Journal of Anxiety Disorders 21 (2007) 955–965

Generalized anxiety disorder and entry into marriage or a marriage-like relationship K. Lira Yoon a,*, Richard E. Zinbarg a,b b

a Northwestern University, United States The Family Institute at Northwestern University, United States

Received 13 March 2006; received in revised form 14 August 2006; accepted 18 October 2006

Abstract Previous studies have suggested that individuals with generalized anxiety disorder report significantly higher levels of marital distress and are at a greater risk for divorce than those without generalized anxiety disorder. Studies also showed that relationship difficulties among those with generalized anxiety disorder predict poor treatment response and long-term outcome. In this study, the relation between a lifetime diagnosis of generalized anxiety disorder and marital history was examined with the data from the National Comorbidity Survey (NCS). Respondents were grouped into those who have no mental disorder, those who have only generalized anxiety disorder, and those who have generalized anxiety disorder and a lifetime history of at least one of the 16 DSM-III-R diagnoses assessed in the NCS. Generalized anxiety disorder was significantly associated with the likelihood of entry into a marriage-like relationship. The results support the continued investigation into the association between couple functioning and the onset, course, and treatment of generalized anxiety disorder, and suggest that couples intervention could be an untapped resource for generalized anxiety disorder treatment. # 2007 Elsevier Ltd. All rights reserved. Keywords: Generalized anxiety disorder; Marriage; National Comorbidity Survey

Generalized anxiety disorder (GAD), one of the more common anxiety disorders, is associated with significant impairment in occupational, interpersonal, and family functioning. Studies have demonstrated that interpersonal issues involving couple and family relationships represent the ˝ st, 1999; Craske, Rapee, most frequent content area of GAD worries (Breitholtz, Johansson, & O Jackel, & Barlow, 1989; Roemer, Molina, & Borkovec, 1997). Relatedly, results from a handful

* Corresponding author at: Department of Psychology, University of Miami, P.O. Box 248185, Coral Gables, FL 33124-0751, United States. Tel.: +1 847 877 0501; fax: +1 305 284 3402. E-mail address: [email protected] (K.L. Yoon). 0887-6185/$ – see front matter # 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.janxdis.2006.10.006

956

K.L. Yoon, R.E. Zinbarg / Journal of Anxiety Disorders 21 (2007) 955–965

of studies over the past decade have consistently revealed elevated levels of marital distress among individuals with GAD. In her pioneering study in this area, McLeod (1994) found that wives with GAD reported significantly higher levels of marital distress than wives who did not have GAD. Their male partners, however, did not report more marital distress than those whose wife did not have GAD. In addition, husbands’ GAD was not associated with poorer marital functioning being reported by either spouse. The zero-order association that marital dissatisfaction was associated with GAD for wives, but not for husbands, was later replicated (Whisman, 1999). Furthermore, a recent study (Whisman, Sheldon, & Goering, 2000) found that the strongest diagnostic correlate of marital dissatisfaction was GAD among nine diagnoses (MDE, mania, dysthymia, social phobia, simple phobia, Agoraphobia, Panic Disorder, GAD and Alcohol dependence/abuse), and this effect was not moderated by gender. Even more noteworthy, using dissatisfaction with relationships with other relatives and with friends as covarites did not attenuate the significant association between marital dissatisfaction and GAD (Whisman et al., 2000). This finding is consistent with the notion that the association between GAD and marital dissatisfaction cannot be accounted for entirely by reporting bias or catastrophizing. Though we would not argue that all relationship difficulties end in divorce, the occurrence of divorce certainly provides an objective indicator of relationship failure. Thus, it is important to note that Kessler, Walters, and Forthofer (1998) found that many psychiatric disorders including GAD were associated with a significantly elevated risk of divorce (the odds ratio for predicting divorce from GAD was significant for both men and women and in the total sample equaled 1.7 with a 95 percent confidence interval of 1.2–2.2). This result strongly suggests that the association between GAD and marital distress is not due entirely to reporting bias or catastrophizing. Four findings suggest that it might be useful to add a couples intervention to standard individual cognitive-behavioral therapy (CBT) for GAD. First, though CBT is arguably the treatment of choice for GAD, there is much room for improvement as only 50 percent of patients treated with CBT achieve high end-state status1 (e.g., Borkovec, Newman, Pincus, & Lytle, 2002; Borkovec & Whisman, 1996; Butler, Fennell, Robson, & Gelder, 1991; Durham, Allan, & Hackett, 1997; Gould, Otto, Pollack, & Yap, 1997). Second, as discussed above, individuals with GAD show elevated marital distress and elevated risk for divorce (e.g., McLeod, 1994; Whisman, 1999). Third, marital problems seem to predict poor treatment outcome in individuals with GAD. For example, Durham et al. (1997) investigated predictors of 12-month post-treatment (either cognitive therapy, analytic psychotherapy, or anxiety management training) GAD status and showed that marital quality (degree of marital tension) was a significant predictor of relapse. Similarly, poor overall life satisfaction and impaired ratings of the GAD individual’s relationships were more highly associated than other predictors with those who continued to suffer from GAD in a study investigating the 16-month post-medication treatment status of GAD patients (Mancuso, Townsend, & Mercante, 1993) and in a 5-year longitudinal study (Yonkers, Dyck, Warshaw, & Keller, 2000). A related study was recently completed in our laboratory and found that partner pre-treatment hostility and the absence of partner pre-treatment non-hostile criticism predicted significantly worse end state functioning (Zinbarg et al., 2007). Finally, major

