Affect and Sexual Behavior in Adolescents: A Review of the Literature and Comparison of Momentary Sampling With Diary and Retrospective Self
http://www.100md.com
《小儿科》
Division of Adolescent/Young Adult Medicine Clinical Research Program
Department of Psychiatry, Children's Hospital Boston and Harvard Medical School, Boston, Massachusetts
ABSTRACT
Objective. Assessment of mental health is important in understanding sexual risk behavior in adolescents, yet few studies have examined how affect is directly related to sexual behavior. Momentary sampling (MS) methods permit real-time assessment of affect in relation to specific events and embed the collected data in the context of the respondent's moment-to-moment life. The objectives of this study were to review the literature on affect and sexual behavior and to compare the feasibility and acceptability of MS with diaries and retrospective self-report as a means of collecting temporally relevant data on affect and sexual behavior in adolescents.
Methods. Sexually active, nondepressed adolescent outpatients who were aged 15 to 18 years were randomly assigned to a schedule of the 3 methods of data collection for 2 weeks each. All participants completed a retrospective self-report by interview at the end of each 2-week period. In the diary arm, participants completed twice-daily paper-and-pencil diary cards, which were returned by mail. In the MS arm, participants used 2-way pagers to respond to several random pages per day. Primary outcomes included rates of completion (diaries vs MS reports) and the participants' tolerance of and preferences for the methods. A secondary outcome was the agreement in means for positive and negative affect and in report of days on which substance use and sexual activity occurred. Associations of affect with contextual factors and with sexual activity were also explored in the MS arm.
Results. Ten youths completed 30 of 30 retrospective self-reports (100%, 3 per participant, by design), 254 of 280 diaries (91%; mean: 25.4 per participant), and 442 of 600 MS reports (74%; mean: 44.2 per participant). Most participants preferred the MS method to the diaries or retrospective self-report. Affect scores and reports of sexual activity and substance use were correlated among the methods. Measured with MS, affect was found to differ by location, companionship, and thoughts when paged; notably, positive affect was highest when participants reported thoughts about sex. There was no difference in affect before versus after coitus.
Conclusions. The results of this study suggest that MS in adolescents is feasible and preferred and provides contextual, temporally relevant, event-level data on affect and sexual activity that are not readily measured with traditional methods. Future research using MS methods will be important in increasing our understanding of the link between affect and sexual behavior and inform the development of improved risk reduction interventions for adolescents.
Key Words: affect sexual behavior substance use momentary sampling diary retrospective self-report
Abbreviations: STI, sexually transmitted infection MS, momentary sampling BDI, Beck Depression Inventory NA negative affect PA, positive affect ESM, Experience-Sampling Method EMA, Ecological Momentary Assessment PANAS, Positive Affect-Negative Affect Schedule
Preventive interventions for sexually active youths have had only modest success in reducing their high risk for HIV/sexually transmitted infection (STI) and unintended pregnancy,1–3 suggesting that new approaches to sexual risk reduction must be considered. Interventions that target specific groups of at-risk youths may be particularly promising. Although affect, depression, and other measures of mental health have been shown to affect sexual behavior,4–8 the nature of the associations has not been fully explored and few interventions have been directed toward depressed youths. In particular, few studies have attempted to capture affect, a highly variable measure of mental state, or to establish the temporal and causal associations of affective states and sexual behaviors. The measurement of affect and sexual behavior presents several methodologic challenges. In this article, we present a review of the research and theories supporting the link between affect and sexual behavior and the methods that have been used to measure these constructs, including retrospective self-report, diaries, and momentary sampling (MS) methods. We then present the results of a feasibility study that directly compares MS with the other 2 methods in a sample of sexually active adolescents.
LINK BETWEEN AFFECT AND SEXUAL RISK BEHAVIOR: REVIEW OF THE LITERATURE
Terminology
The literature is varied in the use of the terms "mood," "affect," and "emotion." Affect and mood both refer to subjectively experienced feeling states (emotions).9 Affect is a momentary emotion that is a variable, fluctuating phenomenon, in contrast to the more pervasive, sustained condition of mood. However, in the review of the literature below, the terms used are those preferred by the investigators for the research cited. Affect and mood are clearly related, and the data from research on mood can inform the study of affect and sexual risk.
Clinical Research on Affect, Mood, and Sexual Behavior
Data from clinical studies support a link between affect and sexual behavior. Depressive symptoms10,11 and mood disorders12,13 are prevalent in adolescents, particularly among those who are at risk for HIV/STI.14–18 In a study of 674 high-risk adolescents and young adults, 11% of young women and 7% of young men reported frequent, severe depressive symptoms (Beck Depression Inventory [BDI] score >20).14 Among 125 clients who attended an STI clinic, 39.2% reported symptoms consistent with probable depression on the General Health Questionnaire.19 Compared with mentally healthy youths, young people with severe mental illness engage in increased sexual risk behavior, including early onset of sexual activity, more sexual partners, and less condom use.15–18 Adolescents with mental illness may engage in high-risk sexual behaviors as a result of poor impulse control, impaired judgment and perception of reality, decreased motivation for self-care,20 poor decision making, lack of knowledge or understanding of the risks, and dysfunctional social interactions.16,17,21 Self-regulatory processes to repair negative affect (NA) may be disrupted22 or self-gratifying in unhealthy ways (eg, substance use, unprotected sexual intercourse).23 Psychiatric distress is also associated with substance use,24–26 which is correlated with increased sexual risk behavior and STIs.6,27–30 Depressive symptoms have been associated with STI in many studies14,15,31,32 but not all.19
Prolonged depressed mood is the hallmark of depression; 90% to 95% of adolescents and young adults with major depressive disorder report depressed mood for at least 2 weeks (the rest meet the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition)9 criteria for major depressive disorder with the symptom of anhedonia).33 However, little is known about the role of mood and affect in sexual risk behavior among depressed adolescents. Among adult men with a propensity for depression, depressed mood was associated with increased sexual interest or response.34 In-depth interviews of gay and heterosexual men revealed that men who experienced increased sexual interest or activity when depressed reported wishing to use sex for mood regulation, to connect with someone and feel validated and/or to directly improve one's mood.34,35 A minority of gay men reported that they were more likely to take sexual risks when depressed because they did not care about the consequences.35
In a few clinical studies, moods in normal adolescents have been associated with sexual behaviors.36,37 Positive mood has been correlated with sexual interest37 and the occurrence of coitus.37 Negative mood has also been associated with sexual interest but, in contrast to positive mood, was correlated with a decreased likelihood of coitus.37 In addition, condom use has been associated with both less positive mood36 and lower levels of negative mood.37 These findings suggest that methods of HIV/STI prevention may conflict with thoughts and behaviors that produce desirable mood. To inform HIV/STI prevention intervention research, it will be important to establish whether antecedent mood and affect influence the occurrence of sexual behavior and the use of measures to prevent HIV/STIs.
Affective changes may follow as well as precede sexual behavior. Sexual intercourse has affect-improving effects. However, little is known about these effects in adolescents and in depressed individuals. As seen in research on the effects of substance use on affect,38 sexual intercourse may be used for both NA reduction and positive affect (PA) enhancement. Depressed individuals who engage in sexual behavior may experience greater increases in PA because they are starting at lower levels of PA than their nondepressed peers. Conversely, the emotional dysregulation seen in depression may inhibit mood elevation after sexual activity or result in a negative affective experience.
Potential Mechanisms for the Association Between Affect and Sexual Behavior
There is no single theory or model of affect and behavior that has been applied specifically to the study of sexual risk behavior. However, research on behavior in general suggests that affect may be associated with sexual behavior through cognitive, personality, behavioral, and interpersonal mechanisms. Several studies have demonstrated that affect is linked to cognition. PA facilitates cognitive flexibility, innovation, and problem solving.39 In safe situations, PA results in exploration and trying new things (risk-prone behavior), but in risky situations, PA is associated with self-protection (risk aversion).39–41 Other research suggests that PA signals that the situation is safe; thus, PA is associated with less forethought, planning, and consideration of risk.42 Whether PA produces more exploration and innovation or heuristic processing, the end result may be increased sexual risk behavior. Affect has also been linked to cognitive distortions,43 which may impair an individual's ability to make safer sex decisions. Personality factors, such as self-esteem44 and impulsiveness,44 may moderate affective influences on cognition and behavior.45
According to several theories, the desire to regulate affect drives much of the behavioral response to affect.46–48 Research in mood repair found that individuals who were induced to feel a sad mood initially recalled a negative memory but then deliberately recruited a positive memory to counteract the negative mood.49 In addition, substance use may be used for self-regulation of affect and can substantially impair risk assessment.39–41
Affect is also correlated with key concepts in psychosocial theories of health behavior. NA is related to decreased self-efficacy to perform health-promoting and illness-alleviating behaviors.50 Thus, individuals with high NA are likely to demonstrate reduced condom use self-efficacy and subsequently be less likely to use condoms.50 PA is associated with lower perceived vulnerability to adverse behavioral outcomes ("optimistic bias").50 Thus, individuals with high PA may perceive low risk of HIV and other STIs and engage in more sexual activity and use condoms less often. This research suggests that both high PA and high NA put individuals at risk for unsafe sexual behavior.
The extent to which affect is associated with behavior may also depend a great deal on the social situation in which the behavior occurs.51 Furthermore, interpersonal processes deriving from social persuasion52 or the relationship with the individuals with whom one is interacting53 or chronic feelings about them54 may influence the effects of mood on social judgments and behaviors. Mood also seems to motivate selective partner choices, with sad individuals preferring social partners with high interpersonal qualities,44 which may be associated with increased sexual behavior.
In a meta-analysis of 34 studies, Crepaz and Marks20 found limited evidence to support an association between negative affective states and HIV sexual risk behaviors. However, in their discussion of the analysis and in the accompanying editorial comment,55 important limitations were raised and directions for future research were suggested. Most studies have had cross-sectional designs and could not determine the temporal nature of the associations. In addition, global measures of affect have been used; thus, NA immediately before sexual thoughts and behaviors as well as indirect effects of NA (eg, through substance use) could not be assessed. In their comment, Kalichman and Weinhardt55 urged the study of NA and sexual behavior at the event level, using methods such as momentary assessments.
MEASURING AFFECT AND SEXUAL BEHAVIOR
The optimal method of measuring the temporal associations of affect and sexual behavior56–58 has not been determined, particularly for adolescent populations. Most studies of mental health and sexual behavior have relied on retrospective self-report using aggregate, cumulative, or composite measures, such as frequency of depressive symptoms over 2 weeks and frequency of condom use in general. This method is simple to administer, but the accuracy of retrospective self-report is not known; it is subject to "recollective interpretation,"59 the process of reconstituting experience in the broader context afforded by reflection, new perspectives, and the accumulation of other experiences, including consequences of the original experience. This recall bias may result in distorted estimates of certain behaviors and a reorganization of the experienced affect.59 Frequency self-reports also may be subject to under- or overreporting to provide socially desirable responses. Finally, retrospective reports have limited utility in relating specific feelings and thoughts to specific behavioral events.
