Intervening with couples:
Assessing Contraceptive Outcomes in a Randomized Pregnancy and HIV/STD Risk Reduction Intervention Trial
Article Outline
Purpose
This study assessed the contraceptive outcomes of the Partners Against Risk-Taking: A Networking, Evaluation and Research Study (PARTNERS). The PARTNERS project developed and evaluated a 3-session intervention to help young women and their male partners reduce their risk for unintended pregnancies, and HIV and other STDs.
Methods
Participating couples were randomly assigned to the 3-session intervention or a 1-session information session for couples. Changes in psychosocial factors related to women’s motivation to use contraception and relationship factors were assessed using analysis of variance with repeated measures. Changes in contraceptive outcomes were assessed using logistic regression with generalized estimating equations.
Results
Comparison of changes from baseline to 6 months among women who participated in the 3-session intervention with those who participated in the information session showed no significant intervention effect on reports of contraceptive use. Instead, contraceptive use increased in both conditions. Both groups exhibited similar changes in the psychosocial variable measuring the importance of avoiding pregnancy and in the relationship variable measuring women’s participation in contraceptive decision making. Members of the intervention group, however, showed greater improvement in the psychosocial variable measuring positive expectations pertaining to partner’s support for contraception.
Conclusion
These findings raise questions for further investigation to better understand couples behavior, and whether and how to intervene with couples.
Unprotected sexual intercourse puts women at risk for unintended pregnancy, human immunodeficiency virus (HIV) infection, and other sexually transmitted diseases (STDs). In 1994 (48%) and again in 2001 (49%), nearly half of all pregnancies in the United States were unintended (Finer & Henshaw, 2006). In 2001, 48% of unintended pregnancies occurred in a month when women were using contraception, suggesting both nonuse and incorrect use of contraception contribute to unintended pregnancy (Finer & Henshaw, 2006). Nonuse of condoms has other negative consequences, including infection with HIV, which has increased among women since the late 1980s (Centers for Disease Control and Prevention [CDC], 1999). Finally, although rates of some STDs among women have declined recently, STDs have negative consequences for women’s reproductive health (e.g., pelvic inflammatory disease; CDC, 2003).
Background
Research describes the benefits of addressing individual psychosocial factors in interventions that promote behavior change to enhance health (e.g., Albarracin et al 2005, Fishbein 2000). Several psychosocial factors motivate behavior change, including skills and confidence in one’s skills (or self-efficacy), positive expectations for outcomes associated with change, social norms, and risk perception. A meta-analysis showed that HIV interventions that addressed these factors were among the more effective interventions (Albarracin et al., 2005).
Recognizing that pregnancy and disease prevention take place within relationships, behavioral and intervention research draws attention to the role of relationship factors in protective behaviors. Results from our formative study indicate that women perceive shared reproductive decision making with their partners and studies with other multiethnic samples found a high degree of joint decision making and control over condom use (Cabral et al 1998, Harvey et al 1999, Harvey et al 2002, Soler et al 2000). Moreover, Grady, Tanfer, Billy, & Lincoln-Hanson (1996) report that men accept responsibility for contraception. Intervention research shows that intervening with couples may be effective (Becker, 1996). Education and counseling interventions for couples were associated with higher rates of contraceptive use than similar interventions for women only, perhaps because they led to agreement about contraception. Further, couples counseling for HIV-discordant couples may increase safer sexual behaviors (Becker, 1996). Despite these findings, most interventions tend to focus on one partner, and most do not address relationship issues (Misovich, Fisher, & Fisher, 1997).
These findings compelled us to design, implement, and evaluate an intervention for heterosexual couples by intervening with both partners. The intervention for the Partners Against Risk-Taking: A Networking, Evaluation and Research Study (PARTNERS Project) was designed to reduce unintended pregnancy, and HIV and other STDs among 18- to 25-year-old women and their primary male partners by increasing the use of preventive strategies (Harvey et al., 2006). As suggested by the literature, the PARTNERS intervention addressed psychosocial factors and relationship factors to encourage couples to adopt preventive strategies. This paper presents results of an examination of intervention effects on contraceptive outcomes 6 months after the intervention. Other papers present disease prevention findings (e.g., Harvey et al., 2004).
The PARTNERS Intervention
We evaluated the intervention in a randomized comparison trial with 2 conditions: the 3-session intervention or a 1-session informational condition. Funding constraints limited the number of comparison groups to 1. Because we were interested in addressing relationship factors, the comparison session was also for couples, but did not address relationship factors.