1 The most widely used strategy for assessing clinically significant change in treatment outcome studies in this area has been to classify patients according to whether they have achieved high end-state functioning (HES) which essentially relates to whether scores on outcome measures fall within the non-clinical range.

K.L. Yoon, R.E. Zinbarg / Journal of Anxiety Disorders 21 (2007) 955–965

957

depressive disorder (MDD) and GAD appear to share a single, common genetic diathesis and behavioral couples therapy is an efficacious treatment for MDD (e.g., Beach & O’Leary, 1986, 1992; Jacobson, Dobson, Fruzzetti, Schmaling, & Salusky, 1991; Leff et al., 2000; O’Leary & Beach, 1990). Before investing significant resources in the development, testing or both of a couples intervention for GAD, however, a remaining question should be answered. That is, do individuals with GAD have more difficulties entering into marriage or a marriage-like relationship than other people because of the interpersonal problems they tend to experience? If individuals with GAD are much less likely than others to be involved in an intimate relationship, investing our limited resources in a couples intervention for GAD might not be the best option to pursue. A contrasting prediction is suggested by consideration of the findings that MDD has been associated with interpersonal dependency and early marriage (e.g., Davila, 2001; Gotlib, Lewinsohn, & Seeley, 1998). If the common genetic diathesis, or other common etiological factors, shared by GAD and MDD also lead to elevated interpersonal dependency in GAD, then people with GAD should be expected to be even more likely to enter into committed, intimate relationships than others. Three previous studies (Blazer, Hughes, George, Swartz, & Boyer, 1991; Forthofer, Kessler, Story, & Gotlib, 1996; Wittchen, Zhao, Kessler, & Eaton, 1994) using large, epidemiological samples provided some insights related to this question by reporting either the rates of GAD in never married, married, and previously married or associations between psychiatric disorders and the probability and timing of first marriage. However, we are more interested in the rates of marriage and a marriage-like relationship in people with GAD compared to those without GAD whereas two of these previous studies reported the rates of GAD within different groups defined by marital status. To be able to answer the question of whether people with GAD are more likely to enter into a committed relationship, we should investigate the association between GAD and those who have ever been married (i.e., married, divorced, separated, and widowed all combined). Thus, past investigations of the rates of GAD in the never married, married and divorced do not directly answer our question. In addition, the previous study (Wittchen et al., 1994) using the National Comorbidity Survey (NCS: Kessler et al., 1994) did not account for being in a marriage-like relationship. Though Forthofer et al. (1996) included GAD in their analyses of associations between psychiatric disorders and probability and timing of first marriage, it is impossible to determine the association between any disorder including GAD and the overall likelihood of entering into marriage based on these analyses (Kessler, personal communication, April 22, 2005). That is, these authors only reported their results in terms of associations with early marriage (i.e., prior to age 19), on-time marriage (i.e., between the ages of 19 and 24) and late marriage (i.e., after age 24) and did not conduct analyses of the associations between disorders and marriage regardless of whether the marriage was early, on-time or late. In addition, whereas their point estimates revealed that GAD had a positive association with early marriage, GAD had negative associations with on-time and late marriages and all of these associations were not statistically significant making it unclear as to what the overall association between GAD and the probability of getting married is. Thus, we analyzed the data from the NCS (Kessler et al., 1994) to determine the relation between a lifetime diagnosis of GAD and entry into marriage or a marriage-like relationship. Because the NCS (Kessler et al., 1994) is an old dataset based on DSM-III-R (American Psychiatric Association, 1987), there might be concerns whether the current study could adequately be applied to individuals with DSM-IV (American Psychiatric Association, 1994) GAD.