Daily diaries have also been used to collect data about specific events, such as coitus,60,61 and there is some support for the validity of event-specific diaries with regard to certain sexual risk behaviors.62 This unobtrusive method reduces the time between the occurrence of an event and its recording, thereby limiting recall bias. Diaries also may be more accurate than retrospective methods for reports of event-specific feelings and behaviors. However, because the diaries tend to be completed at the same times each day, respondents may become habituated to the instruments, develop patterned responses, or otherwise fatigue from the repetitive nature of the task. Some individuals may have difficulty in consistently recording complete data over a period of time or hoard their responses, threatening the reliability of the data. Daily diaries have been used to collect data on mood and substance use in young women63 and on sexual behavior and substance use in adolescents60 and adults.57,63 In 1 study of mood states and substance use, young women completed 90% of daily diary questionnaires over 3 menstrual cycles.63
MS methods, such as the Experience-Sampling Method (ESM) and Ecological Momentary Assessment (EMA), were developed to add ecological validity to the study of the feelings, thoughts, and behaviors of daily life.64–66 With ESM, individuals are asked to log self-reports in response to random signals (sent via wristwatch, beeper, or handheld computer) during the waking hours of a normal week. Completed in "real time," these random reports are not affected by recall bias. EMA extends ESM by including self-monitoring as well as random reports67,68; in addition to responding to random signals, participants record a response as soon as possible after they identify a target behavior. Self-monitoring enriches the random data of ESM by adding more detailed information about events of interest.
MS methods offer several unique features to collecting data on affect and sexual behavior. First, MS involves assessment of a random sample of momentary states that permits examination and description of the topology of daily experience. With MS data, one can construct composite measures, such as average affect, without losing the value associated with understanding the variance and extremes of affect. These data may be used to create measures of affect variability and affect reactivity both within and across participants. Second, MS samples a large number of repeated observations to create a very rich and detailed database. Third, MS collects data on environmental factors, such as location and companionship, that provide a context for the momentary states. Fourth, the electronic technology used in MS data collection uses the same device for recording both responses to random signals and reports prompted by a target event. Fifth, data may be downloaded directly from the electronic devices, eliminating errors acquired during data entry. Sixth, MS overcomes many of the problems associated with other self-report methods, including errors associated with retrieval, telescoping, inference, recency, and salience.68 Extensive experience with MS has minimized many of the methods' disadvantages, including participant burden, retention, and adherence.68,69 Other limitations include small sample size, device expense, reliance on self-report, the potential for self-selection bias, and the possibility that the method influences the phenomena being measured.65,66 MS methods have been used to describe daily mood in children and adolescents,59,70,71 adolescent substance use and mood states,72 and adolescent moods and smoking behavior.69,73,74 Adolescents who have participated in studies that used MS methods have responded to 69% to 90% of signals.63,69,70,73–75
In the following study, we compared 3 methods of collecting timed data on affect, substance use, and sexual behavior in male and female adolescents: a retrospective interview and questionnaire, twice-daily diaries, and MS using a 2-way pager.
FEASIBILITY STUDY
Methods
Patients of a children's hospital-based adolescent clinic who were aged 15 to 18 years and currently sexually active (defined as sexual intercourse, on average, at least once every 2 weeks) were eligible for this study. Patients with a current mood disorder, suicidality, or mental/emotional crisis were excluded. As this was a feasibility study and includes qualitative assessment elicited from participants, it was important that participants were able to provide complete data and participate actively in discussions around the methods of data collection. To this end, we asked clinic providers to suggest motivated, interested patients. Providers invited interested adolescents to meet with a research assistant to discuss the study protocol and provide informed consent.
In a private room in the clinic, each participant was asked to complete a baseline assessment battery using audio computer-assisted self-interview that included demographic information, sexual history, substance use history, history of mood disorder, the Positive Affect-Negative Affect Schedule (PANAS),76 and the BDI.77 The PANAS is based on a 2-factor model of affect, which posits that PA and NA are independent dimensions. Respondents were asked to rate on a 5-point Likert-type scale the extent to which they feel each of 10 positive and 10 negative affective states. Responses are summed separately for PA and NA, each with a possible score of 10 to 50. The PANAS has been shown to be valid and reliable in adolescent samples.76,78 Furthermore, the time frame can be varied, so in the baseline assessment and recall interviews, affect was measured over the previous 2 weeks. For the diary and MS reports, affect was measured for "right now."
On completion of the baseline assessment battery, the research assistant scored the BDI by the summing of the responses. Participants with scores of 16 or more (out of a possible 63) were considered to have a high level of depressive symptoms79,80 and excluded from the study. (A BDI score of 16 or greater has been shown to have a sensitivity of 100% and a specificity of 93.2% for major depressive episode in adolescents.81) Any participant who responded other than, "0, I don't have any thoughts of killing myself," to question 9 on suicidal ideation were also excluded from study participation.
Using a random-numbers list stratified by gender, each participant was randomly assigned to 1 of the 6 possible orders of the 3 data collection methods. The methods, described in detail below, included a retrospective questionnaire and interview, twice-daily diary, and MS using a 2-way pager. Items and scales included in each method are presented in Table 1. Each arm of the study took 2 weeks to complete, with at least 1 week of no data collection between each arm. Total time of study involvement therefore was a minimum of 8 weeks.
Twice-Daily Diary
Each participant was given 14 stamped, addressed diary cards that could be folded over, sealed, and mailed to the principal investigator each day of the 2-week study period. Each card included the PANAS affect scale, to be completed on awakening and before going to bed each day. The card included a space for logging episodes of sexual intercourse, including the time that the sexual activity began, the relationship with the sexual partner (new, occasional, regular, main), and whether a condom was used. The card also included a space for logging alcohol, marijuana, and other substance use, including the time each episode of use began and ended and the type and quantity of substance(s) used. To protect the participant's confidentiality, the diaries did not include the participant's name, and the responses were coded such that their meaning would not be obvious if someone other than the participant or study personnel read the diary card. Each participant was provided with a "cheat sheet," a small card that could be kept in the participant's wallet, explaining the code for the diaries. Completion of the diaries took no more than 10 minutes each day.
MS
Participants were provided with an electronic pager and paged each day for 2 weeks at random times within each of 4 blocks of time: 9:00 AM to 12:00 PM, 12:01 PM to 3:00 PM, 3:01 PM to 6:00 PM, and 6:01 PM to 9:00 PM, with a fifth signal between 9:01 PM and 12:00 AM on Friday and Saturday nights. Paging was suspended from 9:00 PM on weekday nights, 12:00 AM on weekend nights, until 9:00 AM the following morning to avoid sending signals during hours when participants would be most likely to be asleep. At the signal, the participant recorded his or her responses to the PANAS scale and to contextual questions on thoughts, location, companionship, and activity (Table 1). A card on the back of the pager contained the questions, which were the same as for the diary except that the interval for reporting the risk behaviors was "since last page" (Appendix). Participants were instructed to respond immediately to a page but could ignore the page if it occurred during an incompatible activity (eg, taking a test). To protect their own confidentiality, the adolescents controlled whether they carried the pager and whether and when they responded to the signals. They were provided with a business card with the research assistant's contact information, which they could present to curious adults who may be concerned about the adolescents' carrying a pager. Completion of this method took 22 min/day.
Retrospective Report
Participants were asked to recollect their affect, substance use, and sexual behavior for a 2-week period using a timeline followback calendar recall method. This reliable method has been used extensively in substance use research82 and specifically with adolescents for recall of both substance use83 and sexual behavior.83,84 The timeline followback calendar uses recall-enhancing techniques to provide a detailed assessment of substance use and coitus at the event level.
Because it would have been difficult to recall the specifics of affect on any given day, participants were asked to give a general impression of their affect for each of the 14 days, noting in particular days with the best and the worst affective states. The interview took 30 minutes (more or less, depending on the frequency of substance use and sexual behavior). Participants were also asked to complete a brief questionnaire with general frequency measures for affect, substance use, and sexual behavior for the previous 2 weeks. The questionnaire took 10 minutes to complete. Participants completed the retrospective interview and questionnaire at the end of each of the three 2-week periods.
Exit Interview and Study Termination
On completion of all 3 methods, participants were invited to share their impressions about the methods and suggest which method(s) they believed best represented their actual experience and how they felt about the time commitment and tasks involved, issues of confidentiality, and other aspects of the research experience that they wished to share with the investigators. Because we were interested in comparing the 3 methods for each participant, we discontinued an adolescent's participation if he or she was unable to complete all 3 retrospective self-reports and at least 50% of both the diaries and the MS reports. We offered participants remuneration commensurate with the number of reports, diaries, and interviews that they completed, up to a maximum of $61.50 for 6 weeks of active study participation. The hospital institutional review board approved the study protocol.
Data Analysis
Data were dually entered into a Microsoft Excel database, and the analyses were conducted using SAS version 9.0. The outcomes of interest were the rates of completion of the study instruments, adherence to each method's protocol, the adolescent participants' tolerance of and preferences for the methods, and agreement among the methods in mean scores for PA and NA and in the report of days on which substance use and sexual activity occurred. To understand better the advantages of MS, we also explored associations between affect and contextual questions, including location, companionship, and thoughts. For these analyses, we used linear mixed-effects models to account for potential correlations between repeated measurements taken on the same participant. Pairwise comparisons then were conducted to explore patterns of differences. Because these analyses were exploratory, we did not adjust for multiple comparisons. We also evaluated whether and how affect changed from before to after coital events, again by linear mixed-effects models. Because substance use was infrequent in our sample (overall, only 3 participants reported 2 marijuana and 4 alcohol use events), no analyses were performed to evaluate the associations between affect and substance use.
Results
Sample Characteristics
Six girls and 4 boys completed the study protocol. Mean (±SD) age was 16.9 (±1.0) years, and 70% of the sample self-identified as black race. Mean (±SD) age at first sexual intercourse was 14.4 (±1.8) years. Participants reported a median (range) of 3.5 (1–16) sexual partners in their lifetime and 1 (1–3) sexual partner in the previous 3 months. The adolescents estimated that they had sexual intercourse a median (range) of 3 (1–10) times in the previous 2 weeks. Only 4 youths reported any alcohol or marijuana use, and none reported smoking cigarettes in the previous 30 days. Three participants reported feeling sad or hopeless for at least 2 weeks in the previous 12 months. Median (range) BDI score was 5 (1–13).
Completion Rates
Nineteen adolescents were initially enrolled, with 6 of the nine dropouts occurring during the MS arm. Noncompleters were significantly more likely than completers to report on baseline marijuana use in the previous 30 days. By study design, 30 of 30 of the retrospective interviews were completed (100%, 3 per participant). Of the 280 diaries expected, 254 were returned and complete (91%; mean: 25.4 per participants) and three fourths (74%) of the MS reports were completed (442 of 600; mean: 44.2 per participant; completion rate for diaries vs MS, P = .065). The median time to respond to an MS signal was 28 minutes (range: 0–288 minutes).
Tolerance of and Preferences for the Methods
Most adolescents (6 of 10) preferred the MS method because they were asked to "react to a specific moment" and "being paged is fun" and "cool," with the remaining stating a preference for the diaries because they "capture the whole day instead of a specific moment," were "easy," and "didn't interfere with daily activities." Three participants preferred diaries the least, citing that they were "boring," "a pain to mail," and "got in the way of school," and 2 did not like MS, finding it "annoying" to get paged, especially "at awkward times," and that they had to respond to "too many questions." One participant noted, "The pager going off can affect your mood because it might agitate you." Girls (5 of 6) tended to prefer MS, whereas boys (3 of 4) tended to prefer the diaries.
