Background
Infertility is an ongoing public health problem and is estimated to affect 9% of reproductive-aged couples worldwide [
1]. Besides being a medical condition in itself, infertility is a social condition that can lead to considerable social and psychological distress [
2,
3]. Among psychological distress, anxiety is one of the most frequent psychiatric disorders in infertile patients. For example, in two studies conducted in Iran, the prevalence of anxiety and genialized anxiety disorder were 49.6% and 33.0%, respectively, which are considerably higher than what was reported in general population [
4,
5]. One the other hand, it has been frequently reported that infertility and its treatments can cause a negative effect on one’s marital satisfaction [
6,
7]. Also, some previous studies have found that subclinicals anxiety symptoms was negatively associated with marital satisfaction [
8]. Since infertility is a shared condition within the couple [
9], considering both members of the couple in the analysis of their relationships is necessary.
On the other hand, one of the basic assumptions underlying statistical analysis is the independence of observations. However, the dyadic designs (e.g., research on man-woman dyads and caregiver-patient dyads) create observations that are not independent which make the conventional statistical methods as inappropriate analysis tools [
10]. This type of design is usually applicable when researchers want to understand relationship processes between two people. Some of these processes are inherently dyadic processes, and one of the special characteristics of these processes is dyadic nonindependence. Conceptually, dyadic nonindependence is defined as “if the two scores from the two members of the dyad are non-independent, then those two scores are more similar to (or different from) one another than are two scores from two people who are not members of the same dyad” [
10]. Dyadic nonindependence occurs in a variety of contexts that involve two people, such as relationship processes studies in husband-wife, doctor-patient, caregiver-patient, parent-child, siblings and friends. In these conditions, statistical analytical techniques that take non-independence into account is required. To address this issue, Kenny et al. [
10] developed three important classes of dyadic models: the Actor-Partner Interdependence Model (APIM), Mutual Influence Model (MIM), and Common Fate Model (CFM). A detailed description of these models is presented in the Methods section.
Most of the studies with inherently dyadic nature in the infertility context and other behavioral medicine have been analyzed using individual as the unit of analysis rather than dyads. Although valuable, these studies fail to consider the dyadic non-independency of the data resulting in biased estimates of associations [
11]. However, in the last two decades, a great number of researchers have started to develop dyadic data analysis (DDA) techniques that consider dyadic nonindependence as a source of information rather than attempt to control for it [
12,
13]. In the context of infertility, APIM framework is by far the most widely used model to examine many relationship processes, such as effect of severe depressive symptoms on infertility-related distress [
14], attachment patterns on perceived infertility stress [
15], and the role of attachment anxiety and attachment avoidance on the psychosocial well-being [
16]. Although, the DDA are being used more and more in a variety of researches as a framework for investigating dyadic processes [
17‐
19], many of researchers still require more information on how to analyze the data in a way that optimizes its value [
13]. Thus, the aim of this article is twofold. First, we explain how DDA can aid our understanding of dyadic nature of phenomena in medical research. Second, we used three different dyadic models (i.e. APIM, MIM, and CFM) to examine how anxiety is associated with marital satisfaction in infertile couples. In this study, we examine the following research hypothesis:
APIM analysis: (1) One’s level of anxiety is associated with his/her own level of marital satisfaction (actor effects); (2) One’s level of anxiety is associated with his/her spouse’s level of marital satisfaction (partner effects); (3) There is a significant difference between male and female actor effects of anxiety on marital satisfaction; (4) There is a significant difference between male and female partner effects of anxiety on marital satisfaction. MIM analysis: (5) There is a dyadic feedback effect between males and females’ marital satisfaction; (6) There is a significant difference between male and female dyadic feedback effects of marital satisfaction. CFM analysis: (7) Couple’s anxiety is associated with couple’s marital satisfaction.
Discussion
We used the APIM, MIM, and CFM frameworks to explore the associations among husbands’ and wives’ assessments of anxiety and marital satisfaction, using data gathered from 141 infertile couples who seek treatments for infertility in Tehran, Iran. The findings from these analyses all show that anxiety is related to marital satisfaction. The interpretation of that relationship, however, would be different depending on which analysis one performed.
