Human Reproduction, Vol.24, No.11 pp. 2810–2820, 2009 

Advanced Access publication on July 22, 2009 doi:10.1093/humrep/dep257

ORIGINAL ARTICLE Psychology and Counselling 

The impact of social relations on the incidence of severe depressive 

symptoms among infertile 

women and men 

1. Lund1, C.S. Sejbaek, U. Christensen, and L. Schmidt 

The Institute of Public Health, University of Copenhagen, 5 Øster Farimagsgade, PO Box 2099, DK-1014 Copenhagen K, Denmark 1Correspondence address. Tel: þ45-35-32-79-92; Fax: þ45-35-35-11-81; E-mail: r.lund@pubhealth.ku.dk 

background: The present study investigates anticipated support and perceived appreciation as well as perceived negative aspects of social relations as determinants of the incidence of severe depressive symptoms among women and men in unsuccessful fertility treatment. methods: A prospective cohort study of consecutively recruited, new couples in fertility treatment completed questionnaires immedi ately before initiating treatment (T1) and at 1-year follow-up (T2) (response rates 80%, 88%). The study population included 695 participants (355 women and 340 men) who had no severe depressive symptoms at T1, who had no child together at T1 and whose treatment had failed. We measured severe depressive symptoms with Mental Health Inventory 5 from Short-Form 36, functional aspects of general social relations and of infertility-specific social relations. 

results: Fifteen per cent of women and 6% of men in unsuccessful treatment reported severe depressive symptoms at T2. Among men, low emotional support, low appreciation and high excessive demands from the partner were significant determinants of the incidence of severe depressive symptoms. Among women and men, low appreciation from the family, many conflicts and high excessive demands from family, friends and neighbours were significant determinants of severe depressive symptoms. The analyses were controlled for gender. Duration of infertility, infertility diagnosis and socio-economic position did not confound the associations. 

conclusions: More women than men report severe depressive symptoms after 1 year of unsuccessful treatment. It is important to be aware of the possible negative impact of relational strain among fertility patients in unsuccessful treatment. 

Key words: assisted reproduction / depressive symptoms / prospective cohort study / relational strain / social relations 

Introduction 

Infertility, fertility treatment and depression are suggested to be related in complex ways (Williams et al., 2007). Jacob et al. (2007) found that significantly more women with a fertility barrier (e.g. infertility and other fertility problems) scored higher on depression compared with women with no fertility problems. A previous systematic review of women’s emotional adjustment to IVF (Verhaak et al., 2007a) showed that in general women did not differ from norm groups with respect to depression levels before the start of treatment. However, longitudinal studies measuring pre- and post-treatment emotional adjustment reported an increase in depression level after one or more unsuccessful treatment cycles. One study (Verhaak et al., 2005) evaluated anxiety and depression 6 months after the final treatment cycle and found no recovery from enhanced anxiety and depression 

levels after unsuccessful treatment among women. In this study, men showed no change in anxiety and depression either after suc cessful or after unsuccessful treatment. Sydsjo¨ et al. (2005) showed that couples in unsuccessful treatment have a stable relationship from the start of treatment and at 1 year after the last IVF cycle. Despite this, Lechner et al. (2007) reported from a cross sectional study among definite involuntarily childless women and men that the women had more health complaints, more anxiety and depressive symptoms compared with the background popu lations. In a longitudinal study, 3–5 years after start of IVF/ICSI treatment, Verhaak et al. (2007b) found that the depression level among those women who did not give birth to a child did not differ significantly between pretreatment and at follow-up. However, significantly higher levels of depression were found among those women still pursuing a desire for pregnancy compared with those who had abandoned their active pursuit of pregnancy. 

& The Author 2009. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org 

Social relations and depressive symptoms 2811

It seems therefore that depressive symptoms are both a possible consequence of infertility/failed fertility treatment and a risk factor for lower treatment success. Many, but not all, studies found that depressive symptoms may decrease the success rate of fertility treat ment (review in Williams et al., 2007). Further, the hormone treat ment used in IVF treatment is in itself suggested to be associated with depressive symptoms after unsuccessful treatment. de Klerk et al. (2007) found that when mild IVF treatment was compared with standard treatment, the level of depressive symptoms was lower among women in mild IVF-treatment. 

Infertility, social relations and depression Qualitative interviews with fertility patients have shown that infertile people find it severely challenging to learn how to manage the inferti lity in their social interactions with family, friends and co-workers (Schmidt, 1996; Tjørnhøj-Thomsen, 1999). Infertile people report both supportive and unsupportive social interactions (Miall, 1986; San delowski and Jones, 1986; Abbey et al., 1991; Schmidt, 1996; Tjørnhøj-Thomsen, 1999; Akizuki and Kai, 2008). Only few studies have investigated specifically the social interactions impact on the infertile individual’s mental well-being. Mindes et al. (2003) found in cross-sectional analyses among infertile women that infertility-specific unsupportive responses received from other people were associated positively with adjustment problems. After controlling for Time 1 adjustment, their longitudinal subanalyses including 67 women showed that unsupportive social interactions at Time 1 were associ ated positively with depressive symptoms and overall psychological distress among those women who had not achieved a pregnancy at Time 2. Slade et al. (2007) tested a model of stigma, disclosure, social support and distress among new attendees at a fertility clinic. Perceptions of stigma were related to low social support for both genders. Further, social support was negatively related to anxiety, depression and overall infertility distress. Lechner et al. (2007) found among definite involuntarily childless people that dissatisfaction with social support was positively associated with health complaints, depression, anxiety and complicated grief. 

Due et al. (1999) described social relations as a concept divided in structure (e.g. number and type of relations, frequency and duration of contacts) and function (e.g. social support and relational strain). In this study, we focus on the self-reported functional aspects of social relations, i.e. on anticipated social support (emotional support) and on perceived appreciation and perceived relational strain (conflicts and excessive demands), as previous research has shown that studying both the positive and the negative social exchanges is important (Rook, 1998; Newsom et al., 2005). In addition, we focus on social support as well as relational strain specifically in relation to the partici pant’s infertility and treatment, hereafter denoted infertility-specific social relations. Social support is associated with lower levels of depression and anxiety during times of stress and promotes psycho logical adjustment to a broad array of chronically stressful conditions (Taylor, 2007). Relational strain on the contrary is a risk factor for psychological distress and depressive symptoms (Rook, 1998, 2001; Finch et al., 1999). 