Correlations Among the Methods Between Average Daily Affect Scores and Reports of Substance Use and Sexual Activity
Mean daily PA and NA scores that were measured using twice-daily diaries or MS were moderately to highly correlated with scores that were obtained using retrospective report (Table 2). Mean affect scores were consistently higher on retrospective report than with the other 2 methods. There was 100% agreement on report of sexual activity between diary or MS and recall and on report of substance use between MS and recall. One adolescent did not recall 1 episode of substance use that was reported by MS.
Correlations of Momentary Affect With Contextual Variables and With Coitus
Participants were most often in nonschool public places when paged (42% of signals), followed by home (34%) then school (24%; Table 3). PA and NA both differed significantly by location when paged (P = .006 and 0.01, respectively). Pairwise comparisons revealed that PA when in public was higher than PA when at home or in school (P = .014) and NA when in school was higher than NA when in public or at home. Participants were most often with other people when paged (65% of reports). NA differed by companionship (P = .03); pairwise comparisons showed that NA in class was higher than NA with friends. There were no differences in PA according to companionship (P = .72). PA differed by thoughts at the time of the signal (P = .03). Although the adolescents reported thinking about sex on only 5% of reports, their PA scores were highest at those times compared with times when they were thinking about anything else (pairwise comparison P = .047). Eight participants reported 20 coital events (mean: 2.5 coital events per participant; range: 1–6 events). Participants provided data on partner type and condom use for 18 and 16 coital events, respectively; 94% of the coital events (17 of 18) occurred with a main sexual partner, and only 38% (6 of 16) were condom protected. From before to after coitus, PA increased and NA decreased, but the differences were not significant (change in PA: ±2.87 [SE: 1.96; P = .19]; change in NA: –0.39 [SE 1.39; P = .78]).
CONCLUSIONS
A substantial body of theoretical and experimental literature supports the notion that affect or mood is related to risk behavior; far fewer empirical studies have been conducted to examine directly the associations, particularly in adolescents. What studies have been done suggest that both positive and negative mood are correlated with sexual interest,36,37 the occurrence of coitus,36,37 safer sex behaviors such as condom use,36,37 and degree of substance use.38,63
For addressing the large gap between theory and research in affect and risk behavior, methods to measure these constructs need to be refined. MS methods offer several advantages over traditional retrospective self-report and diary methods, including the ability to assess the variability in affect in daily life and to relate affect to contextual factors and specific behavioral and environmental events while avoiding many of the biases of other methods.
In a small feasibility study of methods to measure affect and sexual behavior in adolescents, we found that retrospective self-report, twice-daily diary, and MS methods were comparable for collecting data on the occurrence of coitus and on summary affective states, although data on affect from retrospective self-report may be biased toward the extremes. In contrast to the other 2 methods, MS was better able to detect within-day fluctuations in affect as well as situational factors that may influence affect. Importantly, adolescents in this study preferred MS to diaries and retrospective self-report. Findings from the MS data demonstrate that affect differs by contextual factors such as location, companionship, and thoughts. By collecting temporal and contextual data, MS permits exploration of possible causal mechanisms for the associations among affect and sexual activity in adolescents. The finding that adolescents experience their most PA when thinking about sex suggests that further research into the association between affect and sexual behavior is warranted, particularly with regard to sexual intercourse for the purposes of affect management.
This study has several limitations. The number of adolescents who participated in this intensive protocol was small. The relatively infrequent nature of substance use and sexual events may limit the utility of MS methods in studies of these behaviors. Future studies will need to consider longer or repeated periods of data collection and include assessments prompted by the target event (as has been done in EMA studies67,68). The MS method was associated with decreased compliance compared with diaries, although the completion rate was not significantly different and was similar to that reported in previous studies using MS.63,69,70,73–75 This problem may be remedied in future studies by shortening the MS report and considering use of other electronic devices, such as palm computers.69,74
Despite its limitations, this study adds to the relatively sparse literature exploring the use of MS in adolescent samples and uniquely examines the feasibility of this method to collect data on sexual behavior. If the methodologic challenges can be addressed, then additional research with larger samples will help to determine how MS may complement existing methods of collecting affect, substance use, and sexual behavior data in adolescents. Specifically, these studies should examine how MS can increase our understanding of the role of contextual factors, offer a unique look at adolescents' affect and other salient experience before and after occurrence of the risk event, and permit examination of the influence of affect variability on adolescent risk behavior.
Findings from this line of research may influence HIV/STI prevention interventions in several ways, such as consideration of affect assessment and management before or as part of an intervention, affect self-assessment and increasing an individual's understanding of the relation of affect to risk behavior within an intervention, and identification of risk behavior as a method of affect self-management and the development of alternative response strategies as part of an intervention. Increasing our understanding of the associations of affect and sexual risk behavior among adolescents will be important to enhancing behavioral interventions for this at-risk population and may be particularly salient for interventions to reduce risk among depressed youths.
APPENDIX
Momentary Sampling Report for the Measurement of Affect and Sexual Behavior in Adolescents
As you were beeped...
1. What were you thinking about (a) Family (b) Food (c) Friends (d) School (e) Self (f) Sex (g) Society, Religion, Politics (h) Sports (i) Television (j) Time (k) Other
2. Where were you (a) At Home (b) In School (c) In Public—nonschool
3. What was the MAIN thing you were doing
4. WHY were you doing this particular activity (a) I had to (b) I wanted to (c) I had nothing else to do
Use the numbers to express how you feel right now: 1 = Not at all 2 = A little 3 = Moderately 4 = Quite a bit 5 = Extremely
5. Interested
6. Distressed
7. Excited
8. Upset
9. Strong
10. Guilty
11. Scared
12. Hostile
13. Enthusiastic
14. Proud
15. Irritable
16. Alert
17. Ashamed
18. Inspired
19. Nervous
20. Determined
21. Attentive
22. Jittery
23. Active
24. Afraid
25. When you were beeped, whom were you with (type all the lettered responses that apply.) (a) Alone (b) Friends (after this letter, type how many female [F] and how many male [M]) (c) In class (d) Relative(s) (e) Strangers
Questions 26–30. Since you were last beeped... (Y for yes or N for no)
26. Has anything happened or have you done anything that could have affected the way you feel (a) If Yes, indicate what it was.
27. Have you had sexual intercourse (a) If Yes, indicate the type of sexual partner MP = Main Partner, OP = Other Partner, NP = New Partner (b) If Yes, did you use a condom (Y = Yes, N = No)
28. Have you used alcohol (a) If Yes, what kind B = Beer, W = Wine, L = Liquor (b) If Yes, how much (eg, 2 cans = "2cn") cn = Can, btl = Bottle, gl = Glass, oz = Ounce/Shot
29. Have you used marijuana (a) If Yes, indicate the way you smoked and how many (eg, 2 cigarettes = "2c") c = Cigarette, bl = Blunt, p = Pipe, bg = Bong
30. Have you used any other drugs (a) If Yes, name the drug.
31. Have you had your period [Asked only of female participants]
ACKNOWLEDGMENTS
The study was funded in part by the Charles A. Janeway Award in Child Health Research, Child Health Research Center, Children's Hospital Boston (to L.A.S.); and 5 K23 MH01845, National Institute of Mental Health, National Institutes of Health (to L.A.S.).
We are grateful to J. Dennis Fortenberry, MD, MS; Reed Larson, PhD; and Arthur A. Stone, PhD, for support and guidance; Elizabeth Mariano and Colin Hynes for efforts on the day-to-day research activities, and Laura Hacker for editorial assistance.
FOOTNOTES
Accepted Nov 23, 2004.
No conflict of interest declared.
REFERENCES
Kirby D, Short L, Collins J, et al. School-based programs to reduce sexual risk behaviors: a review of effectiveness. Public Health Rep. 1994;109 :339 –360
DiCenso A, Guyatt G, Willan A, et al. Interventions to reduce unintended pregnancies among adolescents: systematic review of randomised controlled trials. BMJ. 2002;324 :1426
Kirby D, DiCenso A, Guyatt G, et al. Effective approaches to reducing adolescent unprotected sex, pregnancy, and childbearing. J Sex Res. 2002;39 :51 –57
Stiffman AR, Dore P, Earls F, Cunningham R. The influence of mental health problems on AIDS-related risk behaviors in young adults. J Nerv Ment Dis. 1992;180 :314 –320
Joffe A, Radius SM. Self-efficacy and intent to use condoms among entering college freshmen. J Adolesc Health. 1993;14 :262 –268
Shrier LA, Emans SJ, Woods ER, DuRant RH. The association of sexual risk behaviors and problem drug behaviors in high school students. J Adolesc Health. 1997;20 :377 –383
Seal A, Minichiello V, Omodei M. Young women's sexual risk taking behaviour: re-visiting the influences of sexual self-efficacy and sexual self-esteem. Int J STD AIDS. 1997;8 :159 –165
Pao M, Lyon M, D'Angelo LJ, Schuman WB, Tipnis T, Mrazek DA. Psychiatric diagnoses in adolescents seropositive for the human immunodeficiency virus. Arch Pediatr Adolesc Med. 2000;154 :240 –244
American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 4th ed. Washington, DC: American Psychiatric Association; 1994
Centers for Disease Control and Prevention. Youth risk behavior surveillance—United States, 2001. MMWR Recomm Rep. 2002;51(SS4) :1 –64
Pine DS, Cohen E, Cohen P, Brook J. Adolescent depressive symptoms as predictors of adult depression: moodiness or mood disorder Am J Psychiatry. 1999;156 :133 –135
Garrison CZ, Waller JL, Cuffe SP, McKeown RE, Addy CL, Jackson KL. Incidence of major depressive disorder and dysthymia in young adolescents. J Am Acad Child Adolesc Psychiatry. 1997;36 :458 –465
Lewinsohn PM, Hops H, Roberts RE, Seeley JR, Andrews JA. Adolescent psychopathology: I. Prevalence and incidence of depression and other DSM-III-R disorders in high school students. J Abnorm Psychol. 1993;102 :133 –144
Shrier L, Braslins P, Christiansen D, Markowitz LE, Fortenberry JD, Orr D. Gender Differences in the Associations of Mental Health, Sexual Risk Behavior, and Sexually Transmitted Infection Among Adolescents and Young Adults. Paper presented at the International Society for Sexually Transmitted Disease Research; July 30, 2003; Ottawa, Ontario, Canada
