|Home | About | Journals | Submit | Contact Us | Français|
We examined male partners’ influence on the decision to seek medical help for infertility using from the National Survey of Fertility Barriers. Building upon an existing help-seeking framework, we incorporated characteristics of both partners from 219 heterosexual couples who had ever perceived a fertility problem. In logistic regression analyses, we found an association between couple-level attitudes and medical help-seeking even when other predisposing and enabling conditions existed. Overall, the findings highlight that both partners contribute to the infertility help-seeking process, and that different factors may play a role in different stages of help-seeking. Studies of infertility help-seeking need to be more inclusive of the context that these decisions are embedded within to better understand service use.
Medical help-seeking for infertility is a complex family health issue that is demanding increased attention. Since the 1980s there has been growing demand for and availability of infertility services; nevertheless, less than one-half of women who meet criteria for fertility problems seek medical help (Stephen & Chandra, 2000). Despite the growth in the number of women seeking medical help for infertility, researchers remain puzzled why such a large proportion does not seek medical help (Kalmuss, 1987; Stephen & Chandra).
The majority of help-seeking studies for infertility have focused mainly on women despite longstanding acknowledgment that infertility is inherently a couple issue with social dimensions (Greil, 1997; Greil, Leitko, & Potter, 1988; Lorber & Bandlamudi, 1993; Matthews & Matthews, 1986; Stephen & Chandra, 2000). We suggest that medical help-seeking behavior for infertility cannot be understood by studying only individual partners because it is necessarily a dyadic process for couples (Greil et al., 1988). The onset, cognition, and outcome involve not only the individual afflicted with the medical problem but also affect the fulfillment of social roles for both partners.
The health behavior model (Andersen, 1968), which was originally developed and subsequently expanded to address health care utilization, provides a useful framework for understanding medical help seeking for infertility. This model organizes aspects of health care use into three main components: predisposing, enabling, and need factors. Predisposing factors are characteristics that “exist prior to the onset [of an illness episode]” and increase propensity to use health services. Specifically, Andersen pointed to “family composition, health beliefs, and social structure” (p. 15) as key predisposing factors. Enabling factors are the means or conditions through which health services are made available to people, such as individual and family resources like health insurance, or community resources like local availability of services. Finally, need factors refer primarily to individuals’ appraisals of their health or health-related experiences.
Revisiting his original health behavior model, Andersen (1995) observed that social relationships fit conceptually as an enabling factor, acting to either “facilitate or impede” health service use (p. 3). Andersen’s observation is supported by a body of empirical work documenting the function of social relationships, particularly family member involvement in help-seeking and medical decision-making. Bass and Noelker (1987) pointed to the relationship between the family’s ability to provide care and elderly patients’ use of professional services. Looking at help-seeking pathways for patients with brain tumors, Salander, Bergenheim, Hamberg, and Henrikkson (1999) found that spouses played a key role in urging help-seeking and also provided useful information to physicians about symptoms and behavior that patients were not immediately aware of. Schäfer, Putnik, Dietl, Leiberich, Loew, and Kölbl (2006) addressed the often perceived conflict between patient autonomy and family inclusion in decision-making about medical care. Although they noted the prevalent view in the medical profession that “a disease is often regarded as an event, which concerns only one person” (p. 952), they found that the majority of patients in their study favored including family in decision-making. They concluded with a call for the formulation of a family-based decision-making theory.
Recently, White, McQuillan, Greil, and Johnson (2006) conceptualized and tested a model of medical help seeking for infertility based on the health behavior model, particularly emphasizing the social context of health decisions as well as infertility-specific indicators. White and colleagues included four main components: symptom salience, such as the strength of intentions to have children; life course factors, including age or marital status; individual and social cues, such as the importance of parenthood to oneself and one’s partner; and Andersen’s (1968) original predisposing and enabling components. White et al. posited and concluded that these primary factors were largely mediated by perception of oneself as infertile. They tested their model on women, but did not obtain data on men.
