Mood, Sexual Interest and Sexual Activity among Adolescent Women Running Head: Adolescent Mood and Sexual Behavior

Objective: to examine day-to-day associations of coitus, sexual interest, partner emotional support, negative mood and positive mood among adolescent women. Methods: Women (ages 14 – 17 at enrollment; N=146) enrolled from one of three adolescent primary care clinics completed up to five 84-day diaries over a 27-month period. The diaries assessed partner interactions, sexual activity, substance use and mood. Partner-specific measures assessed on each day included partner emotional support (4 items; alpha = 0.94), argument with a partner (no/yes) and coitus (no/yes). Within-day measures assessed marijuana use (no/yes), Positive Mood (3-items; alpha = 0. 86); Negative Mood (3-items; alpha = 0.82) and Sexual Interest (1-item). Lagged measures of mood and sexual activity were included in multivariate models to control for recent mood and sexual behavior effects on current day mood and coitus. Two main analyses were conducted: coitus as a predictor of positive and negative mood; and the role of positive and negative mood as predictors of coitus. Analyses were conducted by multivariate mixed effect regression and mixed effect logistic regression models. Results: Data represent 28,376 days from 146 participants. The average number of diary days was 194 days per participant. Sexual activity was reported on 8.3% of days, with condoms used for 27.0% of these coital events. Marijuana was used on 11% of days. Significant predictors of positive mood on a given day included partner support, marijuana use, and coitus. Negative mood was associated with having an argument with a partner and with prior day coitus. Predictors of coitus on a given day included age (Odds ratio = 1.22), increased coital frequency in previous week (OR = 1.49), coitus on the previous day (1.21), increased same-day sexual interest (OR = 2.8) and decreased same-day negative mood (OR = 0.92). Conclusions: The data demonstrate complex associations of sexual interest, mood, partner interactions and sexual activity. Introduction Studies of temporal factors associated with coitus may be important because condom use (a mainstay of STI prevention) is conditional on the occurrence of coitus and thus may be affected by variable factors associated with coitus. Partner factors and substance are additional influences on both mood and sexual activity. The purpose of this paper is to evaluate partner factors, substance use, mood and sexual interest as factors associated with day-to-day occurrence of coitus among adolescent women. Both lay and professional discourse typically frames adolescent sexuality simply in terms of epidemiologic risk of adverse outcomes such as unintended pregnancy or sexually transmitted infection. Adverse outcomes of sexual activity are common and clearly detrimental to healthy adolescent development. However, disproportionate emphasis on risk contributes to a perspective of adolescent women as sexual victims who are incapable of sexual agency other than refusal.{Fine 1988 112 /id} Several contemporary writers record the importance of sexual interest and desire in discussions of sexuality by adolescent women. These discussions – often interlaced with conflicting themes of romance, power, risk and danger – nonetheless emphasize the importance of sexual interest as a central component of sexual activity of adolescents.{Fine 1988 112 /id}{Thompson 1990 111 /id}{Tolman 2002 4051 /id} However, sexual desire is only one dimension of the complex phenomenon of sex. Mood, characteristics of the interpersonal relationship with a partner, expectations for sex as part of an established pattern of behavior, opportunities for sex, and use of mood-altering substances are additional elements that, taken together, influence the occurrence of sexual activity on a given day. Understanding of these factors is important because adolescent sexual activity is thought to be typically “sporadic” and because prevention of adverse consequences such as sexually transmitted infections requires specific additional behaviors (e.g., condom use) in conjunction with sex. Among these other phenomenological correlates of sexual activity, mood may be most obviously relevant to adolescents. Increased negative mood and increased mood variability are reasonably well documented aspects of adolescent development. Negative affective states, irritability and conflict (especially with parents) increase at the time of puberty.