Valois RF, Bryant ES, Rivard JC, Hinkle KT. Sexual risk-taking behaviors among adolescents with severe emotional disturbance. J Child Fam Stud. 1997;6 :409 –419
Smith MD. HIV risk in adolescents with severe mental illness: literature review. J Adolesc Health. 2001;29 :320 –329
Brown L, Danovsky M, Lourie K, DiClemente R. Adolescents with psychiatric disorders and the risk of HIV. J Am Acad Child Adolesc Psychiatry. 1997;36 :1609 –1615
Baker DG, Mossman D. Potential HIV exposure in psychiatrically hospitalized adolescent girls. Am J Psychiatry. 1991;148 :528 –530
Erbelding EJ, Hummel B, Hogan T, Zenilman J. High rates of depressive symptoms in STD clinic patients. Sex Transm Dis. 2001;28 :281 –284
Crepaz N, Marks G. Are negative affective states associated with HIV sexual risk behaviors A meta-analytic review. Health Psychol. 2001;20 :291 –299
Kalichman SC, Kelly JA, Johnson JR, Bulto M. Factors associated with risk for HIV infection among chronic mentally ill adults. Am J Psychiatry. 1994;151 :221 –227
Leith KP, Baumeister RF. Why do bad moods increase self-defeating behavior Emotion, risk taking, and self-regulation. J Pers Soc Psychol. 1996;71 :1250 –1267
Morris WN, Reilly NP. Toward the self-regulation of mood: Theory and research. Motivation Emotion 1987;11 :215 –249
Rao U, Daley SE, Hammen C. Relationship between depression and substance use disorders in adolescent women during the transition to adulthood. J Am Acad Child Adolesc Psychiatry. 2000;39 :215 –222
Mueser KT, Drake RE, Wallach MA. Dual diagnosis: a review of etiological theories. Addict Behav. 1998;23 :717 –734
Tapert SF, Aarons GA, Sedlar GR, Brown SA. Adolescent substance use and sexual risk-taking behavior. J Adolesc Health. 2001;28 :181 –189
Graves KL, Leigh BC. The relationship of substance use to sexual activity among young adults in the United States. Fam Plan Perspect. 1995;27 :18 –22, 33
Castillo-Mezzich A, Tarter R, Giancola P, Lu S, Kirisci L, Parks S. Substance use and risky sexual behavior in female adolescents. Drug Alcohol Depend. 1997;44 :157 –166
Strunin L, Hingson R. Alcohol, drugs, and adolescent sexual behavior. Int J Addict. 1992;27 :129 –146
Shrier LA, Harris SK, Sternberg M, Beardslee WR. Associations of depression, self-esteem, and substance use with sexual risk among adolescents. Prev Med. 2001;33 :179 –189
Shrier L, Harris S, Beardslee W. Temporal associations between depressive symptoms and self-reported sexually transmitted disease among adolescents. Arch Pediatr Adolesc Med. 2002;156 :599 –606
Lewinsohn PM, Petit JW, Joiner TE Jr, Seeley JR. The symptomatic expression of major depressive disorder in adolescents and young adults. J Abnorm Psychol. 2003;112 :244 –252
Bancroft J, Janssen E, Strong D, Carnes L, Vukadinovic Z, Long J. The relation between mood and sexuality in heterosexual men. Arch Sex Behav. 2003;32 :217 –230
Bancroft J, Janssen E, Strong D, Vukadinovic Z. The relation between mood and sexuality in gay men. Arch Sex Behav. 2003;32 :231 –242
Skiba A, Fortenberry J, Blythe M. Daily mood and behavioral correlates of coitus and condom use. J Adolesc Health. 1997;20 :145
Fortenberry J, Temkit MH, Tu W, Katz B, Orr D. Mood, mood change, sexual activity and condom use among adolescent women. J Adolesc Health. 2003;32 :157
Wills TA, Sandy JM, Shinar O, Yaeger A. Contributions of positive and negative affect to adolescent substance use: test of a bidimensional model in a longitudinal study. Psychol Addict Behav. 1999;13 :327 –338
Isen AM. Missing in action in the AIM: positive affect's facilitation of cognitive flexibility, innovation, and problem solving. Psychol Inq. 2002;13 :57 –65
Isen AM, Patrick R. The effect of positive feelings on risk taking: when the chips are down. Organ Behav Hum Decis Process. 1983;31 :194 –202
Isen AM, Geva N. The influence of positive affect on acceptable level of risk: the person with a large canoe has a large worry. Organ Behav Hum Decis Process. 1987;39 :145 –154
Clore GL, Gasper K, Garvin E. Affect as information. In: Forgas JP, ed. Handbook of Affect and Social Cognition. Mahwah, NJ: Lawrence Erlbaum Associates; 2001:121 –144
Forgas JP, Bower GH, Krantz SE. The influence of mood on perceptions of social interactions. J Exp Soc Psychol. 1984;20 :497 –513
Forgas JP. Affective influences on partner choice: role of mood in social decisions. J Pers Soc Psychol. 1991;61 :708 –720
Rusting C. Personality as a moderator of affective influences on cognition. In: Forgas JP, ed. Handbook of Affect and Social Cognition. Mahwah, NJ: Lawrence Erlbaum Associates; 2001:371 –391
Clark MS, Isen AM. Toward understanding the relationship between feeling states and social behavior. In: Hastorf AH, Isen AM, eds. Cognitive Social Psychology. New York, NY: Elsevier; 1982:73 –108
Erber M, Erber R. The role of motivated social cognition in the regulation of affective states. In: Forgas JP, ed. Handbook of Affect and Social Cognition. Mahwah, NJ: Lawrence Erlbaum Associates; 2002:275 –290
Wegener DT, Petty RE. Mood management across affective states: the hedonic contingency hypothesis. J Pers Soc Psychol. 1994;66 :1034 –1048
Josephson BR, Singer JA, Salovey P. Mood regulation and memory: repairing sad moods with happy memories. Cogn Emotion. 1996;10 :437 –444
Salovey P, Birnbaum D. Influence of mood on health-relevant cognitions. J Pers Soc Psychol. 1989;57 :539 –551
Forgas JP. Mood and judgment: the affect infusion model (AIM). Psychol Bull. 1995;117 :39 –66
Haidt J. "Dialogue between my head and my heart": affective influences on moral judgment. Psychol Inq. 2002;13 :54 –56
Clark MS. We should focus on interpersonal as well as intrapersonal processes in our search for how affect influences judgments and behavior. Psychol Inq. 2002;13 :32 –37
Perrott DA, Bodenhausen GV. The way you make me feel: integral affective influences on interpersonal behavior. Psychol Inq. 2002;13 :84 –86
Kalichman SC, Weinhardt L. Negative affect and sexual risk behavior: comment on Crepaz and Marks (2001). Health Psychol. 2001;20 :300 –301
Coxon AP. Parallel accounts Discrepancies between self-report (diary) and recall (questionnaire) measures of the same sexual behaviour. AIDS Care. 1999;11 :221 –234
Leigh BC. Alcohol consumption and sexual activity as reported with a diary technique. J Abnorm Psychol. 1993;102 :490 –493
Ramjee G, Weber AE, Morar NS. Recording sexual behavior: comparison of recall questionnaires with a coital diary. Sex Transm Dis. 1999;26 :374 –380
Freeman M, Csikszentmihalyi M, Larson R. Adolescence and its recollection: toward an interpretive model of development. Merrill Palmer Q. 1986;32 :167 –185
Fortenberry JD, Orr DP, Katz BP, Brizendine EJ, Blythe MJ. Sex under the influence. A diary self-report study of substance use and sexual behavior among adolescent women. Sex Transm Dis. 1997;24 :313 –319
Fortenberry JD, Orr DP, Zimet GD, Blythe MJ. Weekly and seasonal variation in sexual behaviors among adolescent women with sexually transmitted diseases. J Adolesc Health. 1997;20 :420 –425
Fortenberry J, Cecil H, Zimet G, Orr D. Concordance between self-report questionnaires and coital diaries for sexual behaviors of adolescent women with sexually transmitted infections. In: Bancroft J, ed. Research Sexual Behavior: Methodological Issues. Bloomington, IN: Indiana University Press; 1997:461
Lex BW, Griffin ML, Mello NK, Mendelson JH. Alcohol, marijuana, and mood states in young women. Int J Addict. 1989;24 :405 –424
Csikszentmihalyi M, Larson R, Prescott S. The ecology of adolescent activity and experience. J Youth Adolesc. 1977;6 :281 –294
Larson R, Csikszentmihalyi M. The experience sampling method. In: Reis H, ed. Naturalistic Approaches to Studying Social Interactions. Vol 15. San Francisco, CA: Jossey-Bass; 1983:41–56
Csikszentmihalyi M, Larson R. Validity and reliability of the experience-sampling method. J Nerv Ment Dis. 1987;175 :526 –536
Hufford MR, Shields AL, Shiffman S, Paty J, Balabanis M. Reactivity to ecological momentary assessment: an example using undergraduate problem drinkers. Psychol Addict Behav. 2002;16 :205 –211
Shiffman S. Real-Time Self-Report of Momentary States in the Natural Environment: Computerized Ecological Momentary Assessment. Mahwah, NJ: Lawrence Erlbaum; 2000
Whalen CK, Jamner LD, Henker B, Delfino RJ. Smoking and moods in adolescents with depressive and aggressive dispositions: evidence from surveys and electronic diaries. Health Psychol. 2001;20 :99 –111
Larson R, Csikszentmihalyi M, Graef R. Mood variability and the psychosocial adjustment of adolescents. J Youth Adolesc. 1980;9 :469 –490
Larson R. Beeping children and adolescents: a method for studying time use and daily experience. J Youth Adolesc. 1989;18 :511 –530
Larson R, Csikszentmihalyi M, Freeman M. Alcohol and marijuana use in adolescents' daily lives: a random sample of experiences. Int J Addict. 1984;19 :367 –381
Henker B, Whalen CK, Jamner LD, Delfino RJ. Anxiety, affect, and activity in teenagers: monitoring daily life with electronic diaries. J Am Acad Child Adolesc Psychiatry. 2002;41 :660 –670
Whalen CK, Jamner LD, Henker B, Delfino RJ, Lozano JM. The ADHD spectrum and everyday life: experience sampling of adolescent moods, activities, smoking, and drinking. Child Dev. 2002;73 :209 –227
Larson R, Lampman-Petraitis C. Daily emotional states as reported by children and adolescents. Child Dev. 1989;60 :1250 –1260
Watson D, Tellegen A. Toward a consensual structure of mood. Psychol Bull. 1985;98 :219 –235
Beck A, Ward C, Mendelson M, et al. An inventory for measuring depression. Arch Gen Psychiatry. 1961;4 :561 –571
Watson D, Clark LA, Tellegen A. Development and validation of brief measures of positive and negative affect: the PANAS scales. J Pers Soc Psychol. 1988;54 :1063 –1070
Kendall P, Hollon S, Beck A, Hammen C, Ingram R. Issues and recommendations regarding use of the Beck Depression Inventory. Cognit Ther Res. 1987;11 :289 –299
Gotlib I, Cane D. Self-report assessment of depression and anxiety. In: Kendall PC, Watson D, eds. Anxiety and Depression. New York, NY: Academic Press; 1989:131–169
Barrera M, Garrison-Jones CV. Properties of the Beck Depression Inventory as a screening instrument for adolescent depression. J Abnorm Child Psychol. 1988;16 :263 –273
Carney MA, Tennen H, Affleck G, Del Boca FK, Kranzler HR. Levels and patterns of alcohol consumption using timeline follow-back, daily diaries and real-time "electronic interviews." J Stud Alcohol. 1998;59 :447 –454
Stein MD, Anderson B, Charuvastra A, Friedmann PD. Alcohol use and sexual risk taking among hazardously drinking drug injectors who attend needle exchange. Alcohol Clin Exp Res. 2001;25 :1487 –1493
Weinhardt LS, Carey MP, Maisto SA, Carey KB, Cohen MM, Wickramasinghe SM. Reliability of the timeline follow-back sexual behavior interview. Ann Behav Med. 1998;20 :25 –30(Lydia A. Shrier, MD, MPH,)
Department of Psychiatry, Children's Hospital Boston and Harvard Medical School, Boston, Massachusetts
ABSTRACT
Objective. Assessment of mental health is important in understanding sexual risk behavior in adolescents, yet few studies have examined how affect is directly related to sexual behavior. Momentary sampling (MS) methods permit real-time assessment of affect in relation to specific events and embed the collected data in the context of the respondent's moment-to-moment life. The objectives of this study were to review the literature on affect and sexual behavior and to compare the feasibility and acceptability of MS with diaries and retrospective self-report as a means of collecting temporally relevant data on affect and sexual behavior in adolescents.
