Cross-validation of the Arabic MINI, Module U, for Diagnosis of Premenstrual Dysphoric Disorder and the Arabic Premenstrual Symptoms Screening Tool

Background Premenstrual dysphoric disorder (PMDD) is a severe form of premenstrual syndrome (PMS) that affects the functioning and includes various physical, depressive, and anxiety symptoms. The Mini-International Neuropsychiatric Interview, Module U (MINI-U), assesses the diagnostic criteria for probable PMDD. The Premenstrual Symptoms Screening Tool (PSST) measures the severity of these symptoms and provides scores from which we can diagnose PMS and PMDD. The purpose of this study is to obtain the cut-offs from PSST that would t well with the diagnosis of PMDD using the MINI-U as a gold standard. Methods We administered the Arabic MINI-U and PSST independently to Arab women (N = 194) in Doha, Qatar. The Wilcoxon-Mann-Whitney test was used to compare the PSST ordinal scores between those who answered Yes vs No on the corresponding MINI-U items. These comparisons were followed by Receiver Operating Characteristics (ROC) analyses using the MINI-U answers as the gold standard to determine the cut-off scores on the PSST, in addition to their sensitivity and specicity measures Results The rating of PSST scores for participants who answered Yes in the MINI-U varied from 1.5 (not at all to mild) to 3 (moderate). Receiver Operating Characteristics (ROC) analyses showed that all areas under the curves (AUCs) were signicant with the cut-off scores on the corresponding PSST items. This cross-validation gives reassurance that the severity measures of PSST can recognize patients with moderate/severe PMS and PMDD who would benet from immediate treatment instead of waiting another two months.


Background
Premenstrual syndrome (PMS) is characterized by a collection of mild to severe physical, affective, and behavioral symptoms experienced by about 80% of reproductive age women [1,2]. The symptoms occur cyclically before or during the luteal phase of the menstrual cycle. Premenstrual dysphoric disorder (PMDD) is a more severe form of PMS with a greater emphasis on depressive and anxiety symptoms [3]. The etiology of PMS and PMDD is not clearly understood but the onset of symptoms is associated with hypersensitivity to changes in the ovarian hormonal level during the menstrual cycle [4,5], dysregulated immune function [3], neurotransmitter dysregulation, stress, diet, and lifestyle [6]. Pharmacological interventions include analgesic treatment, combined oral contraceptives [7], and selective serotonin reuptake inhibitors [8,9]. Nonpharmacological treatments are lifestyle interventions and cognitive behavioral therapy (CBT).
Overall, 75-85% of women have experienced PMS symptoms [1,2] whereas PMDD affects 5-8% of reproductive age women worldwide [10]. According to the International Classi cation of Diseases (ICD-10), only one distressing symptom at the time of menstruation is required for PMS diagnosis. It does not consider the severity of the symptoms, and no clear de nition exists when PMS becomes clinically signi cant. Contrarily, diagnosis of PMDD mandates the impairment of functioning by the symptoms [11]. According to the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) [12], the criteria for the diagnosis of PMDD are: (A) at least 5 symptoms must be present in the nal week before the onset of menses and resolve within a few days of the onset of menses and these symptoms must occur in the majority of the menstrual cycles, (B) at least one symptom must be marked affective lability, marked irritability or anger or marked depressed mood or anxiety, (C) one or more of the following symptoms must be present: decreased interest in usual activities, di culty in concentration, increased fatigue, change in appetite, marked change in sleep, feeling overwhelmed or physical symptoms, (D) these symptoms should affect productivity at work or school, relationships, responsibilities, and social activities. These symptoms should not be attributable/resultant to symptoms from (E) another psychiatric disorder or (F) physiological effects of a substance. Finally, (G) these symptoms should be con rmed by prospective daily ratings for at least two symptomatic cycles.
The Mini International Neuropsychiatric Interview (MINI) is a structured interview consisting of several modules to diagnose DSM-IV-TR psychiatric disorders [13][14][15]. Module U (MINI-U) is the corresponding module that measures categorically the presence or absence of symptoms to ful ll diagnostic criteria for PMDD [16]. Prospective daily ratings have to be completed for at least two symptomatic cycles to con rm the diagnosis and are the only way to measure both severity and monitor symptoms over time [17].
The Premenstrual Symptoms Screening Tool (PSST) is an instrument that includes all premenstrual symptoms as well as a measure of impairment as per DSM-IV-TR criteria. It also translates categorical DSM-IV-TR criteria into a dimensional rating scale to assess severity [17]. It is a useful diagnostic tool to capture moderate/severe PMS and PMDD diagnoses in symptomatic women who would bene t from treatment [18].
The MINI-U for diagnosis of PMDD relies mainly on the presence or absence of symptoms, including the impact on functioning, while PSST uses a dimensional scale to measure the severity of symptoms, which ultimately is very important to determine the effects of symptoms on daily activities. There are no studies that compared the diagnostic categorical scales with dimensional measures of severity of PMDD symptoms. Such comparisons would enhance the accuracy of the psychometric measures of the combined approaches when diagnosing and monitoring patients with moderate/severe PMS and PMDD. The availability of valid cut-off scores from PSST tested through answers from MINI-U (DSM criteria) would give more con dence to diagnose PMDD based on the severity measures of PSST. This reassurance would facilitate the initiation of treatment for this group of patients instead of waiting two months, especially that the daily recording of symptoms has proven to be very di cult in practice [19,20]. Thus, the aims of this study are (1) to compare the responses between the dichotomous MINI-U answers and the scores on the PSST items; and (2) to establish the cut-off scores on the dimensional PSST items by using the categorical MINI-U as a gold standard.

