- Email: [email protected]
Is fertility-problem stress different? The dynamics of stress in fertile and infertile couples*†
FERTILITY AND STERILITY
Vol. 57, No.6, June 1992
Copyright" 1992 The American Fertility Society
Printed on O£id·{ree paper in U.S.A.
Is fertility-problem stress different? The dynamics of stress in fertile and infertile couples*t
Frank M. Andrews, Ph.D.:!:§ Antonia Abbey, Ph.D. II L. Jill Halman, Ph.D.:!: University of Michigan, Ann Arbor, and Wayne State University, Detroit, Michigan
Objective: To compare the dynamics of fertility-problem stress experienced by wives and husbands in infertile couples with the dynamics of stress from other sources experienced by members of couples presumed to be fertile. Design: Relationships of stress to four marriage factors and four aspects of life quality (subjective well-being) are examined within a causal modeling framework using data from structured interviews. Setting: Face-to-face interviews were conducted in study participants' homes. Participants: Wives and husbands from 157 couples with primary infertility and from 82 presumed-fertile couples were studied. Main Outcome Measures: Final outcome measures were four multi-item scales assessing life quality with regard to the marriage, own self-efficacy, own health and appearance, and life as a whole. Intervening outcome scales measured four marriage factors: marital conflict, sexual selfesteem, sexual dissatisfaction, and frequency of intercourse. Results and Conclusions: Higher levels of stress, regardless of whether that stress was from attempting to solve a fertility problem or another problem, were related to reduced marital functioning and decreased life quality. For husbands, the strengths of the linkages did not depend on the source of the stress. For wives, however, the causal model suggested that fertility-problem stress had stronger negative impacts on sense of sexual identity and self-efficacy than did stress from other problems (P < 0.05). Stress from any source had more impact on the lives of wives than of husbands, more impact on satisfaction with self and general well-being than on satisfaction with the marriage or health, and affected life quality mostly indirectly through its impacts on the marriage factors. Fertil SteriI1992;57:1247-53 Key Words: Stress, infertility, life quality, subjective well-being, marriage factors, psychosocial factors
Received August 5, 1991; revised and accepted February 18, 1991. * This report is a revised version of a presentation made to the 1991 Annual Meeting of the Society of Obstetricians and Gynecologists of Canada, Toronto, Ontario, Canada, June 12 to 15, 1991. t Supported by grant number HD21240 from the National Institute of Child Health and Human Development, Bethesda, Maryland. :\: Institute for Social Research and Department of Population Planning and International Health, University of Michigan. § Reprint requests: Frank M. Andrews, Ph.D., Institute for Social Research, University of Michigan, Ann Arbor, Michigan 48106-1248. II Department of Community Medicine, Wayne State University, and Institute for Social Research, University of Michigan.
The literature is full of anecdotal evidence about the linkage between infertility and stress and about the deleterious impacts that fertility-problem stress can have on the functioning of a marriage and the life quality ofthe spouses (1-5). Some ofthe themes of that literature stress that: (1) childbearing is extremely important to most couples; (2) the occurrence of a fertility problem, though common, is generally unanticipated by most couples; and (3) the solution of a fertility problem is often difficult, timeconsuming, and expensive. Although the solution usually involves some physiological intervention (6), the psychological stress that many couples experience and its effect on their life quality become im-
Vol. 57, No.6, June 1992
Andrews et al.
Is fertility-problem stress different?
1247
portant problems in their own right. Physicians who treat couples with fertility problems have found they need to be concerned not only with the medical aspects of the problem but also with its psychosocial ramifications. None of the results emerging from our own research lead us to doubt the basic correctness of the anecdotal reports. We find that fertility problems are often linked to substantial amounts of stress, a range of impairments to marital functioning, and reduced life quality (7, 8). By identifying the specific mechanisms by which this psychosocial system works, we hope to be able to suggest to couples and their physicians how they can maintain and enhance life quality while coping with the fertility problem. Although we do not doubt the importance of psychosocial factors to couples with a fertility problem, we also know that life is not always easy for fertile couples. An important question, then, is how different is the stress faced by fertility-problem couples from that of couples presumed to be fertile? In what respects is the stress associated with a fertility problem unique and in what respects does fertilityproblem stress operate much like the stress that arises in attempting to cope with other problems? There is a large psychological literature that deals with general stress-how it arises, what its effects are, and the effectiveness of various means of coping with it (9-12)-and to the extent that fertilityproblem stress functions like other kinds of stress, much of the existing knowledge about general stress becomes relevant. As part of a larger study of infertility, psychosocial factors, and life quality, this report describes an analysis focusing on how stress links to several marriage factors, a domain of obvious relevance for couples coping with an infertility problem. Also examined is how stress and the marriage factors link to several aspects of subjective well-being (i.e., life quality). An important purpose ofthis analysis is to compare the dynamics of stress in fertile and infertile couples. MATERIALS AND METHODS Study Participants
Separate in-person interviews were conducted with both wives and husbands in 275 couples (550 individuals). Couples with primary infertility (n = 185) were principally recruited from infertility specialists. All but one of the major infertility practices in southeastern Michigan agreed to collaborate 1248
Andrews et al.
Is fertility-problem stress different?