Family or couple-based medical decision-making is particularly relevant to infertility because it involves distinctly dyadic elements that may not apply to other medical conditions. First, there may be uncertainty if the problem resides with the male or female partner or both. Second, proper diagnosis and treatment may require disclosure about issues concerning physical intimacy between partners as opposed to simply information about the partner being medically treated. Finally, if the outcome is involuntary childlessness, both partners are prevented from becoming parents.
Even though both partners in an infertile couple are affected, many studies have described infertility as an asymmetric, gendered experience. Scholars have addressed gender differences in terms of the transition to non-parenthood (Ulbrich, Coyle, & Llabre, 1990), infertility-related stress (Greil, 1991a; Greil et al.,1988), the experience of infertility as a unique life problem (Andrews, Abbey, & Halman, 1992), and wives’ distress regardless of whether or not they physically have an impairment (Greil, 1997). Consequently, the female partner may act as the primary decision maker in help-seeking because she disproportionately experiences the effects of infertility (Greil, 1991a; Greil, 1991b; Greil et al., 1988; Lorber & Bandlamudi, 1993). Yet, given the extensive time and resources that infertility services may involve and the fact that certain procedures literally require consent of both partners (e.g., in-vitro fertilization), it seems an oversight not to view male partners as integral to the help-seeking process.
The first step to examining male partners’ influence is to add their characteristics to models of help-seeking. Further, to go beyond the individual level to address couple dynamics, we suggest that it is useful to view infertility help-seeking as a couple decision-making process. Godwin and Scanzoni (1989) introduced a context-process-outcome model for couple decision-making. Context refers to individual and couple characteristics that describe the relationship environment in which a particular issue is being decided; for instance, the couple’s socioeconomic status, partners’ racial identifications, or religious preferences. Process refers to aspects involved with negotiating the outcome. In their work, Godwin and Scanzoni (1989) examined the amount of control by each partner when making decisions. For our purposes, process factors are issues that partners need to communicate about in order to decide to seek medical help. Outcome is the final decision or behavior resulting from the negotiation process.
In the current study, we seek to extend the existing literature by examining whether male partners affect medical help-seeking for infertility. An additional research question focuses on whether couple agreement about family building attitudes is one way that male partners affect the help-seeking decision. We situate the process-context-outcome model in the broader framework for medical help seeking for infertility proposed by White et al. (2006), focusing mainly on the process component of this model. As part of the decision-making process, we address agreement between partners’ attitudes as a proxy for couple communication. Agreement conveys that a particular issue either has already been negotiated to the satisfaction of both partners (Scanzoni & Godwin, 1990), or that partners have similar attitudes if and when the topic arises in the future. We posit that couple communication and negotiation enables medical help-seeking for infertility. We begin by looking at fertility preferences and intentions and the importance of parenthood as these are immediately relevant to infertility help-seeking as a decision about family building.
Fertility preferences and intentions have a direct link to help-seeking behavior. Couples who want (more) children are more likely than couples who do not want (more) children to engage in help-seeking behavior. Complexities arise, however, because intentions may change based on life circumstances (Axinn & Yabiku, 2001; Voas, 2003), and partners may not agree about childbearing desires.
Several authors suggest that personal preferences interact within unions to become couple fertility intentions (Schoen, Astone, Kim, Nathanson, & Fields, 1999; Voas, 2003). Thomson, McDonald, and Bumpass (1990) and Schoen et al. (1999) found partner agreement to be important in achieving intentions; they did not find a gender effect when partners disagreed. Thomson (1997) found that couple disagreement “shifted intentions toward not having a child” (p. 347). Applying this to infertility, couples who agree on desired number of children and their intentions to have children might be more likely to seek medical help than couples with dissimilar preferences or intentions.
Although related to fertility preferences and intentions, the importance of parenthood is a separate dimension in childbearing decisions. This is the value placed on fulfilling the social role of being a parent. White et al. (2006) hypothesized that greater value of motherhood may lead women to “pay more attention to signals of infertility” (p. 1034), which would translate into greater propensity to seek help. Although they did not find direct support for this, we suggest that if both partners agree that being a parent is important they may be more likely to seek help when faced with infertility.