Male and female facilitators with experience in social work, counseling, or a related field facilitated the intervention and comparison sessions. To ensure that the intervention and comparison sessions were conducted systematically and consistently, facilitators participated in training workshops and used structured protocols including a manual with standardized scripts. One male and one female facilitator administered each group session for up to 6 couples per session. To reduce potential bias, facilitators did not recruit or interview participants.
The comparison condition, which lasted 1.5–2 hours, provided education about HIV, STDs, and contraception. Facilitators provided information on HIV and STDs, displayed sample contraceptive methods, and described their correct use and their efficacy for preventing pregnancy and disease. Finally, facilitators led a question-and-answer period.
Couples in the intervention group participated in 3 sessions, each lasting 2.5 hours, over 3 weeks. In addition to the information provided in the comparison session, the intervention included activities and discussion to address such psychosocial factors as perceived risk, positive expectations, norms, skills, and self-efficacy as they relate to preventive strategies. Activities were designed to enhance communication so couples could discuss decisions about preventive strategies, including contraception.
More specifically, the first intervention session focused on increasing perceived risk for unintended pregnancy, and HIV and STDs. The session also introduced 3 preventive strategies: abstinence, condom use, and mutual testing and monogamy. Because monogamy does not prevent pregnancy, facilitators stressed the need for contraception. This session included activities to demonstrate how STDs can be spread in a community (e.g., increase risk perception) and encouraged participants to identify and discuss their attitudes toward the 3 strategies.
The second intervention session focused on the psychosocial factors that motivate use of each strategy. For example, to increase condom use, the session addressed common perceptions of condoms (positive expectations), encouraged participants to discuss others’ beliefs about condoms (norms), and encouraged couples to talk about how to use condoms together. Participants practiced putting condoms on models to build skills and self-efficacy.
The last session provided the information about contraceptives that was provided in the comparison session. To increase risk perception, facilitators also presented information on the risk of pregnancy when not using contraception. They drew attention to the importance of correct and consistent use of each method, and the benefits of some methods for disease prevention. Although all sessions implicitly addressed communication (exercises for couples), this session explicitly focused on communication. Before discussing the preventive strategy they would use, couples watched a video about and practiced using assertive communication strategies. Facilitators encouraged couples to share their attitudes so that partners would know about each other’s support for preventive strategies.
Methods
PARTNERS Study Design
Participants were recruited from the Los Angeles, California, and Oklahoma City, Oklahoma, areas. Within several weeks of participating in baseline interviews, couples attended a group meeting to be randomly assigned to a condition. Both conditions were standardized across the sites with some site-specific tailoring. Three-month follow-up interviews were conducted with women and men. Due to cost constraints, 6-month follow-up interviews were conducted with women only.
PARTNERS Study Participants
Couples were recruited through the women using active and passive recruitment strategies. Each site selected a variety of community (e.g., colleges, universities, housing projects, malls) and clinic (e.g., general clinics, STD clinics, Planned Parenthood) settings frequented by young women. At these locations, recruiters approached women with information about the project. For passive recruitment, recruiters placed printed materials (e.g., posters) that described the project and listed a toll-free number in community locations; we also advertised in local media. Interested women completed short screening interviews to determine eligibility.
Eligibility criteria focused on sexual risk and pregnancy intentions. Women were eligible if they were 18–25 years old, had a male sex partner age ≥18 years old, had sex without a condom at least once in the past 3 months, and met ≥1 of the following criteria: 1) engaged in risky behavior (e.g., had another partner in the past year, ever used intravenous drugs); 2) knew or thought their partners were at risk (e.g., had an STD); or 3) thought they or their partners would have sex with someone else in the next year while they were still together. Women who were pregnant, intended to become pregnant within the year, or reported being HIV positive were ineligible. Women in Los Angeles had to self-identify as Latina or Hispanic; women in Oklahoma could be of any race/ethnicity. We restricted eligibility to Latina or Hispanic women in Los Angeles because individuals of Hispanic ethnicity comprise 47% of Los Angeles’ total population (US Census Bureau, 2001) and because Hispanics are heavily affected by AIDS, comprising >25% of Los Angeles County’s AIDS cases (County of Los Angeles Department of Health and Human Services, 2003).
After completing the screening interview, recruiters asked eligible women to invite primary partners (someone like a spouse or steady boyfriend) to participate. Both partners had to agree to enroll in the study before the couple could complete baseline interviews.