with this study. Couples presumed to be fertile (n = 90) mainly came from two sources-recent applicants for marriage licenses in the same area of southeastern Michigan and patients in the regular gynecological practices of the above specialists. Eligibility criteria for all couples were as follows: (1) married, (2) no previous children by either member of the couple but interested in having a child sometime soon; (3) white; and (4) middle class (defined as having at least a high school education and a 1987 household income in the range $20,000 to $100,000). Only childless couples were included because stresses may differ for couples with children. White, middle-class couples were used because these are the couples most likely to seek treatment for infertility (13), and having a relatively homogeneous group allows more sophisticated analyses with a smaller number of cases. Additional eligibility criteria for infertile couples were that they not have tried in vitro fertilization (IVF) or gamete intrafallopian transfer (GIFT). Patients who had tried IVF or GIFT were omitted because these interventions are generally prescribed for couples who have progressed further in their treatment; we were interested in studying couples who were relatively early in the treatment process. Additional eligibility criteria for couples presumed to be fertile were that the wife have regular menstrual cycles and neither member have a known fertility problem. Eighty-seven percent (n = 162) of the infertile couples included in this study met the standard medical definition of infertility-1 year of regular unprotected sexual intercourse without conceiving or carrying a child to term. The remaining 23 infertile couples had been trying to conceive a child for <1 year (7 months on average) but were being treated by an infertility specialist. Sometimes individuals who are older or who have known physical problems related to infertility will seek treatment before 1 year has elapsed. For the causal modeling analyses reported below, couples are restricted to those in which both the wife and husband had complete data for all measures being analyzed. This is 157 who had fertility problems and 82 who were presumed fertile. Procedures
Couples who fit this study's eligibility criteria were invited to participate and sent a brochure that described the study as a university study of marriage, family, and childbearing issues. Couples were then contacted by a professional interviewer from the Fertility and Sterility
Survey Research Center at the University of Michigan. Separate I-hour, in-person interviews were conducted with each member of the couple. Husbands and wives were usually interviewed on the same day, and neither was able to hear the other's responses. The ethical guidelines of the American Psychological Association were followed throughout the study. Although the words infertile and infertility are used in this report, these words were not used with the couples who participated in the research. Pilot testing indicated that these words connoted a sense of finality that the study participants found unsettling. Instead, the term fertility problem was used. Characteristics of Study Participants
The fertility-problem couples participating in this research were not intended to be a representative sample from any defined population. However, when compared with a representative survey, they proved to be fairly typical of American white middle-class couples who seek to resolve a fertility problem involving a first child. The typical fertility-problem couple in our study had been married for approximately 6 years, had approximately $45,000 in family income, and consisted of a 32-year-old wife (range: 22 to 42 years) and a 34-year-old husband (range: 23 to 44 years). Men averaged 3 years of college education, and women averaged 2.5 years. Ninety-nine percent of the men and 92% of the women were active participants in the labor force. The mean length of time these couples had been trying to have a child was 2.5 years, and the mean number of physicians seen was 2.4 (range: 1 to 6). The couples presumed to be fertile were sociodemographically similar but were approximately 3 years younger and had been married for approximately 3 years less. The fertility-problem couples and couples presumed to be fertile also turned out to be very similar with regard to all of the variables used in the causal modeling analyses described below. Measures
The analyses reported here focus on nine measures, each assessed separately for wives and husbands. These are: (1) stress because of the fertility problem (for infertile couples) or because of a selfidentified biggest problem (for couples presumed to be fertile); (2) four marriage factors-marital conflict, sexual self-esteem, sexual dissatisfaction, and frequency of intercourse; and (3) four aspects of life Vol. 57, No.6, June 1992
quality-assessments of the marriage, own health and appearance, own self-efficacy, and life as a whole. Eight of these nine measures are multi-item scales. Frequency of sexual intercourse is the only measure that is not a multi-item scale. (A document is available from the authors that presents the exact wording of all of the component items, the response categories for each item, and the reliability of each scale.) The scales of stress, marital conflict, sexual selfesteem, and sexual dissatisfaction were constructed by the research team using standard procedures for scale construction (14) by combining questionnaire items that reflected concerns expressed in the literature and pilot interviews and that showed substantial covariation. The four life quality scales are adapted from national surveys conducted by Andrews and Withey (15). The stress scale reflects self-reported disruption, change, and stress that an important problem has caused in people's lives. For members of infertile couples, the problem was explicitly the fertility problem. Members of fertile couples were asked to identify the "biggest problem in your life now" and were then asked about its impacts. There was no necessity that wives and husbands in fertile couples identify the same problem, but in approximately half the couples they did. Causal Modeling Procedures
The main analysis methodology employs the relatively new and powerful form of causal modeling called latent variable multigroup covariance structure modeling (16). Examples of latent variables used in other contexts include the factors of factor analysis and the functions of discriminant function analysis. The general strategy is to specify hypothesized causal linkages among a set of relevant concepts and then to see if the resulting system can come sufficiently close to predicting what is actually observed in the real world to lead one to think the hypothesized system is plausible. Using this new technology, it is possible to test whether each linkage in the system operates the same way for different subgroups-in the present case, for the fertilityproblem couples and the couples presumed to be fertile. The causal model is shown in Figure 1, which also includes the parameter estimates obtained when satisfaction with marriage was used as the subjective well-being variable. Information from wives appears Andrews et al.
Is fertility-problem stress different?
1249
@ Sl' .27
.10' .10·
.1S' .14*
.09 .07
Figure 1 Causal model and estimated parameters (standardized) for wives and husbands in fertility-problem couples and couples presumed fertile when final outcome variable is feelings about own marriage. (First figure is for fertility-problem couples; second for couples presumed fertile). Pairs of parameters within ovals are significantly different between the two groups; all others are equivalent (except for standardization). NOTES: W = wife; H = husband. This exhibit follows conventions for showing causal models: one headed arrows indicate direct effects, two headed arrows indicate covariation for which causal effects are not specified; arrows without a labeled start indicate effects of all other (i.e., residual) causes on the target variable. All variables shown are latent; each one's link to its corresponding observed measure was equal to the square root of its reliability. Covariances among the residual inputs to the eight marriage factors were not restricted to zero and were allowed to differ between the groups. Standardized parameters are shown for ease of interpretability; all calculations were performed using unstandardized variables.
in the upper part of the figure and information from husbands in the lower part. In accord with prevailing theory, the model in Figure 1 says that stress may have direct effects on the four marriage factors (marital conflict, sexual self-esteem, sexual dissatisfaction, and frequency of intercourse) and also a direct effect on subjective well-being, e.g., on evaluations of one's own marriage. In addition, the four marriage variables themselves may have direct effects on well-being. Of course, this means that stress can also have indirect effects (through the marriage factors) on subjective well-being. This set of causal dynamics is hypothesized to apply to both wives and husbands. 1250
Andrews et al.
Is fertility-problem stress different?