As part of a large, consecutive, prospective cohort study of fertility patients, we took the opportunity to study the impact of anticipated and perceived functional aspects of social relations on developing 

severe depressive symptoms during unsuccessful fertility treatment in more detail than in previous studies have achieved. Participants having severe depressive symptoms at recruitment were excluded from the analyses. For each functional aspect (support, appreciation, conflicts and excessive demands), we studied the impact from partner, family, friends, co-workers and neighbours separately. 

The objective of this study was to examine whether anticipated and perceived (i) functional aspects of social relations in general and (ii) infertility-specific social relations were determinants of the incidence of severe depressive symptoms among women and men not having achieved a treatment-related delivery or pregnancy at 1-year follow-up. On the basis of previous research, our assumptions were that developing severe depressive symptoms were associated with relational strain (Finch et al., 1999), infertility-specific unsupportive social interactions (Mindes et al., 2003), longer duration of infertility (Berg and Wilson, 1991), with male infertility diagnosis (Nachtigall et al., 1992), lower socio-economic position (Lorant et al., 2003) and female gender (Leibenhult, 1999). 

Materials and Methods 

Setting 

This study is part of The Copenhagen Multi-centre Psychosocial Infertility (COMPI) Research Programme, which includes a prospective longitudinal cohort study on 2250 participants initiating fertility treatment in the year 2000 – 2001 in four large public hospital-based tertiary fertility clinics and one private clinic (Schmidt, 2006). In Denmark, infertile couples are offered three fully reimbursed IVF or ICSI transfer cycles with fresh embryos, and an unrestricted number of cycles with frozen embryo repla cement and intrauterine insemination with husband or donor semen until the woman reach the age of 41 within the National Health Programme. All types of fertility treatment are available with self-payment at private fertility clinics until the female age of 46. 

Procedure 

In the period from January 2000 to August 2001, all new couples received a questionnaire for each spouse, immediately before their first treatment attempt at the clinic (T1). A follow-up questionnaire was sent 12 months later (T2, from January 2001 to August 2002). All questionnaires were returned to the author (L.S.) who was not employed at any of the clinics. Participants received up to two written reminders. 

The study was assessed by the Scientific Ethical Committee of Copen hagen and Frederiksberg Municipalities and no objections were noted. The study was approved by the Danish Data Protection Agency. 

Study population 

In total, 2812 people (1406 couples) received a questionnaire at T1, and 2250 (80.0%) responded. Forty-four patients were lost to follow-up: 38 participants (19 couples) whose identity was not registered at T1; 2 men and 2 women whose address could not be traced; 1 man had died; and 1 woman suffered a sever brain injury. In total, 2206 participants received the 1-year follow-up questionnaire (T2) and 1934 (87.7%) responded. In this study, we excluded: (i) participants who had had a child during the period from T1 to T2 or who were pregnant/partner pregnant at T2 (n ¼ 1106); (ii) those who had a child together at inclusion (n ¼ 58); (iii) those not having responded at T2 (n ¼ 284); (iv) those outside the social class classification (n ¼ 7), (v) those having severe depressive symptoms at T1 (n ¼ 66); and (vi) those who responded to 

2812 Lund et al.

,50% of the items in the Mental Health Inventory 5 (MHI-5) measuring depression (n ¼ 15). The final study population for this study comprised 355 women and 340 men (n ¼ 695) (Fig. 1). The total number included in the different analyses (Tables II–IV) varies according to missing responses on the specific items. 

Measurements 

Outcome 

‘Severe depressive symptoms’ were measured by the five-item MHI-5 from the 36-item Short-Form Health Survey (SF-36), the Medical 

Figure 1 Flow diagram of the study population. 

Social relations and depressive symptoms 2813

Outcomes Study (MOS) (Bjorner et al., 1998a, b). The MHI-5 was measured both at T1 and at T2. The MHI-5 at T1 was used to exclude those participants who had severe depressive symptoms at inclusion in COMPI from the longitudinal analyses in this study. MHI-5 measures how much of the time during the past 4 weeks the participant has: (a) been a very nervous person, (b) felt so down in the dumps that nothing could cheer them up, (c) felt calm and peaceful, (d) felt downhearted and blue, and (e) been a happy person. Response key was a six-point scale ranging from (1) all of the time to (6) none of the time. After reverse scoring for items c and e, the answers based on all five items were summed up for each participant into a raw score ranging from 5 to 30 (Bjørner et al., 1997). The raw score was transformed to a scale ranging from 0 to 100. The scale was dichotomized with the cut-point at 52 according to previous studies showing that participants scoring  52 are categorized as having severe depressive symptoms (Holmes, 1998; Strand et al., 2003). Participants with severe depressive symptoms were coded 1 and those without were coded 0. 

Determinants 

Social relations in general 

This measurement which is based on Due et al.’s (1999) conceptual fra mework on social relations has been used in several Danish study popu lations. We measured anticipated support (i.e. the possibility to talk to other people when support is needed) as well as perceived appreciation from other people and perceived relational strain (i.e. conflicts with other people; excessive demands from other people). Four items were included: (a) Are you able to talk to the following persons if you need support? (b) Do you experience that one or more of the following persons appreciate you? (c) Do you in your everyday life have conflicts with the following persons? (d) Do you think in your everyday life that some persons are demanding too much from you? For each item, five different social relations were listed—partner, family, friends, colleagues and neighbours. The response key for each item and for each group was ‘have none’ and a five-point scale from (1) always to (5) never. Test – retest reliability was estimated by weighted k-values which ranged between 0.40 and 0.80. According to Landis and Kock (1977), this corre sponds to moderate to excellent agreement. 