Interview Procedures
Couples were enrolled in the study when they completed baseline interviews. Baseline interviews were conducted from January 2000 through June 2002. Six-month interviews were conducted 6 months after the last intervention session and 6 months and 2 weeks after the comparison session; 6-month outcome data collection was completed by January 2003. Interviews were 60 minutes long (at baseline, partners were interviewed individually). Women received $30 and $50 for participating in the baseline and 6-month interviews, respectively, and were compensated for travel and/or child care costs. The Institutional Review Boards of the institutions responsible for each site and the CDC approved this research. Participants signed written consent forms.
Participants and interviewers were matched by gender. In Los Angeles, interviewers were Hispanic, and participants could be interviewed in Spanish or English. For most questions, interviewers asked questions and entered participants’ responses directly into a computer using a computer-assisted survey interview (CASI). The CASI software we used, QDS, allowed us to insert a partner’s name in the text of questions so that we could create partner-specific measures (see Appendix I; Nova Research Company, 2003). For the more sensitive sexual behavior questions, participants could choose to listen to sexual behavior questions on headphones and enter their responses (62.2% did so), rather than having the interviewer read questions to them.
Randomization
After both members of a couple completed baseline interviews, they were scheduled to attend a group meeting with up to 12 other couples. Four facilitators (2 each for the comparison and intervention sessions) met the couples at the group meeting. Facilitators prepared an equal number of slips of paper indicating assignment to the intervention or the comparison condition and placed them in a jar. Couples picked slips of paper to determine which condition they were assigned and, thus, assignment was not blinded.
Participant Flow
Of women screened, approximately half (49% in Oklahoma City, 51% in Los Angeles) were eligible. Although most eligible women agreed to participate, fewer than half of the women and their partners completed baseline interviews (26% in Oklahoma City, 41% in Los Angeles). Of the 435 couples who completed baseline interviews, 301 (69.2%) were randomized to a condition, and the overwhelming majority (all in the comparison and just >90% in the intervention) completed all sessions. Two hundred thirty-four (77.7%) of the 301 women completed 6-month interviews. Women in the both conditions were equally likely to complete 6-month interviews.
Evaluating Contraceptive Outcomes
We focus on longer term contraceptive outcomes among women who completed baseline and 6-month interviews. We relied on women’s reports because many participants indicated they used female-controlled methods and women’s reports of using such methods may be more reliable than men’s reports. It is important to note that men’s and women’s reports did not differ significantly; at baseline, 77% of male and female partners were classified the same on our contraceptive measure (see Measures). Analyses (not presented) showed similar results for women’s and men’s 3-month outcomes.
Outcomes and Hypotheses
The primary outcome is contraceptive use. Secondary outcomes include individual psychosocial and relationship factors addressed in the intervention. Specifically, we hypothesized that women in the intervention group would show greater improvement in consistent use of an effective contraception than women in the comparison group. We hypothesized that women in the intervention group would also show greater improvement on 3 psychosocial measures related to motivation to use contraception and 1 relationship measure (see Measures).
The Analytic Sample
Of the 234 women who completed baseline and 6-month interviews, we excluded 8 women because they did not have sex in the 3 months prior to the 6-month interview (and so we did not ask them about contraception) and 3 women because they had missing data on the primary outcome. Thus, the sample size for the primary outcome was 233. This sample size is lower than anticipated. Power analyses suggested we needed 150 per condition to detect a 10% point difference in the primary outcome with a power of .85 at the .05 level.
Measures
Psychosocial and Relationship FactorsThe questionnaire included measures of individual psychosocial factors pertaining to contraceptive use and relationship factors. Questions used to measure these factors had 5-point response scales, where higher scores indicated higher levels of a construct (see Appendix 1). Most measures were partner specific, asking about the primary or last partner (at the 6-month interview some women no longer had a primary partner but had sex in the past 3 months, so we asked about their “last partner”). For constructs measured with multi-item scales, we computed the average (alphas or correlations in Table 2).
Table 2. Baseline means on psychosocial factors that motivate women to use contraception and relationship factors, and α or correlation coefficients for scales
| Psychosocial Variables Pertaining to Contraceptive Use | N | α | Correlation | Mean (SD) |
|---|---|---|---|---|
| Perceived risk for pregnancy | 223 | N/A | N/A | 4.44 (0.93) |
| Importance of not becoming pregnant | 222 | N/A | N/A | 4.35(0.92) |
| Positive expectations pertaining to partner’s support for contraception | 222 | N/A | .60 | 3.68(1.12) |
| Participation in contraceptive decision making | 222 | .68 | N/A | 3.92(.96) |
The psychosocial measures focused on motivations to use contraception. We measured a woman’s perceived risk of pregnancy and the importance of being safe from pregnancy (single items; see Appendix 1). We also measured positive expectations pertaining to her partner’s support for contraception, asking about the likelihood that he would think it is good to talk about unplanned pregnancies and to use condoms to prevent pregnancy. Because norms and self-efficacy for using contraception may depend on the method used and because we had to limit the instrument, we did not measure these constructs. Finally, we assessed participation in contraceptive decision making by asking women how much they participated in making decisions about when to get pregnant, whether to use birth control, and whether to use condoms.