The model allows for some linkages between the members of a couple. How much stress one member experiences is linked (by an amount that is to be determined by the empirical analyses) to how much stress the other member experiences. Also, the model explicitly allows for each member's feelings about life quality to have a direct effect on the other member's feelings. (These reciprocal effects were constrained to be equal.) In addition, but not drawn in the figure, are linkages among all possible pairs of residuals for the marriage factors; thus factors other than stress that influence a marriage factor for one spouse may be related to factors (other than stress) that influence the marriage factors for the other spouse. This model was estimated separately but simultaneously for the two types of couples: (1) fertilityproblem couples in whom the stress being assessed was specifically related to the fertility problem and (2) couples presumed to be fertile, for whom a parallel set of questions inquired about stress associated with a self-identified biggest problem. The models were estimated using Bentler's EQS program (17) . This same model was estimated for each of four different aspects of life quality: (1) self evaluations of own marriage; (2) own health; (3) own self-efficacy; and (4) own life as a whole. The models proved to fit well: the Comparative Fit Indexes were always >0.97.
RESULTS The statistical results when stress and the marriage factors are linked to the marital aspect of life quality appear in Figure 1. Comparable results for the other three aspects of life quality (health/appearance, self-efficacy, and life as a whole) are available from the authors. Although the data are from a one-time survey and cannot be used to prove causality, because this analysis uses a causal model that assumes a causal structure, causal terminology is used when describing results from the model. Of key interest for this report are the comparisons between the strength of the linkages estimated for the fertility-problem couples and the estimates for the couples presumed to be fertile. In Figure 1, all direct effects are statistically equivalent for the two groups except for the pairs of parameters within enclosures. Direct Effects for Wives and Husbands
For wives it makes a big difference whether the problem is a fertility problem or something else. Although all stresses tend to increase wives' reports Fertility and Sterility
of marital conflict, stress from fertility problems has especially strong negative effects on wives' own sexual identity (sexual self-esteem and sexual satisfaction). These differences, each of which is statistically significant at the P < 0.05 level, are shown by the pairs of enclosed parameters in Figure 1. Furthermore (as shown in Table 1), the direct effect of stress on wives' sense of self-efficacy is much stronger for wives with fertility problems (-0.49) than for wives with other problems (-0.04). Thus it appears that fertility problems are interpreted as carrying a special negative message for wives' sense of sexual- and self-efficacy. At the most general level, the pattern of effects for husbands is broadly similar to that for wives. A striking and perhaps surprising difference, however, is that for husbands the causal dynamics of stress on the marriage factors and on life quality do not differ for men coping with fertility problems and for men with other problems. It is not the case that husbands' lives are unaffected by fertility problems and the stresses that may accompany them. Rather, for husbands the impact of a fertility problem is not fundamentally different from the impact of other problems. This represents an important difference from wives, for whom, as described above, fertility-problem stress carries special threats to sense of sexual identity and self-efficacy.
Total and Indirect Effects of Stress on Well-Being
In addition to the direct effects of stress, the model allows computation of indirect and total effects of stress on life quality. The indirect effects come about through stress's effects on the marriage factors and in turn their effects on life quality. Total effects are the sum of direct plus indirect effects. Summaries of the direct, indirect, and total effects of stress are in Table 1 for wives and husbands of fertility-problem couples and couples presumed to be fertile and for all four aspects of life quality. If we ask whether the total effects of stress on people's life quality differs between those in fertilityproblem couples and those in couples presumed to be fertile, the general finding shown in Table 1 is that fertility-problem stress tends to have at least as much, and sometimes more, total impact on wellbeing as does the stress experienced by other couples who have a different "biggest problem." (The only exception in Table 1, which we regard as trivial, occurs for wives' evaluations of life as a whole.) As analyzed here, total effects offertility-problem stress on well-being usually exceed the effects of stress experienced by wives and husbands in other couples who face other significant problems. Table 1 is also interesting for the relative magnitudes of several other comparisons. Table 1 also shows that: (1) wives tend to be more affected by stress than husbands, whatever the source of the
Table 1 Direct, Indirect, and Total Effects of Stress From Fertility Problem or Biggest Problem for Wives and Husbands in Fertility-Problem Couples or Couples Presumed Fertile* Aspect of life quality evaluated
Couples with a fertility problem Wives Direct Indirect Total Husbands Direct Indirect Total Couples presumed fertile Wives Direct Indirect Total Husbands Direct Indirect Total
Marriage
Health
Self
Life as a whole
-0.11 -0.20 -0.31
-0.20 -0.32 -0.52
-0.49 -0.12 -0.61
-0.32 -0.33 -0.65
0.09 -0.26 -0.17
-0.18 -0.14 -0.32
-0.11 -0.32 -0.43
0.00 -0.37 -0.37
-0.10 -0.19 -0.29
-0.05 -0.33 -0.38
-0.04 -0.50 -0.54
-0.37 -0.33 -0.70
0.07 -0.20 -0.13
-0.03 -0.24 -0.27
-0.09 -0.25 -0.34
0.00 -0.21 -0.21
* Figures show magnitude of effects from standardized solution. Indirect effects go via marital conflict, sexual self-esteem, sexual dissatisfaction, and/or frequency of intercourse (see Fig. 1). Vol. 57, No.6, June 1992
Andrews et at.
Is fertility-problem stress different?
1251
stress; (2) stress, whatever its source, tends to have bigger impacts on people's sense of self-efficacy and general well-being than it does on their satisfaction with their marriages or health; and (3) much of the impact of stress is indirect because stress affects the marriage factors that then affect life quality. DISCUSSION
At the most general level, this model of the causal dynamics of stress suggests that stress (from any source) has important effects for both wives and husbands on all four of the marriage factors. It increases marital conflict and sexual dissatisfaction and decreases sexual self-esteem and frequency of intercourse. In addition, stress has negative direct effects on life quality (with some minor exceptions for husbands). The marriage factors in turn have their own direct effects on life quality, and most of these are in expected directions. Marital conflict and sexual dissatisfaction tend to decrease most aspects of life quality, and sexual self-esteem and the frequency of sexual intercourse (as reported by men but not women) tend to increase life quality. The total effects of stress on life quality tend to be substantially stronger for wives than for husbands, as is discussed below. What is the answer to the question posed in the title? Is fertility-problem stress different? Does the stress experienced by some infertile couples as they cope with their fertility problem work in the same way as the stresses that may be experienced by fertile couples as they cope with other kinds of problems? The answer has considerable interest from both theoretical and practical perspectives. If all stresses have similar dynamics, a single psychology of stress will be sufficient to understand the phenomenon and can be applied in attempts to help people deal with stress. Alternatively, if fertility-problem stress has special aspects, those who experience it and professionals who try to help those who experience it can benefit by recognizing how it is distinctive. The results presented here suggest that the answer to the question Is fertility-problem stress different? differs for men and women. The answer for men is easy to describe: No. For men, the dynamics of fertility-problem stress are very similar to the dynamics of stress from other problems. For women, however, fertility-problem stress has some special features. It shows significantly stronger and more deleterious effects on wives' sexual self-esteem, sexual dissatisfaction, and sense of self-efficacy than does stress from other problems. However, fertility-problem 1252
Andrews et al.
