This study’s results support the validity of a Brazilian Portuguese version of the “Knowledge of Gestational Diabetes (GDM)” questionnaire for women with GDM, indicating that it could be reliably used in future studies.
The reliability of the instrument was assessed using several methods. The internal consistency (Cronbach’s α) found in our study was 0.81, which is adequate [19], suggesting that despite the modifications, the scale remained internally consistent. However, the Cronbach’s α of the Knowledge of Gestational Diabetes domains, if analyzed individually, was < 0.7 for each domain, indicating the unreliability of the subscales when used separately. Reliability reduction when separately analyzing domains has also been observed in other studies [13].
Most questions in the original instrument have one correct answer and are scored as either correct or incorrect. Four questions had more than one correct answer and were scored as correct (all correct answers identified) or incorrect (did not identify all correct answers). Consequently, since partially correct responses were treated as incorrect in the scoring process, this criterion resulted in a substantial number of responses being categorized as incorrect in the questions with more than one correct answer, leading to a notably high percentage of inaccuracies. Despite the high prevalence of incorrect responses, these questions were retained for analysis. Some potential solutions could include considering a partially correct answer as acceptable or reformulating the questions to ensure that they offer only one correct option. Question 22, about exercise intensity, also showed a high percentage of incorrect responses, despite having only one correct option. One potential explanation is that it pertains to physical activity, and pregnant women often lack professional guidance on the appropriate intensity of exercise. Consequently, there is confusion regarding what constitutes “mild” (“leve”) or “moderate” (“moderado”) exercise.
In the item-total correlation analysis, alpha values after the exclusion of a particular item showed no significant change, with the satisfactory alpha value remaining unchanged. We did not consider question 2, with no recorded correct answers, to perform the confirmatory factor analysis. According to the confirmatory factor analysis, the test indicated that the model was well fitted. Furthermore, the two independent raters agreed on the total scores.
The results of internal reliability, item-total correlation, and intraclass correlation are consistent with studies that have also validated instruments [20–22], despite the difficulty in finding studies assessing knowledge about GDM. This fact reinforces the importance of noninvasive assessment instruments aimed at the interviewed public. However, the original questionnaire does not indicate a score to determine good knowledge. Therefore, our study also did not categorize the obtained scores.
Given the extensive and multifaceted consequences of a GDM diagnosis, it is challenging to assess adequate control using only a single capillary glucose monitoring measure. Continuous glucose monitoring, a new method addressed in some studies, offers an alternative for accurately assessing glycemic control. This method provides more precise values and has shown promising results [23, 24]. We recognize that analyzing the relationship between scale scores and glycemic control (predictive validity) should be useful; however, we cannot accurately assess glycemic control.
There are several limitations to consider in this study. First, the test-retest evaluation period ranged from 1 to 12 weeks, aligning with the typical frequency of multidisciplinary consultations for pregnant women. The extended interval was due to women missing appointments or not having available time to respond to the questionnaire. This timeline may explain the enhanced learning and discrepancies in responses between the two interviews. Additionally, this period coincided with various physiological changes in the participants’ bodies, concerns about the baby, and the need to adapt to numerous new situations, all of which may have impacted their willingness to participate in the questionnaire. Within this demographic, the onset of diabetes represents only one aspect of the multitude of changes experienced during pregnancy.