Infertility-specific social relations 

This measurement was developed from previous qualitative interviews with fertility patients (Schmidt, 1996). We used two items: (a) Do you get support and understanding from some people in relation to your child lessness or treatment? (b) Do you experience that some people react negatively to your childlessness or treatment? For each item, four different social relations—family, family-in-law, friends and colleagues—were listed. The response key was ‘have none’ and a five-point scale from (1) always to (5) never. 

Each variable for the infertility-specific and general social relations was either dichotomized or trichotomized based on sensitivity analyses against severe depressive symptoms testing different categorizations of the measures of social relations. Also the distribution of the responses in each single response category (i.e. from never to always) was taken into account. 

Control variables 

‘Occupational social class’ was measured in a standardized way, including seven items about school education, vocational training and occupation. Socio-economic position was categorized in a descending scale of occu pational social class from social class I (high) to social class V (low) (Hansen, 1984). Further, social class VI comprised participants who received social welfare benefits, and social class VII comprised participants 

still under education. In the analyses, occupational social class was recoded into: social class I þ II (professionals, executives, medium-level white collar employees), social class III þ IV (low-level white-collar employees, skilled workers), social class V þ VI (unskilled and semi-skilled workers, participants receiving social welfare benefits) and social class VII (students). Reference group was social class I þ II. 

‘Duration of infertility’ at inclusion (T1) was measured by the item: how many years have you and your partner tried to have a child? Categorized in ,2, 2, 3, 4 – 5, .5 years. Reference group was ,2 years. 

‘Infertility diagnosis’ was measured by one item: what did the medical examination show? The response key was (a) my/my partners tube(s) is/are blocked, (b) I/my partner have/has irregular ovulation or anovula tion, (c) I/my partner have decreased semen quality, (e) no cause was proven, (f) anything else. For this study, responses were categorized into female infertility, male infertility, both female and male infertility and no cause proven. Reference group was female infertility. 

Determinants and control variables were all measured at baseline (T1) and the outcome (incident depressive symptoms) was measured at 1-year follow-up. 

Data analyses 

Participants who responded ‘have none’ (n ¼ 1 – 50) when questioned about family/family-in-law/friends/colleagues/neighbours were excluded from the analyses. The association between general and infertility-specific functions of social relations and each of the control variables and, sub sequently, the association between the control variables and severe depressive symptoms were studied. As this was an exploratory study, associations with a significance level ,0.10 based on the x2 test and the gamma-coefficient were included in the further logistic regression ana lyses. We conducted univariate logistic regression analyses with calculation of crude odds ratios (ORs) for women and men combined. In the multi variate logistic regression analyses, the determinants of social relations in general and infertility-specific social relations were included for each source separately (e.g. partner and family). Only the control variables gender and occupational social class were included in all analyses, as dur ation of infertility and infertility diagnosis were not associated with the inci dence of severe depressive symptoms. In addition, the ORs for the final model were calculated by eliminating stepwise backwards the insignificant ORs. Initially, the interaction terms between the determinant (i.e. each measure of social relations separately) and gender as well as occupational social class were included. If a significant interaction between gender and a specific measurement for social relations was identified by including an interaction term, the analyses were conducted for women and men sep arately as well. There was no significant interaction between any of the measures of social relations and occupational social class. When only determinants associated with severe depressive symptoms at a significance level P , 0.05, the model was considered final. 

Results 

Table I shows the frequency distribution of the outcome incidence of severe depressive symptoms and of the control variables investigated. Almost 11% of the study population reported severe depressive symp toms at follow-up. Women reported severe depressive symptoms more often than men; 14.8% of the women and 5.7% of the men (data not shown). There was an equal distribution by gender throughout our study population (i.e. 51% women and 49% men). A little more than half of the population was categorized in social class III and IV and around a fifth in social class I þ II and V þ VI. At inclusion, the participants were in their mid-30s with men being older (mean age ¼ 34.7 years, 

2814 Lund et al.

Table I Frequency of depressive symptoms and distribution by gender, occupational social class, duration of infertility and infertility diagnosis in the study population (n 5 695) 

n % 

……………………………………………………………………………. Incidence of severe depressive symptoms (T2) 

No symptoms 622 89.5 Symptoms 73 10.5 Gender 

Female 355 51.1 Male 340 48.9 Occupational social class (T1) 

High, I þ II 150 21.6 Medium, III þ IV 367 52.8 Low, V þ VI 137 19.7 Students 38 5.5 Missing 3 0.4 Duration of infertility (T1) 

,2 years 28 4.0 2 years 111 16.0 3 years 171 24.6 4– 5 years 219 31.5 .5 years 166 23.9 Infertility diagnosis (T1) 

Female infertility 209 30.1 Male infertility 189 27.2 Female and male 74 10.6 Other cause 51 7.3 No cause proven 162 23.3 Missing 10 1.4 

T1, baseline; T2, follow-up. 

SD ¼ 5.3 years) than women (mean age ¼ 32.1 years, SD ¼ 3.6 years). At inclusion, the participants had in average lived together with their partner around 8 years (mean ¼ 7.6 years, SD ¼ 3.7 years). Around one-quarter of the study population had been infertile for more than 5 years. Nearly one-third reported female infertility and 27.2% reported male infertility. At 1-year follow-up, the participants had, on average, 2.6 treatment attempts (SD ¼ 1.3). Only 10.4% reported at 1-year follow-up that they have decided to stop treatment. 

Table II shows the distribution of the variables for general and infertility-specific social relations in the study population. Concerning the general social relations, the majority of our study population always expects to get emotional support from (83%) and to be appreciated by (76%) their partner. Almost half of the study popu lation reported always to expect emotional support from and to be appreciated by family and friends. Considerably fewer expect the same amount of support and appreciation from their colleagues and neighbours (i.e. between 6% and 24%). 

The major part of the study population never or rarely experienced conflicts with, or excessive demands from, their partner, family, friends, colleagues and neighbours (67 –97%). 