Methods. Sexually active, nondepressed adolescent outpatients who were aged 15 to 18 years were randomly assigned to a schedule of the 3 methods of data collection for 2 weeks each. All participants completed a retrospective self-report by interview at the end of each 2-week period. In the diary arm, participants completed twice-daily paper-and-pencil diary cards, which were returned by mail. In the MS arm, participants used 2-way pagers to respond to several random pages per day. Primary outcomes included rates of completion (diaries vs MS reports) and the participants' tolerance of and preferences for the methods. A secondary outcome was the agreement in means for positive and negative affect and in report of days on which substance use and sexual activity occurred. Associations of affect with contextual factors and with sexual activity were also explored in the MS arm.
Results. Ten youths completed 30 of 30 retrospective self-reports (100%, 3 per participant, by design), 254 of 280 diaries (91%; mean: 25.4 per participant), and 442 of 600 MS reports (74%; mean: 44.2 per participant). Most participants preferred the MS method to the diaries or retrospective self-report. Affect scores and reports of sexual activity and substance use were correlated among the methods. Measured with MS, affect was found to differ by location, companionship, and thoughts when paged; notably, positive affect was highest when participants reported thoughts about sex. There was no difference in affect before versus after coitus.
Conclusions. The results of this study suggest that MS in adolescents is feasible and preferred and provides contextual, temporally relevant, event-level data on affect and sexual activity that are not readily measured with traditional methods. Future research using MS methods will be important in increasing our understanding of the link between affect and sexual behavior and inform the development of improved risk reduction interventions for adolescents.
Key Words: affect sexual behavior substance use momentary sampling diary retrospective self-report
Abbreviations: STI, sexually transmitted infection MS, momentary sampling BDI, Beck Depression Inventory NA negative affect PA, positive affect ESM, Experience-Sampling Method EMA, Ecological Momentary Assessment PANAS, Positive Affect-Negative Affect Schedule
Preventive interventions for sexually active youths have had only modest success in reducing their high risk for HIV/sexually transmitted infection (STI) and unintended pregnancy,1–3 suggesting that new approaches to sexual risk reduction must be considered. Interventions that target specific groups of at-risk youths may be particularly promising. Although affect, depression, and other measures of mental health have been shown to affect sexual behavior,4–8 the nature of the associations has not been fully explored and few interventions have been directed toward depressed youths. In particular, few studies have attempted to capture affect, a highly variable measure of mental state, or to establish the temporal and causal associations of affective states and sexual behaviors. The measurement of affect and sexual behavior presents several methodologic challenges. In this article, we present a review of the research and theories supporting the link between affect and sexual behavior and the methods that have been used to measure these constructs, including retrospective self-report, diaries, and momentary sampling (MS) methods. We then present the results of a feasibility study that directly compares MS with the other 2 methods in a sample of sexually active adolescents.
LINK BETWEEN AFFECT AND SEXUAL RISK BEHAVIOR: REVIEW OF THE LITERATURE
Terminology
The literature is varied in the use of the terms "mood," "affect," and "emotion." Affect and mood both refer to subjectively experienced feeling states (emotions).9 Affect is a momentary emotion that is a variable, fluctuating phenomenon, in contrast to the more pervasive, sustained condition of mood. However, in the review of the literature below, the terms used are those preferred by the investigators for the research cited. Affect and mood are clearly related, and the data from research on mood can inform the study of affect and sexual risk.
Clinical Research on Affect, Mood, and Sexual Behavior
Data from clinical studies support a link between affect and sexual behavior. Depressive symptoms10,11 and mood disorders12,13 are prevalent in adolescents, particularly among those who are at risk for HIV/STI.14–18 In a study of 674 high-risk adolescents and young adults, 11% of young women and 7% of young men reported frequent, severe depressive symptoms (Beck Depression Inventory [BDI] score >20).14 Among 125 clients who attended an STI clinic, 39.2% reported symptoms consistent with probable depression on the General Health Questionnaire.19 Compared with mentally healthy youths, young people with severe mental illness engage in increased sexual risk behavior, including early onset of sexual activity, more sexual partners, and less condom use.15–18 Adolescents with mental illness may engage in high-risk sexual behaviors as a result of poor impulse control, impaired judgment and perception of reality, decreased motivation for self-care,20 poor decision making, lack of knowledge or understanding of the risks, and dysfunctional social interactions.16,17,21 Self-regulatory processes to repair negative affect (NA) may be disrupted22 or self-gratifying in unhealthy ways (eg, substance use, unprotected sexual intercourse).23 Psychiatric distress is also associated with substance use,24–26 which is correlated with increased sexual risk behavior and STIs.6,27–30 Depressive symptoms have been associated with STI in many studies14,15,31,32 but not all.19
Prolonged depressed mood is the hallmark of depression; 90% to 95% of adolescents and young adults with major depressive disorder report depressed mood for at least 2 weeks (the rest meet the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition)9 criteria for major depressive disorder with the symptom of anhedonia).33 However, little is known about the role of mood and affect in sexual risk behavior among depressed adolescents. Among adult men with a propensity for depression, depressed mood was associated with increased sexual interest or response.34 In-depth interviews of gay and heterosexual men revealed that men who experienced increased sexual interest or activity when depressed reported wishing to use sex for mood regulation, to connect with someone and feel validated and/or to directly improve one's mood.34,35 A minority of gay men reported that they were more likely to take sexual risks when depressed because they did not care about the consequences.35
In a few clinical studies, moods in normal adolescents have been associated with sexual behaviors.36,37 Positive mood has been correlated with sexual interest37 and the occurrence of coitus.37 Negative mood has also been associated with sexual interest but, in contrast to positive mood, was correlated with a decreased likelihood of coitus.37 In addition, condom use has been associated with both less positive mood36 and lower levels of negative mood.37 These findings suggest that methods of HIV/STI prevention may conflict with thoughts and behaviors that produce desirable mood. To inform HIV/STI prevention intervention research, it will be important to establish whether antecedent mood and affect influence the occurrence of sexual behavior and the use of measures to prevent HIV/STIs.
Affective changes may follow as well as precede sexual behavior. Sexual intercourse has affect-improving effects. However, little is known about these effects in adolescents and in depressed individuals. As seen in research on the effects of substance use on affect,38 sexual intercourse may be used for both NA reduction and positive affect (PA) enhancement. Depressed individuals who engage in sexual behavior may experience greater increases in PA because they are starting at lower levels of PA than their nondepressed peers. Conversely, the emotional dysregulation seen in depression may inhibit mood elevation after sexual activity or result in a negative affective experience.
Potential Mechanisms for the Association Between Affect and Sexual Behavior
There is no single theory or model of affect and behavior that has been applied specifically to the study of sexual risk behavior. However, research on behavior in general suggests that affect may be associated with sexual behavior through cognitive, personality, behavioral, and interpersonal mechanisms. Several studies have demonstrated that affect is linked to cognition. PA facilitates cognitive flexibility, innovation, and problem solving.39 In safe situations, PA results in exploration and trying new things (risk-prone behavior), but in risky situations, PA is associated with self-protection (risk aversion).39–41 Other research suggests that PA signals that the situation is safe; thus, PA is associated with less forethought, planning, and consideration of risk.42 Whether PA produces more exploration and innovation or heuristic processing, the end result may be increased sexual risk behavior. Affect has also been linked to cognitive distortions,43 which may impair an individual's ability to make safer sex decisions. Personality factors, such as self-esteem44 and impulsiveness,44 may moderate affective influences on cognition and behavior.45
According to several theories, the desire to regulate affect drives much of the behavioral response to affect.46–48 Research in mood repair found that individuals who were induced to feel a sad mood initially recalled a negative memory but then deliberately recruited a positive memory to counteract the negative mood.49 In addition, substance use may be used for self-regulation of affect and can substantially impair risk assessment.39–41
Affect is also correlated with key concepts in psychosocial theories of health behavior. NA is related to decreased self-efficacy to perform health-promoting and illness-alleviating behaviors.50 Thus, individuals with high NA are likely to demonstrate reduced condom use self-efficacy and subsequently be less likely to use condoms.50 PA is associated with lower perceived vulnerability to adverse behavioral outcomes ("optimistic bias").50 Thus, individuals with high PA may perceive low risk of HIV and other STIs and engage in more sexual activity and use condoms less often. This research suggests that both high PA and high NA put individuals at risk for unsafe sexual behavior.
The extent to which affect is associated with behavior may also depend a great deal on the social situation in which the behavior occurs.51 Furthermore, interpersonal processes deriving from social persuasion52 or the relationship with the individuals with whom one is interacting53 or chronic feelings about them54 may influence the effects of mood on social judgments and behaviors. Mood also seems to motivate selective partner choices, with sad individuals preferring social partners with high interpersonal qualities,44 which may be associated with increased sexual behavior.
In a meta-analysis of 34 studies, Crepaz and Marks20 found limited evidence to support an association between negative affective states and HIV sexual risk behaviors. However, in their discussion of the analysis and in the accompanying editorial comment,55 important limitations were raised and directions for future research were suggested. Most studies have had cross-sectional designs and could not determine the temporal nature of the associations. In addition, global measures of affect have been used; thus, NA immediately before sexual thoughts and behaviors as well as indirect effects of NA (eg, through substance use) could not be assessed. In their comment, Kalichman and Weinhardt55 urged the study of NA and sexual behavior at the event level, using methods such as momentary assessments.
MEASURING AFFECT AND SEXUAL BEHAVIOR
The optimal method of measuring the temporal associations of affect and sexual behavior56–58 has not been determined, particularly for adolescent populations. Most studies of mental health and sexual behavior have relied on retrospective self-report using aggregate, cumulative, or composite measures, such as frequency of depressive symptoms over 2 weeks and frequency of condom use in general. This method is simple to administer, but the accuracy of retrospective self-report is not known; it is subject to "recollective interpretation,"59 the process of reconstituting experience in the broader context afforded by reflection, new perspectives, and the accumulation of other experiences, including consequences of the original experience. This recall bias may result in distorted estimates of certain behaviors and a reorganization of the experienced affect.59 Frequency self-reports also may be subject to under- or overreporting to provide socially desirable responses. Finally, retrospective reports have limited utility in relating specific feelings and thoughts to specific behavioral events.