{Arnett 1999 2860 /id} Older adolescent girls (i.e., 9 graders) report significantly lower average daily happiness, cheerfulness, and friendliness compared to pre-teens and younger adolescents. Moreover, older girls show significantly greater day-to-day variation in mood.{Larson & Lampman-Petraitis 1989 42 /id} A number of studies show that depressed mood peaks in mid-adolescence. (See {Petersen, Compas, et al. 1993 3923 /id} for a review). These developmental changes in affect temporally coincide with increased levels of adolescent sexual activity. Higher levels of depression are seen in sexually experienced adolescent women, compared to those without sexual experience.{Kowaleski-Jones & Mott 1998 4065 /id}{Ramrakha, Caspi, et al. 2000 4066 /id} Self-esteem may decrease following sexual initiation among adolescent women.{Spencer, Zimet, et al. 2002 3833 /id} The studies discussed above evaluated global associations of negative affect and adolescent sexual activity rather than the day-to-day phenomena of mood and sex. Moreover, none of the studies appear to include assessments of positive mood. In common usage, the phrase “in the mood” connects positive mood and sexual interest with sexual activity. Strong relations are found between positive mood and sexual desire among adult women.{Warner & Bancroft 1988 3919 /id} Conversely, “not in the mood” suggests negative affect and lack of sexual interest that may return if mood changes. Among older adult women, clinical depression is usually associated with decreased sexual interest, arousal and pleasure.{Kennedy, Dickens, et al. 1999 4048 /id} Among collegeaged women, depressive symptoms (i.e., Beck Depression Inventory scores of 20 or greater) are associated with decreased sexual satisfaction and pleasure, although no differences in sexual desire or partnered sexual activity were noted.{Frolich & Meston 2002 4049 /id} This study noted, however, higher levels of interest in masturbation among the depressed women. This suggests that mood may not have consistent effects on all aspects of sexual expression. Recently, negative affective states common among adolescents (e.g., depression and anxiety) were linked to sexual behaviors associated with increased risk for sexually transmitted infections although these findings applied mainly to males.{Gold & Skinner 1992 3915 /id}{Shrier, Harris, et al. 2002 3917 /id}{Shrier, Harris, et al. 2001 3918 /id} Negative mood is aversive and may motivate sexual behavior that is believed to be ameliorate negative mood.{Thayer, Newman, et al. 1994 3995 /id}{Frolich & Meston 2002 4049 /id} However, a meta-analysis of 34 studies found no evidence of association between depressive symptoms and increased sexual risk behavior.{Crepaz & Marks 2001 3994 /id} Kalichman and Weinhardt pointed out in an accompanying editorial that such results were unsurprising given the paucity of event-level longitudinal studies.{Kalichman & Weinhardt 2001 3993 /id} The possibility of reciprocal effects of sexual activity on mood has also received little research attention. Feelings of satisfaction, relaxation and intimacy are well known psycho-physiologic correlates of coitus.{Heiman 1998 4069 /id} In such circumstances, coitus might be expected to increase positive mood for some period of hours or one or more subsequent days. Some contextual aspects of a given coital event, however, could be associated with increased negative subsequent mood. The possibility of mood elevating properties of sexual activity is important since increased levels of depressed mood are associated with lower levels of condom use and increased self-report of STI among adolescents.{Shrier, Harris, et al. 2002 3917 /id}{Shrier, Harris, et al. 2001 3918 /id} On the other hand, unplanned or coerced coitus, or coitus subsequently interpreted as ill-advised (e.g., because of partner characteristics, non-use of condoms, lack of contraception) might contribute to lingering negative mood. Given the psychological and social complexities of adolescent sexual activity, it is possible to imagine effects of coitus on both positive and negative mood. The characteristics of the interpersonal relationship with sex partners may also affect the likelihood of coitus on any given day. Sexual activity itself serves important functions in relationship building and maintenance, even though the structure of adolescent romantic relationships may differ from that of adults.