Concerning the infertility-specific social relations, the majority of the study population reported always or often to receive emotional support in relation to their infertility from their family, family-in-law, friends (56 –85%) and the vast majority never experience any negative reactions from their family, family-in-law, friends and colleagues in relation to their infertility (83 –90%). 

Only the associations between determinants and the incidence of severe depressive symptoms that were significant in the initial analyses, as described above in the Data analyses section, are included in the following analyses. 

Support: social relations in general Table III shows the results from the multivariate logistic regression analyses for the incidence of severe depressive symptoms by emotion al support from partner and by appreciation from partner and from family. 

There was surprisingly no association between occupational social class and severe depressive symptoms in this population. Conse quently, the following multivariate analyses are only adjusted for gender. 

Emotional support from partner 

Before stratification for gender, there was no association between receiving emotional support from partner and developing severe depressive symptoms, crude OR ¼ 0.97 (0.50–1.87). A significant interaction between gender and emotional support was found (data not shown). Analyses performed by gender showed that men who did not always receive emotional support from their partner had a more than two-fold higher OR for severe depressive symptoms com pared with men always receiving emotional support from their partner when needed. There was no association between emotional support from the partner and depressive symptoms among women. 

Appreciation from other people 

Likewise, in relation to appreciation from one’s partner, a significant interaction was found between appraisal and gender (data not shown), and hence these analyses were performed separately for women and men. Among men, there was a significant risk of severe depressive symptoms when comparing men who less than always experienced appreciation from their partner with men who reported always being appreciated (OR ¼ 3.52, 95% CI 1.41– 8.79). There was no association between appreciation from the partner and depressive symptoms among women. 

Among men and women, those who less than often experienced appreciation from their family were at increased risk of severe depressive symptoms compared with those who reported always to be appreciated by the family (OR ¼ 2.39 95% CI 1.21 –4.71) after adjustment for gender. 

Strain: social relations in general 

Conflicts with others 

Having conflicts with either family, friends or neighbours were signifi cant determinants of the incidence of severe depressive symptoms in gender-adjusted analyses, the ORs were two to three times higher in the exposed group. In contrast, conflicts with either partner or col leagues were not determinants of severe depressive symptoms in this 

Social relations and depressive symptoms 2815

Table II Frequency distribution (n and %) of general and infertility-specific social relations at baseline in the study population (n 5 695) 

Social relations in general 

……………………………………………………………………………………………………………………………………………………………………… Emotional support Conflicts 

……………………………………………………………………………… …………………………………………………………………………….. 

n % n % 

……………………………………………………………………………………………………………………………………………………………………… Partner (Missing: 15) Always 564 82.9 Partner (Missing: 5) Always 2 0.3 Often 96 14.1 Often 12 21.7 

Sometimes 18 2.7 Sometimes 213 30.9 

Rarely 2 0.3 Rarely 366 53.0 

Never 0 0.0 Never 97 14.1 

Family (Missing: 15) Always 352 51.8 Family (Missing: 5) Always 0 0.0 Often 199 29.3 Often 13 1.9 

Sometimes 98 14.4 Sometimes 119 17.3 

Rarely 24 3.5 Rarely 422 61.2 

Never 7 1.0 Never 136 19.7 

Friends (Missing: 16) Always 338 49.8 Friends (Missing: 5) Always 1 0.1 Often 194 28.6 Often 1 0.1 

Sometimes 111 16.3 Sometimes 62 9.0 

Rarely 29 4.3 Rarely 376 54.5 

Never 7 1.0 Never 250 36.2 

Colleagues (Missing: 39) Always 118 17.9 Colleagues (Missing: 27) Always 2 0.3 Often 141 21.5 Often 26 3.9 

Sometimes 201 30.6 Sometimes 193 28.9 

Rarely 103 15.7 Rarely 303 45.4 

Never 93 14.2 Never 144 21.6 

Neighbours (Missing: 27) Always 43 6.4 Neighbours (Missing: 17) Always 1 0.2 Often 40 6.0 Often 2 0.3 

Sometimes 80 12.0 Sometimes 20 3.0 

Rarely 100 15.0 Rarely 158 23.3 

Never 405 60.6 Never 497 73.3 

Appreciation Excessive demands 

Partner (Missing: 3) Always 523 75.6 Partner (Missing: 14) Always 12 1.8 Often 151 21.8 Often 23 3.4 

Sometimes 18 2.6 Sometimes 190 27.9 

Rarely 0 0.0 Rarely 267 39.2 

Never 0 0.0 Never 189 27.8 

Family (Missing: 3) Always 352 50.9 Family (Missing: 14) Always 4 0.6 Often 261 37.7 Often 22 3.2 

Sometimes 64 9.3 Sometimes 133 19.5 

Rarely 12 1.7 Rarely 287 42.1 

Never 4 0.4 Never 235 34.5 

Friends (Missing: 5) Always 278 40.3 Friends (Missing: 17) Always 2 0.3 Often 292 42.3 Often 7 1.0 

Sometimes 103 15.0 Sometimes 74 10.9 

Rarely 12 1.7 Rarely 312 46.0 

Never 5 0.7 Never 283 41.7 

Colleagues (Missing: 27) Always 160 24.0 Colleagues (Missing: 36) Always 7 1.1 Often 311 46.6 Often 66 10.0 

Continued 

2816 Lund et al.

Table II Continued 

Social relations in general 

……………………………………………………………………………………………………………………………………………………………………… Emotional support Conflicts 

……………………………………………………………………………… …………………………………………………………………………….. 

n % n % 

……………………………………………………………………………………………………………………………………………………………………… Sometimes 153 22.9 Sometimes 185 28.1 

Rarely 27 4.0 Rarely 194 29.4 

Never 17 2.5 Never 207 31.4 

Neighbours (Missing: 22) Always 77 11.4 Neighbours (Missing: 26) Always 0 0.0 Often 131 19.5 Often 8 1.2 