Methods
This cross-sectional study is part of a project to validate the Arabic version of the PSST [citation deleted to maintain integrity of the review process]. This article reports a secondary analysis of the relationship between answers on the Arabic MINI-U with the corresponding items (see Table 1) in the Arabic PSST.

Study Setting and Subjects
The study took place in Doha, Qatar where Arab women were recruited at two Primary Healthcare Centers between October 2013 and March 2014.
Participants were eligible to join the study if they were Arab females between 18 and 45 years old and with a regular menstrual cycle of 24 to 32 days. The following exclusion criteria were adopted to control for other confounding conditions: (1) taking oral contraceptive pills, hormonal therapy, psychotropic medication, and suspected of being pregnant or in menopause, (2) women with endometriosis, acute thyroid or pituitary disorders or any other acute medical problem and (3) women with a history of drug and alcohol abuse or with an active psychiatric disorder (other than PMDD) diagnosed in the previous six months.
During the recruitment period, a total of 430 women were approached to join the study, and among them, 280 women were eligible for the study. However, only 194 women agreed toparticipate. This sample size was su cient to detect the projected sensitivity or speci city of 85 percent and an estimated prevalence of severe PMS/PMDD of 20 percent, within a margin of error of 10 percent and a 95 percent con dence interval.

Research Design
This study is cross-sectional with no interventions, and all participants provided written consent before enrollment. The Institutional Review Boards of Hamad Medical Corporation and Weill Cornell Medicine in Doha, Qatar approved this study. A licensed physician or nurse interviewed participants to con rm their eligibility. The psychiatrists then administered the Arabic Mini International Neuropsychiatric Interview Plus version 6 (MINI-Plus 6) to screen for any psychiatric disorders including PMDD (MINI-U) as per DSM-IV-TR criteria [16]. An independent second rater, blinded to the results of the MINI, collected sociodemographic information, past medical and psychiatric history, smoking and exercise patterns, and administered the PSST. The independent raters were medical students or nurses who were formally trained on how to administer and rate the PSST. Good inter-rater agreement was established before the collection of data. A pilot sample of 20 women were rated independently by more than two raters and the interclass coe cient was 0.89.

Procedures
Recruitment for this study commenced shortly after the introduction of DSM-5, no diagnostic instruments were available at the time to diagnose PMDD according to DSM-5 criteria; hence we used MINI-U that followed DSM-IV-TR criteria. DSM-5 adopted the same criteria for the diagnosis of PMDD as DSM-IV-TR except for minor modi cations [12,13].
Module U in the MINI is a screening and diagnostic tool for PMDD. It is composed of 3 main dichotomous questions (U1, U2, and U3) with the possibility of answering "yes" or "no". The rst two questions respectively assess mood changes before menstruation and if the subject experienced any di culty at work or in usual activities and relationships during these periods. The last question determines the presence of affective, behavioral, and physical symptoms using lettered questions U3-A to U3-K, as indicated in Table 1.
A diagnosis of probable PMDD is reached if the rst two questions are answered positively together with at least one affective symptom from U3-A to U3-D and a further four emotional, behavioral, or physical symptoms from U3-E to U3-K [15].
Translation of the PSST was approved by the original author [17] and McMaster University. The PSST was translated to Arabic using the repeated forward-backward procedure. All concerns were resolved by modifying the Arabic version of PSST until the original author approved the English back-translated version.
Please refer to the study by [citation deleted] for further details on the translation and validation procedures for the Arabic versions. The PSST is composed of two sections. The rst section includes a list of premenstrual symptoms followed by a second section that measures impairment following DSM-IV-TR criteria for PMDD. Responses are reported on a severity scale of "not at all," "mild," "moderate" or "severe." The instrument provides the diagnosis of moderate to severe PMS and PMDD to identify patients who would bene t from pharmacological treatments [17].

Statistical Analysis
All analyses were performed using IBM Statistical Package for Social Sciences (SPSS) for Mac version 24 [21]. The level of signi cance was set at 5%. Sociodemographic characteristics and clinical features were reported as means and standard deviations (SD) for continuous measures such as age, and as frequency and percentage for categorical measures such as education level. The ordinal PSST scores were reported as the median and interquartile range (IQR). The Wilcoxon-Mann-Whitney test was used to compare the PSST ordinal scores between those who answered Yes vs No on the corresponding MINI-U items. These comparisons were followed by receiver operating characteristics (ROC) analyses using the MINI-U answers as the gold standard to determine the cut-off scores on the PSST, in addition to their sensitivity and speci city measures. We used the highest Youden indices (J) to determine the best cut-off scores on each item in PSST and the corresponding sensitivity and speci city.

