Cross-cultural validation of the Chinese version of Disorder-Specific Intolerance of Uncertainty (DSIU): psychometric properties in female college students with premenstrual syndrome (PMS).

Recruitment of Chinese female undergraduate students from three comprehensive universities in Jinzhou, Liaoning Province (Jinzhou Medical University, Bohai University, and Liaoning University of Technology) who meet the diagnostic criteria for premenstrual syndrome (PMS) through convenience sampling. In addition, the inclusion criteria for the study were PMS score ≥ 6 informed consent, and voluntary participants. The exclusion criteria: (1) Current or past diagnoses of organic mental disorders (such as brain injuries, neurodegenerative diseases) or other mental disorders (such as depression, anxiety, bipolar disorder) made by a psychiatrist following the International Classification of Diseases (ICD-11). (2) Patients with a clear diagnosis of thyroid dysfunction (TSH levels outside the normal range: 0.4-4.0 mIU/L) or diabetes (fasting blood glucose ≥ 7.0 mmol/L or HbA1c ≥ 6.5%) based on self-reports or medical records; those who have received thyroid hormone replacement therapy, antithyroid drugs, or hypoglycemic agents (including insulin) in the past three months. (3) Patients with a history of organic diseases diagnosed through pelvic ultrasound or gynecological examinations (such as uterine fibroids ≥ 5 cm, endometriosis, polycystic ovary syndrome) or malignant tumors of the reproductive system are excluded. (4) Patients who have been using medications such as hormones and contraceptives for the treatment of PMS within the past three months. The DSIU scale underwent a rigorous process of cross-cultural adaptation, the steps of which are illustrated in Fig.  1 . Fig. 1 DSIU scale cross-cultural adaptation flowchart DSIU scale cross-cultural adaptation flowchart Permission to use, translate, and modify the instrument was obtained from the original authors via e-mail. This study used Brislin’s model and Delphi method for translation and cross-cultural adaptation. The Brislin model (forward translation → back-translation → expert synthesis) provides an operationalized chain of conceptual equivalence; The Delphi method specifically addresses the problem of identifying culture-specific manifestations by aggregating the consensus of cross-domain experts through multiple iterations (3 rounds). First, the English version of the DSIU was independently translated into two different Chinese versions by two bilingual medical graduate students with expertise in English. Then, a third nursing graduate student who was not a participant in the translation compared the differing versions and brought together all translators and researchers to discuss and synthesize the comparisons into a first draft of the Chinese version (A). When putting the scale back, we invited two native Chinese speakers, (one with a master’s degree in nursing from the UK and the other with a master’s degree in linguistics from Australia) to reverse translate the first draft of the Chinese version of (B). Finally, it was reviewed by a panel of experts. Finally, two psychiatric educators (one a local psychiatric educator with 10 years of experience and the other a psychiatric educator with study abroad experience) and two psychoeducators (one a psychologist with local experience and the other a psychoeducator with study abroad experience), as well as two psychiatrists familiar with the concepts of empowerment and with relevant research experience (one a local chief physician with extensive experience, the other is a supervising physician with study abroad experience) formed an expert panel to compare the content equivalence of the original scale, the first draft of the Chinese version, and the revised English version, revise controversial items, and make rigorous cross-cultural adjustments. The final version of the C-DSIU (C) is more compatible with the Chinese language environment. A convenience sample of 40 PMS female college students was selected for the pretest to test the comprehensibility of the items. The scale took about 8 to 15 min to complete. At this point, a Chinese version of the DSIU has been developed, which was found to be fluent and easy to understand. The first part is the Premenstrual Syndrome Scale (PMSS), which was developed by John Bancroft [ 31 ]. The diagnosis is based on the presence of 5 out of 12 symptoms within 14 days before the last menstrual period and up to the time of the period, 1 of which must be a symptom of emotional abnormality (e.g., nervousness, depression, anxiety, fidgetiness, inattentiveness, agitation). Each symptom was scored on a 4-point scale. The total score was the sum of the scores of each item, ranging from 0 ~ 36, with a total score < 6 indicating no symptoms, and a total score ≥ 6 determining the presence of PMS, i.e., positive for PMS. The overall internal consistency reliability of the questionnaire was good, with a Cronbach’αof 0.92 and a retest reliability of 0.73. The second section included demographic questions as demographic information such as education level, gender, age, ethnicity, and the Disease-Specific Intolerable Uncertainty Scale (DSIU) developed by Thibodeau et al. in 2015 [ 30 ]. The DSIU consists of 24 elements in eight dimensions: generalized anxiety disorder (IU-GAD), obsessive-compulsive disorder (IU-OCD), health anxiety (IU-HA), social anxiety (IU-SA), posttraumatic stress disorder (IU-PTSD), panic disorder (IU-PD), major depressive disorder (IU-MDD) and specific phobia (IU- Phobia). The aim is to evaluate which uncertainties result in intolerance. A 5-point Likert scale (0 = “not at all”, 4 = “Extremely”) was used with higher ratings indicative of increased disease-specific uncertainty tolerance. To enhance the reliability of self-reported data, this study employed Attention Check Items (ACIs) to identify careless or invalid responses. This approach is grounded in empirical evidence indicating that participants who fail attention checks are more likely to exhibit random or patterned responding, significantly compromising data quality [ 28 , 29 , 32 ]. Direct-instruction ACIs: Required strict adherence to textual instructions to select a specified option (e.g., “Please select ‘Not at all’ ”). Logic-based ACIs: Assessed basic cognitive consistency through commonsense questions (e.g., “I am certain which country I am from”). All ACIs were embedded within DSIU scale items, using the same 5-point Likert scale as the DSIU (0 = “Not at all”, 4 = “Extremely”) to prevent participant forewarning. Exclusion criterion: Data from participants failing any one ACI were flagged as invalid and excluded from analysis. This study complied with the Declaration of Helsinki of the World Medical Association [ 33 ]. And it was approved by the Ethics Committee of Jinzhou Medical University (ID: JZMULL2025039). Each participant gave their free and informed consent to participate in the study. Furthermore, all participant data was treated with the utmost confidentiality and utilized exclusively for this research. Data were collected from March through October 2024. The researcher was trained in the use of uniform instructions before data collection. The researchers