In our analysis, we map the contextual aspect of the context-process-outcome model onto main components from White et al.’s (2006) framework: symptom salience, life course factors, individual and social cues, and predisposing and enabling conditions. Attitudinal elements of these components could feasibly be a part of both the context and process of the decision. For example, health beliefs might fit as contextual factors that predispose one to using services, but beliefs could also emerge from or come into negotiation during the decision-making process. Godwin and Scanzoni (1989) explicitly recognize such feedback in their context-process-outcome model of couple decision-making.
Finally, the outcome in this analysis is whether a couple decides to seek medical help, which can range from simply talking to a doctor to pursuing tests and treatment. These different degrees of medical help likely have different implications for partner involvement. If a woman is simply talking to her doctor, her partner may not even have knowledge of this action. Medical tests or treatments, however, are more likely to require the involvement and consent of both partners as they entail greater investment of both time and money.
In summary, this analysis focuses on (1) fertility preferences and intentions and (2) the importance of parenthood as two factors in the decision to seek medical help for infertility. We expect both will be positively related to medical help-seeking. We further posit that agreement between partners acts as a proxy for couple communication and negotiation of childbearing attitudes, which may enable medical help-seeking. In other words, couples who agree about childbearing attitudes and have stronger intentions or place greater value on parenthood should be more likely to seek help. Given the gendered experiences of infertility and help-seeking, we expect male partners’ attitudes, and consequently, couple attitudes and agreement, to have greater relevance when help-seeking involves medical tests or treatment compared to simply talking to a doctor. We address two potential dynamics of agreement: (1) couple agreement may facilitate help-seeking, while disagreement may be a barrier to help-seeking, or (2) couple disagreement may be present, but since previous research has shown the female partner to be the primary decision-maker (Greil, 1991a; Greil, 1991b; Greil et al., 1988; Lorber & Bandlamudi, 1993); if her attitudes are stronger, then help-seeking may still occur.
We used data from the first wave of the National Study of Fertility Barriers (NSFB), which is a nationally representative telephone survey, obtained using random digit dialing (RDD). The NSFB was designed to understand the social and psychological aspects of infertility. The overall survey response rate was 37.2%. Although this was relatively low (but typical of recent national RDD telephone studies; Keeter, Kennedy, Dimock, Best & Craighill, 2006), observed differences were small in a comparison of similar items about fertility, infertility, and help-seeking in the National Survey of Family Growth (NSFG). For example, 85.1% of women surveyed in the NSFB had ever been pregnant compared to 83.4% of women surveyed in the NSFG. Among women who had ever been pregnant, 31.5% of women in the NSFB had ever miscarried compared to 32.5% of women in the NSFG. Most relevant to our current study, 15.6% of women in the NSFB reported seeing a doctor for help with a fertility problem or miscarriage compared to 16.2% in the NSFG. For a more extended comparison of select measures between the NSFB and NSFG, see Johnson and White (2009).
Sample selection for the current study was based on having both partner interviews completed as well as several cues and conditions from White et al.’s (2006) framework, such as being in a marital union, having health insurance, and perceiving a fertility problem. To minimize issues of timing and irrelevance given our use of lifetime definitions of infertility and help-seeking, we eliminated couples where either partner was surgically sterile, couples where the fertility problem occurred with a previous partner, and couples where neither partner wanted children. The final sample included 219 couples. This more homogeneous sample may limit the generalizability of findings but also helps to eliminate the potential effects of other barriers to help-seeking (e.g., the lack of health insurance) and focuses more on the addition of male partners and couple measures.