958

K.L. Yoon, R.E. Zinbarg / Journal of Anxiety Disorders 21 (2007) 955–965

There were four major changes in the diagnostic criteria from DSM-III-R to DSM-IV: (1) the A criterion that defines the excessive worry has been simplified, (2) a new criterion was added stating that the worry must be difficult to control, (3) the associated symptom criterion has been focused, deleting most of the previous autonomic symptoms and leaving only the symptoms that reflect hyper-vigilance and motor tension, and (4) a new criterion regarding clinical significance was added. The simplification of the A criterion could increase the prevalence of those defined as excessive worriers, but as for the diagnosis of GAD, the more stringent associated symptom criterion and the two additional criteria could result in lower prevalence rate (Carter, Wittchen, Pfster, & Kessler, 2001). Several studies (e.g., Abel & Borkovec, 1995; Brown, Marten, & Barlow, 1995; Carter et al., 2001), however, have demonstrated that those who met DSM-III-R GAD generally met DSM-IV criteria for GAD. For example, when 40 patients with DSM-III-R GAD and 36 non-anxious controls were interviewed using both the DSM-III-R and DSM-IV criteria, all 40 patients also met DSM-IV GAD whereas those who did not meet GAD based on the former version also failed to meet GAD based on DSM-IV (Abel & Borkovec, 1995). Thus, the results based on the NCS data with DSM-III-R appear to have potential to shed light on DSM-IV GAD. As mentioned earlier, we addressed one specific question: is a lifetime diagnosis of GAD associated with the likelihood of being in a committed relationship? 1. Method 1.1. Respondents The NCS is a nationally representative survey based on a stratified, multistage probability sample of non-institutionalized civilian individuals aged 15–54 living in the coterminous 48 United States (Kessler et al., 1994). In Part I 8,098 respondents completed face-to-face structured interview conducted by lay interviewers to assess diagnoses based on the revised third edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-III-R; American Psychiatric Association, 1987) between September of 1990 and February of 1992. A probability subsample of 5,877 respondents consisted of (a) all Part I respondents aged 15– 24 years (99 percent of whom completed Part II), (b) all Part I respondents who were positive on initial questions in one or more psychiatric diagnostic sections of the interview (98 percent of whom completed Part II), and (c) a one-in-six random subsample of all remaining Part I respondents (99 percent of whom completed Part II). Cases were weighted to adjust for differential probabilities of selection into the survey, and weighted data were then poststratified to approximate the cross-classification of the distributions of age, sex, race, and several other demographic variables in the 1989 U.S. National Health Interview Survey. The demographic distribution of the Part II subsample was highly representative of the total population of the United States (Kessler et al., 1994). The analyses reported here are based on the Part II subsample, weighted to be representative of the population. Respondents who were unlikely to be married at the time of the interview because of their relative youth were excluded from our analyses. That is, respondents younger than 18 of age were excluded from our analyses.2

2

The pattern of results reported here remained the same when respondents whose age was less than 18 were included in our analyses.