had one-on-one contact with all students and received verbal and written information on the purpose of this survey and instructions on how to complete the questionnaire. Female undergraduates were asked to complete the Premenstrual Syndrome Scale and the Disease-Specific Intolerance of Uncertainty Scale if the total PMS score was ≥ 6 and 1 item had to be a symptom of an emotional abnormality (e.g., nervousness, depression, anxiety, fidgetiness, inattentiveness, agitation). Female college students were asked to complete the questionnaire in 5 min to reflect their disease-specific intolerance to uncertainty. In addition, to assess the reliability of the retest, 50 students were randomly selected to add contact information (WeChat) with their consent. These 50 students were contacted again two weeks later through the medium of the WeChat medium to complete the questionnaire. SPSS 26.0 was used to enter and analyze the data. Each test had two tails and the α = 0.05 were two-sided. The following specific statistical techniques were used: means and standard deviations were used to communicate quantitative data and frequencies and percentages were used to express qualitative data. Item distribution, critical ratio, and Spearman’s correlation coefficient methods were used for item analysis of the scale. Reliability was evaluated using the Cronbach α coefficient, split-half reliability, and the two-week retest intracorrelation coefficient (ICC). Validity was assessed using the content validity index (CVI), exploratory factor analysis (EFA), confirmatory factor analysis (CFA), and criterion-related validity assessment. The scale elements were analyzed using correlation analysis and the crucial ratio decision value (CR) technique. The entire DSIU scale score was ranked from high to low using the CR value technique, with the top 27% representing a high group and the bottom 27% representing a low group. Items that had a CR value of less than three and no statistically significant differences were removed [ 34 ]. The t-value obtained by the t-test of independent samples of high and low subgroups was used as the CR to evaluate the differentiation of the items. Pearson’s correlation analysis was used to determine the correlation coefficient between the total score on the scale and the score of each item to assess the consistency of its measured attributes and exclude items with correlation coefficients less than 0.40 [ 35 ]. Six senior clinical experts were invited to rate each item of the scale on a 4-component scale (1–4 indicating “no correlation”, “low correlation”, “strong correlation”, and “very strong correlation”). The content validity index (CVI) [ 36 ] was used to evaluate the content validity of the scale, including the item-level content validity index (I-CVI) and the universal agreement scale level content validity index (S-CVI). When I-CVI ≥ 0.78 and S-CVU ≥ 0.80, the scale has good content validity [ 37 ]. The structural validity of the Chinese version of the self-rating scale was examined using the first dataset ( N  = 641). This dataset was randomly divided into Sample 1 (N1 = 320) and Sample 2 (N 2  = 321) using Excel’s RAND, RANK, and ROUNDUP functions, designated for Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA), respectively. The suitability of factor analysis for Sample 1 was assessed using Bartlett’s test of sphericity [ 36 ] and the Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy [ 37 ]. A KMO value greater than 0.60 and a statistically significant Bartlett’s test ( p  < 0.05) indicated that the data were appropriate for factor analysis. Exploratory Factor Analysis (EFA) was used to explore the dimensional structure of depressive symptoms in target cultural groups. EFA was conducted on Sample 1 using Principal Component Analysis (PCA) extraction [ 38 ] to evaluate the scale’s internal structure. The extracted factors were rotated orthogonally using the varimax method. Varimax rotation, the most widely used orthogonal rotation strategy, maximizes the variance of factor loadings while simplifying factor interpretation by minimizing their complexity [ 39 ]. The number of factors to retain was determined through a comprehensive judgment based on the scree plot, cumulative percentage of variance explained, and eigenvalues exceeding 1.0. The CFA was used to assess the fit of the structural model in sample 2 (n2 = 321 [ 38 ]. The CFA was validated for cross-cultural equivalence by means of a multicohort invariance test. In this study, the following indicators were assessed: The degree of freedom (df), the chi-square test (χ 2 ), the tucks-lewis index (TLI), the comparative fit index (CFI), the incremental fit index (IFI), the goodness of fit index (GFI), and RMSEA are all mentioned in reference [ 39 ]. The models deemed reasonably suited were those with χ 2 /df < 3, RMSEA and RMR  0.90 [ 41 ]. Next, the average variance extracted (AVE) and the CR values are calculated using the standardized factor load. The composite reliability value (CR) and AVE indicators assess the reliability and validity of the scale within the framework of the CFA, which is more adaptable to the complexity of multidimensional measurement models than traditional methods. AVE measures the proportion of variation in the question items explained by the factors, and √AVE is used to test discriminant validity. An AVE value of greater than 0.50 indicates strong convergence validity, while a CR value of greater than 0.70 indicates strong composite reliability [ 42 ]. Ultimately, the discriminant validity of the model was evaluated based on the √AVE output of each project and the AMOS 24.0 correlation coefficient. The discriminant validity of the model is good if √AVE is larger than the correlation coefficient between items. Split-half reliability, Cronbach α coefficient, and McDonald’s Omega coefficient were used to assess the scale’s internal consistency. Cronbach α coefficient was computed after the removal of each item to determine whether the scale’s internal consistency was enhanced by doing so. Based on the item order, the reliability half-break is computed and the scale is split into two sections. The correlation coefficient between the two components is determined using Spearman’s correlation analysis. One of the better methods for determining reliability is the McDonald’s Omega coefficient [ 43 , 44 ], where a value of 0.80 or higher indicates good internal consistency [ 45 ]. Good internal consistency is defined as having a Cronbach α coefficient and a half-fold coefficient ≥ 0.70 [ 46 ]. To assess the stability of the scale, the reliability of the retest, or the intragroup correlation coefficient (ICC), was employed. The ICC is used to test the reliability of the scale in the time dimension and between raters, providing data support for clinical follow-up. Fifty female college students were given the opportunity to retake the DSIU in Chinese two weeks after they finished the first test, providing them with contact details. Measurement of the difference between the two tests and evaluation of the correlation between the two tests are done using Spearman correlation analysis. The reliability of the retest is good if the correlation coefficient is greater than 0.70 [ 37 ].