Sometimes 136 20.2 Sometimes 10 1.5 

Rarely 96 14.3 Rarely 100 15.0 

Never 233 34.6 Never 551 82.1 

Infertility-specific social relations 

Emotional support Negative reactions 

Partner (Missing: 26) Always 409 61.1 Partner (Missing: 20) Always 4 0.6 Often 100 14.9 Often 3 0.4 

Sometimes 73 10.9 Sometimes 21 3.1 

Rarely 35 5.2 Rarely 40 5.9 

Never 52 7.8 Never 607 89.9 

Family-in-law (Missing: 32) Always 340 51.3 Family-in-law (Missing: 28) Always 3 0.5 Often 109 16.4 Often 5 0.8 

Sometimes 94 14.2 Sometimes 30 4.5 

Rarely 41 6.2 Rarely 37 5.6 

Never 79 11.9 Never 592 88.8 

Friends (Missing: 20) Always 329 48.7 Friends (Missing: 18) Always 1 0.2 Often 175 25.9 Often 1 0.2 

Sometimes 87 12.9 Sometimes 25 3.7 

Rarely 32 4.7 Rarely 62 9.2 

Never 52 7.7 Never 588 86.9 

Colleagues (Missing: 50) Always 193 29.9 Colleagues (Missing: 42) Always 2 0.3 Often 167 25.9 Often 4 0.6 

Sometimes 113 17.5 Sometimes 26 4.0 

Rarely 55 8.5 Rarely 79 12.1 

Never 117 18.1 Never 542 83.0 

study population. The closer the relationship was (family versus friends versus neighbours) the higher were the significant ORs (Table IV). 

Excessive demands 

In the analyses of excessive demands from the partner and severe depressive symptoms, we found a significant interaction between demands and gender (P ¼ 0.02, data not shown), and these analyses were hence conducted for women and men separately. Among men, excessive demands from the partner were significant determinants of the incidence of severe depressive symptoms (OR ¼ 5.85, 95% CI 1.30 –26.31). Always and often to experience excessive demands from the partner was also associated with an increased risk of severe depressive symptoms among women, although not significant. Among men and women, excessive demands from family, friends or neighbours were significant predictors of severe depressive symptoms (Table IV). 

Infertility-specific social relations 

Neither social support nor negative reactions in relation to the infer tility from family, family-in-law, friends or colleagues were in this study determinants of the incidence of severe depressive symptoms. 

Discussion 

Incidence of severe depressive symptoms In our study, 14.8% of the women in unsuccessful treatment devel oped severe depressive symptoms at 1-year follow-up. This is com parable with a Dutch longitudinal cohort study reporting that 20% of the women after a final unsuccessful treatment cycle scored above the threshold for subclinical forms of depression measured by the Beck Depression Inventory (Verhaak et al., 2005). The reported 

Social relations and depressive symptoms 2817

Table III ORs (95% CI) of the incidence of severe depressive symptoms by emotional support from partner and by appreciation from partner and from family (n 5 695) 

Table IV ORs (95% CI) for the incidence of severe depressive symptoms by conflicts with other people and by excessive demands from partner and other people (n 5 695) 

Total n/n with severe depressive symptoms 

OR (95% CI) 

Total N OR (95%CI) 

……………………………………………………………………………. 

……………………………………………………………………………. Emotional support from: 

Partner 

Women (n ¼ 354) 

Conflicts with: 

Family (n ¼ 687) 

Often/sometimes 132/31 3.35 (1.55–7.23)* Rarely 422/32 0.96 (0.46– 2.03) 

Often/ 

sometimes/ rarely/never 

286/46 0.51 (0.21 –1.24) 

Never 136/10 1.00 Friends (n ¼ 690) 

Always 68/6 1.00 Men (n ¼ 326) 

Always/often/sometimes 64/13 2.84 (1.31–6.19)* Rarely 376/40 1.43 (0.80– 2.54) 

Often/ 

sometimes/ rarely/never 

278/14 2.69 (0.98 –7.40) 

Never 250/19 1.00 Neighbours (n ¼ 678) 

Always/often/sometimes/rarely 181/26 1.76 (1.04–2.96)* 

Always 48/6 1.00 Appreciation from: 

Partner 

Women (n ¼ 354) 

Never 497/47 1.00 Excessive demands from: 

Partner 

Women (n ¼ 354) 

Often/ 

sometimes/ rarely/never 

76/9 0.73 (0.34 –1.58) 

Always/often 12/4 2.87 (0.82– 10.00) Sometimes 86/10 0.76 (0.36– 1.59) Rarely/never 256/38 1.00 

Always 278/43 1.00 Men (n ¼ 338) 

Men (n ¼ 327) 

Always/often 23/3 5.85 (1.30–26.31) 

Often/ 

sometimes/ rarely/never 

93/11 3.52 (1.41–8.79) 

Sometimes 104/12 5.09 (1.74–14.86) Rarely/never 200/5 1.00 

Always 245/9 1.00 Family (n 5 689) 

Family (n ¼ 681) 

Always/often/sometimes 159/26 1.85 (1.10–3.12)* 

Sometimes/ rarely/never 

79/15 2.39 (1.21–4.71)* 

Rarely/never 522/47 1.00 Friends (n ¼ 678) 

Often 261/26 1.17 (0.67 –2.02) Always 352/32 1.00 

OR’s in bold: P , 0.05. *Adjusted for gender. 

prevalence in this Dutch study was, however, not adjusted for the prevalence of depression when starting fertility treatment. In total, 5.7% of the men in our study developed severe depressive symptoms during follow-up. The prevalence of depression among men initiating fertility treatment was 5.1.% in an on-going longitudinal cohort study from Sweden (Volgsten et al., 2008). We have not been able to ident ify any previous longitudinal study reporting incidence of depression among men in fertility treatment and therefore regard this new knowledge. 