Contraceptive useWe measured whether women used birth control consistently. We asked women what they used to prevent pregnancy with their primary or last partner in the past 3 months and follow-up questions to assess “consistency” of use. For example, we asked women who used the pill if they missed >2 pills in any of the past 3 months and we asked women who used the male condom if they had sex with their partners at least once without using a condom in the past 3 months.
Based on data from Hatcher, Rinehart, Blackburn, and Geller (1997), we classified methods as effective if fewer than 10% of women become pregnant in a year of consistent use of the method. Hatcher et al. (1997) classified methods as (a) always very effective (e.g., Norplant); (b) effective as commonly used, very effective when used correctly and consistently (e.g., the pill); or (c) somewhat effective as commonly used, effective when used correctly and consistently (e.g., condom). Accordingly, we classified women’s reports of sterilization, intrauterine device, Depo Provera, or Norplant, and consistent use of birth control pills, diaphragms, and/or male condoms as “effective.” We combined women’s reports of withdrawal, rhythm, inconsistent use of abstinence, spermicides, or other methods and nonuse into an “other” category. When women reported >1 method, we gave priority to the most effective method used.
Analytic Procedures
To evaluate the effectiveness of the intervention on the psychosocial and relationship factors we used a series of 2 (group: intervention vs. comparison) × 2 (time: baseline vs. 6-month) analysis of variance models with repeated measures (for time). A main effect of time would indicate that change occurred in both groups. A statistically significant group by time interaction would indicate more change in one group. To evaluate the effectiveness of the intervention on contraceptive use, we used repeated measures logistic regression using generalized estimating equations.
Results
Description of the Sample
Characteristics of the sample are presented in Table 1. Almost half were Hispanic, nearly one third were non-Hispanic white, and the rest were African American or of another race/ethnicity. One fourth had not finished high school and most were employed. Analyses not shown indicated that there were no significant differences in these demographic variables between this sample and women who were lost to follow-up (not assigned to a condition, not interviewed at 6 months). In addition, no statistical differences between this sample and women lost to follow-up were found on sexual risk variables (e.g., consistent condom use, having had another partner in the past 3 months). Finally, in this sample, there were no statistical differences in demographic or risk variables by group assignment.
Table 1. Percentage distributions or means on baseline background characteristics for women in the PARTNERS project who completed baseline and 6-month interviews, and who had valid data on the primary outcome (contraception; n = 223)
| Race/ethnicity | |
| Latina | 48.2% |
| White, not Hispanic | 30.0% |
| African American | 12.6% |
| Other | 8.1% |
| Age | 20.91 (SD=2.29) |
| Martial status/living arrangements | |
| Married | 17.5% |
| Living together, not married | 37.2% |
| Single | 45.3% |
| Education | |
| Mean | 12.61 (SD=2.29) |
| <12 | 24.3% |
| 12 | 24.3% |
| >12 years | 51.4% |
| Employed | |
| Yes | 57.4% |
| No | 42.2% |
Baseline Data
Participants had high baseline levels on the psychosocial and relationship variables (Table 2). For instance, the average score of 3.92 (standard deviation [SD] = .96) on the decision making variable indicates that, on average, women believed that they participated “a lot” in decisions about contraception. Despite the fact that all of the women reported not wanting to become pregnant in the next year, most women were not using an effective method of contraception consistently at baseline. Only one third (33.2%) of the women reported using an effective method of birth control consistently in the past 3 methods (Table 3). Most women in this category used the pill or Depo Provera. Among women who were not consistent users of an effective method, 12.1% reported using nothing and 53.7% reported using condoms inconsistently.