Because male partner interviews were completed for only 43% of women meeting the selection criteria, we assessed possible differences between these groups of women. In a logistic regression analysis we related 18 female partner and union characteristics to whether a male partner completed an interview (yes/no). Women with a partner interview were more likely to be White (odds ratio = 2.19, p < .01) and had greater fertility preferences and intentions (odds ratio = 1.18, p < .001). They were less likely to have talked to a doctor about a fertility problem (odds ratio = .36, p < .01) and their partners were less likely to have had children from a previous relationship (odds ratio = .50, p < .05). Two of these variables were central to our study, so we created a sample weight using the inverse probability of male partner response and incorporated this into the final analytic weight that adjusted for the NSFB sampling design. All analyses were weighted.
The outcome of interest was help-seeking behavior. Help seeking was measured using multiple yes/no items: (1) Have you ever been to a doctor or a clinic to talk about ways to help you have a baby? (2) Did you or your partner get medical tests to determine the nature of the problem? (3) Did you ever seek treatment to get pregnant? Responses to these were used to categorize couples into one of three mutually exclusive categories: (1) no services sought, (2) talked to a doctor only, and (3) had infertility-related tests or treatment. If either partner responded ‘yes’ to categories two or three, help-seeking was considered to have occurred for the couple. From these categories, we constructed two dichotomous measures: first, to compare those who had sought no medical services to those who had sought any and, second, to compare those who had at least talked to a doctor to those who went on for tests or treatment. Combining tests and treatment into one category better reflects the possibility that these are not discrete stages (some tests may also act as treatments) and do not always occur in an intuitive order (doctors often prescribe relatively inexpensive fertility drugs before running any tests).
Although one or both partners of each couple in our sample self-identified as having a fertility problem, following Greil and McQuillan (2004) and White et al. (2006), we further distinguished between those who were infertile with intent (they had an infertility episode while trying to have a baby) and those who were infertile without intent (they had unprotected sex for a year or more without conception but were not actively trying to get pregnant). Demographic studies on infertility and help-seeking generally have not accounted for episode-specific intentions (see, for example, Kalmuss, 1987; Stephen & Chandra, 2000). As Greil and McQuillan and White et al. (2006) showed, the distinctions between intentions at the time of the episode are strongly associated with help-seeking behavior because couples actively trying to conceive will most likely notice and respond to failed attempts.
Like White et al. (2006), we treated episode-specific intentions as distinct from the process factor, fertility preferences and intentions (White and colleagues use a similar measure labeled intensity of childbearing plans), because it captures intentions at a particular time as opposed to current, and future, attitudes about childbearing. We positioned fertility preferences and intentions as more general attitudes about family building that couples would need to negotiate prior to intentions during a specific episode. Additionally, fertility preferences and intentions indicate the strength of intentions, while episode-specific intentions are a simple categorical measure of whether or not the couple was trying to get pregnant (yes/no). As such, we suggest that these measures are associated with one another but have distinct relationships with help-seeking. Ideally, we would include both measures as part of couple agreement during the decision-making process. However, although both partners were asked about more general intentions, we were limited to deriving episode-specific intentions according to female partner responses because female partners were asked more extensively about their pregnancy history.
We also controlled for parental status of both partners (i.e., biological motherhood for women, which the current male partner may or may not be the biological father; for men, an indication of children from a previous relationship) and whether or not they were Catholic (yes/no). We included the latter control because of the official stance the Catholic Church has taken against assisted reproductive technologies (Congregation for the Doctrine of the Faith, 1987). Due to the small sample size, in couple level models, we combined religious preference into one measure indicating if either partner was Catholic. Most tests and treatment occur on the woman even for male factor infertility (Greil, 1991b), so we included a few contextual controls just for the female partner. These were race (White/Non-white), employment status (“Keeping House” versus all other employment statuses), and age (years). Union characteristics included years living together and total family income, which was measured in 12 categories ranging from ‘under $5,000′ to ‘$100,000 or more.’
We examined two main factors for the decision-making process: fertility preferences and intentions and the importance of parenthood. We constructed both individual partner and couple level scales to use in individual and couple models. Given that individual attitudes within dyads are generally dependent on one another (Kenny, Kashy & Cook 2006), the couple level measures do differ from the individual measures because they account for the shared variance between partners. As such, we conceive of the couple level measure as theoretically more complete than either of the individual measures. The individual measures do not acknowledge the influence of the relationship on personal attitudes and attribute all of the explanatory power to the individual, an issue that Kenny et al. referred to as pseudo-unilaterality.