Is fertility-problem stress different?
stress has about the same impact as other stresses in increasing wives' reports of marital conflict and decreasing their evaluations of the marriage, their own health, and their life as a whole. Is there a surprise in this pattern of results? We are aware of no previous studies that have examined this matter in a reasonably broad empirical way, so these results neither contradict nor confirm earlier findings. That stress from a fertility problem is specially virulent for a woman's sexual identity and sense of self-efficacy does not seem unreasonable and concurs with some prior indications in the literature (18). Perhaps the surprise is the finding for men that the dynamics of fertility-problem stress are not distinctive. In addition to comparing the effects of stress from a fertility problem with stress from other problems, this analysis allows us to compare the impacts of stress for wives and husbands. As noted, the total effects of stress, regardless of its source, on life quality were stronger for wives. In addition, the direct effects of stress on three of the marriage factors were also stronger for wives than husbands. This applied to marital conflict, sexual self-esteem, and sexual dissatisfaction. However, stress affected the fourth marriage factor, reported frequency of intercourse, more for husbands than it did for wives. Although stress had stronger and more wideranging impacts on wives than husbands, it does not follow that husbands were unaffected. Any attempt to improve life quality by helping couples cope with stress can usefully focus on both members of a couple. Our model suggests that stress, including stress from fertility problems, affects life quality not only directly but also substantially through its impacts on marriage factors. This has important implications for professionals who attempt to help infertile couples cope with stress. It suggests that there are two different types of interventions available to health professionals and to infertile couples themselves in their attempts to improve life quality. One approach is the traditional one of trying to reduce the stress being experienced directly. In terms of our model, this amounts to focusing on the original source ofthe problem. Some ofthe interventions likely to be effective here include the following: (1) providing valid information about treatments and their likely outcomes; (2) attempting to minimize pain, inconvenience, and other hassles associated with treatments; (3) ensuring the availability of appropriate psychological support; and (4) letting paFertility and Sterility
tients know they have substantial control over decisions about their treatment (19). In addition, however, these results suggest there are potentially effective indirect approaches that aim to assist in maint~ining the marriage factors at satisfactory levels. Even if stress cannot be eliminated, its indirect effects may be greatly reduced. In terms of our model, this amounts to cutting some of the paths by which stress impacts life quality. For example, minimizing marital conflicts and maintaining wives' and husbands' own sexual self-esteem and satisfaction with their sexual lives should be effective in ameliorating the typical indirect effects of stress on life quality. These are endeavors that couples can be encouraged to address themselves and for which health professionals can attempt to provide counseling, support, and helpful information. Of course, the conclusions presented in this report are a function of the particular couples examined, the set of concepts that have been considered, and the appropriateness of our causal model. Within the constraints of the present analysis, the internal validity of the conclusions seems high. These fertilityproblem couples are typical of many of those who seek help for infertility; the marriage and life-quality factors that have been considered are certainly relevant to the issue; and the causal model is both theoretically plausible and empirically effective in fitting the data. However, this research should be recognized as being an early exploration in this area. Confidence in our conclusions will be enhanced if in subsequent research they are confirmed using broader samples of couples, wider ranges of concepts, and longitudinal analysis. REFERENCES 1. Dunkel-Schetter C, Lobel M. Psychological reactions to infertility. In: Stanton A, Dunkel-Shetter C, editors. Infertility. New York: Plenum, 1991:29-57.
Vol. 57, No.6, June 1992
2. Lasker IN, Borg S. In search of parenthood. Boston: Beacon, 1987. 3. Menning BE. Infertility: a guide for childless couples. Englewood Cliffs (NJ): Prentice-Hall, 1977. 4. Menning BE. The emotional needs of infertile couples. Fertil Steril 1980;34:313-9. 5. Seibel MM, Taymor ML. Emotional aspects of infertility. Fertil Steril 1982;37:137-45. 6. Moghissi KS, Wallach EE. Unexplained infertility. Fertil Steril 1983;39:5-21. 7. Abbey A, Andrews FM, Halman LJ. The importance of social relationships for infertile couples' well-being. In: Stanton A, Dunkel-Shetter C, editors. Infertility. New York: Plenum, 1991:61-86. 8. Abbey A, Andrews FM, Halman, LJ. Gender's role in responses to infertility. Psychol Women Quart 1991;15: 295-316. 9. Fisher S, Reason J., editors. Handbook of life stress cognition and health. New York: Wiley, 1988. 10. Claus KE, Bailey JT, editors. Living with stress and promoting well-being: a handbook for nurses. St. Louis (MO): Mosby, 1980. 11. Sethi AS, Schuler RS. Handbook of organizational stress coping strategies. Cambridge (MA): Ballinger, 1984. 12. Kutash IL, Schlesinger LB. Handbook on stress and anxiety. San Francisco: Jossey-Bass, 1980. 13. Henshaw SK, Orr MT. The need and unmet need for infertility services in the United States. Fam Plann Perspect 1987;19:180-6. 14. McKennell AC. Attitude scale construction. In: O'Muircheartaigh CA, Payne C, editors. Analysis of survey data. New York: Wiley, 1977;1:183-220. 15. Andrews FM, Withey SB. Social indicators of well-being: Americans' perceptions of life quality. New York: Plenum, 1976. 16. Bollen KA. Structural equations with latent variables. New York: Wiley, 1989. 17. Bentler PM. EQS: structural equations program manual. Los Angeles: BMDP Statistical Software, 1989. 18. Freeman EW, Boxer AS, Rickels K, Tureck R, Mastroianni L Jr. Psychological evaluation and support in a program of in vitro fertilization and embryo transfer. Fertil SterilI985;43: 48-53. 19. Abbey A, Halman LJ, Andrews FM. Psychosocial, treatment, and demographic predictors ofthe stress associated with infertility. Fertil SterilI992;57:122-8.