The survey included 641 female college students of PMS. The average age was 21.46 years. See Table  1 for additional details. Table 1 Demographic characteristics ( N  = 641) Variable N % Age 22 21.6 33.4 Ethnicity Han 405 63.33 Man 204 31.83 Others 31 4.84 Education Junior female college students 138 21.53 Undergraduates 432 67.39 Postgraduates 71 11.08 Religious beliefs Without 590 92.04 With 51 7.96 Major Medical science 346 53.98 Non-medical 295 46.02 menstrual regularity Regular 407 63.49 Irregular 234 36.51 Number of menstrual days ≤ 3 89 13.88 4 ~ 7 495 77.22 >7 57 8.90 Dysmenorrhea No 240 37.44 Yes 401 62.56 Demographic characteristics ( N  = 641) Based on expert opinions and feedback from university students who initially completed the scale, this study revised the scale with the original authors’ permission. Unlike English, the complexity of Chinese translation is not only reflected in the structure of the language but also involves multiple aspects such as culture, habits, and expressions. Therefore, adjustments were made to the word order. The second item in the IU-SAD, “When I am unsure of myself in a social situation, I am uncomfortable and unable to be myself,” was changed to “I can’t be myself in social situations when I’m not sure if I’m going to embarrass myself”. Secondly, the first item on IU-Phobia “I should avoid my fears and stay safe rather than to face them and be unsure of the consequences” was changed to “For me, it’s better to avoid fear and stay safe than to face fear and be uncertain about the consequences”. And change the IU-Phobia preface, “The next series of items is about the things or situations you fear the most. (heights, snakes, flying, elevators). Please answer the following questions, which require you to visualize your greatest fear even if you are not very afraid of anything” to “Assume that one of heights, snakes, or flying is your greatest fear (even if you aren’t very afraid of anything, you need to visualize your greatest fear). Depending on your situation at the time, answer the following questions.” After cross-cultural debugging and pre-study, none of the entries in the scale were deleted. Table  2 displays the participants’ mean scores (SD) for each dimension in the updated Chinese version of the DSIU. Table 2 Descriptive statistics of the Chinese version of the DSIU scale and reliability and validity analysis of each dimension Total ( n  = 641) EFA (n1 = 320) CFA (n2 = 321) M SD SK K Cronbach α Spill-half reliability Cronbach α Spill-half reliability AVE CR √AVE IU- GAD 5.79 2.18 -0.11 -0.15 0.783 0.772 0.780 0.781 0.563 0.794 0.750 IU- SAD 4.44 2.21 -0.20 -0.47 0.793 0.788 0.770 0.759 0.590 0.811 0.768 IU- OCD 4.76 2.66 0.11. -0.58 0.859 0.842 0.847 0.818 0.694 0.872 0.833 IU- HA 3.96 2.15 -0.06 -0.27 0.786 0.803 0.760 0.804 0.581 0.807 0.762 IU-PTSD 3.43 2.14 0.30 -0.27 0.802 0.798 0.804 0.792 0.574 0.801 0.758 IU- PD 1.98 1.72 0.83 1.05 0.638 0.666 0.547 0.583 0.441 0.699 0.664 IU- MDD 2.70 1.78 1.07 2.02 0.737 0.774 0.696 0.743 0.516 0.761 0.718 IU- Phobia 3.88 2.53 0.24 -0.68 0.784 0.773 0.751 0.754 0.613 0.824 0.783 Descriptive statistics of the Chinese version of the DSIU scale and reliability and validity analysis of each dimension The disparities in the Chinese version of DSIU scores between the high and low groups were statistically significant ( p  < 0.001), with CR ranging from 6.044 to 21.360, indicating good discriminatory power to assess responses from various subjects effectively. The correlation coefficient between the item and the total score of the scale ranged from 0.416 to 0.661 ( p  < 0.001), all of which exceeded the threshold of 0.40 except for item 16 [ 33 ], see Table  3 . Table 3 Item analysis of the C-DSIU (N 1 = 641) CR Correlation coefficient between item and total score P -value Cronbach’s Alpha if the item is deleted T1 12.186 0.546 <0.001 0.883 T2 14.304 0.599 <0.001 0.882 T3 14.709 0.571 <0.001 0.883 T4 15.793 0.613 <0.001 0.882 T5 14.906 0.568 <0.001 0.883 T6 17.307 0.605 <0.001 0.882 T7 21.360 0.661 <0.001 0.880 T8 20.037 0.632 <0.001 0.881 T9 18.987 0.585 <0.001 0.883 T10 12.900 0.539 <0.001 0.884 T11 10.17 0.485 <0.001 0.885 T12 11.497 0.537 <0.001 0.884 T13 12.678 0.523 <0.001 0.884 T14 12.733 0.532 <0.001 0.884 T15 10.322 0.464 <0.001 0.886 T16 6.044 0.344 <0.001 0.888 T17 7.907 0.416 <0.001 0.887 T18 9.593 0.492 <0.001 0.885 T19 8.808 0.467 <0.001 0.885 T20 7.791 0.482 <0.001 0.885 T21 7.344 0.472 <0.001 0.885 T22 10.892 0.452 <0.001 0.887 T23 15.228 0.565 <0.001 0.883 T24 14.091 0.531 <0.001 0.884 Item analysis of the C-DSIU (N 1 = 641) The content validity of the Chinese version of the DSIU (C-DSIU) was evaluated through an expert evaluation [ 47 ]. Analysis of the content validity indicated an I-CVI of 0.92 to 1.00 and an S-CVI of 0.83 for the Chinese DSIU. The consensus among experts was that the scale’s elements and overall content align well with its intended measurement objectives. Exploratory factor analysis (EFA) The EFA in Table  2 (N1 = 320), we initially assessed factoriality and Cronbach’s α. Bartlett’s sphericity test for sphericity yielded a statistically significant result (χ 2  = 2482.532, p  < 0.001). The KMO value exceeded the minimum acceptable threshold of 0.60 [ 48 ], suggesting strong correlations between the variables. Cronbach α stood at 0.90(N1 = 320), and the dimensional Cronbach α (refer to Table  2 for EFA) was considered suitable for factor analysis. EFA employed principal APC and maximum-variance orthogonal rotation, unveiling eight common factors with eigenvalues surpassing 1. These factors collectively contributed to 69.99% of the total variance (Table  4 ), with all items loading above 0.50 on each dimension (Table  5 ). This implies that every item in the Chinese version of DSIU aligns with each of the eight dimensions, affirming the construct validity of the eight-dimensional