Turkish cross-cultural adaptation and construct validation of the HyperEmesis Level Prediction (HELP) Score in pregnancy: a multicenter cross-sectional study

The HyperEmesis Level Prediction (HELP) Score was developed to assess symptom severity and broader clinical burden in nausea and vomiting of pregnancy (NVP) and hyperemesis gravidarum (HG). We translated and culturally adapted HELP into Turkish (TR-HELP) and evaluated construct validity in a multicenter pregnancy cohort. This cross-sectional study included 453 pregnant women from seven Turkish centers between February and July 2026. Following forward–backward translation, independent expert appraisal, and cognitive debriefing, participants completed TR-HELP followed by the Turkish Pregnancy-Unique Quantification of Emesis and Nausea-24 (PUQE-24) during the same visit. Convergent construct validity was assessed using pooled Spearman and center-adjusted partial rank correlations. Internal consistency was estimated descriptively using Cronbach’s alpha. Category agreement and cross-sectional associations with NVP/HG-related hospitalization, emergency department attendance, and intravenous (IV) fluid therapy during the current pregnancy up to assessment were supportive analyses. Post-hoc receiver operating characteristic (ROC) comparisons were repeated among first-trimester participants. Of 453 participants, 240 (53.0%), 84 (18.5%), and 129 (28.5%) were assessed in the first, second, and third trimesters; 36 (7.9%) reported no NVP onset by assessment. Windsor-defined HG was recorded in 166 (36.6%). Pooled and center-adjusted correlations were 0.863 (95% confidence interval [CI] 0.832–0.886) and 0.758 (95% CI 0.709–0.799), respectively. Excluding PUQE-24 floor scores yielded ρ = 0.788 (95% CI 0.742–0.827). Cronbach’s α was 0.915 (95% CI 0.904–0.925). Exploratory weighted κ between non-equivalent categories was 0.627 (95% CI 0.566–0.685). For cumulative hospitalization, areas under the ROC curves (AUCs) were 0.864 for TR-HELP and 0.816 for PUQE-24 (ΔAUC 0.048, 95% CI 0.012–0.086; p = 0.013). In the first trimester ( n = 240; 46 hospitalizations), corresponding AUCs were 0.829 and 0.797 (ΔAUC 0.032, 95% CI − 0.021 to 0.086; p = 0.246). The findings support convergent construct validity of TR-HELP with PUQE-24 in a cohort including women with and without HG. Cumulative healthcare-utilization associations do not establish HG diagnostic validity, prospective prediction, or clinical superiority. Structural validity, test–retest reliability, measurement error, responsiveness, and clinically meaningful change require further evaluation.