In contrast to a previous meta-analysis (Lorant et al., 2003) report ing evidence for social inequality in depression, we identified no associ ation between occupational social class and the incidence of severe depressive symptoms in this cohort of infertile men and women. The distribution of participants according to occupational social class was fairly similar to the background population, as demonstrated in a recent survey based on a random sample of 40 –50 years old in Denmark (Christensen et al., 2004). In contrast to most other 

Always/often/sometimes 83/15 3.54 (1.17–7.48)* Rarely 312/38 2.04 (1.13–3.68) Never 283/18 1.00 

Neighbours (n ¼ 669) 

Often/sometimes/rarely 118/19 1.96 (1.10–3.49)* Never 551/54 1.00 

OR’s in bold: P , 0.05 *Adjusted for gender. 

health outcomes, infertility has been shown to be equally distributed over the social classes in a representative population-based study from Denmark (Schmidt et al., 1995). However, even in a country with free access to reimbursed fertility treatment, there was social inequality in medical help-seeking, with women with shorter education seeking treatment to a lower extent (Schmidt et al., 1995). Appar ently, the social inequality in depression in population studies is not present in our patient population. It could be that the stressful life event of not becoming parents may harm unequally over the social strata in contrast to many other health outcomes. Perhaps the inferti lity and its treatment is one of the first times people from the higher social classes experience a serious loss of control over their life, which 

2818 Lund et al.

again influences their mental well-being negatively. Another possible explanation of the severe depressive symptoms being socially equal in the COMPI study could be due to selection bias. It is reasonable to believe that the most vulnerable infertile people from lower social classes either do not seek fertility treatment or participate in a longitudinal cohort study with repeated self-reported measure ments. Therefore, we might have underestimated the effect of occu pational class in this study. 

Infertility-specific social relations 

In contrast to previous research (Mindes et al., 2003), we identified no significant associations between either anticipating support and under standing or perceiving negative reactions in relation to the infertility problem from other people. 

Very few individuals in our population reported experiencing nega tive reactions sometimes or more often (3 –4%). The problem is prob ably so rare and the effect too small to be of any relevance in this population. This low level of negative reactions is comparable with Mindes et al.’s study (2003) where most infertile women reported relatively low levels of unsupportive interactions; a level that was com parable to results based on a sample of college students. A recent qualitative interview study from Japan (Akizuki and Kai, 2008) ident ified nine different negative social interactions, e.g. prying on the topic of children, showing a negative attitude towards infertility and reproductive medicine, being criticized for not having children. Prob ably our measurement of unsupportive social interactions in COMPI could have been expanded and hence been able to capture more of the different negative social interactions infertile people sometimes experience. 

A larger proportion of our study population reported receiving support and understanding in relation to the infertility ‘rarely’ or ‘never’ (12 –27%). We found the lowest degree of support from col leagues and family-in-law. This is in line with the findings of van Balen et al. (1996) who found that family and friends, persons close to the individual, are almost always told about the infertility problem, whereas persons less close to the individual such as colleagues and neighbours are told less often. 

Gender differences 

In our study, men seemed more vulnerable to lack of emotional support, lack of appreciation and high level of demands from their partner than women did. Thus, women were found to have a greater impact on their partners’ incident depressive symptoms than men on their female partners’ mental health. It has been suggested that men are more dependent on emotional support from their partner than women are. Further, it has been suggested that women may be more nurturing, and therefore men might have an advantage in relation to receiving support in the relationship compared with women (Wareham et al., 2007). Greil’s (1997) review indicated that women in infertile couples are more satisfied with the support they receive from their partner, which could explain why emotional support from the partner was not associated with severe depressive symptoms among women in our study. 

The data in the present study were collected from couples, and the paired data may not be independent. There is a possibility that levels of depression and perceptions of social support or negative aspects of 

social relations could be related. Hypothetically, there is a risk that both partners agree on their social relations being unsupportive and it may consequently increase the risk of depression in both. Since knowledge about the influence of social relations on the risk of devel opment of depression is scarce, we chose in the present study to focus on individual level analyses. However, in future studies, multile vel models taking the effect of the couple into account may be of great value. 

Conflicts 

It was surprising that conflicts in relation to a partner were not associ ated with developing severe depressive symptoms, but conflicts with family, friends and neighbours were significant determinants. Conflicts are inevitable between an individual and significant others, as suggested by Abbey et al. (1995), and must be considered a normal part of a close relationship. In their longitudinal study among infertile couples and presumed fertile couples, Abbey et al. (1995) reported that increased perceived disregard was related to a decrease in marital life quality for both partners in both groups. It could be assumed that as long as conflicts with the partner are not the predo minating state, the risk of depression may not be increased. We assume that couples continuing infertility treatment will not have long lasting conflicts as going through treatment also requires a supportive relationship (Schmidt, 1996; Tjørnhøj-Thomsen, 1999). In contrast, ongoing conflicts with more distant relations may be more strenuous if the conflict is not resolved or if the relationship does not also contain supportive elements. 

Excessive demands 

Excessive demands from both partners, family, friends and neighbours were predictive of later severe depressive symptoms. In contrast to conflicts, excessive demands may not necessarily be part of the relationship with significant others, at least not for extended periods. In a stressful situation like being unsuccessful in fertility treat ment, the reaction to excessive demands may be stronger than in the general population. Excessive demands have been shown to be a pre dictor of the incidence of angina pectoris in population studies (Lund and Christensen, 2008). 

Supportive and unsupportive interactions The present study clearly shows that perceived relational strain among women and men is a significant determinant for developing severe depressive symptoms. Previous studies in populations not undergoing fertility treatment support this (Rook, 1998; Finch, 1999; Rook, 2001). Our findings put an emphasis on how important it is to measure social support and relational strain at the same time when investigating the impact of social relations on mental well-being. 

Measurement of depressive symptoms MHI-5 is usually used to measure mental health in general. However, previous studies comparing clinical diagnosis of depression with MHI-5 have shown that this scale when dichotomized  52 point is a valid measurement of severe depressive symptoms (Rumpf et al., 2001). Cronbach’s a for MHI-5 has been estimated at 0.80, in a general Danish population (Bjørner et al., 1997) and in an earlier testing, it exceeded 0.90 (Ware et al., 1994). 