K.L. Yoon, R.E. Zinbarg / Journal of Anxiety Disorders 21 (2007) 955–965

959

1.2. Assessment of GAD and other mental disorders The NCS assessed 17 DSM-III-R disorders, including GAD. Most diagnoses were derived from a modified version of the Composite International Diagnostic Interview (CIDI; World Health Organization, 1990), which is a structured interview that was designed to be used by trained interviewers who are not clinicians to make diagnoses based on the DSM-III-R. Diagnoses of conduct disorder, adult antisocial behavior, and antisocial personality disorder (all not assessed in the CIDI), as well as posttraumatic stress disorder (PTSD; Kessler, 1995, p. 141), were derived from the Diagnostic Interview Schedule (DIS; Kessler, Borges, & Walters, 1999). 1.3. Assessment of marital history A question that assessed respondents’ marital status and a question that assessed whether respondents were in a marriage-like relationship were used to classify respondents as ever in a marriage-like relationship (EML) or never married (NM). EML was defined as those who were either married, divorced, separated, or widowed (i.e., ever married) or those who were never married but currently in a marriage-like relationship. Ninety-one respondents who were never married were currently in a marriage-like relationship. 1.4. Data analysis Two sets of analyses were conducted. For the first set of analysis, we created a GAD diagnostic status variable reflecting those who have no mental disorder (NMD), those who have only GAD (Pure GAD), and those who have GAD and a lifetime history of at least one of the following 16 DSM-III-R diagnoses: panic disorder, agoraphobia, simple phobia, social phobia, PTSD, major depressive episode, dysthymia, bipolar disorder, alcohol abuse, alcohol dependence, drug abuse, drug dependence, conduct disorder, adult antisocial behavior, antisocial personality disorder, and nonaffective psychosis (Comorbid GAD). We used a lifetime prevalence of these psychiatric disorders. In keeping with the original reports on the NCS (Kessler et al., 1994), diagnoses were made without exclusions for DSM-III-R hierarchy rules. A 3 (Diagnostic Status)  2 (Relationship Status) x2 test was conducted to investigate whether the diagnostic status variable (NMD, Pure GAD, Comorbid GAD) is associated with entry into marriage and marriage-like relationships. To investigate whether the effects were unique vis-a`-vis demographic variables, logistic regressions were conducted. For logistic regressions, relevant demographic variables including sex were entered first, then the participants’ diagnostic status (i.e., NMD, Pure GAD, Comorbid GAD) was entered. Finally, the cross-products of centered sex and diagnostic status scores were entered into the regression analyses to examine the presence of sex differences in the associations between GAD and the relationship status. Significant main effects of diagnostic status in the logistic regressions were followed by pairwise comparisons among the three diagnostic groups (i.e., NMD vs. Pure GAD, NMD vs. Comorbid GAD, Pure GAD vs. Comorbid GAD). There were only 28 respondents in the Pure GAD group, which could lead to a lack of power. Therefore, we created slightly different four diagnostic status groups for the second set of analyses. That is, those who have no mental disorder (NMD), those who have GAD (All GAD: Pure GAD and Comorbid GAD combined), those who have one of the 16 diagnoses measured in the NCS interview but do not have GAD (Other Psychopathology), and those who have at least

960

K.L. Yoon, R.E. Zinbarg / Journal of Anxiety Disorders 21 (2007) 955–965

two of the 16 diagnoses without GAD (Comorbid Other). The same series of analyses described in the previous paragraph were carried out with these four groups. For all the analyses, we applied NCS recommended sampling weights (Kessler et al., 1994). 2. Results Table 1 shows the percentages of participants who were ever married, never married but currently in a marriage-like relationship and never married nor currently in a marriage-like relationship as a function of GAD diagnostic status. We first conducted a x2 test to examine whether individuals’ GAD status is associated with ones’ relationship status. Because the diagnostic status variable (NMD, Pure GAD, and Comorbid GAD) could be considered as an ordinal measure with increasing severity of psychopathology, we followed Agresti (1996) and looked at the chi-square for linearity. The results suggested that GAD is associated with being in an intimate relationship, as the Linear-by-Linear Association measure equaled 16.08, p < .001. Taken together with the pattern revealed in Table 1 that the percentage of Pure GADs (92.9 percent) who ever entered into marriage was even somewhat higher than the corresponding percentage of the Comorbid GAD group (80.8 percent) with the lowest rate being observed among the NMD group (67.4 percent), this indicates that GAD per se was associated with a higher rate of entry into an intimate relationship rather than the association all being accounted for by the disorders that GAD is comorbid with. Next, we conducted a logistic regression to examine whether the effect is unique vis-a`-vis demographic variables that were found to be associated with relationship status. To determine the background variables that are associated with relationship status, a series of x2 tests and t tests were conducted. The results indicated that age, sex, religion, race, urbanicity, number of years in formal education, and the total family income were significantly different between the respondents who have ever been married or in a marriage-like relationship and those who have never been married (all ps < .001). Therefore, these background variables were entered first, then the respondents’ GAD status, and finally the cross-products of centered sex and GAD status scores were entered into the regression analysis. The interaction between sex and diagnostic status was not significant, B = .11, ns, indicating no significant difference between men and women in a GAD-intimate relationship association. Therefore, the results presented hereafter, and shown in Table 2, are based on the analyses conducted across men and women. Respondents’ diagnostic status was significantly associated with entry into a committed, intimate relationship, B = .31, p < .006, indicating that respondents’ GAD status predicts above Table 1 Association between a lifetime history of generalized anxiety disorder (GAD) and entry into marriage Marital status

Ever married Never married, but currently in a marriage-like relationship Never married

GAD status NMD (n = 2,814)

Pure GAD (n = 28)

Comorbid GAD (n = 266)

1,897 (67.4 percent) 76 (2.7 percent)

26 (92.9 percent) 0 (0 percent)

215 (80.8 percent) 15 (5.6 percent)

2 (7.1 percent)

36 (13.5 percent)

841 (29.9 percent)

Note. The values represent percentage of respondents for each marital status category within each GAD status. NMD = no mental disorder.