Daily diaries have also been used to collect data about specific events, such as coitus,60,61 and there is some support for the validity of event-specific diaries with regard to certain sexual risk behaviors.62 This unobtrusive method reduces the time between the occurrence of an event and its recording, thereby limiting recall bias. Diaries also may be more accurate than retrospective methods for reports of event-specific feelings and behaviors. However, because the diaries tend to be completed at the same times each day, respondents may become habituated to the instruments, develop patterned responses, or otherwise fatigue from the repetitive nature of the task. Some individuals may have difficulty in consistently recording complete data over a period of time or hoard their responses, threatening the reliability of the data. Daily diaries have been used to collect data on mood and substance use in young women63 and on sexual behavior and substance use in adolescents60 and adults.57,63 In 1 study of mood states and substance use, young women completed 90% of daily diary questionnaires over 3 menstrual cycles.63
MS methods, such as the Experience-Sampling Method (ESM) and Ecological Momentary Assessment (EMA), were developed to add ecological validity to the study of the feelings, thoughts, and behaviors of daily life.64–66 With ESM, individuals are asked to log self-reports in response to random signals (sent via wristwatch, beeper, or handheld computer) during the waking hours of a normal week. Completed in "real time," these random reports are not affected by recall bias. EMA extends ESM by including self-monitoring as well as random reports67,68; in addition to responding to random signals, participants record a response as soon as possible after they identify a target behavior. Self-monitoring enriches the random data of ESM by adding more detailed information about events of interest.
MS methods offer several unique features to collecting data on affect and sexual behavior. First, MS involves assessment of a random sample of momentary states that permits examination and description of the topology of daily experience. With MS data, one can construct composite measures, such as average affect, without losing the value associated with understanding the variance and extremes of affect. These data may be used to create measures of affect variability and affect reactivity both within and across participants. Second, MS samples a large number of repeated observations to create a very rich and detailed database. Third, MS collects data on environmental factors, such as location and companionship, that provide a context for the momentary states. Fourth, the electronic technology used in MS data collection uses the same device for recording both responses to random signals and reports prompted by a target event. Fifth, data may be downloaded directly from the electronic devices, eliminating errors acquired during data entry. Sixth, MS overcomes many of the problems associated with other self-report methods, including errors associated with retrieval, telescoping, inference, recency, and salience.68 Extensive experience with MS has minimized many of the methods' disadvantages, including participant burden, retention, and adherence.68,69 Other limitations include small sample size, device expense, reliance on self-report, the potential for self-selection bias, and the possibility that the method influences the phenomena being measured.65,66 MS methods have been used to describe daily mood in children and adolescents,59,70,71 adolescent substance use and mood states,72 and adolescent moods and smoking behavior.69,73,74 Adolescents who have participated in studies that used MS methods have responded to 69% to 90% of signals.63,69,70,73–75
In the following study, we compared 3 methods of collecting timed data on affect, substance use, and sexual behavior in male and female adolescents: a retrospective interview and questionnaire, twice-daily diaries, and MS using a 2-way pager.
FEASIBILITY STUDY
Methods
Patients of a children's hospital-based adolescent clinic who were aged 15 to 18 years and currently sexually active (defined as sexual intercourse, on average, at least once every 2 weeks) were eligible for this study. Patients with a current mood disorder, suicidality, or mental/emotional crisis were excluded. As this was a feasibility study and includes qualitative assessment elicited from participants, it was important that participants were able to provide complete data and participate actively in discussions around the methods of data collection. To this end, we asked clinic providers to suggest motivated, interested patients. Providers invited interested adolescents to meet with a research assistant to discuss the study protocol and provide informed consent.
In a private room in the clinic, each participant was asked to complete a baseline assessment battery using audio computer-assisted self-interview that included demographic information, sexual history, substance use history, history of mood disorder, the Positive Affect-Negative Affect Schedule (PANAS),76 and the BDI.77 The PANAS is based on a 2-factor model of affect, which posits that PA and NA are independent dimensions. Respondents were asked to rate on a 5-point Likert-type scale the extent to which they feel each of 10 positive and 10 negative affective states. Responses are summed separately for PA and NA, each with a possible score of 10 to 50. The PANAS has been shown to be valid and reliable in adolescent samples.76,78 Furthermore, the time frame can be varied, so in the baseline assessment and recall interviews, affect was measured over the previous 2 weeks. For the diary and MS reports, affect was measured for "right now."
On completion of the baseline assessment battery, the research assistant scored the BDI by the summing of the responses. Participants with scores of 16 or more (out of a possible 63) were considered to have a high level of depressive symptoms79,80 and excluded from the study. (A BDI score of 16 or greater has been shown to have a sensitivity of 100% and a specificity of 93.2% for major depressive episode in adolescents.81) Any participant who responded other than, "0, I don't have any thoughts of killing myself," to question 9 on suicidal ideation were also excluded from study participation.
Using a random-numbers list stratified by gender, each participant was randomly assigned to 1 of the 6 possible orders of the 3 data collection methods. The methods, described in detail below, included a retrospective questionnaire and interview, twice-daily diary, and MS using a 2-way pager. Items and scales included in each method are presented in Table 1. Each arm of the study took 2 weeks to complete, with at least 1 week of no data collection between each arm. Total time of study involvement therefore was a minimum of 8 weeks.
Twice-Daily Diary
Each participant was given 14 stamped, addressed diary cards that could be folded over, sealed, and mailed to the principal investigator each day of the 2-week study period. Each card included the PANAS affect scale, to be completed on awakening and before going to bed each day. The card included a space for logging episodes of sexual intercourse, including the time that the sexual activity began, the relationship with the sexual partner (new, occasional, regular, main), and whether a condom was used. The card also included a space for logging alcohol, marijuana, and other substance use, including the time each episode of use began and ended and the type and quantity of substance(s) used. To protect the participant's confidentiality, the diaries did not include the participant's name, and the responses were coded such that their meaning would not be obvious if someone other than the participant or study personnel read the diary card. Each participant was provided with a "cheat sheet," a small card that could be kept in the participant's wallet, explaining the code for the diaries. Completion of the diaries took no more than 10 minutes each day.
MS
Participants were provided with an electronic pager and paged each day for 2 weeks at random times within each of 4 blocks of time: 9:00 AM to 12:00 PM, 12:01 PM to 3:00 PM, 3:01 PM to 6:00 PM, and 6:01 PM to 9:00 PM, with a fifth signal between 9:01 PM and 12:00 AM on Friday and Saturday nights. Paging was suspended from 9:00 PM on weekday nights, 12:00 AM on weekend nights, until 9:00 AM the following morning to avoid sending signals during hours when participants would be most likely to be asleep. At the signal, the participant recorded his or her responses to the PANAS scale and to contextual questions on thoughts, location, companionship, and activity (Table 1). A card on the back of the pager contained the questions, which were the same as for the diary except that the interval for reporting the risk behaviors was "since last page" (Appendix). Participants were instructed to respond immediately to a page but could ignore the page if it occurred during an incompatible activity (eg, taking a test). To protect their own confidentiality, the adolescents controlled whether they carried the pager and whether and when they responded to the signals. They were provided with a business card with the research assistant's contact information, which they could present to curious adults who may be concerned about the adolescents' carrying a pager. Completion of this method took 22 min/day.
Retrospective Report
Participants were asked to recollect their affect, substance use, and sexual behavior for a 2-week period using a timeline followback calendar recall method. This reliable method has been used extensively in substance use research82 and specifically with adolescents for recall of both substance use83 and sexual behavior.83,84 The timeline followback calendar uses recall-enhancing techniques to provide a detailed assessment of substance use and coitus at the event level.
Because it would have been difficult to recall the specifics of affect on any given day, participants were asked to give a general impression of their affect for each of the 14 days, noting in particular days with the best and the worst affective states. The interview took 30 minutes (more or less, depending on the frequency of substance use and sexual behavior). Participants were also asked to complete a brief questionnaire with general frequency measures for affect, substance use, and sexual behavior for the previous 2 weeks. The questionnaire took 10 minutes to complete. Participants completed the retrospective interview and questionnaire at the end of each of the three 2-week periods.
Exit Interview and Study Termination
On completion of all 3 methods, participants were invited to share their impressions about the methods and suggest which method(s) they believed best represented their actual experience and how they felt about the time commitment and tasks involved, issues of confidentiality, and other aspects of the research experience that they wished to share with the investigators. Because we were interested in comparing the 3 methods for each participant, we discontinued an adolescent's participation if he or she was unable to complete all 3 retrospective self-reports and at least 50% of both the diaries and the MS reports. We offered participants remuneration commensurate with the number of reports, diaries, and interviews that they completed, up to a maximum of $61.50 for 6 weeks of active study participation. The hospital institutional review board approved the study protocol.
Data Analysis
Data were dually entered into a Microsoft Excel database, and the analyses were conducted using SAS version 9.0. The outcomes of interest were the rates of completion of the study instruments, adherence to each method's protocol, the adolescent participants' tolerance of and preferences for the methods, and agreement among the methods in mean scores for PA and NA and in the report of days on which substance use and sexual activity occurred. To understand better the advantages of MS, we also explored associations between affect and contextual questions, including location, companionship, and thoughts. For these analyses, we used linear mixed-effects models to account for potential correlations between repeated measurements taken on the same participant. Pairwise comparisons then were conducted to explore patterns of differences. Because these analyses were exploratory, we did not adjust for multiple comparisons. We also evaluated whether and how affect changed from before to after coital events, again by linear mixed-effects models. Because substance use was infrequent in our sample (overall, only 3 participants reported 2 marijuana and 4 alcohol use events), no analyses were performed to evaluate the associations between affect and substance use.
Results
Sample Characteristics
Six girls and 4 boys completed the study protocol. Mean (±SD) age was 16.9 (±1.0) years, and 70% of the sample self-identified as black race. Mean (±SD) age at first sexual intercourse was 14.4 (±1.8) years. Participants reported a median (range) of 3.5 (1–16) sexual partners in their lifetime and 1 (1–3) sexual partner in the previous 3 months. The adolescents estimated that they had sexual intercourse a median (range) of 3 (1–10) times in the previous 2 weeks. Only 4 youths reported any alcohol or marijuana use, and none reported smoking cigarettes in the previous 30 days. Three participants reported feeling sad or hopeless for at least 2 weeks in the previous 12 months. Median (range) BDI score was 5 (1–13).
Completion Rates
Nineteen adolescents were initially enrolled, with 6 of the nine dropouts occurring during the MS arm. Noncompleters were significantly more likely than completers to report on baseline marijuana use in the previous 30 days. By study design, 30 of 30 of the retrospective interviews were completed (100%, 3 per participant). Of the 280 diaries expected, 254 were returned and complete (91%; mean: 25.4 per participants) and three fourths (74%) of the MS reports were completed (442 of 600; mean: 44.2 per participant; completion rate for diaries vs MS, P = .065). The median time to respond to an MS signal was 28 minutes (range: 0–288 minutes).
Tolerance of and Preferences for the Methods
Most adolescents (6 of 10) preferred the MS method because they were asked to "react to a specific moment" and "being paged is fun" and "cool," with the remaining stating a preference for the diaries because they "capture the whole day instead of a specific moment," were "easy," and "didn't interfere with daily activities." Three participants preferred diaries the least, citing that they were "boring," "a pain to mail," and "got in the way of school," and 2 did not like MS, finding it "annoying" to get paged, especially "at awkward times," and that they had to respond to "too many questions." One participant noted, "The pager going off can affect your mood because it might agitate you." Girls (5 of 6) tended to prefer MS, whereas boys (3 of 4) tended to prefer the diaries.