Social relations and depressive symptoms 2819

Study design 

The prospective, longitudinal study design was useful in investigating the development of severe depressive symptoms after failed fertility treatment. By having repeated identical measurement of MHI-5, it was possible to exclude participants from the analyses who had severe depressive symptoms at baseline. Although we restricted our study population to those without severe depressive symptoms at baseline, there still remains a small risk of reverse causality between the functional measures of social relations and development of severe depression at follow-up. Individuals with light and moderate depressive symptoms at baseline may have been included in our study population, and these individuals may both be at higher risk of developing severe depressive symptoms and be more liable to report lack of support and high levels of strain from the social relations. In other words, we are not able to determine whether light/moderate depressive symptoms reported at T1 are the result of negative aspects of the social relations at baseline, for example, or the other way round. Longitudinal studies of the possible effect of social relations on the risk of developing severe depressive symp toms among infertile are few, and we believe that our study design is suited to investigate the problem. 

A particular strength of this study was the use of a validated instru ment, previously used in several other study populations in Denmark, for measuring the functional aspects of social relations (Lund and Christensen, 2008). The study sample was large and response rates both at T1 and at T2 were high. 

The COMPI project is designed with fixed period of data collection with 12 months between baseline data collection and first follow-up. As a consequence, the participants are at different stages regarding the treatment process at follow-up. It is possible that the development of depressive symptoms is related, in the short term, to the time since last unsuccessful treatment, as Verhaak et al. (2007b) showed in their long-term follow-up study. However, a design with data collection intervals linked to the treatment process would not necessarily solve the problem of patients being in different stages of emotional reactions to unsuccessful treatment. 

In the analyses of risk factors of the incidence of severe depressive symptoms, we included factors shown in previous empirical studies to be risk factors. It is, however, possible that other factors not pre viously identified could be important risk factors as well. We have later investigated age, duration of partnership, number of unsuccessful treatment attempts and decision-making about future treatment, but none of these factors was associated with the incidence of severe depressive symptoms in this study (data not shown). 

Clinical implications 

It is important for clinical staff to be aware that some fertility patients develop severe depressive symptoms after unsuccessful treatment. Further, it is important to be aware of the negative impact of relational strain among fertility patients being in a vulnerable period of life. 

Acknowledgement 

The Infertility Cohort is part of The Copenhagen Multi-centre Psycho social Infertility (COMPI) Research Programme initiated by Dr Lone Schmidt, University of Copenhagen, 2000. The programme is a 

collaboration between the Danish public Fertility Clinics at Braedstrup Hospital; Herlev University Hospital; The Juliane Marie Centre, Rig shospitalet; Odense University Hospital. 

Funding 

This study has received support from the Danish Health Insurance Fund (J.nr. 11/097-97), the Else and Mogens Wedell-Wedellsborgs Fund, the manager E. Danielsens and Wife’s Fund, the merchant L.F. Foghts Fund, the Jacob Madsen and Wife Olga Madsen’s Fund, and the Engineer K.A. Rohde and Wife’s Fund. 

References 

Abbey A, Andrews FM, Halman LJ. The importance of social relationships for infertile couples’ well-being. In: Stanton AL, Dunkel-Schetter C (eds). Infertility: Perspectives from Stress and Coping Research. New York: Plenum Press, 1991, 61 – 86. 

Abbey A, Andrews FM, Halman LJ. Provision and receipt of social support and disregard: what is their impact on the marital life quality of infertile and fertile couples? J Pers Soc Psychol 1995;68:455 – 469. 

Akizuki Y, Kai I. Infertile Japanese women’s perception of positive and negative social interactions within their social network. Hum Reprod 2008;23:2737 – 2743. 

Berg BJ, Wilson JF. Psychological functioning across stages of treatment for infertility. J Behav Med 1991;14:11– 26. 

Bjorner JB, Damsgaard MT, Watt T, Groenvold M. Tests of data quality, scaling assumptions, and reliability of the Danish SF-36. J Clin Epidemiol 1998a;51:1001– 1011. 

Bjorner JB, Thunedborg K, Kristensen TS, Modvig J, Bech P. The Danish SF-36 health survey: translation and preliminary validity studies. J Clin Epidemiol 1998b;51:991– 999. 

Bjørner JB, Damsgaard MT, Watt T, Bech P, Rasmussen NK, Kristensen TS, Modvig J, Thunedborg K. Dansk manual til SF-36. Et spørgeskema om helbredsstatus. [A Danish Manual to SF-36. A Questionnaire Regarding Health]. Copenhagen: Lif, 1997. 

Christensen U, Lund R, Damsgaard MT, Holstein BE, Ditlevsen S, Diderichsen F, Due P, Iversen L, Lynch J. Cynical hostility, socioeconomic position, health behaviors, and symptom load: a cross-sectional analysis in a Danish population-based study. Psychosom Med 2004;66:572 – 577. 

de Klerk C, Macklon NS, Heijnen EMEW, Eijkemans MJC, Fauser BCJM, Passchier J, Hunfeld JAM. The psychological impact of IVF failure after two or more cycles of IFV with a mild versus standard treatment strategy. Hum Reprod 2007;22:2554– 2558. 

Due P, Holstein B, Lund R, Modvig J, Avlund K. Social relations: network, support and relational strain. Soc Sci Med 1999;48:661 – 673. Finch JF, Okun MA, Pool GJ, Ruehlman LS. A comparison of the influence of conflictual and supportive social interactions on psychological distress. J Pers 1999;67:581– 621. 

Greil AL. Infertility and psychological distress: a critical review of the literature. Soc Sci Med 1997;45:1679 – 1704. 

Hansen EJ. Social grupper i Danmark. [Social Classes in Denmark]. Copenhagen: The Danish National Institute of Social Research, 1984. Holmes WC. A short, psychiatric, case-finding measure for HIV 

seropositive outpatients: performance characteristics of the 5-item mental health subscale of the SF-20 in a male, seropositive sample. Med Care 1998;36:237 – 243. 