K.L. Yoon, R.E. Zinbarg / Journal of Anxiety Disorders 21 (2007) 955–965

961

Table 2 Summary of logistic regression analyses for variables predicting the likelihood of being in an intimate relationship B Age Race Religion Sex Total family income Number of years in formal education Urbanicity GAD status (NMD, Pure GAD, Comorbid GAD) *

p < .05;

**

p < .01;

***

.19 .08 .01 .72*** .05* .13*** .50*** .31**

SE

Odds ratio

.01 .07 .06 .12 .02 .03 .08 .11

1.21 .94 .98 2.19 1.03 .88 1.71 1.45

***

p < .001.

and beyond the prediction provided by demographic variables. Follow-up logistic regression analyses using the same covariates as mentioned above were conducted to compare (a) NMD versus Pure GAD, (b) Pure GAD versus Comorbid GAD, and (c) NMD versus Comorbid GAD. These analyses indicated that individuals with Comorbid GAD were significantly more likely to enter into a committed, intimate relationship than individuals with NMD, B = .31, p < .007, but were not significantly different from those with Pure GAD, B = .03, ns. The difference between individuals with NMD and those with Pure GAD was not significant, B = .45, ns. Considering that the odds ratio for the difference between NMD and Pure GAD was greater than the odds ratio for the difference between NMD and Comorbid GAD (i.e., 1.57 vs. 1.36), it seems likely that the reason we did not obtain a significant result for the difference between NMD and Pure GAD was a lack of power resulting from the fact that there were only a total of 28 respondents with Pure GAD. To further examine whether the findings reported above are mainly due to having GAD or to having any mental disorder and to increase power, a second set of analyses were carried out. In this second set of analyses all those respondents with GAD (i.e., Pure GAD and Comorbid GAD) were combined to form one group (All GAD). In addition, those who have one mental disorder other than GAD formed a group (Other Psychopathology), and those who have at least two mental disorders but do not have GAD formed another group (Comorbid Other). Table 3 shows the percentages of participants who were ever married, never married but currently in a marriage-like relationship and never married nor currently in a marriage-like relationship as a function of this new diagnostic status. We first conducted a x2 test to examine Table 3 Association between diagnostic status and the marital status Marital status

Ever married Never married, but currently in a marriagelike relationship Never married

Diagnostic status NMD (n = 2,544)

All GAD (n = 292)

Other Psychopathology (n = 923)

Comorbid Other (n = 1,634)

1,893 (74.41 percent) 77 (3.02 percent)

241 (82.53 percent) 16 (5.48 percent)

688 (74.54 percent) 37 (4.01 percent)

1,158 (70.87 percent) 76 (4.65 percent)

574 (22.56 percent)

35 (11.99 percent)

198 (21.45 percent)

400 (24.48 percent)

Note. The values represent percentage of respondents for each marital status category within each diagnostic status. NMD = no mental disorder.

962

K.L. Yoon, R.E. Zinbarg / Journal of Anxiety Disorders 21 (2007) 955–965

Table 4 Summary of logistic regression analyses for variables predicting the likelihood of being in an intimate relationship B Age Number of years in formal education Race Religion Sex Total family income Urbanicity Diagnostic status (NMD, All GAD, Other Psychopathology, Comorbid Other) *

p < .05;

***

.18 .14*** .09 .06 .60*** .06*** .48*** .06*

SE

Odds ratio

.01 .02 .05 .04 .08 .01 .06 .03

1.20 .87 .92 .94 1.82 1.06 1.62 1.07

***

p < .001.