Table 3. Baseline percentage distributions on contraceptive use: Consistent use of effective contraception and specific types of contraceptives used
| Percent | |
|---|---|
| All women (N=223) | |
| Use effective method of contraception consistently | 33.2 |
| Nonuse, ineffective method, effective method inconsistently | 66.8 |
| Women who use an effective method consistently (n=74) | |
| Sterilization (woman or her partner) | 4.1 |
| Depo Provera | 40.5 |
| Norplant | 1.4 |
| IUD | 5.4 |
| Consistent use of birth control pills | 40.5 |
| Consistent use of a diaphragm | 1.4 |
| Consistent use of the male condom | 8.1 |
| Women who do not use a method, use an ineffective method, or use an effective method inconsistently (n =149) | |
| Nonuse | 12.1 |
| Use ineffective method | 22.8 |
| Withdrawal | 1.3 |
| Rhythm | |
| Use effective method inconsistently | |
| Birth control pills | 18.1 |
| Abstain | 5.4 |
| Male condom | 53.7 |
| Spermicide | 5.4 |
| Female condom | 0.7 |
Effects of the Intervention
We observed few effects on the secondary outcomes. At the 6-month follow-up, women in the intervention group did not differ significantly from those in the comparison group on perceived risk for pregnancy (Table 4). Women’s reports of the importance of not becoming pregnant and their participation in contraceptive decision making changed over time, but those changes were the same for women in both groups. Specifically, women in both groups reported statistically significant decreases in the importance of not becoming pregnant (F = 8.79, p = .003) and statistically significant increases in their participation in contraceptive decision making (F = 27.15, p = .001). However, the intervention was associated with changes in positive expectations pertaining to partner’s support for contraception. Specifically, there was a small mean increase in the intervention group and a small mean decrease in the comparison group (F = 4.83, p = .029).
Table 4. Mean change overtime for intervention and comparison group: Unadjusted 2 × 2 analyses of variance models with repeated measures for time for psychosocial and relationship factors
| Baseline Mean | 6-Month Mean | |
|---|---|---|
| Perceived risk for pregnancy (n=221) | ||
| Intervention group | 4.46 | 4.47 |
| Comparison group | 4.41 | 4.47 |
| Importance of not becoming pregnant (n=220) | ||
| Intervention group | 4.37 | 4.11 |
| Comparison group | 4.35 | 4.10 |
| Time, F=8.79, p=.003 | ||
| Positive expectations pertaining to partner’s support for contraception (n=220) | ||
| Intervention group | 3.53 | 3.64 |
| Comparison group | 3.83 | 3.57 |
| Time by condition, F=4.83, p=.029 | ||
| Participation in contraceptive decision making (n=190) | ||
| Intervention group | 3.82 | 4.29 |
| Comparison group | 3.93 | 4.20 |
| Time, F=27.15, p=.001 |
Women in the intervention group were no more likely than women in the comparison group to use effective contraception consistently at the 6-month follow-up (data not presented). Consistent use of an effective method increased over time in both groups (likelihood estimate = .566, p < .01), from 31.3% to 48.2% in the intervention group and from 35.1% to 46.8% in the comparison group. Given these differences, a sample size of 300 (see The Analytic Sample) would not have resulted in statistically significant intervention effects.
Discussion
Synopsis
To address the prevention needs of young couples, we developed and evaluated a 3-session intervention to help couples modify behaviors that place them at risk for unintended pregnancy and disease. The intervention addressed individual psychosocial factors known to motivate behavior change (Albarracin et al 2005, Fishbein 2000) and relationship factors associated with couples’ contraception (Becker 1996, Becker and Robinson 1998). In our 6-month evaluation, we found no significant intervention effects on women’s contraceptive use. Intervention effects were limited to an effect on positive expectations for taking steps to avoid pregnancy. However, we found significant increases in contraceptive use and women’s participation in contraceptive decision making in both groups over time. We also found significant decreases in the importance of not becoming pregnant in both groups.
The changes in women’s reports of the importance of not becoming pregnant were unexpected. Analyses suggested that this finding may be due to the inclusion of 20 women who were pregnant at the 6-month interview. Women in both groups were equally likely to become pregnant, and we found no change in this variable when we excluded pregnant women.
It is encouraging that the intervention increased positive expectations pertaining to a partner’s support for contraception. The intervention promoted discussion and taught communication skills. This finding may indicate that partners communicated more about contraception, and thus had more accurate perceptions of each other’s support for pregnancy prevention.
Limitations
Despite the strengths of the study (e.g., 6-month follow-up and partner-specific data), several limitations may have produced the observed changes in both groups. First, the large number of eligible women who did not participate suggests that some women would or could not persuade their partners to participate. For instance, couples who were not already motivated to change may have declined to participate, and so couples in both groups may have already been motivated to make the changes we assessed. Alternatively, women in new or less stable relationships may have been less comfortable talking to their partners about the study. Although women had concerns about convincing their partners to participate, we do not have data to determine if less motivated or less stable couples were less likely to participate, or whether the intervention would have had different effects on them.