Andrews et a1.
Is fertility-problem stress different?
1253
Vol. 57, No.6, June 1992
Copyright" 1992 The American Fertility Society
Printed on O£id·{ree paper in U.S.A.
Is fertility-problem stress different? The dynamics of stress in fertile and infertile couples*t
Frank M. Andrews, Ph.D.:!:§ Antonia Abbey, Ph.D. II L. Jill Halman, Ph.D.:!: University of Michigan, Ann Arbor, and Wayne State University, Detroit, Michigan
Objective: To compare the dynamics of fertility-problem stress experienced by wives and husbands in infertile couples with the dynamics of stress from other sources experienced by members of couples presumed to be fertile. Design: Relationships of stress to four marriage factors and four aspects of life quality (subjective well-being) are examined within a causal modeling framework using data from structured interviews. Setting: Face-to-face interviews were conducted in study participants' homes. Participants: Wives and husbands from 157 couples with primary infertility and from 82 presumed-fertile couples were studied. Main Outcome Measures: Final outcome measures were four multi-item scales assessing life quality with regard to the marriage, own self-efficacy, own health and appearance, and life as a whole. Intervening outcome scales measured four marriage factors: marital conflict, sexual selfesteem, sexual dissatisfaction, and frequency of intercourse. Results and Conclusions: Higher levels of stress, regardless of whether that stress was from attempting to solve a fertility problem or another problem, were related to reduced marital functioning and decreased life quality. For husbands, the strengths of the linkages did not depend on the source of the stress. For wives, however, the causal model suggested that fertility-problem stress had stronger negative impacts on sense of sexual identity and self-efficacy than did stress from other problems (P < 0.05). Stress from any source had more impact on the lives of wives than of husbands, more impact on satisfaction with self and general well-being than on satisfaction with the marriage or health, and affected life quality mostly indirectly through its impacts on the marriage factors. Fertil SteriI1992;57:1247-53 Key Words: Stress, infertility, life quality, subjective well-being, marriage factors, psychosocial factors
Received August 5, 1991; revised and accepted February 18, 1991. * This report is a revised version of a presentation made to the 1991 Annual Meeting of the Society of Obstetricians and Gynecologists of Canada, Toronto, Ontario, Canada, June 12 to 15, 1991. t Supported by grant number HD21240 from the National Institute of Child Health and Human Development, Bethesda, Maryland. :\: Institute for Social Research and Department of Population Planning and International Health, University of Michigan. § Reprint requests: Frank M. Andrews, Ph.D., Institute for Social Research, University of Michigan, Ann Arbor, Michigan 48106-1248. II Department of Community Medicine, Wayne State University, and Institute for Social Research, University of Michigan.
The literature is full of anecdotal evidence about the linkage between infertility and stress and about the deleterious impacts that fertility-problem stress can have on the functioning of a marriage and the life quality ofthe spouses (1-5). Some ofthe themes of that literature stress that: (1) childbearing is extremely important to most couples; (2) the occurrence of a fertility problem, though common, is generally unanticipated by most couples; and (3) the solution of a fertility problem is often difficult, timeconsuming, and expensive. Although the solution usually involves some physiological intervention (6), the psychological stress that many couples experience and its effect on their life quality become im-
Vol. 57, No.6, June 1992
Andrews et al.
Is fertility-problem stress different?
1247
portant problems in their own right. Physicians who treat couples with fertility problems have found they need to be concerned not only with the medical aspects of the problem but also with its psychosocial ramifications. None of the results emerging from our own research lead us to doubt the basic correctness of the anecdotal reports. We find that fertility problems are often linked to substantial amounts of stress, a range of impairments to marital functioning, and reduced life quality (7, 8). By identifying the specific mechanisms by which this psychosocial system works, we hope to be able to suggest to couples and their physicians how they can maintain and enhance life quality while coping with the fertility problem. Although we do not doubt the importance of psychosocial factors to couples with a fertility problem, we also know that life is not always easy for fertile couples. An important question, then, is how different is the stress faced by fertility-problem couples from that of couples presumed to be fertile? In what respects is the stress associated with a fertility problem unique and in what respects does fertilityproblem stress operate much like the stress that arises in attempting to cope with other problems? There is a large psychological literature that deals with general stress-how it arises, what its effects are, and the effectiveness of various means of coping with it (9-12)-and to the extent that fertilityproblem stress functions like other kinds of stress, much of the existing knowledge about general stress becomes relevant. As part of a larger study of infertility, psychosocial factors, and life quality, this report describes an analysis focusing on how stress links to several marriage factors, a domain of obvious relevance for couples coping with an infertility problem. Also examined is how stress and the marriage factors link to several aspects of subjective well-being (i.e., life quality). An important purpose ofthis analysis is to compare the dynamics of stress in fertile and infertile couples. MATERIALS AND METHODS Study Participants
Separate in-person interviews were conducted with both wives and husbands in 275 couples (550 individuals). Couples with primary infertility (n = 185) were principally recruited from infertility specialists. All but one of the major infertility practices in southeastern Michigan agreed to collaborate 1248
Andrews et al.
Is fertility-problem stress different?