{Quirk, Heiman, et al. 2002 4068 /id}{Graber, Britto, et al. 1999 4002 /id} A majority of adolescent women characterize their sexual partners in terms of an ongoing affiliation such as friend, boyfriend or fiancé.{Graber, Britto, et al. 1999 4002 /id} These socially constructed relationships also provide important linkages to peers as well as families.{Brown 1999 3877 /id} Among older adolescents, relationships characterized by affection and intimacy were associated with increased coital frequency.{Rostosky, Galliher, et al. 2000 3754 /id} Coital frequency may increase in frequency and regularity among adolescent women with more intense or more enduring close relationships.{Katz, Fortenberry, et al. 2000 3198 /id} In general, closeness within adolescence develops with age and is characterized by reciprocity, interdependence and frequency of interactions.{Adams, Laurensen, et al. 2001 3700 /id} Therefore, it is reasonable to hypothesize that coitus might be more or less likely on a given day within the context of specific interactions with a partner. After an initial coital experience, most adolescents continue sexual activity, although not necessarily on a frequent basis. Median frequencies of about two coital events per month are reported but ranges are from a few annual coital events per to daily or almost daily coitus.{Seidman, Mosher, et al. 1994 629 /id} Up to one-third of adolescent women with any lifetime coital experience report no coitus within the previous 30 days and 10% 15% report no coitus in the previous two months.{Aral & Cates 1989 813 /id} Although coitus is slightly more likely to occur on weekends than weekdays, sexual activity of adolescent women occurs throughout the week.{Fortenberry, Orr, et al. 1997 1837 /id}{Katz, Fortenberry, et al. 2000 3391 /id} This suggests that adolescent sexual activity is associated with other temporal phenomena and is not solely conditional on opportunities associated with decreased supervision.{Cohen, Farley, et al. 2002 4028 /id} The most widely evaluated phenomenon associated with coitus is use of alcohol or drugs before coitus (see {Fortenberry 1995 281 /id} for a review). Intoxicating substances are assumed to impair judgment, leading to unplanned or unwanted sex and non-use of condoms and contraception. Almost none of the numerous studies of this issue assessed associated sexual interest, mood, usual behaviors and partner interactions.{Leigh & Stall 1993 80 /id} In a previous diary study of adolescent women (assessing condom use, partner change and coitusassociated substance use), we showed that fewer than 30% of adolescent women ever paired substance use and coitus, and (among those who did) only about 25% of coital events were substance-associated. Coitus-specific condom use was predicted by the prior level of condom use but not by event-specific substance use.{Fortenberry, Orr, et al. 1997 1537 /id} A recent meta-analysis showed that drinking was unrelated condom use in recent sexual encounters.{Leigh 2002 3960 /id} However, wide-spread use of substances among adolescents and continued interest in their role in unplanned or risky sexual activity warrants additional research. A final relevant issue stems from the nature of the relationships of mood and sexual activity: coitus can serve a variety of purposes, leading to associations with either positive or negative mood, conceivably for the same person. Retrospective reports and single “critical event” studies cannot provide sufficiently detailed data to disentangle these complex effects.{Shiffman & Stone 1998 2177 /id} We chose daily, pre-printed diaries with weekly pickup because prior experience with adolescents suggested low levels of dropout, high levels of daily completion and relatively low levels of item-level missing data, even for reports of sensitive behaviors.{Skiba, Fortenberry, et al. 1997 2059 /id}{Howard, Fortenberry, et al. 1999 2791 /id}{Fortenberry, Tu W., et al. 2002 3475 /id}{Morrison, Leigh, et al. 1999 2626 /id} Data from retrospective reports and aggregated diary records (for the same time period) show good concordance.{Fortenberry, Cecil, et al. 1997 1684 /id} Diary methods place substantial demand on participants but Gillmore et al. found no evidence of respondent fatigue among young adults completing daily self-reports of multiple health behaviors over a 56-day period.{Gillmore, Gaylord, et al. 2001 3602 /id} Thus, the diary methodology provides a useful tool to address the complex phenomenology of adolescent women’s sexual activity.