligibility criteria were invited to participate in the study and gave their consent by signing the informed consent form.

Pregnant women were considered eligible if they were 18 years of age or older, had a diagnosis of diabetes during pregnancy or had been diagnosed with GDM at least 1 month before, identified with a Brazilian cultural identity, and had consulted with at least one healthcare professional. For the diagnosis of GDM, the criteria were fasting blood glucose levels between 92 and 125 mg/dL, 1-h plasma glucose levels of 180 mg/dL or higher, or 2-h plasma glucose levels between 153 and 199 mg/dL. For overt diabetes, the criteria were a fasting glucose levels greater than 126 mg/dL or a random blood glucose level greater than 200 mg/dL [1]. Exclusion criteria included known pre-pregnancy diabetes, previous GDM, pregnant women without blood glucose records, and any communication or comprehension barriers, such as mental disorders or illiteracy. Interested eligible patients signed an informed consent form. Inclusion and exclusion criteria were defined by the investigators LBP, BDS and PMB.

Clinical and demographic data, including the women’s age, place of residence, level of education, current gestational age, self-monitoring of capillary glucose results, insulin use, current medications, diagnostic test results (transcribed if performed elsewhere or in the hospital), and family history of diabetes were obtained from the patient’s medical records.

The authors of the original questionnaire were asked for permission by email, and the study began only after their approval. The GDMKQ, which was originally validated by Carolan-Olah and Vasilevsli in 2021, is a structured interview designed to assess pregnant women’s knowledge of GDM (Additional file 1). The questionnaire consists of 33 questions covering the following domains: 1. Knowledge of GDM, including normal blood glucose levels and the effects of GDM on mothers and babies; 2. Knowledge of nutritional values and food choices; 3. Knowledge of principles of GDM self-management, including blood glucose monitoring, physical activity, and dietary habits. Answers are numerically coded, and each correct answer is scored as 1 point. Most questions have one correct answer and are scored as either true or false. Four questions have more than one correct answer and are scored as correct (all correct answers identified) or incorrect (all correct answers not identified) [10].

The GDMKQ underwent a multi-step process, including translation and content validity assessment. After administration to participants, internal consistency, item-total correlation, and intraclass correlation were assessed. Confirmatory factor analysis was also used to ensure validity.

The initial translation of the original instrument into Brazilian Portuguese was done by two independent translators who were native speakers of Brazilian Portuguese, fluent in both languages, and from different professional backgrounds than the researchers. In the synthesis process, a consensus was reached from the resulting versions, and words and phrases that showed differences were readapted [11]. Content validity was assessed by experts in the field and by the target audience.

A committee of experts in the field was assembled to produce a final version of the modified instrument. This committee included a Pharm.D. pharmacist, a Ph.D. pharmacist and a Ph.D. endocrinologist, all of whom are bilingual [12]. The committee carefully reviewed each item of the instruments to better adapt them, ensure that the translation was fully comprehensible, and verify cross-cultural equivalence between the source and final versions. Semantic, idiomatic, empirical, and conceptual aspects were considered [12]. The synthetic version of the instrument was back-translated by a translator who is a native speaker of the source language and fluent in the target language. This translator had no prior knowledge of the instrument. Based on the translation and back-translation, a final version of the instrument was obtained [13].

The final stage of the adaptation process was the assessment of content validity by the target population, also known as cognitive debriefing or pretesting, in which the instrument was administered to 30 women diagnosed with GDM or with diabetes detected during pregnancy. A sample size of 30 to 40 individuals is recommended for pretesting in questionnaire validation and cultural adaptations studies [11, 13]. Participants completed the questionnaire and were interviewed to assess the comprehensibility, interpretability, and cultural relevance of the translation. Cognitive debriefing was used to obtain feedback from participants and confirm the acceptability of the translation [11,12,13]. The content validity coefficient (CVC) was determined by calculating the items related to the clarity, appropriateness, and comprehensibility of the questions, which were rated on a scale of 1 to 5. A CVC ≥ 0.70 for each item and for the instrument as a whole was used as the cut-off point to determine satisfactory levels of language clarity and appropriateness [14]. The instrument was administered to pregnant women at least 4 weeks after their diagnosis and with an expected delivery date of at least 1 month. Participants’ privacy was protected during the interviews.

Upon completion of this cross-cultural adaptation, and based on a suggestion regarding validation sample size [15], a further 125 patients were selected to complete the final version in order to determine the reliability and validity of the instrument. Intraclass correlation was established by two independent interviews of 57 participants within 1–12 weeks of the previous assessment. The number of second interviews was adequate according to previous instrument validation studies [16, 17].

The Statistical Package for Social Science Professional program version 20.0 (IBM Corp., Armonk, NY, USA) and the R package lavaan (version 0.6–17) were used to perform the analyses.

Interrater agreement was measured using the intraclass correlation coefficient. Intraclass correlation values below 0.5 indicate poor reliability, values between 0.5 and 0.75 indicate moderate reliability, values between 0.75 and 0.9 indicate good reliability, and values above 0.90 indicate excellent reliability [18]. Cronbach’s alpha (α) was calculated to measure internal consistency, and values above 0.7 were considered acceptable [12]. In addition, the effect of removing each question on the Cronbach’s alpha value, which may result in the exclusion of a particular item after confirmatory factor analysis, was examined. Item-total correlations were calculated. The inclusion of items with values above 0.2 is considered appropriate, and negative values, indicating a negative correlation, may represent a question that is inversely proportional to the others [19, 20]. The confirmatory factor analysis model was fitted using diagonal weighted least squares with robust standard errors. The comparative fit index (CFI) was used to analyze the model fit by examining the discrepancy between the data and the hypothesized model, taking into account sample size issues inherent in the chi-squared test and the normed fit index [21].