Study design and context
This cross-sectional study looked at the relationship between psychological factors and premenstrual symptoms in female students at Ilam universities.
Setting and sample size
The study sample consisted of 624 students. Participants were recruited from all universities in Ilam City between April and September 2024. Data were collected online. Questionnaires were created on the Porsline website (https://porsline.ir/), and then we shared the link with university teachers and asked them to explain the research objectives and share the link in students’ social groups.
Participants
Inclusion criteria
Female students who were above 18, enrolled in a university program, currently students, and having regular menstrual cycles or the absence of gynecological diseases.
Exclusion criteria
The exclusion criteria included a history of major psychiatric disorders, hormonal therapies in use currently, or incomplete survey responses.
Data collection tools and ethical considerations
Ethical approval and consent to participate
As participants completed questions online, according to national regulations of “Research Ethics Committees of Ilam University of Medical Sciences,Ilam,Iran”, informed consent form to participate was deemed unnecessary, However, we requested an ethical code conformation of “Research Ethics Committees of Ilam University of Medical Sciences,Ilam,Iran”. Research Ethics Committees of Ilam University of Medical Sciences confirmed our study and gave this code to our study:
IR.MEDILAM.REC.1403.256: https://ethics.research.ac.ir/form/tph1hzssz1om5wny.pdf.
Participants were fully informed about the survey’s purpose, how their data will be used, and any potential risks involved. Clear consent forms were provided, allowing participants to voluntarily agree to participate without any coercion. We wrote an informed consent form according to the Helsinki Declaration (https://www.wma.net/policies-post/wma-declaration-of-helsinki/) form. Participants did not need to write their name, phone number, and personal data, if they requested about suicidal thought severity or PMS severity, we gave an option to participants, they could write their email, we sent suicidal thought and PMS severity by email, if they requested. Writing email was arbitrary. Participants without fear of repercussions completed the questionnaires. Secure platforms and encryption methods were used in online forms. Participants communicated openly about who was conducting the survey, its purpose, and how the data would be used. This transparency builds trust and encourages participation.
Demographic information questionnaire
Demographic factors included age, marital status, school year, education, height, weight, and sleep patterns.
PMS
The Premenstrual Symptoms Screening Questionnaire (PSST)
Premenstrual syndrome (PMS) and premenstrual dysphoric disorder (PMDD) intensity are frequently assessed with the PSST. It is a standardized test with 19 items broken down into two halves. Five items in the second portion measure how these symptoms affect day-to-day functioning, while fourteen symptom-based questions in the first section evaluate emotional, physical, and behavioral problems. Every question receives a score of 0, 1, 2, or 3. Questions do not have a reverse score. Mild, moderate, and severe PMS were classified in most studies based on a total score that ranged from 0 to 19, 20 to 28, and more than 28, respectively (scoring 2) [31, 32]. In some studies Mild, moderate, and severe PMS were classified on a total score that ranged from 0 to 19, 20 to 38, and more than 38 resoectively (scoring 1) [33, 34]. The DSM-5 states that a diagnosis of moderate-to-severe PMS requires the presentation of the following criteria:
(1) According to PSST inventory [35], at least one of questions 1, 2, 3, and 4 is moderate to severe; (2) at least four of questions 1 through 14 are also moderate to severe; and (3) at least one of questions 15 through 19 (A, B, C, D, and E) is moderate to severe.
To classify participants as having mild, moderate, or severe PMS, we create a program in STATA and call it PMS_class.
The PSST has proved valid and reliable in various populations. In the study, Steiner et al. [31] reported Cronbach’s alpha to be 0.91. The reliability of the Persian version of this questionnaire was estimated 0.93 [36] and 0.91 [35]. The questionnaire’s Content Validity Ratio and Content Validity Index were estimated as 0.7 and 0.8, respectively [36]. Cronbach’s Alpha was calculated for our investigation at 0.914, which demonstrated excellent reliability.
Depression and anxiety
The Depression, Anxiety, and Stress Scale (DASS-42):
It has 42 items total, broken down into three subscales: stress, anxiety, and depression. Each subscale has 14 items. It was created especially to gauge how severe these three aspects of emotional anguish are. The validity and reliability of the DASS-42 have been established globally. In the original study by Lovibond [37], it showed a Cronbach’s alpha of 0.94 for depression, 0.87 for anxiety, and 0.91 for stress. In Iran, a study by Sahebi et al. [38] found a Cronbach’s alpha of 0.87 for an overall scale, showing it is also reliable for this group of people. Cronbach’s Alpha was calculated for our investigation at 0.92, which demonstrated excellent reliability. Every instrument was evaluated and standardized in Persian for the intended audience. To maintain anonymity and lessen response bias, questionnaires were sent out electronically and filled out anonymously.