Key findings from the APIM analyses were that (a) men’s and women’s report of their anxiety predicted changes in their marital satisfaction (actor effects). Therefore, the Hypothesis 1 was accepted; (b) men’s and women’s anxiety was associated with their wife’s and husband’s marital satisfaction, respectively (partner effect) (based on Model 3 and 4). Therefore, the Hypothesis 2 was accepted; (c) actor effects as well as partner effects of anxiety on marital satisfaction were similar between men and women. Therefore, the Hypotheses 3 and 4 were rejected. The key feature of this APIM analysis which cannot be examined by individual model analytical approach is the explore the interpersonal effects (i.e. partner effects).
Findings of the MIM framework indicated that husbands’ and wives’ marital satisfaction were considerably reciprocally correlated (thus the Hypothesis 5 was accepted), suggesting feedback, and the magnitude of this association was similar (thus the Hypothesis 6 was rejected). The MIM also can test whether husbands’ and wives’ anxiety are indirectly related with their partners’ marital satisfaction via their own marital satisfaction (i.e., mediation). Based on the MIM findings, the indirect effects of spouses’ anxiety on their partners’ marital satisfaction imply that marital satisfaction in infertile patients was influenced by not only their own anxiety, but also their spouses’ anxiety.
The CFM framework provided a valuable insight with regard to the couple-level process. The key finding from this framework was that anxiety symptoms was related with marital satisfaction at the dyadic level. In other words, higher couple anxiety scores predicted lower couple marital satisfaction. Therefore, the Hypothesis 7 was accepted. In the data, two of the key indicators for selecting CFM as a reasonable model were met. First, the interpartner (dyadic) correlations were greater than 0.2 and 0.4 for anxiety and marital satisfaction scores, respectively, which approximately satisfied the recommendations in the literature [
28,
31] for variables that are thought to represent dyadic-level variation. Second, the result that actor and partner effects between anxiety and marital satisfaction were same in both direction and strength warrants the use of these variables in a CFM framework. Theoretical and methodological issues regarding the selecting of model are discussed in Ref [
31,
37].
Regarding preliminary analysis, women’s anxiety was significantly higher than their husbands, indicating that women may be more considerably affected than men by infertility problem. This result is in line with previous research [
38,
39]. Consistent with previous studies [
40‐
42], marital satisfaction was unrelated to gender. This finding is also in line with two previous studies on infertile couples’ quality of life, measured by the Fertility Quality of Life (FertiQoL), which found that the FertiQoL-Relational scores did not differ across partners [
43,
44]. However, in a study performed among infertile couples in Poland, women’s marital satisfaction was lower than their husbands [
45].
Of all dyadic data analytic approaches, the APIM framework has been—and is likely to continue to be—favored by researchers. However, when measurements are taken on both members of a dyad, researchers must decide whether the process is best represented as interdependent or common fate. The APIM and MIM would be appropriate dyadic models that represent interdependent processes. In APIM framework, interdependence of the outcome variables is hypothesized to come about because of the predictor variables, whereas in MIM framework, interdependence of the outcome variables may come about because these outcome variables simultaneously affect each other [
32]. The CFM is a dyadic model that is useful when both dyad members are affected by the common factor.
The current study offers a number of important contributions to the literature. Researches regarding the relationship between anxiety and marital satisfaction tend to focus on individuals, despite the obviously dyadic nature of marital satisfaction, particularly in the infertility context. Moreover, given the paucity of prior research on couples involved in committed marital relationships and shared health condition like infertility problem, the current research conducted among infertile married couples from a dyadic perspective. The results of this study have potentially important clinical implications. First, confirmation of the significance of the studied relationship at the level of the dyad members (individuals) and at the level of the dyads highlights the necessity of considering both members of the dyad in assessment. Second, therapists working with infertile couples should be aware of the different dyadic effects; therefore, psychological interventions that target an improvement in marital satisfaction should treat the couple as a unit.
Several limitations of the study should be noted. First, this research was carried out only in one infertility clinic and therefore limits the generalizability of our findings. Second, due to the cross-sectional nature of the study design, casual inference between anxiety and marital satisfaction cannot be made. Further studies, particularly longitudinal research, are required to disentangle the complex associations between spouses’ anxiety and marital satisfaction, including direction of causality and existence of dyadic feedback over time. Other limitations of this study were relatively small sample size, the lack of control for some important clinical variables (i.e., duration of infertility, cause of infertility, and failure of previous treatment), and the lack of specific infertility stress measures.