with this study. Couples presumed to be fertile (n = 90) mainly came from two sources-recent applicants for marriage licenses in the same area of southeastern Michigan and patients in the regular gynecological practices of the above specialists. Eligibility criteria for all couples were as follows: (1) married, (2) no previous children by either member of the couple but interested in having a child sometime soon; (3) white; and (4) middle class (defined as having at least a high school education and a 1987 household income in the range $20,000 to $100,000). Only childless couples were included because stresses may differ for couples with children. White, middle-class couples were used because these are the couples most likely to seek treatment for infertility (13), and having a relatively homogeneous group allows more sophisticated analyses with a smaller number of cases. Additional eligibility criteria for infertile couples were that they not have tried in vitro fertilization (IVF) or gamete intrafallopian transfer (GIFT). Patients who had tried IVF or GIFT were omitted because these interventions are generally prescribed for couples who have progressed further in their treatment; we were interested in studying couples who were relatively early in the treatment process. Additional eligibility criteria for couples presumed to be fertile were that the wife have regular menstrual cycles and neither member have a known fertility problem. Eighty-seven percent (n = 162) of the infertile couples included in this study met the standard medical definition of infertility-1 year of regular unprotected sexual intercourse without conceiving or carrying a child to term. The remaining 23 infertile couples had been trying to conceive a child for <1 year (7 months on average) but were being treated by an infertility specialist. Sometimes individuals who are older or who have known physical problems related to infertility will seek treatment before 1 year has elapsed. For the causal modeling analyses reported below, couples are restricted to those in which both the wife and husband had complete data for all measures being analyzed. This is 157 who had fertility problems and 82 who were presumed fertile. Procedures
Couples who fit this study's eligibility criteria were invited to participate and sent a brochure that described the study as a university study of marriage, family, and childbearing issues. Couples were then contacted by a professional interviewer from the Fertility and Sterility
Survey Research Center at the University of Michigan. Separate I-hour, in-person interviews were conducted with each member of the couple. Husbands and wives were usually interviewed on the same day, and neither was able to hear the other's responses. The ethical guidelines of the American Psychological Association were followed throughout the study. Although the words infertile and infertility are used in this report, these words were not used with the couples who participated in the research. Pilot testing indicated that these words connoted a sense of finality that the study participants found unsettling. Instead, the term fertility problem was used. Characteristics of Study Participants
The fertility-problem couples participating in this research were not intended to be a representative sample from any defined population. However, when compared with a representative survey, they proved to be fairly typical of American white middle-class couples who seek to resolve a fertility problem involving a first child. The typical fertility-problem couple in our study had been married for approximately 6 years, had approximately $45,000 in family income, and consisted of a 32-year-old wife (range: 22 to 42 years) and a 34-year-old husband (range: 23 to 44 years). Men averaged 3 years of college education, and women averaged 2.5 years. Ninety-nine percent of the men and 92% of the women were active participants in the labor force. The mean length of time these couples had been trying to have a child was 2.5 years, and the mean number of physicians seen was 2.4 (range: 1 to 6). The couples presumed to be fertile were sociodemographically similar but were approximately 3 years younger and had been married for approximately 3 years less. The fertility-problem couples and couples presumed to be fertile also turned out to be very similar with regard to all of the variables used in the causal modeling analyses described below. Measures
The analyses reported here focus on nine measures, each assessed separately for wives and husbands. These are: (1) stress because of the fertility problem (for infertile couples) or because of a selfidentified biggest problem (for couples presumed to be fertile); (2) four marriage factors-marital conflict, sexual self-esteem, sexual dissatisfaction, and frequency of intercourse; and (3) four aspects of life Vol. 57, No.6, June 1992
quality-assessments of the marriage, own health and appearance, own self-efficacy, and life as a whole. Eight of these nine measures are multi-item scales. Frequency of sexual intercourse is the only measure that is not a multi-item scale. (A document is available from the authors that presents the exact wording of all of the component items, the response categories for each item, and the reliability of each scale.) The scales of stress, marital conflict, sexual selfesteem, and sexual dissatisfaction were constructed by the research team using standard procedures for scale construction (14) by combining questionnaire items that reflected concerns expressed in the literature and pilot interviews and that showed substantial covariation. The four life quality scales are adapted from national surveys conducted by Andrews and Withey (15). The stress scale reflects self-reported disruption, change, and stress that an important problem has caused in people's lives. For members of infertile couples, the problem was explicitly the fertility problem. Members of fertile couples were asked to identify the "biggest problem in your life now" and were then asked about its impacts. There was no necessity that wives and husbands in fertile couples identify the same problem, but in approximately half the couples they did. Causal Modeling Procedures
The main analysis methodology employs the relatively new and powerful form of causal modeling called latent variable multigroup covariance structure modeling (16). Examples of latent variables used in other contexts include the factors of factor analysis and the functions of discriminant function analysis. The general strategy is to specify hypothesized causal linkages among a set of relevant concepts and then to see if the resulting system can come sufficiently close to predicting what is actually observed in the real world to lead one to think the hypothesized system is plausible. Using this new technology, it is possible to test whether each linkage in the system operates the same way for different subgroups-in the present case, for the fertilityproblem couples and the couples presumed to be fertile. The causal model is shown in Figure 1, which also includes the parameter estimates obtained when satisfaction with marriage was used as the subjective well-being variable. Information from wives appears Andrews et al.
Is fertility-problem stress different?
1249
@ Sl' .27
.10' .10·
.1S' .14*
.09 .07
Figure 1 Causal model and estimated parameters (standardized) for wives and husbands in fertility-problem couples and couples presumed fertile when final outcome variable is feelings about own marriage. (First figure is for fertility-problem couples; second for couples presumed fertile). Pairs of parameters within ovals are significantly different between the two groups; all others are equivalent (except for standardization). NOTES: W = wife; H = husband. This exhibit follows conventions for showing causal models: one headed arrows indicate direct effects, two headed arrows indicate covariation for which causal effects are not specified; arrows without a labeled start indicate effects of all other (i.e., residual) causes on the target variable. All variables shown are latent; each one's link to its corresponding observed measure was equal to the square root of its reliability. Covariances among the residual inputs to the eight marriage factors were not restricted to zero and were allowed to differ between the groups. Standardized parameters are shown for ease of interpretability; all calculations were performed using unstandardized variables.
in the upper part of the figure and information from husbands in the lower part. In accord with prevailing theory, the model in Figure 1 says that stress may have direct effects on the four marriage factors (marital conflict, sexual self-esteem, sexual dissatisfaction, and frequency of intercourse) and also a direct effect on subjective well-being, e.g., on evaluations of one's own marriage. In addition, the four marriage variables themselves may have direct effects on well-being. Of course, this means that stress can also have indirect effects (through the marriage factors) on subjective well-being. This set of causal dynamics is hypothesized to apply to both wives and husbands. 1250
Andrews et al.
Is fertility-problem stress different?
The model allows for some linkages between the members of a couple. How much stress one member experiences is linked (by an amount that is to be determined by the empirical analyses) to how much stress the other member experiences. Also, the model explicitly allows for each member's feelings about life quality to have a direct effect on the other member's feelings. (These reciprocal effects were constrained to be equal.) In addition, but not drawn in the figure, are linkages among all possible pairs of residuals for the marriage factors; thus factors other than stress that influence a marriage factor for one spouse may be related to factors (other than stress) that influence the marriage factors for the other spouse. This model was estimated separately but simultaneously for the two types of couples: (1) fertilityproblem couples in whom the stress being assessed was specifically related to the fertility problem and (2) couples presumed to be fertile, for whom a parallel set of questions inquired about stress associated with a self-identified biggest problem. The models were estimated using Bentler's EQS program (17) . This same model was estimated for each of four different aspects of life quality: (1) self evaluations of own marriage; (2) own health; (3) own self-efficacy; and (4) own life as a whole. The models proved to fit well: the Comparative Fit Indexes were always >0.97.