Fertility preferences and intentions was a summated scale of the following three items: Like to have a baby (four categories, ‘definitely no’ to ‘definitely yes’); Intend to have a baby (no/maybe/yes); and Certainty of intentions (six categories ranging from ‘very sure, do not intend’ to ‘very sure, intend’). Because these items did not have similar response categories, male and female scores were standardized together before being summed (α = 0.93, 0.94, and 0.93 for female and male partners, individually, and the couple level scale, respectively). For ease of interpretation, we rescaled the final measure to range between 0 (minimal intentions to have (more) children) and 10 (highest intentions). A fourth item, desired number of children, was initially included, but appeared to measure a separate dimension of childbearing attitudes. We used this as a dichotomous measure of whether partners agreed on their desired number of children, which was included in models along with the other measures of couple agreement, discussed below.
The importance of parenthood scale was composed of six items, each with four response categories ranging from ‘strongly disagree’ to ‘strongly agree’: 1) Having children is important to me feeling complete as a woman/man, 2) I always thought I would be a parent; 3) Life will be or is more fulfilling with children; 4) It is important for me to have children; 5) It is important to my partner or spouse that we have children, and 6) It is important to my own parents that I have children. Higher scores indicated a greater value placed on parenthood (α = 0.82, 0.82, and 0.87 for female and male partners, individually, and the couple scale, respectively). In separate exploratory factor analyses for women and men, a single factor explained approximately 46% of the variance for women as well as for men. The lowest factor loading was .47 across indicators for both men and women; the majority of loadings were .5 and higher. According to Brown (2006), factor loadings of .3 to .4 or higher suggest a meaningful relationship between an indicator and a factor.
We used two types of difference scores to assess couple agreement: absolute and directional. First, we took the absolute value of the difference between partners’ scores multiplied by negative one. A value of zero indicated perfect agreement and negative values indicated a continuum of disagreement between partners. Then, for the directional measures we subtracted the male partner score from the female partner score. A positive score indicated a female partner effect, which is consistent with the expectation from previous studies that the female partner is the primary decision-maker (Greil, 1991a; Greil 1991b; Greil et al., 1988; Lorber & Bandlamudi, 1993).
Some scholars have addressed issues with the use and interpretation of difference scores (Glass & Polisar, 1987; Griffin, Murray, & Gonzalez, 1999). Griffin et al. stressed that too much substantive weight is often placed on the effects of difference scores rather than the main effects of the original component measures. Because we were interested in both the main effects (i.e., the individual and couple level attitudes) as well as the differences between partners, and we controlled for both partners’ contributions, we believe that concerns about the use of difference scores were minimized in this study.
Missing data were present under three circumstances in the NSFB: 1) user missing from ‘don’t know’ and ‘refuse’ responses, 2) missing due to survey skip patterns, and 3) a planned missing survey design for scale items. We recoded user missing data and skip patterns when considered appropriate. For example, respondents who answered ‘don’t know,’ or ‘leave it up to God or nature’ for fertility intentions were recoded as ‘neutral.’ The two items with the highest amount of user missing data were family income (4.6%) and asking men about their ideal number of children (3.2%). All other items had either 0% or less than 1% missing due to respondent refusals. The remaining missing data were from the planned missing design, which was used in the NSFB to shorten interviews by omitting items on scales. The planned missing design is specifically relevant to the importance of parenthood scale that we used in this study. Respondents were asked two-thirds of the items on each multi-item scale—one third of the items were omitted at random. Because planned missing was a randomized feature of the survey design, these values were missing completely at random (Allison, 2002) and do not bias the results (see Johnson, McQuillan, Shreffler, and Johnson (2006) for more information on and validation of this planned missing design).
