Setting
The site for this study is the In Vitro Fertilization (IVF) Center at the King Abdullah University Hospital within the Jordan University of Science and Technology in Irbid which is one of approximately 25 IVF clinics in the country. Most IVF clinics in Jordan are private. The study clinic is public; however, most services are not covered by insurance. Some insurance plans cover a percentage of the diagnostic testing, but not the cost of the visits, medications, or the IVF procedures themselves. Individuals can approach the clinic for services directly or through a referral. Approximately 1500 patients use the IVF center yearly.
Sample
Figure 1 depicts how the final analytic sample was developed from eligible couples. Couples residing in Jordan who had been married at least 2 years and who were seeking services for primary infertility (having unprotected sex for at least 1 year without conception) or secondary infertility (having unprotected sex for at least 1 year subsequent to a birth or abortion) at the study clinic were eligible to participate. Names of eligible patients were provided to the research team by a physician practicing in the clinic. The research team then contacted patients and invited them to participate in the study to avoid coercion or expectations that participation would affect the treatment they received at the clinic. Psychologists administered the baseline survey individually to each member of the couples. This was deemed safer for participants, as the psychologists had professional training in dealing with distress, should it occur. Potentially eligible individuals subsequently underwent an intake evaluation by a psychologist to assess mental health needs. Thirty-eight couples were deemed in need of mental health support and therefore eligible for the intervention. These problems that the individuals presented with differed and ranged from abuse to distress. Individuals in particularly acute need were offered individual sessions prior to the start of the groups (described below). Eligible couples were recruited into group therapy by the research team. Eight couples were not able to be subsequently contacted after 3 attempts (phone disconnected, no answer), and six couples refused to participate. Of the 24 couples who agreed to participate in the group sessions, 16 attended more than 1 session. The analytic sample includes baseline and endline data on all wives and 15 husbands as 1 husband was unavailable for the endline survey.
Intervention
Eleven gender-segregated CBT sessions were delivered per group (2 groups per gender). All four groups convened simultaneously on the same day. The sessions were held once a week with each session lasting between one and a half to two hours. Session focus areas included topics in line with the hypothesized theory of change (Fig. 2): (1) group norms and goal-setting; (2) understanding psychosocial stress; (3) identification of stressors and impacts on day-to-day life; (4) the impact of self-perception of day-to-day life; (5) expression and safe management of feelings (6) overcoming obstacles to communication; (7) effective communication strategies; (8) questioning assumptions and self-perceptions about problems; (9) problem-solving and strategies for restructuring relationships; (10) development of new approaches to life by removing unfounded anxieties; and (11) highlights of prior sessions and reminders for application in day-to-day life. Prior to the group therapy, 6 participants were deemed in need of individual therapy (range of 4 to 6 sessions) before entering the group sessions. After the cessation of the intervention, 8 participants continued individual therapy for 4 additional sessions. All sessions were led by psychologists experienced in CBT and were gender-matched to the participants.
Intervention theory of change
Data
Face-to-face interviews were conducted with patients at the clinic site in data collection sessions prior to and 16 weeks after the cessation of group therapy (baseline and endline assessments, respectively). Survey content at each timepoint included socio-demographics, reproductive history, mental health treatment history, and outcomes in alignment with the study’s theory of change (Fig. 2), including: symptoms of depression, anxiety, and PTSD; fertility-related quality of life and norms; risk factors for and experience of IPV and in-law abuse (women only); and social support and coping. We also measured program participation (at endline) and reactions to survey participation (at both timepoints) to assess self-reported exposure to the program along with adherence to study ethics. Interviews were conducted in Arabic, by a therapist of the same gender and in a private space.
Primary outcomes
Intimate partner violence ever and in the past 12 months was assessed with an augmented version of the World Health Organization’s Multi-Country Study on Women’s Health and Domestic Violence [37] which had been used in prior research in Jordan [38]. Participants reported on their experience of 8 psychological items, 6 physical items and 2 sexual items using a 4-point Likert scale (Never/Once/A few times/Many times). Separate dichotomous variables were created for each subtype with endorsement of any experience within that subset as indicative of abuse. The Cronbach’s alpha for the scale was good (0.92).