RESULTS The statistical results when stress and the marriage factors are linked to the marital aspect of life quality appear in Figure 1. Comparable results for the other three aspects of life quality (health/appearance, self-efficacy, and life as a whole) are available from the authors. Although the data are from a one-time survey and cannot be used to prove causality, because this analysis uses a causal model that assumes a causal structure, causal terminology is used when describing results from the model. Of key interest for this report are the comparisons between the strength of the linkages estimated for the fertility-problem couples and the estimates for the couples presumed to be fertile. In Figure 1, all direct effects are statistically equivalent for the two groups except for the pairs of parameters within enclosures. Direct Effects for Wives and Husbands
For wives it makes a big difference whether the problem is a fertility problem or something else. Although all stresses tend to increase wives' reports Fertility and Sterility
of marital conflict, stress from fertility problems has especially strong negative effects on wives' own sexual identity (sexual self-esteem and sexual satisfaction). These differences, each of which is statistically significant at the P < 0.05 level, are shown by the pairs of enclosed parameters in Figure 1. Furthermore (as shown in Table 1), the direct effect of stress on wives' sense of self-efficacy is much stronger for wives with fertility problems (-0.49) than for wives with other problems (-0.04). Thus it appears that fertility problems are interpreted as carrying a special negative message for wives' sense of sexual- and self-efficacy. At the most general level, the pattern of effects for husbands is broadly similar to that for wives. A striking and perhaps surprising difference, however, is that for husbands the causal dynamics of stress on the marriage factors and on life quality do not differ for men coping with fertility problems and for men with other problems. It is not the case that husbands' lives are unaffected by fertility problems and the stresses that may accompany them. Rather, for husbands the impact of a fertility problem is not fundamentally different from the impact of other problems. This represents an important difference from wives, for whom, as described above, fertility-problem stress carries special threats to sense of sexual identity and self-efficacy.
Total and Indirect Effects of Stress on Well-Being
In addition to the direct effects of stress, the model allows computation of indirect and total effects of stress on life quality. The indirect effects come about through stress's effects on the marriage factors and in turn their effects on life quality. Total effects are the sum of direct plus indirect effects. Summaries of the direct, indirect, and total effects of stress are in Table 1 for wives and husbands of fertility-problem couples and couples presumed to be fertile and for all four aspects of life quality. If we ask whether the total effects of stress on people's life quality differs between those in fertilityproblem couples and those in couples presumed to be fertile, the general finding shown in Table 1 is that fertility-problem stress tends to have at least as much, and sometimes more, total impact on wellbeing as does the stress experienced by other couples who have a different "biggest problem." (The only exception in Table 1, which we regard as trivial, occurs for wives' evaluations of life as a whole.) As analyzed here, total effects offertility-problem stress on well-being usually exceed the effects of stress experienced by wives and husbands in other couples who face other significant problems. Table 1 is also interesting for the relative magnitudes of several other comparisons. Table 1 also shows that: (1) wives tend to be more affected by stress than husbands, whatever the source of the
Table 1 Direct, Indirect, and Total Effects of Stress From Fertility Problem or Biggest Problem for Wives and Husbands in Fertility-Problem Couples or Couples Presumed Fertile* Aspect of life quality evaluated
Couples with a fertility problem Wives Direct Indirect Total Husbands Direct Indirect Total Couples presumed fertile Wives Direct Indirect Total Husbands Direct Indirect Total
Marriage
Health
Self
Life as a whole
-0.11 -0.20 -0.31
-0.20 -0.32 -0.52
-0.49 -0.12 -0.61
-0.32 -0.33 -0.65
0.09 -0.26 -0.17
-0.18 -0.14 -0.32
-0.11 -0.32 -0.43
0.00 -0.37 -0.37
-0.10 -0.19 -0.29
-0.05 -0.33 -0.38
-0.04 -0.50 -0.54
-0.37 -0.33 -0.70
0.07 -0.20 -0.13
-0.03 -0.24 -0.27
-0.09 -0.25 -0.34
0.00 -0.21 -0.21
* Figures show magnitude of effects from standardized solution. Indirect effects go via marital conflict, sexual self-esteem, sexual dissatisfaction, and/or frequency of intercourse (see Fig. 1). Vol. 57, No.6, June 1992
Andrews et at.
Is fertility-problem stress different?
1251
stress; (2) stress, whatever its source, tends to have bigger impacts on people's sense of self-efficacy and general well-being than it does on their satisfaction with their marriages or health; and (3) much of the impact of stress is indirect because stress affects the marriage factors that then affect life quality. DISCUSSION
At the most general level, this model of the causal dynamics of stress suggests that stress (from any source) has important effects for both wives and husbands on all four of the marriage factors. It increases marital conflict and sexual dissatisfaction and decreases sexual self-esteem and frequency of intercourse. In addition, stress has negative direct effects on life quality (with some minor exceptions for husbands). The marriage factors in turn have their own direct effects on life quality, and most of these are in expected directions. Marital conflict and sexual dissatisfaction tend to decrease most aspects of life quality, and sexual self-esteem and the frequency of sexual intercourse (as reported by men but not women) tend to increase life quality. The total effects of stress on life quality tend to be substantially stronger for wives than for husbands, as is discussed below. What is the answer to the question posed in the title? Is fertility-problem stress different? Does the stress experienced by some infertile couples as they cope with their fertility problem work in the same way as the stresses that may be experienced by fertile couples as they cope with other kinds of problems? The answer has considerable interest from both theoretical and practical perspectives. If all stresses have similar dynamics, a single psychology of stress will be sufficient to understand the phenomenon and can be applied in attempts to help people deal with stress. Alternatively, if fertility-problem stress has special aspects, those who experience it and professionals who try to help those who experience it can benefit by recognizing how it is distinctive. The results presented here suggest that the answer to the question Is fertility-problem stress different? differs for men and women. The answer for men is easy to describe: No. For men, the dynamics of fertility-problem stress are very similar to the dynamics of stress from other problems. For women, however, fertility-problem stress has some special features. It shows significantly stronger and more deleterious effects on wives' sexual self-esteem, sexual dissatisfaction, and sense of self-efficacy than does stress from other problems. However, fertility-problem 1252
Andrews et al.