We conducted multiple imputation using the ICE module in Stata v. 9 and created 10 datasets. Data were imputed based on the larger sample of women who met the sample selection criteria without necessarily having had a male partner interviewed. The imputation model included all variables used in the final analyses. We constructed scales and difference scores after imputing. Following the imputation of the missing values, only the women who had a responding male partner were retained for the analyses reported here. Analyses were conducted in each of the 10 datasets, and the estimates were then combined using the rules developed by Rubin (1987).
We used logistic regression to make two comparisons. First, we compared those who had not sought help to those who had sought any medical help. Then we compared those who had only talked to a doctor to those who had tests or treatment. We examined the effects of individual partners distinct from one another and from the contributions of union and couple agreement measures. For brevity, we present the results from the final ‘union’ model plus the two models measuring absolute and directional couple agreement. Because we used weighted data, the models were estimated with robust standard errors.
Descriptive information is shown in Table 1. In our sample, 30.9% of couples did not seek any medical services; 56.8% pursued tests or treatment. Slightly more than two-thirds (67.8%) were intending to get pregnant when they experienced an infertility episode. On average, women had slightly higher scores than men for fertility preferences and intentions and importance of parenthood. Less than half of the couples (44.7%) agreed on their ideal number of children.
We first look at the comparison between those who did not seek medical help and those who sought any form of medical help. The main predictors were intentions to get pregnant at the time of the infertility episode, number of years living together, and total family income. Couples who were intending to get pregnant were much more likely to seek some form of medical help as opposed to those who were not intending to get pregnant at the time of the infertility episode (odds ratios ranged from 10.66 to 11.57). Because these effect sizes were fairly large, we conducted a bivariate analysis between intentions and help-seeking. Estimates were similarly high in this analysis (odds ratio = 12.85, p < .000) suggesting that the multivariate effects did not result from violating any assumptions of logistic regression models, particularly collinearity, or sparse cell counts.
Longer duration living together and higher family income were also associated with increased odds of medical help seeking for infertility. Each additional year living together was associated with an 11-12% increase in the odds of seeking medical help across the three models. With each unit increase in family income, the odds of help-seeking increased by 24-25% in the two models of couple agreement. None of the couple level scales or agreement measures were significantly related to seeking any help compared to no medical help.
We next explored differences among couples who had sought any form of medical help, comparing those who only talked to a doctor to those who went on to having tests or treatment. Similar to the previous models, being infertile with intent and the number of years living together were both significant predictors of having tests or treatment compared to simply talking to a doctor. Those who were infertile with intent were approximately 5 to 6 times as likely to seek tests or treatment compared to couples who were not intending to get pregnant at the time of their infertility episode. For each additional year living together, couples were 19-21% more likely to seek tests or treatment. In contrast to the previous models, family income was not a significant predictor of help-seeking behavior.
In these models, couple level attitudes about the importance of parenthood were significant predictors of seeking tests or treatment compared to simply talking to a doctor: For each unit increase in importance of parenthood ratings, couples were 11-14% more likely to seek tests or treatment. In the ‘Agree Absolute’ model, the absolute difference between partners’ ratings of the importance of parenthood was also significantly related to help-seeking behavior. Partners who had more similar ratings about the importance of parenthood were more likely to seek tests or treatment (odds ratio = 1.26). Although the agreement measure did decrease the effect size for the couple level scale, both were significant in the same model, suggesting that they were related but not identical aspects of couple attitudes. There did not appear to be a gender effect (‘Agree Direction’ model) in situations where partners disagreed.
We also examined individual level models as well as models where we controlled for both partners attitudes separately rather than in combined couple scales (results available upon request). Individual measures of the importance of parenthood were significant predictors of help-seeking behavior in separate models for women and men (odds ratio for women was 1.22, p < .05; odds ratio for men was 1.15, p < .01), but when we simultaneously included both individual level measures (as opposed to combining them as a couple-level scale), the individual partner effects were no longer significant because they were highly correlated. This affirms the appropriateness of conceptualizing these attitudes at the couple level.