whether relationship status in individuals with GAD is different from other groups. There was a significant difference between the diagnostic groups in terms of their relationship status, x2(3) = 22.76, p < .001. The follow-up analyses revealed that the GAD group was significantly different from all the other three groups (all ps < .001), suggesting individuals with GAD are more likely to be in an intimate, committed relationship than individuals in the other groups (i.e., NMD, Other Psychopathology, Comorbid Other). No other comparisons yielded significant results. This indicates that GAD per se was at least partly associated with a higher rate of entry into a committed relationship rather than just having any mental disorder. Next, we conducted a logistic regression to examine whether the effect is unique vis-a`-vis demographic variables that were found to be associated with relationship status. Again the background variables related to relationship status were entered first, then diagnostic status, and finally the cross-products of centered sex and diagnostic status scores were entered into the regression analysis. The interaction between sex and diagnostic status was not significant, B = .07, ns, indicating that there was not a significant difference in a diagnostic status-relationship association between men and women. Therefore, the results presented hereafter, and shown in Table 4, are based on the analyses conducted across men and women. Respondents’ diagnostic status was significantly associated with one’s relationship status, B = .06, p < .02, indicating that respondents’ diagnostic status predicts above and beyond the prediction provided by demographic variables. Follow-up logistic regression analyses using the same covariates as mentioned above were conducted to compare (a) NMD versus All GAD, (b) NMD versus Other Psychopathology, (c) NMD versus Comorbid Other, (d) All GAD versus Other Psychopathology, (e) All GAD versus Comorbid Other, and (f) Other Psychopathology versus Comorbid Other. The analysis comparing NMD with All GAD was significant, B = .60, p < .006, indicating that individuals with GAD were significantly more likely to have been in a committed relationship than individuals with NMD above and beyond demographic variables. The analyses comparing NMD with Other Psychopathology, B = .10, p < .07, and with Comorbid Other, B = .06, p < .06, approached significance, suggesting that those with psychopathology were more likely to have been in a committed relationship than those with NMD. Respondents in the Other Psychopathology group, B = .52, p < .02, and those in the Comorbid Other, B = .22, p < .04, were significantly less likely to be in an intimate, committed relationship than those in the All GAD group above and beyond demographic variables, suggesting that GAD per se was at least partly associated with a higher rate of entry into a committed relationship rather than just having a mental

K.L. Yoon, R.E. Zinbarg / Journal of Anxiety Disorders 21 (2007) 955–965

963

disorder. There was no significant difference between Other Psychopathology and Comorbid Other, B = .004, ns. Considering a previous study with a treatment-seeking sample reported 75 percent of GAD patients having onset at age 26 or younger with the mean onset age of 18, one might infer that GAD increases the likelihood of getting into marriage or involved in a marriage-like relationship (rather than entry into marriage or a marriage-like relationship increasing the likelihood of developing GAD). In the current NCS sample, a similar pattern emerged in terms of onset age. That is, 60 percent of respondents with a lifetime diagnosis of GAD had onset at age 26 or younger. However, a better way to examine the potential directionality of the effect is to consider the temporal order of marriage and GAD onset. Establishing the effect of GAD on entry into marriage requires at a minimum that GAD onset would precede the marriage. In the current sample, of 241 GAD respondents who had ever been married, 6.4 percent (15) of the respondents married in the same year of their GAD onset. Approximately 31 percent (74) of the respondents reported onset of GAD followed by marriage, whereas twice as many respondents (i.e., 62.8 percent or 151 respondents) reported the pattern of marriage followed by GAD onset. Thus, it was more likely that respondents reported developing GAD after they got married in this sample, x2(1) = 20.15, p < .001. 3. Discussion Results with this epidemiological sample suggest that GAD is associated with entry into a committed, intimate relationship. More specifically, individuals with GAD are more likely to get married or enter into a marriage-like relationship than those who do not have GAD. Because we also take into consideration of being in a marriage-like relationship, the results seems to rule out the possibility that people without GAD are merely more willing to live together without the official societal sanction of marriage than people with GAD. In other words, the results seemed to be related to the tendency to form committed, intimate relationships rather than conformity to convention. Thus, these results are consistent with the notion that elevated interpersonal dependency is characteristic of GAD. More importantly, the results indicate that the association between GAD and relationship status cannot be explained by simply having a mental disorder. That is, individuals with GAD were still significantly more likely to enter into marriage or a marriage-like relationship than individuals who have mental disorders other than GAD. Thus, despite that fact that GAD is highly comorbid with other psychopathology, having GAD seems to have at least some unique association with one’s relationship status. Furthermore, the association between GAD and one’s relationship status was significant above and beyond the effects of demographic variables associated with one’s relationship status. Regarding the direction of association between entry into an intimate relationship and GAD, the results of our analyses of the temporal order of marriage and GAD onset suggest that in this sample it was typically not the expression of GAD symptoms that led to a higher likelihood of getting married. Of course, it is still possible that a third variable, such as being high on interpersonal dependency is a common cause of both a higher than normal chance of getting married and of developing GAD. The present findings taken together with those showing elevated marital distress (e.g., McLeod, 1994; Whisman, 1999; Whisman et al., 2000) and marital dissolution (Kessler et al., 1998) have implications for future treatment research on GAD. Given that individuals with GAD are more likely to get married but also more likely to report marital dissatisfaction and end their marriages combined with the fact that only 50 percent of patients treated with cognitive-behavior