Is fertility-problem stress different?
stress has about the same impact as other stresses in increasing wives' reports of marital conflict and decreasing their evaluations of the marriage, their own health, and their life as a whole. Is there a surprise in this pattern of results? We are aware of no previous studies that have examined this matter in a reasonably broad empirical way, so these results neither contradict nor confirm earlier findings. That stress from a fertility problem is specially virulent for a woman's sexual identity and sense of self-efficacy does not seem unreasonable and concurs with some prior indications in the literature (18). Perhaps the surprise is the finding for men that the dynamics of fertility-problem stress are not distinctive. In addition to comparing the effects of stress from a fertility problem with stress from other problems, this analysis allows us to compare the impacts of stress for wives and husbands. As noted, the total effects of stress, regardless of its source, on life quality were stronger for wives. In addition, the direct effects of stress on three of the marriage factors were also stronger for wives than husbands. This applied to marital conflict, sexual self-esteem, and sexual dissatisfaction. However, stress affected the fourth marriage factor, reported frequency of intercourse, more for husbands than it did for wives. Although stress had stronger and more wideranging impacts on wives than husbands, it does not follow that husbands were unaffected. Any attempt to improve life quality by helping couples cope with stress can usefully focus on both members of a couple. Our model suggests that stress, including stress from fertility problems, affects life quality not only directly but also substantially through its impacts on marriage factors. This has important implications for professionals who attempt to help infertile couples cope with stress. It suggests that there are two different types of interventions available to health professionals and to infertile couples themselves in their attempts to improve life quality. One approach is the traditional one of trying to reduce the stress being experienced directly. In terms of our model, this amounts to focusing on the original source ofthe problem. Some ofthe interventions likely to be effective here include the following: (1) providing valid information about treatments and their likely outcomes; (2) attempting to minimize pain, inconvenience, and other hassles associated with treatments; (3) ensuring the availability of appropriate psychological support; and (4) letting paFertility and Sterility
tients know they have substantial control over decisions about their treatment (19). In addition, however, these results suggest there are potentially effective indirect approaches that aim to assist in maint~ining the marriage factors at satisfactory levels. Even if stress cannot be eliminated, its indirect effects may be greatly reduced. In terms of our model, this amounts to cutting some of the paths by which stress impacts life quality. For example, minimizing marital conflicts and maintaining wives' and husbands' own sexual self-esteem and satisfaction with their sexual lives should be effective in ameliorating the typical indirect effects of stress on life quality. These are endeavors that couples can be encouraged to address themselves and for which health professionals can attempt to provide counseling, support, and helpful information. Of course, the conclusions presented in this report are a function of the particular couples examined, the set of concepts that have been considered, and the appropriateness of our causal model. Within the constraints of the present analysis, the internal validity of the conclusions seems high. These fertilityproblem couples are typical of many of those who seek help for infertility; the marriage and life-quality factors that have been considered are certainly relevant to the issue; and the causal model is both theoretically plausible and empirically effective in fitting the data. However, this research should be recognized as being an early exploration in this area. Confidence in our conclusions will be enhanced if in subsequent research they are confirmed using broader samples of couples, wider ranges of concepts, and longitudinal analysis. REFERENCES 1. Dunkel-Schetter C, Lobel M. Psychological reactions to infertility. In: Stanton A, Dunkel-Shetter C, editors. Infertility. New York: Plenum, 1991:29-57.
Vol. 57, No.6, June 1992
2. Lasker IN, Borg S. In search of parenthood. Boston: Beacon, 1987. 3. Menning BE. Infertility: a guide for childless couples. Englewood Cliffs (NJ): Prentice-Hall, 1977. 4. Menning BE. The emotional needs of infertile couples. Fertil Steril 1980;34:313-9. 5. Seibel MM, Taymor ML. Emotional aspects of infertility. Fertil Steril 1982;37:137-45. 6. Moghissi KS, Wallach EE. Unexplained infertility. Fertil Steril 1983;39:5-21. 7. Abbey A, Andrews FM, Halman LJ. The importance of social relationships for infertile couples' well-being. In: Stanton A, Dunkel-Shetter C, editors. Infertility. New York: Plenum, 1991:61-86. 8. Abbey A, Andrews FM, Halman, LJ. Gender's role in responses to infertility. Psychol Women Quart 1991;15: 295-316. 9. Fisher S, Reason J., editors. Handbook of life stress cognition and health. New York: Wiley, 1988. 10. Claus KE, Bailey JT, editors. Living with stress and promoting well-being: a handbook for nurses. St. Louis (MO): Mosby, 1980. 11. Sethi AS, Schuler RS. Handbook of organizational stress coping strategies. Cambridge (MA): Ballinger, 1984. 12. Kutash IL, Schlesinger LB. Handbook on stress and anxiety. San Francisco: Jossey-Bass, 1980. 13. Henshaw SK, Orr MT. The need and unmet need for infertility services in the United States. Fam Plann Perspect 1987;19:180-6. 14. McKennell AC. Attitude scale construction. In: O'Muircheartaigh CA, Payne C, editors. Analysis of survey data. New York: Wiley, 1977;1:183-220. 15. Andrews FM, Withey SB. Social indicators of well-being: Americans' perceptions of life quality. New York: Plenum, 1976. 16. Bollen KA. Structural equations with latent variables. New York: Wiley, 1989. 17. Bentler PM. EQS: structural equations program manual. Los Angeles: BMDP Statistical Software, 1989. 18. Freeman EW, Boxer AS, Rickels K, Tureck R, Mastroianni L Jr. Psychological evaluation and support in a program of in vitro fertilization and embryo transfer. Fertil SterilI985;43: 48-53. 19. Abbey A, Halman LJ, Andrews FM. Psychosocial, treatment, and demographic predictors ofthe stress associated with infertility. Fertil SterilI992;57:122-8.
Andrews et a1.
Is fertility-problem stress different?
1253