Depression and anxiety were assessed with the Hopkins Symptoms Checklist-25 (15 items assessing depression, 10 items assessing anxiety) which has been translated and validated in Lebanon [39, 40] and used widely in Jordan e.g. [41,42,43]. Items measured the frequency with which the respondent was bothered by each symptom in the past week on a 4-point Likert scale (not at all/a little/quite a bit/extremely) and then averaged together, with higher scores indicating more depression or anxiety symptoms. The scale had a good Cronbach’s alpha for depression (0.86) and anxiety (0.85). We additionally assessed depression and anxiety dichotomously using the established cutoff of M = 1.75 [41, 43].
PTSD was measured by study team-generated items from a scale that was developed and validated (although unpublished) by the therapeutic team in Jordan based on the DSM-IV. Respondents were asked to report whether they experienced 22 symptoms using a 5-point Likert scale (never/rarely/sometimes/often/always), where higher scores represent more severe symptoms. The Cronbach’s alpha for the measure was strong (0.95).
Secondary outcomes
Quality of life related to fertility was assessed with the 24 core items of the Fertility Quality of Life Scale [44, 45]. The scale includes two additional context questions assessing self-rated health and degree of satisfaction with the respondent’s quality of life. Sub-scales, each with 6 items assess negative emotions (emotional: 6 items), physical symptoms or negative cognitive or behavioral disruptions (mind–body: 6 items), and the impact of infertility on the marital/partner relationship (relational: 6 items) and social interactions (social: 6 items). Each item is scored on a 5-point Likert scale with response scales differing by item; for example, some items used a “very poor” to “very good” scale whereas others used a “not at all” to “completely” scale. Items with a negative valence were reverse coded and averaged together to create the four subscales and total scores, such that higher scores indicated higher quality of life. The Cronbach’s alpha was strong for the total scale (0.92) and was acceptable or good for the emotional (0.90), mind–body (0.84), relational (0.80), and social (0.75) subscales.
Social support was measured with the Arabic version of the Multidimensional Scale of Social Support (Arabic MSPSS) [46]. Respondents were asked their level of agreement on a 7-point Likert scale (very strongly disagree/strongly disagree/mildly disagree/neutral/mildly agree/strongly agree/very strongly agree) to 12 items assessing agreement about support received from family, friends and significant others, such that higher scores indicated more support. Cronbach’s alpha for the scale was very good (0.97).
Coping was measured with the Brief Resilient Coping Scale [47]. Participants were asked how well each of the 4 items described their situation on a 5-point Likert scale (Does not describe me at all/Does not describe me/Neutral/Describes me/Describes me very well). Higher scores indicated greater coping. The Cronbach’s alpha of the scale was good (0.85).
Fear of spouse was measured with a study-generated single item asking respondents, “How often are you afraid of your spouse?” Response options were on a 5-point Likert scale (never, rarely, sometimes, often, always).
Other variables of interest
Socio-demographics assessed included age, date of birth, educational level, participation in paid employment, financial distress, year of marriage, whether the marriage was their first marriage, consanguinity, polygamy, and residence status categorized as nuclear or extended family residence.
The reproductive history module examined history of pregnancy (yes/no), count of total prior pregnancies, whether the respondent is currently trying to become pregnant (yes/no), duration in months of trying to become pregnant, current pregnancy status, number of children alive, type of infertility (primary/secondary), duration of infertility (years), duration of infertility treatment seeking (years), treatments received (IVF, ovulation induction, surgery, intrauterine insemination), and outcome of prior treatment (failed to produce pregnancy, pregnancy that ended in miscarriage, currently pregnant, had a baby).
History of mental health treatment was assessed with 2 items measuring discussions with persons other than family or friends about emotional or psychological issues (yes/no) ever and in the past 12 months (yes/no). If yes, the respondent was asked to identify the person from a list of 9 options (psychologist/psychiatrist/case manager, case workers or outreach worker/social worker/nurse/physician/religious leader/support group/other).