964

K.L. Yoon, R.E. Zinbarg / Journal of Anxiety Disorders 21 (2007) 955–965

therapy achieve high end-state status (e.g., Borkovec et al., 2002; Borkovec & Whisman, 1996; Butler et al., 1991; Durham et al., 1997; Gould et al., 1997), couples intervention could be an untapped resource for GAD treatment. To develop an effective couples treatment for GAD, however, we need to understand the dyadic processes that lead to marital dissatisfaction, dissolution or both in GAD couples and that might impact on treatment response. Thus, there is a need for research that examines couple functioning in GAD that goes beyond simply measuring marital satisfaction or even marital dissolution. Behavioral observations of the interactions of GAD couples as predictors of treatment response and as compared with the interactions of appropriate control couples might be especially useful in this regard. In summary, the results from the current investigation suggest that individuals with GAD are more likely to enter into marriage or a marriage-like relationship. Taken together with earlier reports that individuals with GAD also report elevated marital distress (e.g., McLeod, 1994; Whisman, 1999; Whisman et al., 2000) and are at elevated risk for marital dissolution (Kessler et al., 1998), these results support the continued investigation into the association between couple functioning and the onset, course, and treatment of GAD. Acknowledgements Preparation of this article was supported by the Patricia M Nielsen Research Chair of the Family Institute at Northwestern University. The data reported herein come from the NCS. The NCS is a collaborative epidemiologic investigation of the prevalence, causes, and consequences of psychiatric morbidity and comorbidity in the United States supported by the National Institute of Mental Health (grants R01 MH46376, R01 MH49098, and R01 MH52861), with supplemental support from the National Institute on Drug Abuse, Bethesda (through a supplement to grant MH46376) and supplement grant 90135190 from the W.T. Grant Foundation, New York, NY (Dr. Kessler, principal investigator). References Abel, J. L., & Borkovec, T. D. (1995). Generalizability of DSM-III-R generalized anxiety disorders to proposed DSM-IV criteria and cross-validation of proposed changes. Journal of Anxiety Disorders, 9, 303–315. Agresti, A. (1996). An introduction to categorical data analysis. New York: Wiley. American Psychiatric Association. (1987). Diagnostic and statistical manual of mental disorders (3rd ed., Rev.). Washington, DC: Author. American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author. Beach, S. R., & O’Leary, K. D. (1992). Treating depression in the context of marital discord: outcome and predictors of response of marital therapy versus cognitive therapy. Behavior Therapy, 23, 507–528. Beach, S. R., & O’Leary, K. D. (1986). The treatment of depression occurring in the context of marital discord. Behavior Therapy, 17, 43–49. Blazer, D. G., Hughes, D., George, L. K., Swartz, M., & Boyer, R. (1991). Generalized anxiety disorder. In: L. N. Robin & D. A. Regier (Eds.), Psychiatric disorders in America: The epidemiologic catchment area study (pp. 180–203). New York: The Free Press. Borkovec, T. D., Newman, M. G., Pincus, A. L., & Lytle, R. (2002). A component analysis of cognitive-behavioral therapy for generalized anxiety disorder and the role of interpersonal problems. Journal of Consulting and Clinical Psychology, 70, 288–298. Borkovec, T. D., & Whisman, M. A. (1996). Psychosocial treatment for generalized anxiety disorder. In: M. R. Mavissakalian & R. F. Prien (Eds.), Long-term treatments of anxiety disorders (pp. 171–199). Washington, DC: American Psychiatric Press.