variable. The EFA results for DSIU are depicted in Fig.  2 , illustrating the factor structure that reinforces the plausibility of the dimensional makeup. Notably, the decline in the steepness of the scree plot becomes less prominent after the 8th point (Fig.  2 ). Table 4 Total variance explained Component Initial Eigenvalues Extraction Sums of Squared Loadings Rotation Sums of Squared Loadings Total % of Variance Cumulative % Total % of Variance Cumulative % Total % of Variance Cumulative % 1 5.605 23.355 23.355 5.605 23.355 23.355 2.312 9.633 9.633 2 2.399 9.996 33.350 2.399 9.996 33.350 2.266 9.443 19.076 3 2.206 9.193 42.544 2.206 9.193 42.544 2.232 9.300 28.376 4 1.596 6.651 49.195 1.596 6.651 49.195 2.215 9.228 37.604 5 1.390 5.792 54.988 1.390 5.792 54.988 2.115 8.814 46.419 6 1.315 5.480 60.468 1.315 5.480 60.468 2.106 8.774 55.193 7 1.225 5.105 65.573 1.225 5.105 65.573 1.997 8.322 63.515 8 1.060 4.417 69.990 1.060 4.417 69.990 1.554 6.475 69.990 Extraction Method: Principal Component Analysis Total variance explained Extraction Method: Principal Component Analysis Table 5 Results of the exploratory factor analysis of the Chinese version of the DSIU ( n  = 320) Component 1 2 3 4 5 6 7 8 T8 0.831 T9 0.825 T7 0.764 T14 0.858 T15 0.829 T13 0.776 T12 0.800 T11 0.783 T10 0.730 T3 0.787 T2 0.763 T1 0.755 T5 0.829 T4 0.755 T6 0.727 T21 0.790 T20 0.756 T19 0.745 T23 0.811 T24 0.806 T22 0.711 T17 0.778 T16 0.656 T18 0.641 Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization Results of the exploratory factor analysis of the Chinese version of the DSIU ( n  = 320) Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization Fig. 2 Scree plot of exploratory factor analysis for the Chinese version of DSIU (N1 = 320) Scree plot of exploratory factor analysis for the Chinese version of DSIU (N1 = 320) Confirmatory factor analysis (CFA) Before the CFA was performed in Table  2 (n2 = 321) to validate the EFA-derived model. The model fit indicators satisfied the goodness-of-fit criteria. The results are presented in Table  6 . The data model for the eight-dimensional structure fits well, indicating the model’s validity in measuring the variables. The graphical representation of the CFA can be seen in Fig.  3 . The factor loadings for each item ranged from 0.55 to 0.89. Table 6 Goodness-of-fit indexes of the one-factor model for the Chinese version of DIUS (n2 = 321) χ 2 df χ 2 /df RMR GFI RFI IFI TLI CFI RMSEA 376.468 224 1.681 0.033 0.913 0.927 0.955 0.944 0.955 0.046 Goodness-of-fit indexes of the one-factor model for the Chinese version of DIUS (n2 = 321) Fig. 3 Hypothesized confirmatory factor analysis model of the C-DSIU Hypothesized confirmatory factor analysis model of the C-DSIU Results of the measurement invariance testing across age, ethnicity, education level, and major using multi-group CFA are presented in Table  6 . The baseline model demonstrated acceptable fit across all groups: age (TLI = 0.884, RMSEA = 0.040), ethnicity (TLI = 0.932, RMSEA = 0.041), education level (TLI = 0.848, RMSEA = 0.046), and major (TLI = 0.935, RMSEA = 0.040). These results indicate equivalent factor structures across key demographic variables and establish a baseline model for further invariance testing. Following established criteria [ 49 , 50 ], measurement invariance was confirmed when ΔTLI < 0.010 and ΔRMSEA < 0.015. The Chinese DSIU demonstrated full measurement invariance (metric, scalar, and residual) across all four sociodemographic characteristics, confirming equivalent factor structures, factor loadings, and item intercepts across age, ethnicity, education level, and major. This supports the scale’s validity for assessing and comparing the influence of these demographic variables on illness-related intolerance of uncertainty among female university students with PMS (See Table  7 ). Table 7 Measurement invariance of the Chinese version of the DSIU across age, ethnicity, education, and major (n2 = 321) Model χ2 Δχ2 df Δdf P RMSEA ΔRMSEA TLI ΔTLI Age Unconstrained 1010.023 672 0.040 0.884 Measurement weights 1047.179 37.156 704 32 0.244 0.039 0.001 0.888 0.004 Structural covariances 1175.974 165.951 776 104 0.001 0.040 0.000 0.882 0.002 Measurement residuals 1263.293 253.270 824 152 0.000 0.041 0.001 0.877 0.007 Ethnicity Unconstrained 691.356 448 0.041 0.932 Measurement weights 710.050 18.694 464 16 0.283 0.041 0.000 0.931 0.001 Structural covariances 759.694 68.338 500 52 0.064 0.040 0.001 0.927 0.005 Measurement residuals 790.539 99.183 524 76 0.039 0.040 0.001 0.924 0.008 Education Unconstrained 1122.802 672 0.046 0.848 Measurement weights 1174.006 51.204 704 32 0.017 0.046 0.000 0.849 0.001 Structural covariances 1264.351 141.549 776 104 0.009 0.045 0.001 0.857 0.009 Measurement residuals 1322.770 199.968 824 152 0.006 0.044 0.002 0.856 0.008 Major Unconstrained 678.232 448 0.040 0.935 Measurement weights 691.370 13.138 464 16 0.663 0.039 0.001 0.936 0.001 Structural covariances 732.927 54.695 500 52 0.373 0.038 0.002 0.933 0.002 Measurement residuals 758.507 80.275 524 76 0.347 0.038 0.002 0.933 0.002 Measurement invariance of the Chinese version of the DSIU across age, ethnicity, education, and major (n2 = 321) As shown in Table  2 of CFA , the AVE values of the model were 0.44–0.69, which were greater than 0.5 except for the IU-PD dimensions. The CR values were 0.70–0.87, indicating that the model had good convergent validity and composite reliability [ 42 ]. Regarding discriminant validity, the square root of AVE was calculated for each structure and compared with the correlation between the structures. The √AVE values were 0.66–0.83, greater than the scale items’ correlation coefficients (0.26–0.58, p  < 0.001), indicating that the model has good discriminant validity (See Table  8 ). Table 8 Discriminant validity of C-DSIU IU- GAD IU- SAD IU- OCD IU- HA IU-PTSD IU- PD IU- MDD