Norms about infertility were assessed with 12 items (6 individual statements about men and women) developed by the study team based on formative research. Items assessed the extent to which the participant believed that people in their community would agree with gendered statements about acceptable behaviors or beliefs regarding infertility (e.g., “It is acceptable for a [woman/man] to marry someone else if [her husband/his wife] does not give [her/him] children” and “When a couple cannot have children, blame is usually placed on the [woman/man].). Items were assessed on a 3-point Likert scale (most [people in my community] would agree/half would agree and half would disagree/most would disagree) and averaged across the six statements separately for woman- and man-referencing items, with higher scores indicating more accepting perceived infertility norms. Cronbach’s alpha for this measure was adequate (0.76).
Abuse from other family members was assessed with 3-items derived from the IPV scale and used in prior research for this purpose [6]. Items assessed the occurrence (yes/no) of emotional violence, physical violence, and encouragement of the respondent’s spouse to use violence against her. For each affirmative response, the respondent was asked to indicate which family member(s) perpetrated the act with 13 options across marital and natal family members.
Survey participation
Reactions to survey participation were assessed with 6 items from the Respondents Reactions to Participation Questionnaire to assess patient comfort and perceived benefit during the survey-administration process [48]. Items assessed voluntary participation, ability to stop at any time, experience of intense emotions, meaningfulness of the study to themselves and to others using a yes/no format.
Intervention participation and feedback
Degree of participation (less than half, about half, most of the session, all of the sessions) was self-reported by the participant and spouse along with reasons for less than full participation (not interested, could not miss work, too busy, financial burden, social commitment, spouse refused, and other). The number of sessions that the respondent attended was also reported by the therapist. Open-ended questions were included on the participant survey to obtain feedback on the most helpful intervention content and suggestions for improvement. Reasons provided for participation refusal was systematically documented and feedback was requested from the therapists who delivered the intervention, which was included in a post-intervention report. Finally, the results were shared with a participating clinician for feedback on their relevance to practice in Jordan.
Analysis
Descriptive statistics by gender and time period were calculated. We also examined missing data in the analytic sample due to item skipping and found no systematic skipping patterns (e.g., a particular item that several participants skipped, a particular participant that skipped several items). The highest rate of item-level missing data was 6% (two participants) on one PTSD scale item (“I avoid people associated with the traumatic event”). Whereas several scales (e.g., the Fertility Quality of Life Scale, MPSS) are typically reported with sum/total score scores, we report participants’ average scores for all scales and subscales to avoid downward score bias in the total scores for the few participants missing data on individual items.
To address research question 3.4, independent-samples t tests and Fisher’s exact tests were conducted to determine similarities at baseline between the full sample (those who provided data at baseline but may or may not have provided data at endline) and analytic sample (those who provided data at endline as well as baseline). Tests were stratified by gender and reported with descriptive statistics in Table 1. To address research questions 3.1–3.3, changes between baseline and endline for primary and secondary outcomes were examined using Wilcoxon signed-rank tests and McNemar’s tests and are reported with descriptive statistics in Table 2. The Wilcoxon signed-rank test is a nonparametric alternative to a paired-samples t test appropriate for small samples, and McNemar’s test is a similar nonparametric test appropriate for paired, dichotomous outcomes. Textual responses to the open-ended survey questions, feedback from the therapists administering the intervention and from the clinician providing the fertility services was summarized thematically to provide insight to improve future administration of the intervention.
Ethics
The study was approved by the Institutional Review Boards at Jordan University of Science and Technology (6/141/2021, 6/1/2021) and Emory University (0000321, 9/3/2021). All participants provided written informed consent and the study followed international standards on research involving violence against women, including offering immediate professional assistance for violence or distress [49]. In addition, a special hotline was established by the institute delivering the CBT to support participants and set up a WhatsApp group for ongoing group support which continues to function to this day.
Add Comment