In sum, the main predictors of seeking any medical help compared to no medical help were intentions to get pregnant at the time of the infertility episode, years living together, and total family income. Among those who sought any medical help, intentions to get pregnant and time living together remained significant factors in having tests or treatment compared to just talking to a doctor. Attitudes about the importance of parenthood, specifically couple level attitudes, also played a significant role as well as absolute agreement between partners’ ratings of the importance of parenthood. This suggests that different factors are relevant to different stages of help-seeking. While contextual factors such as resources and immediate intentions to get pregnant may cue the initial use of medical services, couples most likely reflect on other issues when deciding to move forward with tests or treatment.
The majority of studies on infertility help-seeking have relied heavily on women’s characteristics to predict service use (see e.g. Kalmuss, 1987; Stephen & Chandra, 2000). Many of these studies conclude with the perplexing finding that less than half of women who fulfill the criteria for fertility problems seek help. We suggested that this finding might be better understood by bringing male partners into the analysis. In multivariate analyses, we found that couple level attitudes and agreement between partners was significantly associated with seeking tests or treatment among those who sought any form of medical help. Implicit in our findings is that different factors relate to different stages of help-seeking, and that male partners become more involved when help-seeking actions move beyond simply talking to a doctor.
We incorporated both partners by drawing on Godwin and Scanzoni’s (1989) context-process-outcome framework for couple decision-making. We examined attitudes about fertility preferences and intentions and the importance of parenthood in relation to medical help-seeking for infertility using two main comparisons: couples who had sought any type of medical help versus none and those who had talked to a doctor versus those who had tests or treatment. While couple attitudes and agreement about the importance of parenthood were not significant in the former comparison, they were in the latter. This supported our hypothesis that male partner attitudes, working via a couple dynamic, were more relevant to help-seeking that involved actions beyond just talking to a doctor.
We also concluded that the couple level scales and agreement between partners were related but distinct dimensions of couple attitudes about the importance of parenthood. We interpret the scale effect as a substantive relationship between the importance of parenthood ratings and help-seeking behavior and the agreement measure as a couple dynamic related to this scale. As such, the agreement measure cannot stand on its own but gains meaning in relation to the scale. Both agreement measures, however, provide important supplementary information about the couple level scale. Specifically, in examining the directional measures, we did not find that one partner’s attitudes were more influential on help-seeking behavior. The effect of importance of parenthood, then, was more relevant to partners having similar attitudes. This, however, may not be the case for all issues that couples communicate about regarding medical help-seeking. There may be certain issues that one partner feels more strongly about such as ethical issues, especially if partners have different religious backgrounds. Further research should explore agreement dynamics among various constructs that relate to both help-seeking and decision-making in order to further bridge these frameworks.
We did not find a relationship between fertility preferences and intentions and help-seeking behavior. We suggest that the two factors we examined are differentially affected by infertility although conceptually linked to childbearing decisions: Being confronted with infertility may not alter more abstract attitudes about the importance of being a parent, but it alters the realities of actually becoming a parent, which intentions are meant to reflect (Axinn & Yabiku, 2001). Additionally, the lifetime measures of infertility and help-seeking pick up those who have completed their childbearing, so their current attitudes and future intentions are no longer relevant. The issue of timing needs to be better distinguished in future studies, and longitudinal data would be particularly helpful here.
Interestingly, in their earlier study of infertility help-seeking, White et al. (2006) did not find a relationship between importance of parenthood attitudes and help-seeking in a sample of women with fertility problems. Their measure of help-seeking, however, was similar to our first comparison between any form of medical help and no medical help. Implicit in our findings is that different factors relate to different stages of help-seeking: Resources and immediate intentions may be more relevant to earlier stages, but other issues and attitudes may be more relevant to later stages. This points to the need for a more nuanced account of infertility and help-seeking.