IU- Phobia IU- GAD 0.750 IU- SAD 0.557 0.768 IU- OCD 0.475 0.549 0.833 IU- HA 0.409 0.454 0.403 0.762 IU-PTSD 0.308 0.302 0.378 0.305 0.758 IU- PD 0.368 0.422 0.313 0.360 0.452 0.664 IU- MDD 0.357 0.412 0.332 0.509 0.409 0.528 0.718 IU- Phobia 0.344 0.359 0.351 0.298 0.369 0.255 0.290 0.783 Discriminant validity of C-DSIU Reliability analysis demonstrated good internal consistency for the Chinese version of the DSIU (24 items). Both McDonald’s omega coefficient and Cronbach’s alpha coefficient ( N  = 641) were 0.89. Additionally, Table  3 shows that deleting any single item would result in a decrease in the overall Cronbach’s alpha. The scale exhibited a split-half reliability coefficient of 0.69 ( N  = 641), indicating relatively high internal consistency. The 46 female university students with PMS were invited to complete the questionnaire again after a two-week interval. The results indicated a Spearman correlation coefficient of 0.947 between the two tests, which exceeds 0.70. This demonstrates that the Chinese version of the DSIU possesses good long-term stability.

This study represents the first introduction of the Disease-Specific Intolerance of Uncertainty Scale (DSIU) within the Chinese cultural context to assess disorder-specific intolerance of uncertainty characteristics in female university students with Premenstrual Syndrome (PMS). Through a rigorous cross-cultural adaptation process, including standardized translation and psychometric property evaluation, the Chinese version of the DSIU (C-DSIU) was ultimately developed. This version retains all 24 original items of the parent scale and demonstrates good internal consistency reliability (Cronbach’s α = 0.89). Validity testing revealed that exploratory factor analysis confirmed an eight-factor structure, establishing the scale’s sound construct validity. The findings of this study indicate that the psychological construct of Disease-Specific Intolerance of Uncertainty (DSIU) demonstrates fundamental structural consistency between Chinese and Western cultural contexts, while also revealing potential cultural specificities in the expression of certain symptoms and cognitive coping styles. These results support and extend the theoretical model proposed by Thibodeau et al. The study also shows that the DSIU is capable of capturing context-specific cognitive and emotional responses to uncertainty related to health and psychosocial functioning among female university students with PMS. This capacity provides a methodological foundation for more accurate screening of high-risk individuals, facilitates early diagnosis, and contributes to the design of tailored cognitive-behavioral interventions aimed at alleviating uncertainty-driven distress. Translation involves applying information from the source language to the target language. The main challenge in accurately assessing the target population using foreign assessment tools is ensuring correct translation and implementing culturally appropriate tools [ 51 ]. As the translation of the items may not convey the same meaning in the target language [ 52 , 53 ], experts are consulted to evaluate the restoration from a semantic point of view. Taking into account the suggestions of experts, minor adjustments were made to the meaning of the survey items. For instance, the language arrangements were adjusted for Item 2 of IU-SAD from “I can’t be myself in social situations when I’m not sure whether or not I will be embarrassed” to “When I’m not sure if I’m going to embarrass myself, I can’t be myself in social situations.”; The second is to replace IU-Phobia’s item 1, “It is better for me to avoid fear and stay safe than to face it and suffer uncertain consequences”, with “It is better for me to avoid what I fear and stay safe than to face it and suffer uncertain consequences”. And change the IU-Phobia preface, “The next series of items is about the things or situations you fear the most. (heights, snakes, flying, elevators). Please answer the following questions, which require you to visualize your greatest fear even if you are not very afraid of anything” to “Suppose that one of the items High Places, Snakes, and Flying is your greatest fear (even if you are not very afraid of anything, which requires you to visualize your greatest fear). Depending on your situation at the time, answer the following questions”. The results of the cross-cultural adaptation stage of the scales in this study showed that the Chinese translation of the source scales was satisfactory. Content and apparent validity were evaluated by experts to ensure that construction interests were accurately represented in all eight dimensions of the site projects. Six experts reviewed each item for relevance and appropriateness. The findings showed that the instrument’s elements were deemed highly relevant and thorough in measuring the intended construct. The expert review confirmed that the DSIU scale effectively encompassed the new domain and validated the instrument’s content. As with the original scale structure, the eight-factor structure was extracted through the EFA, explaining 69.99% of the total data variance. The factor loadings of the 18 items in this dimension ranged from 0.64 to 0.86. This suggests that all of the items in the Chinese version of the DSIU point to eight different latent traits, suggesting that the scale is eight-dimensional and can examine the units of information involved in the different questions, rather than allowing only one situation to be identified. The CFA showed that the model fit indices all met acceptable criteria. Compared to previous studies, the validation of this scale within the Chinese cultural context not only strengthens the cross-cultural applicability of the theoretical model of Disease-Specific Intolerance of Uncertainty (DSIU) proposed by Thibodeau et al., but also expands its