Statistical analysis
To guarantee adequate representation of the universities and academic levels included, cluster random sampling was used. Assuming a 10% non-response rate, the sample size of 624 was pre-calculated with error (d = 2.4%), P = 0.15 with the sample size calculation formula for prevalence [39].
The severity of premenstrual symptoms, which were divided into three groups—none/mild, moderate, and severe—was the primary outcome variable. Potential confounding factors included independent variables such as anxiety, depression, age, marital status, height, weight, and the average number of hours spent sleeping at night. Chi-square tests were used to find an association between PMS severity and categorical data. After adjusting for demographic confounders, the ORs and 95% CIs for the psychological components and premenstrual symptoms were determined using a multivariate ordinal logistic regression model. We employed proportional odds in cumulative odds and ordinal logistic regression. Four presumptions for ordinal logistic regression are examined, including:
-
1)
The dependent variable was measured at the ordinal scale (PMS measured as mild, moderate, and severe).
-
2)
continuous, ordinal, or categorical independent variables
-
3)
the independent variables did not exhibit multicollinearity
-
4)
For ordinal regression models, proportional odds were satisfied (chi² [4] = 5.71, P = 0.2215). ROC (Receiver Operating Characteristic) Curve and AUC (Area Under the Curve) were used to evaluate the performance of classification models, particularly in binary or ordinal outcomes [40, 41]. AUC is a scalar value (between 0 and 1) quantifying the overall ability of a model to distinguish between classes.
Interpretation:
AUC = 0.5: No discrimination (random model).
AUC = 0.7–0.8: Acceptable discrimination.
AUC = 0.8–0.9: Excellent discrimination.
AUC > 0.9: Outstanding discrimination.
AUC is a widely used measure for evaluating prediction models, such as PMS models, since it calculates the model’s ability to distinguish between controls (non-PMS) and cases (students with PMS) at all possible thresholds, instead of a single cut-off point([42]. This threshold-independence is specially useful in relation to measures such as accuracy, precision, or F1-score, which quote only performance at a single threshold and can be misleading, especially in imbalanced datasets where one class (PMS cases, in this study) is by far less frequent. AUC is also statistically more consistent and discriminative than accuracy, i.e., it distinguishes between better and worse models more often even when accuracy is equal [40, 41].
The ROC curve is generated by plotting the True Positive Rate (TPR) (y-axis) against the False Positive Rate (FPR) (x-axis) at varying classification thresholds (e.g., probability cutoffs for predicting a class). A curve closer to the top-left corner indicates better model performance. A diagonal line (AUC = 0.5) represents a model with no discriminative power (equivalent to random guessing) [40, 41].
Sensitivity analyses were also performed to evaluate how reliable the results were. The study’s significance level was set at P < 0.05. STATA version 17 and SPSS version 27 were used for statistical analysis. Using a STATA program, PMS total scores were divided into three categories: mild, moderate, and severe this commends in STATA:
STATA program for classification PMS.
Generate PMS_category =.
* category based on PMS definition in DSM-5.
replace PMS_category = 0 if (Q1 < 2 & Q2 < 2 & Q3 < 2 & Q4 < 3)//No/Mild PMS.
replace PMS_category = 1 if (Q1 > = 2| Q2 > = 2| Q3 > = 2| Q4 > = 3) &///.
(Q5 + Q6 + Q7 + Q8 + Q9 + Q10 +///.
Q11 + Q12 + Q13 + Q14 > = 3) &///.
(Q15 > = 2| Q16 > = 2| Q17 > = 2|///.
Q18 > = 2| Q19 > = 2)//Moderate PMS.
replace PMS_category = 2 if (Q1 = = 3| Q2 = = 3| Q3 = = 3| Q4 = = 3) &///.
(Q5 + Q6 + Q7 + Q8 + Q9 + Q10 +///.
Q11 + Q12 + Q13 + Q14 > = 5) &///.
(Q15 = = 4| Q16 = = 4| Q17 = = 4|///.
Q18 = = 4| Q19 = = 4)//Severe PMS/PMDD.
label define pms_labels 0 “No/Mild PMS” 1 “Moderate PMS” 2 “Severe PMS/PMDD”.
label values PMS_category pms_labels.