Overall, our findings highlight that both partners contribute to the infertility help-seeking process, and that different factors may play a role in different stages of help-seeking. Studies of infertility help-seeking need to be more inclusive of the context that these decisions are embedded within and the complexity of the help-seeking process in order to gain a better understanding of service use. Frameworks such as the one proposed by White et al. (2006) help to integrate more social and attitudinal factors beyond purely demographic characteristics, but there is still a need to more thoroughly develop the role of both partners in relation to these frameworks. Given the extensive time and resources that infertility tests and treatment may involve and the fact that certain procedures literally require the consent of both partners (e.g. in vitro fertilization), it seems an oversight not to view male partners and couple communication as necessary to the decision-making process.
Andersen (1995) pointed to the methodological difficulties of including social relationships in models of health service use. More recent calls to view dyadic and relational frameworks as integral elements of social life rather than nuisance factors in research provide valuable resources for incorporating these social influences (Kenny et al., 2006). Although we have focused on integrating the male partner in studies of infertility and help-seeking, this framework will hopefully be of use more broadly in terms of other health service decisions that may be affected by dyadic or family relationships. Integrating dyadic and other family relationship models with medical help-seeking frameworks can provide a more holistic view of health service utilization and draw in the influence of social context on the individual.
This study filled gaps of prior research by including male partners and couples who did not seek medical help, however, there are limitations that need to be addressed in further research. A better understanding of couple dynamics and help-seeking behavior could be gained from data over multiple time points as opposed to the cross-sectional data we used here (the NSFB is in the process of collecting a second wave). The generalizability of the findings is also limited by the restrictions we applied in selecting our sample, although we believe that by eliminating other possible barriers to service use, we were able to focus more closely on couple dynamics. We also recognized methodological issues with difference scores, although we suggested that our study minimized these concerns. Further research may seek to expand or fine tune our knowledge by using alternate methods. Finally, we focused primarily on fertility-related dimensions, but there are many factors that couples may consider in the decision to seek medical help. These should be explored in order to provide a more holistic account of infertility help-seeking.
This study reaffirms the characterization of infertility in couples as a “dyadic process” (Greil et al., 1988): the experience, including the decision to seek help, is inherently social. The overt message following from this is that family practitioners should seek to engage with both female and male partners. The barrier we have proposed in this study, however, is particularly problematic: Couples who disagree or do not communicate may not seek any services at all. This may result in asymmetrically unsatisfactory outcomes, given the highly gendered experiences of infertility that Greil et al. (1988) observed two decades ago and is still salient today (Throsby & Gill, 2004).
So what is the role of practitioners if couple disagreement potentially leads to not seeking any services? A small but growing body of literature investigating health information online points to some possible solutions. In a prospective study, Weissman, Gotlieb, Ward, Greenblatt, and Casper (2000) investigated use of the Internet by fertility patients and the extent of available, relevant information. They concluded that the Internet could be an innovative tool for practitioners to educate patients and called for the establishment of “coherent information resources” online (p. 1182). Cousineau et al. (2008) examined more interactive possibilities for Internet health resources, developing and testing an online psycho-educational program for infertile women. Their proposed program was a self-directed, low cost intervention to help guide decision-making about treatment and provide an adjunct to other in-house support services.
Cousineau et al.’s (2008) program was aimed at couples who had sought medical help, however, we suggest this approach could be particularly promising for situations where there is no other point of contact between the would-be patient and the practitioner. This approach might feasibly educate and guide couples through help-seeking decisions as well as offer some support to individual partners for gender-specific aspects of infertility distress. Weissman et al. (2000) did find that women were more likely to use the Internet for fertility-related resources, but this may also have to do with framing of the sources, which are often presented as “women’s issues” (Carmeli & Carmeli, 1994; Throsby & Gill, 2004). Cousineau et al. tested their program on women, noting that future evaluations would be done on male partners and couples. This may be a promising avenue, working in tandem with research addressing couple decision-making in the help-seeking process.
This study was supported in part by grant # R01 HD044144 from the National Institute of Child Health and Development (NICHD), David R. Johnson, Principal Investigator. We would like to thank Larry Greil, P. Rafail, and Mary Casey Jacob for contributing time and energy to reading and providing feedback on successive drafts.