scope in practical applications. The Chinese version of the DSIU items fits well into the 8-factor structural model. The Chinese version of the DSIU is a valid measurement tool for making initial judgments about the impact of uncertainty tolerance on the occurrence of PMS through the disease version. With good scale retest reliability, the Chinese version of the DSIU can assess the disease-specific uncertainty intolerance of Chinese female undergraduates meeting diagnostic criteria for PMS. Due to the independence of the data involved in EFA and CFA, it is necessary to perform the reliability tests separately. The Cronbach coefficient and the KMO test are a prior [ 46 ], and to ensure reliability, we satisfied the Cronbach coefficient and the KMO requirements before proceeding with the EFA model validation. Cronbach α calculations were performed in the EFA sample (N1 = 320), and the results of the study showed that the dimensions and their total scales were reliable. In addition, the total scale ( N  = 641) reported reliability relative to two studies based on the combinatorial reliability (CR) formula proposed by Raykov and Marcoulides in 2015 [ 54 ]. CR has been recommended as a reliability test for CFA measurement models rather than Cronbach’s α [ 55 ]. In the present study, the CR reliability values for all DSIU subscales except IU-PD were satisfactory and above the recommended value of 0.70 [ 42 ]. In this study, all DSIU subscales demonstrated satisfactory average variance extracted (AVE) values exceeding the recommended threshold of 0.50 [ 44 ]. While the AVE for the IU-PD dimension warrants cautious interpretation regarding precision at the individual level, it retains discriminant validity at the group level. Although the AVE was relatively low, the composite reliability (CR = 0.699) of IU-PD approached the acceptable standard of 0.70 (see Table  2 ). This indicates sufficient internal consistency among items, confirming the subscale’s reliability in reflecting the underlying construct. AVE is more sensitive to measurement error [ 56 ]. Lower AVE values may occur despite adequate CR when subscales contain fewer items or exhibit moderate but stable inter-item correlations. Further analysis revealed that all standardized factor loadings for IU-PD items ranged between 0.537 and 0.738 (all > 0.50, p  < 0.001). This confirms that each item effectively reflects its target construct, providing foundational support for convergent validity [ 57 ]. Although the squared loadings (i.e., variance explained by items) yielded an average AVE slightly below 0.50, no individual item exhibited substantial issues. IU-PD constitutes a core subdimension within the Intolerance of Uncertainty theoretical framework, capturing unique aspects of an individual’s prolonged negative affective responses to uncertain situations. Removing this dimension would compromise the scale’s ability to fully represent core theoretical tenets, thereby undermining its content validity and theoretical relevance. Thus, we ultimately retained the IU-PD dimension. Future studies should incorporate clinical interviews with PMS patients to verify the alignment between PD items and patient-reported illness concerns. Nonetheless, given the traditional value of Cronbach’s alpha coefficient in assessing the reliability of total scales in psychology and its ability to visually reflect the level of mean covariance among items, the Cronbach’s alpha coefficient for the simultaneous reporting of total scales was 0.89 to balance methodological rigor with compatibility with academic practice. However, we found large variations in total Cronbach α and Spill-half reliability. Uncertainty itself is a composite concept whose characteristics vary according to anxiety and concern about different situations. The items used to assess uncertainty tolerance include different aspects of uncertainty, and this simple scale contains eight aspects. For PMS female college students with strong beliefs and no uncertainty, the correlation between these items may be high [ 58 ]. However, the relationship between some items with different uncertainties breaks the high correlation of tolerance to uncertainty among PMS female college students. For example, it is possible that for some female college students with a low tolerance to uncertainty, there may be signs of uncertainty in one or more dimensions (i.e., social anxiety, health anxiety) but no signs of uncertainty in other dimensions (i.e., depression, posttraumatic stress). These scales can tell users which types of uncertainty are distressing, which can directly inform case conceptualization and treatment planning. For example, DSIU about social anxiety symptoms can explain social anxiety relapses, and this form of IU can be used to inform case conceptualization and treatment planning by directly identifying uncertainty about the potential onset or consequences of unexplained symptoms. In this study, female college students with PMS scored relatively high in the three dimensions of IU-GAD, IU-SAD, and IU-OCD, with scores of (5.79 ± 2.18), (4.44 ± 2.21), and (4.76 ± 2.66), respectively. The IU-PD score was the lowest, at (1.98 ± 1.72). we discovered that the mean values of the dimensions other than IU-GAD, IU-OAD, and IU-PTSD were higher than the original scale [ 31 ]. The reason for this is that our study was conducted on female university students with PMS. The PMS symptoms themselves may cause physical discomfort and life disturbances to female university students, such as fatigue, pain, and mood swings [ 59 , 60 ], which may increase their psychological stress. Prolonged exposure to this state may lead to cognitive biases [ 61 ], such as excessive worry about one’s health (health anxiety disorder) and hypersensitivity to negative evaluations in social situations (social anxiety disorder). At the same time, such