K.L. Yoon, R.E. Zinbarg / Journal of Anxiety Disorders 21 (2007) 955–965

965

˝ st, L. G. (1999). Cognitions in generalized anxiety disorder and panic disorder patients: Breitholtz, E., Johansson, B., & O a prospective approach. Behaviour Research and Therapy, 37, 533–544. Brown, T. A., Marten, P. A., & Barlow, D. H. (1995). Discriminant validity of the symptoms constituting the DSM-III-R and DSM-IV associated symptom criterion of generalized anxiety disorder. Journal of Anxiety Disorders, 9, 317–328. Butler, G., Fennell, M., Robson, P., & Gelder, M. (1991). Comparison of behavior therapy and cognitive behavior therapy in the treatment of generalized anxiety disorder. Journal of Consulting and Clinical Psychology, 59, 167–175. Carter, R. M., Wittchen, H.-U., Pfster, H., & Kessler, R. C. (2001). One-year prevalence of subthreshold and threshold DSM-IV generalized anxiety disorder in a nationally representative sample. Depression and Anxiety, 13, 78–88. Craske, M. G., Rapee, R. M., Jackel, L., & Barlow, D. H. (1989). Qualitative dimensions of worry in DSM-III-R generalized anxiety disorder subjects and nonanxious controls. Behaviour Research and Therapy, 27, 397–402. Davila, J. (2001). Refining the association between excessive reassurance seeking and depressive symptoms: the role of related interpersonal constructs. Journal of Social and Clinical Psychology, 20, 538–559. Durham, R. C., Allan, T., & Hackett, C. A. (1997). On predicting improvement and relapse in generalized anxiety disorder following psychotherapy. British Journal of Clinical Psychology, 36, 101–119. Forthofer, M. S., Kessler, R. C., Story, A. L., & Gotlib, I. H. (1996). The effects of psychiatric disorders on the probability and timing of first marriage. Journal of Health and Social Behavior, 37, 121–132. Gotlib, I. H., Lewinsohn, P. M., & Seeley, J. R. (1998). Consequences of depression during adolescence: marital status and marital functioning in early adulthood. Journal of Abnormal Psychology, 107, 686–690. Gould, R. A., Otto, M. W., Pollack, M. H., & Yap, L. (1997). Cognitive behavioral and pharmacological treatment of generalized anxiety disorder: A preliminary meta-analysis. Behavior Therapy, 28, 285–305. Jacobson, N. S., Dobson, K., Fruzzetti, A. E., Schmaling, K. B., & Salusky, S. (1991). Marital therapy as a treatment for depression. Journal of Consulting and Clinical Psychology, 59, 547–557. Kessler, R. C. (1995). The National Comorbidity Survey: Preliminary results and future directions. International Journal of Methods in Psychiatric Research, 5, 139–151. Kessler, R. C., Borges, G., & Walters, E. E. (1999). Prevalence of and risk factors for lifetime suicide attempts in the National Comorbidity Survey. Archives of General . Kessler, R. C., McGonagle, K. A., Zhao, S., Nelson, C. B., Hughes, M., Eshleman, S., et al. (1994). Lifetime and 12month prevalence of DSM-III-R psychiatric disorders in the United States: Results form the National Comorbidity Survey. Archives of General Psychiatry, 51, 8–19. Kessler, R. C., Walters, E. E., & Forthofer, M. S. (1998). The social consequences of psychiatric disorders. III. Probability of marital stability. American Journal of Psychiatry, 155, 1092–1096. Leff, J., Vearnals, S., Brewin, C. R., Wolff, G., Alexander, B., Asen, E., et al. (2000). The London Depression Intervention Trial: Randomised controlled trial of antidepressants vs. couple therapy in the treatment and maintenance of people with depression living with a partner: clinical outcome and costs. British Journal of Psychiatry, 177, 95–100. Mancuso, D. M., Townsend, M. H., & Mercante, D. E. (1993). Long-term follow-up of generalized anxiety disorder. Comprehensive Psychiatry, 34, 441–446. McLeod, J. D. (1994). Anxiety disorders and marital quality. Journal of Abnormal Psychology, 103, 767–776. O’Leary, K. D., & Beach, S. R. (1990). Marital therapy: A viable treatment for depression and marital discord. American Journal of Psychiatry, 147, 183–186. Roemer, L., Molina, S., & Borkovec, T. D. (1997). An investigation of worry content among generally anxious individuals. Journal of Nervous and Mental Disease, 185, 314–319. Whisman, M. A. (1999). Marital dissatisfaction and psychiatric disorders: results from the national comorbidity survey. Journal of Abnormal Psychology, 108, 701–706. Whisman, M. A., Sheldon, C. T., & Goering, P. (2000). Psychiatric disorders and dissatisfaction with social relationships: Does type of relationship matter? Journal of Abnormal Psychology, 109, 803–808. Wittchen, H., Zhao, S., Kessler, R. C., & Eaton, W. (1994). DSM-III-R generalized anxiety disorder in the National Comorbidity Survey. Archives of General Psychiatry, 51, 355–364. World Health Organization. (1990). Composite international diagnostic interview. Geneva, Switzerland: Author. Yonkers, K. A., Dyck, I. R., Warshaw, M., & Keller, M. B. (2000). Factors predicting the clinical course of generalised anxiety disorder. British Journal of Psychiatry, 174, 544–549. Zinbarg, R. E., Lee, J. E., & Yoon, K. L. (2007). Dyadic predictors of outcome in a cognitive-behavioral program for patients with generalized anxiety disorder in committed relationships: A ‘‘Spoonful of Sugar’’ and a dose of nonhostile criticism may help. Behaviour Research and Therapy, 45, 699–713.