Correlations Among the Methods Between Average Daily Affect Scores and Reports of Substance Use and Sexual Activity
Mean daily PA and NA scores that were measured using twice-daily diaries or MS were moderately to highly correlated with scores that were obtained using retrospective report (Table 2). Mean affect scores were consistently higher on retrospective report than with the other 2 methods. There was 100% agreement on report of sexual activity between diary or MS and recall and on report of substance use between MS and recall. One adolescent did not recall 1 episode of substance use that was reported by MS.
Correlations of Momentary Affect With Contextual Variables and With Coitus
Participants were most often in nonschool public places when paged (42% of signals), followed by home (34%) then school (24%; Table 3). PA and NA both differed significantly by location when paged (P = .006 and 0.01, respectively). Pairwise comparisons revealed that PA when in public was higher than PA when at home or in school (P = .014) and NA when in school was higher than NA when in public or at home. Participants were most often with other people when paged (65% of reports). NA differed by companionship (P = .03); pairwise comparisons showed that NA in class was higher than NA with friends. There were no differences in PA according to companionship (P = .72). PA differed by thoughts at the time of the signal (P = .03). Although the adolescents reported thinking about sex on only 5% of reports, their PA scores were highest at those times compared with times when they were thinking about anything else (pairwise comparison P = .047). Eight participants reported 20 coital events (mean: 2.5 coital events per participant; range: 1–6 events). Participants provided data on partner type and condom use for 18 and 16 coital events, respectively; 94% of the coital events (17 of 18) occurred with a main sexual partner, and only 38% (6 of 16) were condom protected. From before to after coitus, PA increased and NA decreased, but the differences were not significant (change in PA: ±2.87 [SE: 1.96; P = .19]; change in NA: –0.39 [SE 1.39; P = .78]).
CONCLUSIONS
A substantial body of theoretical and experimental literature supports the notion that affect or mood is related to risk behavior; far fewer empirical studies have been conducted to examine directly the associations, particularly in adolescents. What studies have been done suggest that both positive and negative mood are correlated with sexual interest,36,37 the occurrence of coitus,36,37 safer sex behaviors such as condom use,36,37 and degree of substance use.38,63
For addressing the large gap between theory and research in affect and risk behavior, methods to measure these constructs need to be refined. MS methods offer several advantages over traditional retrospective self-report and diary methods, including the ability to assess the variability in affect in daily life and to relate affect to contextual factors and specific behavioral and environmental events while avoiding many of the biases of other methods.
In a small feasibility study of methods to measure affect and sexual behavior in adolescents, we found that retrospective self-report, twice-daily diary, and MS methods were comparable for collecting data on the occurrence of coitus and on summary affective states, although data on affect from retrospective self-report may be biased toward the extremes. In contrast to the other 2 methods, MS was better able to detect within-day fluctuations in affect as well as situational factors that may influence affect. Importantly, adolescents in this study preferred MS to diaries and retrospective self-report. Findings from the MS data demonstrate that affect differs by contextual factors such as location, companionship, and thoughts. By collecting temporal and contextual data, MS permits exploration of possible causal mechanisms for the associations among affect and sexual activity in adolescents. The finding that adolescents experience their most PA when thinking about sex suggests that further research into the association between affect and sexual behavior is warranted, particularly with regard to sexual intercourse for the purposes of affect management.
This study has several limitations. The number of adolescents who participated in this intensive protocol was small. The relatively infrequent nature of substance use and sexual events may limit the utility of MS methods in studies of these behaviors. Future studies will need to consider longer or repeated periods of data collection and include assessments prompted by the target event (as has been done in EMA studies67,68). The MS method was associated with decreased compliance compared with diaries, although the completion rate was not significantly different and was similar to that reported in previous studies using MS.63,69,70,73–75 This problem may be remedied in future studies by shortening the MS report and considering use of other electronic devices, such as palm computers.69,74
Despite its limitations, this study adds to the relatively sparse literature exploring the use of MS in adolescent samples and uniquely examines the feasibility of this method to collect data on sexual behavior. If the methodologic challenges can be addressed, then additional research with larger samples will help to determine how MS may complement existing methods of collecting affect, substance use, and sexual behavior data in adolescents. Specifically, these studies should examine how MS can increase our understanding of the role of contextual factors, offer a unique look at adolescents' affect and other salient experience before and after occurrence of the risk event, and permit examination of the influence of affect variability on adolescent risk behavior.
Findings from this line of research may influence HIV/STI prevention interventions in several ways, such as consideration of affect assessment and management before or as part of an intervention, affect self-assessment and increasing an individual's understanding of the relation of affect to risk behavior within an intervention, and identification of risk behavior as a method of affect self-management and the development of alternative response strategies as part of an intervention. Increasing our understanding of the associations of affect and sexual risk behavior among adolescents will be important to enhancing behavioral interventions for this at-risk population and may be particularly salient for interventions to reduce risk among depressed youths.
APPENDIX
Momentary Sampling Report for the Measurement of Affect and Sexual Behavior in Adolescents
As you were beeped...
1. What were you thinking about (a) Family (b) Food (c) Friends (d) School (e) Self (f) Sex (g) Society, Religion, Politics (h) Sports (i) Television (j) Time (k) Other
2. Where were you (a) At Home (b) In School (c) In Public—nonschool
3. What was the MAIN thing you were doing
4. WHY were you doing this particular activity (a) I had to (b) I wanted to (c) I had nothing else to do
Use the numbers to express how you feel right now: 1 = Not at all 2 = A little 3 = Moderately 4 = Quite a bit 5 = Extremely
5. Interested
6. Distressed
7. Excited
8. Upset
9. Strong
10. Guilty
11. Scared
12. Hostile
13. Enthusiastic
14. Proud
15. Irritable
16. Alert
17. Ashamed
18. Inspired
19. Nervous
20. Determined
21. Attentive
22. Jittery
23. Active
24. Afraid
25. When you were beeped, whom were you with (type all the lettered responses that apply.) (a) Alone (b) Friends (after this letter, type how many female [F] and how many male [M]) (c) In class (d) Relative(s) (e) Strangers
Questions 26–30. Since you were last beeped... (Y for yes or N for no)
26. Has anything happened or have you done anything that could have affected the way you feel (a) If Yes, indicate what it was.
27. Have you had sexual intercourse (a) If Yes, indicate the type of sexual partner MP = Main Partner, OP = Other Partner, NP = New Partner (b) If Yes, did you use a condom (Y = Yes, N = No)
28. Have you used alcohol (a) If Yes, what kind B = Beer, W = Wine, L = Liquor (b) If Yes, how much (eg, 2 cans = "2cn") cn = Can, btl = Bottle, gl = Glass, oz = Ounce/Shot
29. Have you used marijuana (a) If Yes, indicate the way you smoked and how many (eg, 2 cigarettes = "2c") c = Cigarette, bl = Blunt, p = Pipe, bg = Bong
30. Have you used any other drugs (a) If Yes, name the drug.
31. Have you had your period [Asked only of female participants]
ACKNOWLEDGMENTS
The study was funded in part by the Charles A. Janeway Award in Child Health Research, Child Health Research Center, Children's Hospital Boston (to L.A.S.); and 5 K23 MH01845, National Institute of Mental Health, National Institutes of Health (to L.A.S.).
We are grateful to J. Dennis Fortenberry, MD, MS; Reed Larson, PhD; and Arthur A. Stone, PhD, for support and guidance; Elizabeth Mariano and Colin Hynes for efforts on the day-to-day research activities, and Laura Hacker for editorial assistance.
FOOTNOTES
Accepted Nov 23, 2004.
No conflict of interest declared.
REFERENCES
Kirby D, Short L, Collins J, et al. School-based programs to reduce sexual risk behaviors: a review of effectiveness. Public Health Rep. 1994;109 :339 –360
DiCenso A, Guyatt G, Willan A, et al. Interventions to reduce unintended pregnancies among adolescents: systematic review of randomised controlled trials. BMJ. 2002;324 :1426
Kirby D, DiCenso A, Guyatt G, et al. Effective approaches to reducing adolescent unprotected sex, pregnancy, and childbearing. J Sex Res. 2002;39 :51 –57
Stiffman AR, Dore P, Earls F, Cunningham R. The influence of mental health problems on AIDS-related risk behaviors in young adults. J Nerv Ment Dis. 1992;180 :314 –320
Joffe A, Radius SM. Self-efficacy and intent to use condoms among entering college freshmen. J Adolesc Health. 1993;14 :262 –268
Shrier LA, Emans SJ, Woods ER, DuRant RH. The association of sexual risk behaviors and problem drug behaviors in high school students. J Adolesc Health. 1997;20 :377 –383
Seal A, Minichiello V, Omodei M. Young women's sexual risk taking behaviour: re-visiting the influences of sexual self-efficacy and sexual self-esteem. Int J STD AIDS. 1997;8 :159 –165
Pao M, Lyon M, D'Angelo LJ, Schuman WB, Tipnis T, Mrazek DA. Psychiatric diagnoses in adolescents seropositive for the human immunodeficiency virus. Arch Pediatr Adolesc Med. 2000;154 :240 –244
American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 4th ed. Washington, DC: American Psychiatric Association; 1994
Centers for Disease Control and Prevention. Youth risk behavior surveillance—United States, 2001. MMWR Recomm Rep. 2002;51(SS4) :1 –64
Pine DS, Cohen E, Cohen P, Brook J. Adolescent depressive symptoms as predictors of adult depression: moodiness or mood disorder Am J Psychiatry. 1999;156 :133 –135
Garrison CZ, Waller JL, Cuffe SP, McKeown RE, Addy CL, Jackson KL. Incidence of major depressive disorder and dysthymia in young adolescents. J Am Acad Child Adolesc Psychiatry. 1997;36 :458 –465
Lewinsohn PM, Hops H, Roberts RE, Seeley JR, Andrews JA. Adolescent psychopathology: I. Prevalence and incidence of depression and other DSM-III-R disorders in high school students. J Abnorm Psychol. 1993;102 :133 –144