stress and cognitive biases may be amplified when facing multiple life scenarios and challenges, further increasing the likelihood of developing psychological disorders such as generalized anxiety disorder. Moreover, emotional instability and physical discomfort during PMS may make college women more likely to recall negative experiences, creating a cycle of negative thinking, exacerbating depression, and increasing the likelihood of MDD [ 62 ]. Furthermore, uncertainty has become an important feature of modern social development and will remain prominent for a long time; the factors of uncertainty are increasing, while the difficulty of controlling them and the risks associated with them are also increasing; and uncertainty increases risks of human development and brings about complex effects [ 63 , 64 ]. In our study, most female college students are between the ages of 18–24, which is the transition from late adolescence to early adulthood according to Erikson’s Eight Stages of psychological development [ 58 ]. Faced with various psychological conflicts such as independence and dependence, ideal and reality, psychological closure, and seeking understanding, coupled with the impact of the epidemic and the pressure of academics, employment, life, and other aspects, female college students are prone to various psychological uncertainties. However, PMS is influenced by multiple factors such as hormones and psychosocial elements, and alleviating uncertainty-related anxiety in female university students with PMS cannot be achieved overnight [ 12 , 13 ]. It requires in-depth consideration of specific uncertainties—such as those related to social anxiety, health anxiety, and depression—to develop effective anxiety reduction strategies. It is essential to ensure that students with PMS can access information easily and efficiently to manage their anxiety. For example, a mobile application could be developed to guide premenstrual knowledge, offering personalized plans and real-time consultation. Educational institutions such as universities could establish psychological clubs or other campus activities specifically for students with PMS, creating an interactive forum where mutual support can help reduce uncertainty-driven anxiety. Medical institutions should strengthen collaboration with schools to integrate menstrual health knowledge and psychological counseling into regular curricula, providing systematic education and guidance starting from adolescence. Early and continuous education can empower women with PMS to make informed coping decisions. Only through multi-stakeholder efforts can we enhance women’s accurate understanding and positive attitudes toward PMS and promote psychological well-being. Our research contributes to a better understanding of the cross-cultural applicability of the C-DSIU and offers valuable insights for guiding the physical and mental health of Chinese women with PMS. Through in-depth analysis of DSIU in female university students with PMS, this study clarifies the current status and underlying reasons for the average levels across the eight dimensions of DSIU and proposes concrete measures for improvement. This not only helps identify which specific uncertainties play a central role in influencing the severity of PMS and mental health in this population but also provides new perspectives and directions for further expanding and refining the understanding of PMS development and intervention strategies. This study utilized convenience sampling to recruit female undergraduates with PMS from three universities in Jinzhou. While operationally feasible, this approach carries significant methodological limitations: (1) Restricted Sample Representativeness : The sample is confined to the university population in Jinzhou. It cannot reflect PMS characteristics across urban-rural divides, women in different occupations, or a broader age spectrum. (2) Risk of Selection Bias : Voluntary participants may disproportionately represent individuals with more pronounced symptoms or higher health awareness, potentially leading to an overestimation of prevalence rates. (3) Insufficient Control for Confounding Factors : Factors specific to the undergraduate population (e.g., cyclical academic stress, homogeneous lifestyles) may obscure the influence of other risk factors. These limitations constrain the external validity of the findings, hindering their generalizability to the broader female population with PMS in China. Future research should employ multistage stratified sampling to enhance sample diversity (encompassing different geographic regions, educational/occupational backgrounds, and women of reproductive age). Additionally, although we validated the internal structure and reliability of the measurement model through confirmatory factor analysis (CFA) (Composite Reliability) and assessed convergent/discriminant validity using the Fornell-Larcker criterion and Average Variance Extracted (AVE) (see Table  2 ), empirical examination of criterion-related validity was not incorporated. Due to the cross-sectional design, longitudinal data required for predictive validity analysis were unavailable, nor were synchronous external criteria directly linked to the collected core constructs. Future research should further verify the scale’s predictive power against real-world criteria through longitudinal designs.

The Chinese DSIU scale consists of 24 items and demonstrates an eight-factor structure with adequate reliability and validity. Therefore, the validated Disorder-Specific Intolerance of Uncertainty scale is better suited for the Chinese population. Future research should be directed toward examining its applicability to other populations and analyzing potential influences of demographic, social, physiological, and psychological factors on DSIU among female college students with PMS.