Jacob MC, McQuillan J, Greil AL. Psychological distress by type of fertility barrier. Hum Reprod 2007;22:885 –894. 

2820 Lund et al.

Landis JR, Kock GG. The measurement of observer agreement for categorical data. Biometrics 1977;33:159– 174. 

Lechner L, Bolman C, van Dalen A. Definite involuntary childlessness: associations between coping, social support and psychological distress. Hum Reprod 2007;22:288 – 294. 

Leibenhult E. Gender differences in major depressive disorder and bipolar disorder. CNS Spectr 1999;4:25 – 33. 

Lorant V, Delie`ge D, Eaton W, Robert A, Philippot P, Ansseau M. Socioeconomic inequalities in depression: a meta-analysis. Am J Epidemiol 2003;157:98 – 112. 

Lund R, Christensen U. Distress from social relations and angina pectoris: a 6-year follow-up study of middle-aged Danish men and women. J Epidemiol Community Health 2008;62(Suppl 1):A38. 

Miall CE. The stigma of involuntary childlessness. Soc Probl 1986; 33:268 – 282. 

Mindes EJ, Ingram KM, Kliewer W, James CA. Longitudinal analyses of the relationship between unsupportive social interactions and psychological adjustment among women with fertility problems. Soc Sci Med 2003; 56:2165 – 2180. 

Nachtigall RD, Becker G, Wozny M. The effects of gender-specific diagnosis on men’s and women’s response to infertility. Fertil Steril 1992;57:113 – 121. 

Newsom T, Rook KS, Nishishiba M, Sorkin DH, Mahan TL. Understanding the relative importance of positive and negative social exchanges: examining specific domains and appraisals. J Gerontol B Psychol Sci Soc Sci 2005;60:P304–P312. 

Rook KS. Emotional health and positive versus negative sides of personal relationships: through a glass darkly? In: Spitzberg BH, Cupach WR (eds). The Dark Side of Close Relationships. Mahwah, NJ: Lawrence A Erlbaum Associates, 1998, 369 –393. 

Rook KS. Emotional health and positive versus negative social exchanges: a daily diary analysis. Appl Dev Sci 2001;5:86 – 97. 

Rumpf H-J, Meyer C, Hapke U, John U. Screening for mental health: validity of the MHI-5 using DSM-IV Axis I psychiatric disorders as gold standard. Psychol Res 2001;105:243 – 253. 

Sandelowski M, Jones LC. Social exchanges of infertile women. Issues Ment Health Nurs 1986;40:173– 189. 

Schmidt L. Psykosociale konsekvenser af infertilitet og behandling [Psychosocial Consequences of Infertility and Treatment]. Copenhagen: FADL’s Publishers, 1996. 

Schmidt L. Infertility and Assisted Reproduction in Denmark. Epidemiology and Psychosocial Consequences. Copenhagen: Lægeforeningen’s Publishers, 2006. 

Schmidt L, Mu¨nster K, Helm P. Infertility and the seeking of infertility treatment in a representative population. Br J Obstet Gynaecol 1995; 102:978– 984. 

Slade P, O’Neill C, Simpson AJ, Lashen H. The relationship between perceived stigma, disclosure patterns, support and distress in new attendees at an infertility clinic. Hum Reprod 2007;22:2309– 2317. 

Strand BH, Dalgard OS, Tambs K, Rognerud M. Measuring the mental health status of the Norrwegian population: a comparison of the instruments SCL-25, SCL-10, SCL-5 and MHI-5 (SF-36). Nord J Psychiatry 2003;57:113 –118. 

Sydsjo¨ G, Ekholm K, Wadsby M, Kjellberg S, Sydsjo¨ A. Relationships in couples after failed IVF treatment: a prospective follow-up study. Hum Reprod 2005;20:1952 – 1957. 

Taylor SE. Social support. In: Friedman HS, Silver RC (eds). Foundations of Health Psychology. New York: Oxford University Press, 2007, 145 – 171. 

Tjørnhøj-Thomsen T. Tilblivelseshistorier. Barnløshed, slægtskab og forplantningsteknologi i Danmark [Genesis. Childlessness, Kinship, and Reproductive Technology in Denmark]. Copenhagen: University of Copenhagen, 1999. 

van Balen F, Trimbos-Kemper T, Verdurmen J. Perception of diagnosis and openness of patients about infertility. Patient Educ Couns 1996; 28:247 – 252. 

Verhaak CM, Smeenk JMJ, van Minnin A, Kremer JAM, Kraaimaat FW. A longitudinal, prospective study on emotional adjustment before, during and after consecutive fertility treatment cycles. Hum Reprod 2005; 20:2253 – 2260. 

Verhaak CM, Smeenk JMJ, Evers AWM, Kremer JAM, Kraaimaat FW, Braat DD. Women’s emotional adjustment to IVF: a systematic review of 25 years of research. Hum Reprod Update 2007a;13: 27 – 36. 

Verhaak CM, Smeenk JMJ, Nahuis MJ, Kremer JAM, Braat DDM. Long-term psychological adjustment to IVF/ICSI treatment in women. Hum Reprod 2007b;22:305– 308. 

Volgsten H, Skoog Svanberg A, Ekselius L, Lundkvist O¨ , Sundstro¨m Poromaa I. Prevalence of psychiatric disorders in infertile women and men undergoing in vitro fertilization treatment. Hum Reprod 2008; 23:2056 – 2063. 

Ware JE, Kosinski M, Keller SD. SF-36 Physical and Mental Health Summary Scales: A User’s Manual. Boston, MA: The Health Institute, 1994. Wareham S, Fowler K, Pike A. Determinants of depression severity and duration in Canadian adults: the moderating effects of gender and social support. J Appl Soc Psychol 2007;37:2951– 2979. 

Williams KE, Marsh WK, Rasgon NL. Mood disorders and fertility in women: a critical review of the literature and implications for future research. Hum Reprod Update 2007;13:607– 616. 

Submitted on February 20, 2009; resubmitted on May 4, 2009; accepted on May 20, 2009