Results
Sociodemographic and clinical characteristics (Tables 2 and  3)  According to the PSST, 14% of participants had a diagnosis of PMDD, and 35% had a diagnosis of PMS.
However, according to MINI-U, 49% of participants had a diagnosis of probable PMDD. A minority of participants had been diagnosed in the past with depression (5%) or other psychiatric illness (3.3%).
Frequency of symptoms as per MINI-U and PSST (Table 4)  (1) anger or irritability, (2) anxiety or tension, (3) decreased interest in home activities, and (4) physical symptoms where the median rating was 2 (mild). Participants who answered Yes had a median rating from 1.5 (not at all to mild) to 3 (moderate). Out of the 14 symptoms assessed, nine had a median rating of 3 (moderate), four symptoms had a median rating of 2 (mild), and one symptom had a median rating of 1.5 (not at all to mild) ( Table 5). The median rating of the interference of these symptoms on work or productivity, relationship with family, relationship with co-workers, relationship with family, on social life activities and home responsibilities was 2 (mild).
Cut-off scores on PSST items by MINI-U dichotomous responses (Table 6)  Table 6 The cut-off scores of the PSST items with the corresponding MINI-U items

Discussion
The rst aim of this study was to compare the responses between the dichotomous MINI-U responses and the scores on the PSST items. Our study showed a discrepancy in the prevalence of the diagnosis of PMDD between the MINI criteria (46.7%) and PSST criteria (13.9%). The discrepancy between the two could be attributed to the dichotomous nature of the questions in MINI-U that assess only the presence or absence of symptoms, while those in PSST focuses more on the severity of symptoms to establish the diagnosis of PMDD. The high prevalence of PMDD is also higher than that reported worldwide (5-8%) [10]. Other countries such as Iran [22], Jordan [1], India [23] and Brazil [24] reported a similarly high prevalence of PMDD suggesting that there are ethnic variations in the prevalence of PMDD. It also highlights the need for an e cient and valid diagnosis of PMDD to recognize these patients and initiate treatment as early as needed. In a comparison among the participants who answered positively vs. negatively on the MINI questions, we found that all PSST symptom ratings were signi cantly higher among those who answered positively. Most symptoms on PSST had a median rating of "moderate" indicating clinical signi cance (  (Table 4). These were also found to be common complaints among Jordanian and Emirati women [26,27]. One of the major concerns with the MINI and PSST is the requirement to have daily ratings of symptoms for a minimum of two cycles as per DSM criteria to con rm the cyclical presence of symptoms for moderate/severe PMS and PMDD. Keeping a daily diary before initiating treatment may cause resistance for women to seek treatment. In research settings, an epidemiological study found that 30% of women refused to participate in a study because they did not want to ll daily ratings and the latter is usually associated with a high dropout rate [28]. Our results suggest that the severity measures of PSST can capture the PMDD cases who would bene t from treatment initiation.
The second aim of the study was to establish the cut-off scores on the dimensional PSST items by using the categorical MINI-U as a gold standard. All the cut-off scores showed signi cant differentiation and ranged from 1.5 to 2.5 with adequate sensitivity and speci city ( Table 6). The MINI-U is a diagnostic instrument whereas the PSST is both a diagnostic and dimensional instrument [17]. However, both scales are based on DSM-IV-TR criteria for the diagnosis of PMDD and are therefore assessing the same symptoms (Table 1). Cross-validating these instruments showed that most symptoms corresponding to a Yes in the MINI-U had a cut-off score of 1.5 or a rating of at least Mild on the corresponding PSST items.
A rmative answers to anger/irritability, anxiety/tension, decreased interest in home activities and physical symptoms in the MINI-U had a corresponding cut-off score of 2.5 or at least moderate symptoms in the PSST, meaning that the latter captured mainly the moderate to severe cases. However, the challenge is to distinguish which women need treatment from those whose symptoms are not clinically relevant [28].
Moderate/severe PMS and PMDD are poorly diagnosed and mostly untreated conditions [29]. Furthermore, women with moderate/severe PMS symptoms have a higher rate of work absences and increased medical expenses [2]. These women can, therefore, bene t from a prompt referral and timely treatment [2].

Conclusion
In conclusion, our results showed a signi cant relationship between the responses on the Arabic MINI-U and PSST. Participants who answered positively on the MINI had signi cantly higher ratings and relevant cut-off scores on the corresponding PSST items. This cross-validation gives reassurance that the MINI-U provides an adequate assessment for the probable diagnosis of PMDD and that the severity measures of the PSST can recognize patients with moderate/severe PMS and PMDD who would de nitely bene t from immediate treatment. Thus, there is a clear advantage of using PSST to early identify these patients with moderate/severe symptoms who should not wait for the daily measures of MINI-U.

Consent for publication
Not applicable.

Availability of data and materials
The datasets analysed during the current study are not available publicly available due to privacy or ethical restrictions but are available from the corresponding author on reasonable request.