Anxiety is fundamentally a state of uncertainty, as it centers on an unrealized threat [ 1 ]. Intolerance of Uncertainty (IU) is defined as “a dispositional incapacity to endure the aversive response triggered by the perceived absence of salient, key, or sufficient information, and sustained by the associated perception of uncertainty” [ 2 ]. A recent prospective study found that trait of the IU’s predicted a variety of anxiety disorder symptoms [ 3 ]. Furthermore, a growing body of literature supports the transdiagnostic role of IU across multiple conditions, including: Generalized Anxiety Disorder (GAD) [ 4 ], Social Anxiety Disorder (SAD) [ 5 ], Obsessive-Compulsive Disorder (OCD) [ 6 ], Health Anxiety [ 7 ], and Major Depressive Disorder (MDD) [ 8 ]. Given this evidence, theorists conclude that IU is not specific to any single disorder but is transdiagnosti c [ 9 ]. It represents a salient feature common to anxiety disorders and major depressive disorder and is potentially a trait predisposition underlying these conditions [ 10 ]. Premenstrual syndrome (PMS) is a condition that occurs during the luteal phase of the menstrual cycle and subsides on its own after menstruation begins [ 11 ]. Typically, symptoms begin after day 13 of the menstrual cycle. Studies have shown that premenstrual syndrome, with levels and fluctuations of estrogen and progesterone, affects 3% to 8% of women with PMS report very severe PMS symptoms (PMDD) [ 12 ]. As a result, PMDD is classified as a psychiatric disorder in the DSM-5 [ 13 ]. In recent years, many new developments have challenged this position. Research has confirmed PMDD as a legitimate biological disorder requiring medical care and intervention [ 14 ]. In 2019, the World Health Organization announced the inclusion of PMDD in the eleventh revision of the International Classification of Diseases and Related Health Problems (ICD-11) [ 15 ]. As can be seen, PMDD is both a psychiatric disorder in DSM-5 and a medical disorder in ICD-11, illustrating the complexity of distinguishing between physical and mental health conditions. Although the theoretical framework surrounding premenstrual uncertainty remains incompletely established, its existence has been documented across various studies. Researchers posit that this phenomenon is associated with the levels and fluctuations of estrogen and progesterone in Premenstrual Syndrome (PMS) [ 16 ]. Existing evidence indicates that oscillations in estrogen and progesterone modulate GABA-A receptor subunit expression, reduce cortical inhibitory tone, and impair the prefrontal cortex’s regulatory efficacy over the amygdala [ 17 ]. This neuroendocrine dysregulation propels individuals into a state of anticipatory anxiety concerning the timing, intensity, and physical consequences of symptoms, establishing a “vicious cycle of uncertainty” [ 18 , 19 ]. Prospective studies confirm that for some women with PMS, catastrophic cognitions regarding symptom unpredictability induce emotional vulnerability (depression/anxiety), intense feelings of panic, negative perceptions of interpersonal relationships (social anxiety), and a sense of incapacitation in handling daily affairs. In the most extreme cases, women with PMS report suicidal ideation and attempts [ 20 , 21 ]. It has been shown that this is related to the fact that during the luteal phase, reduced GABAergic transmission in individuals with high PMS leads to attentional control [ 18 ]. Attentional control constitutes a key behavioral strategy for managing and regulating emotional responses to uncertainty [ 5 ]. It is characterized by the ability to focus attention, rapidly reorient it, and adeptly modulate focus, thereby enabling individuals to avoid excessive preoccupation with threatening information [ 22 ]. High-PMS individuals exhibit reduced attentional control capacity specifically during the luteal phase [ 18 ]. When confronted with uncertainty, PMS individuals high in Intolerance of Uncertainty (IU) display heightened focus and vigilance toward environmental cues [ 23 ]. This hypervigilance may impede their ability to disengage from negative stimuli, subsequently exacerbating anxiety and depression [ 24 ]. Individuals who exhibit high IU are less able to disengage from threatening information [ 25 ]. Research has shown that these individuals often employ maladaptive coping strategies to manage their emotional experiences [ 26 ], which leads to increased anxiety symptoms. Premenstrual uncertainty confers risk for multiple affective disorders. Notably, Premenstrual Dysphoric Disorder (PMDD) exhibits more complex clinical trajectories when comorbid with depression, anxiety disorders, panic disorder, social phobia, obsessive-compulsive disorder, and suicidal ideation [ 27 , 28 ]. While Intolerance of Uncertainty (IU) holds significant transdiagnostic relevance for understanding affective disturbances in PMS, conceptualizing IU traits may be insufficient to fully explain the theoretically diverse manifestations of symptoms [ 29 ]. There is an urgent need for a novel measurement tool to identify specific situations of uncertainty that are intolerable for women with PMS. Such a tool would elucidate the sources of anxiety experienced in uncertain contexts and precisely delineate which uncertain scenarios precipitate distress. The theoretical model proposed by Thibodeau et al. (2015) posits that disease-specific intolerance of uncertainty (DSIU) may function as a causal mediator variable between trait-like intolerance of uncertainty (trait IU) and subsequent symptoms. Specifically, while trait IU increases the overall likelihood of developing DSIU, it is the DSIU itself that exerts a differential impact on distinct symptoms (e.g., social anxiety symptoms vs. health anxiety symptoms) [ 30 ]. Consequently, DSIU should exhibit a stronger association with symptoms in its specific disease domain than other symptom domains. To operationalize this, Thibodeau et al. developed eight scales measuring Disease-Specific Intolerance of Uncertainty (DSIU) related to various anxiety disorders and major depressive disorder. The aim was to capture IU that manifests both disorder-specifically (i.e., explaining unique variance specific to a particular disorder) and transdiagnostically (i.e., explaining variance shared across disorders) [ 30 ]. However, currently in China, there is no measurement tool specifically designed to assess the intolerance of uncertainty of a specific nature. Therefore, the objective of our study is to culturally adapt and validate the DSIU developed by Thibodeau et al. Importantly, identifying DSIU in this population is highly relevant for both clinical practice and psychological intervention. By measuring PMS-specific IU, we can gain a better understanding of how uncertainty exacerbates physical and emotional symptoms in affected individuals. This knowledge can inform the development of targeted cognitive-behavioral interventions aimed at reducing uncertainty-related distress, thereby alleviating PMS severity and improving daily functioning. Moreover, validating a culturally-adapted DSIU scale will enable more accurate screening and individualized treatment planning, ultimately supporting the psychological well-being and academic adaptation of female university students with PMS.