Shrier L, Braslins P, Christiansen D, Markowitz LE, Fortenberry JD, Orr D. Gender Differences in the Associations of Mental Health, Sexual Risk Behavior, and Sexually Transmitted Infection Among Adolescents and Young Adults. Paper presented at the International Society for Sexually Transmitted Disease Research; July 30, 2003; Ottawa, Ontario, Canada
Valois RF, Bryant ES, Rivard JC, Hinkle KT. Sexual risk-taking behaviors among adolescents with severe emotional disturbance. J Child Fam Stud. 1997;6 :409 –419
Smith MD. HIV risk in adolescents with severe mental illness: literature review. J Adolesc Health. 2001;29 :320 –329
Brown L, Danovsky M, Lourie K, DiClemente R. Adolescents with psychiatric disorders and the risk of HIV. J Am Acad Child Adolesc Psychiatry. 1997;36 :1609 –1615
Baker DG, Mossman D. Potential HIV exposure in psychiatrically hospitalized adolescent girls. Am J Psychiatry. 1991;148 :528 –530
Erbelding EJ, Hummel B, Hogan T, Zenilman J. High rates of depressive symptoms in STD clinic patients. Sex Transm Dis. 2001;28 :281 –284
Crepaz N, Marks G. Are negative affective states associated with HIV sexual risk behaviors A meta-analytic review. Health Psychol. 2001;20 :291 –299
Kalichman SC, Kelly JA, Johnson JR, Bulto M. Factors associated with risk for HIV infection among chronic mentally ill adults. Am J Psychiatry. 1994;151 :221 –227
Leith KP, Baumeister RF. Why do bad moods increase self-defeating behavior Emotion, risk taking, and self-regulation. J Pers Soc Psychol. 1996;71 :1250 –1267
Morris WN, Reilly NP. Toward the self-regulation of mood: Theory and research. Motivation Emotion 1987;11 :215 –249
Rao U, Daley SE, Hammen C. Relationship between depression and substance use disorders in adolescent women during the transition to adulthood. J Am Acad Child Adolesc Psychiatry. 2000;39 :215 –222
Mueser KT, Drake RE, Wallach MA. Dual diagnosis: a review of etiological theories. Addict Behav. 1998;23 :717 –734
Tapert SF, Aarons GA, Sedlar GR, Brown SA. Adolescent substance use and sexual risk-taking behavior. J Adolesc Health. 2001;28 :181 –189
Graves KL, Leigh BC. The relationship of substance use to sexual activity among young adults in the United States. Fam Plan Perspect. 1995;27 :18 –22, 33
Castillo-Mezzich A, Tarter R, Giancola P, Lu S, Kirisci L, Parks S. Substance use and risky sexual behavior in female adolescents. Drug Alcohol Depend. 1997;44 :157 –166
Strunin L, Hingson R. Alcohol, drugs, and adolescent sexual behavior. Int J Addict. 1992;27 :129 –146
Shrier LA, Harris SK, Sternberg M, Beardslee WR. Associations of depression, self-esteem, and substance use with sexual risk among adolescents. Prev Med. 2001;33 :179 –189
Shrier L, Harris S, Beardslee W. Temporal associations between depressive symptoms and self-reported sexually transmitted disease among adolescents. Arch Pediatr Adolesc Med. 2002;156 :599 –606
Lewinsohn PM, Petit JW, Joiner TE Jr, Seeley JR. The symptomatic expression of major depressive disorder in adolescents and young adults. J Abnorm Psychol. 2003;112 :244 –252
Bancroft J, Janssen E, Strong D, Carnes L, Vukadinovic Z, Long J. The relation between mood and sexuality in heterosexual men. Arch Sex Behav. 2003;32 :217 –230
Bancroft J, Janssen E, Strong D, Vukadinovic Z. The relation between mood and sexuality in gay men. Arch Sex Behav. 2003;32 :231 –242
Skiba A, Fortenberry J, Blythe M. Daily mood and behavioral correlates of coitus and condom use. J Adolesc Health. 1997;20 :145
Fortenberry J, Temkit MH, Tu W, Katz B, Orr D. Mood, mood change, sexual activity and condom use among adolescent women. J Adolesc Health. 2003;32 :157
Wills TA, Sandy JM, Shinar O, Yaeger A. Contributions of positive and negative affect to adolescent substance use: test of a bidimensional model in a longitudinal study. Psychol Addict Behav. 1999;13 :327 –338
Isen AM. Missing in action in the AIM: positive affect's facilitation of cognitive flexibility, innovation, and problem solving. Psychol Inq. 2002;13 :57 –65
Isen AM, Patrick R. The effect of positive feelings on risk taking: when the chips are down. Organ Behav Hum Decis Process. 1983;31 :194 –202
Isen AM, Geva N. The influence of positive affect on acceptable level of risk: the person with a large canoe has a large worry. Organ Behav Hum Decis Process. 1987;39 :145 –154
Clore GL, Gasper K, Garvin E. Affect as information. In: Forgas JP, ed. Handbook of Affect and Social Cognition. Mahwah, NJ: Lawrence Erlbaum Associates; 2001:121 –144
Forgas JP, Bower GH, Krantz SE. The influence of mood on perceptions of social interactions. J Exp Soc Psychol. 1984;20 :497 –513
Forgas JP. Affective influences on partner choice: role of mood in social decisions. J Pers Soc Psychol. 1991;61 :708 –720
Rusting C. Personality as a moderator of affective influences on cognition. In: Forgas JP, ed. Handbook of Affect and Social Cognition. Mahwah, NJ: Lawrence Erlbaum Associates; 2001:371 –391
Clark MS, Isen AM. Toward understanding the relationship between feeling states and social behavior. In: Hastorf AH, Isen AM, eds. Cognitive Social Psychology. New York, NY: Elsevier; 1982:73 –108
Erber M, Erber R. The role of motivated social cognition in the regulation of affective states. In: Forgas JP, ed. Handbook of Affect and Social Cognition. Mahwah, NJ: Lawrence Erlbaum Associates; 2002:275 –290
Wegener DT, Petty RE. Mood management across affective states: the hedonic contingency hypothesis. J Pers Soc Psychol. 1994;66 :1034 –1048
Josephson BR, Singer JA, Salovey P. Mood regulation and memory: repairing sad moods with happy memories. Cogn Emotion. 1996;10 :437 –444
Salovey P, Birnbaum D. Influence of mood on health-relevant cognitions. J Pers Soc Psychol. 1989;57 :539 –551
Forgas JP. Mood and judgment: the affect infusion model (AIM). Psychol Bull. 1995;117 :39 –66
Haidt J. "Dialogue between my head and my heart": affective influences on moral judgment. Psychol Inq. 2002;13 :54 –56
Clark MS. We should focus on interpersonal as well as intrapersonal processes in our search for how affect influences judgments and behavior. Psychol Inq. 2002;13 :32 –37
Perrott DA, Bodenhausen GV. The way you make me feel: integral affective influences on interpersonal behavior. Psychol Inq. 2002;13 :84 –86
Kalichman SC, Weinhardt L. Negative affect and sexual risk behavior: comment on Crepaz and Marks (2001). Health Psychol. 2001;20 :300 –301
Coxon AP. Parallel accounts Discrepancies between self-report (diary) and recall (questionnaire) measures of the same sexual behaviour. AIDS Care. 1999;11 :221 –234
Leigh BC. Alcohol consumption and sexual activity as reported with a diary technique. J Abnorm Psychol. 1993;102 :490 –493
Ramjee G, Weber AE, Morar NS. Recording sexual behavior: comparison of recall questionnaires with a coital diary. Sex Transm Dis. 1999;26 :374 –380
Freeman M, Csikszentmihalyi M, Larson R. Adolescence and its recollection: toward an interpretive model of development. Merrill Palmer Q. 1986;32 :167 –185
Fortenberry JD, Orr DP, Katz BP, Brizendine EJ, Blythe MJ. Sex under the influence. A diary self-report study of substance use and sexual behavior among adolescent women. Sex Transm Dis. 1997;24 :313 –319
Fortenberry JD, Orr DP, Zimet GD, Blythe MJ. Weekly and seasonal variation in sexual behaviors among adolescent women with sexually transmitted diseases. J Adolesc Health. 1997;20 :420 –425
Fortenberry J, Cecil H, Zimet G, Orr D. Concordance between self-report questionnaires and coital diaries for sexual behaviors of adolescent women with sexually transmitted infections. In: Bancroft J, ed. Research Sexual Behavior: Methodological Issues. Bloomington, IN: Indiana University Press; 1997:461
Lex BW, Griffin ML, Mello NK, Mendelson JH. Alcohol, marijuana, and mood states in young women. Int J Addict. 1989;24 :405 –424
Csikszentmihalyi M, Larson R, Prescott S. The ecology of adolescent activity and experience. J Youth Adolesc. 1977;6 :281 –294
Larson R, Csikszentmihalyi M. The experience sampling method. In: Reis H, ed. Naturalistic Approaches to Studying Social Interactions. Vol 15. San Francisco, CA: Jossey-Bass; 1983:41–56
Csikszentmihalyi M, Larson R. Validity and reliability of the experience-sampling method. J Nerv Ment Dis. 1987;175 :526 –536
Hufford MR, Shields AL, Shiffman S, Paty J, Balabanis M. Reactivity to ecological momentary assessment: an example using undergraduate problem drinkers. Psychol Addict Behav. 2002;16 :205 –211
Shiffman S. Real-Time Self-Report of Momentary States in the Natural Environment: Computerized Ecological Momentary Assessment. Mahwah, NJ: Lawrence Erlbaum; 2000
Whalen CK, Jamner LD, Henker B, Delfino RJ. Smoking and moods in adolescents with depressive and aggressive dispositions: evidence from surveys and electronic diaries. Health Psychol. 2001;20 :99 –111
Larson R, Csikszentmihalyi M, Graef R. Mood variability and the psychosocial adjustment of adolescents. J Youth Adolesc. 1980;9 :469 –490
Larson R. Beeping children and adolescents: a method for studying time use and daily experience. J Youth Adolesc. 1989;18 :511 –530
Larson R, Csikszentmihalyi M, Freeman M. Alcohol and marijuana use in adolescents' daily lives: a random sample of experiences. Int J Addict. 1984;19 :367 –381
Henker B, Whalen CK, Jamner LD, Delfino RJ. Anxiety, affect, and activity in teenagers: monitoring daily life with electronic diaries. J Am Acad Child Adolesc Psychiatry. 2002;41 :660 –670
Whalen CK, Jamner LD, Henker B, Delfino RJ, Lozano JM. The ADHD spectrum and everyday life: experience sampling of adolescent moods, activities, smoking, and drinking. Child Dev. 2002;73 :209 –227
Larson R, Lampman-Petraitis C. Daily emotional states as reported by children and adolescents. Child Dev. 1989;60 :1250 –1260
Watson D, Tellegen A. Toward a consensual structure of mood. Psychol Bull. 1985;98 :219 –235
Beck A, Ward C, Mendelson M, et al. An inventory for measuring depression. Arch Gen Psychiatry. 1961;4 :561 –571
Watson D, Clark LA, Tellegen A. Development and validation of brief measures of positive and negative affect: the PANAS scales. J Pers Soc Psychol. 1988;54 :1063 –1070
Kendall P, Hollon S, Beck A, Hammen C, Ingram R. Issues and recommendations regarding use of the Beck Depression Inventory. Cognit Ther Res. 1987;11 :289 –299
Gotlib I, Cane D. Self-report assessment of depression and anxiety. In: Kendall PC, Watson D, eds. Anxiety and Depression. New York, NY: Academic Press; 1989:131–169
Barrera M, Garrison-Jones CV. Properties of the Beck Depression Inventory as a screening instrument for adolescent depression. J Abnorm Child Psychol. 1988;16 :263 –273
Carney MA, Tennen H, Affleck G, Del Boca FK, Kranzler HR. Levels and patterns of alcohol consumption using timeline follow-back, daily diaries and real-time "electronic interviews." J Stud Alcohol. 1998;59 :447 –454
Stein MD, Anderson B, Charuvastra A, Friedmann PD. Alcohol use and sexual risk taking among hazardously drinking drug injectors who attend needle exchange. Alcohol Clin Exp Res. 2001;25 :1487 –1493
Weinhardt LS, Carey MP, Maisto SA, Carey KB, Cohen MM, Wickramasinghe SM. Reliability of the timeline follow-back sexual behavior interview. Ann Behav Med. 1998;20 :25 –30(Lydia A. Shrier, MD, MPH,)