Second, attrition combined with our data collection and retention efforts may have contributed to the observed changes. Although there was less attrition in some study phases (e.g., randomization to follow-up) than others (e.g., baseline to randomization), we cannot rule out the possibility that women not interviewed at the 6-month follow-up were less likely to change. Further, the interviews and intensive efforts to increase retention may have sensitized women to their risk, contributing to behavior change or to socially desirable responses in both groups.
The content of the intervention may have contributed to its apparent lack of effectiveness relative to the comparison condition. Although the intervention had characteristics of effective HIV interventions for individuals (e.g., addressed psychosocial factors), there is little research on characteristics of effective couples interventions. Our intervention addressed psychosocial and relationship factors in the same way for both partners. Interventions may need to address different psychosocial factors for men and women (e.g., Sheeran, Abraham, & Orbell, 1999). For example, because many contraceptive methods are female controlled, women may need more time devoted to building skills and self-efficacy to increase contraceptive use beyond the increases observed in the comparison group. In addition, interventions might need to address other aspects of relationship power, not just contraceptive decision making (Pulerwitz et al 2000, Wingood and DiClemente 1998, Zamboni et al 2000).
Finally, something in the social environment may have contributed to changes in both groups. Participants may have been exposed to information related to unintended pregnancy in the media or other interventions. If so, those programs may have affected participants in both groups.
Possible Mechanisms of Effects and Other Relevant Findings
The improvements in contraceptive use and in participation in decision making in both groups, however, may suggest that even a short educational intervention with couples (such as the comparison session) prompts changes that lead to safer sexual behaviors (e.g., Becker, 1996). Perhaps the information provided to both members of the couple helped couples who had already made a decision to delay pregnancy, as many couples in our sample may have done. Showing and describing how to use different contraceptive methods may have provided the information that Becker and Robinson (1998) suggest is needed for shared contraception (e.g., aware of methods, brand, dosage, and backup methods) or doing so may have increased skills and self-efficacy for contraception. Alternatively, both conditions may have given couples who had already decided to delay a pregnancy the opportunity to support each other (e.g., men could remind their partners to take the pill).
Unfortunately, we could not include additional conditions in our study and, thus, we cannot say whether a shorter or a longer couples intervention is effective. However, research is beginning to show that intervening with couples and addressing relationship factors is effective for some groups. El-Bassel et al. (2003) found that, compared to a standard educational intervention for women only, their interventions for women only or for couples that addressed relationship factors reduced risk behavior among 18- to 55-year-olds sampled from outpatient clinics in New York. Compared with the women in our sample, the women in their sample appeared to be older (e.g., <10% were <25 years old), less likely to be employed, less likely to have ever been married, and more likely to be HIV infected. These differences in findings may suggest that intensive interventions focused on relationship factors might be more important for less stable couples or more “at-risk” couples than we recruited for our study.
In conclusion, the absence of an effect of the PARTNERS intervention on contraceptive use is an important null finding. The differences in intensity, time, and resources devoted to the 2 conditions were substantial. This finding raises important research questions. Questions pertain more broadly to methods to enhance typical use (e.g., only half of the women were consistently using an effective method by the 6-month interview) as well as methods that provide dual protection. Our findings also raise questions about interventions to promote behavior change. Future research should also identify conditions under which couples interventions might be effective by exploring the optimum mix of individual psychosocial and relationship factors to address for different types of couples, whether to explicitly address relationship factors, and whether to target individuals or couples.
First, behavioral studies should identify the mix of individual psychosocial and relationship factors to address in interventions for different types of couples. Much research focuses on how relationship type (e.g., primary versus other) influences risk behavior. Little research, however, explores whether different influences on behavior are more or less important for different types of couples. The null findings from our study suggest variation within primary partnerships. For example, more stable couples may just need more information, skills, and self-efficacy building exercises or an opportunity to talk about contraception, something that even our short comparison session provided. Less stable or less motivated couples than we reached may require more attention to a broader range of psychosocial and relationship factors, focusing on risk perception, positive expectations, and norms earlier in interventions and skills and support for change later in interventions.
Once the mix of influencing factors is better understood, questions about how to address them for different subgroups of couples should be addressed. For example, although addressing relationship factors may be effective when targeting women only or couples (El-Bassel et al., 2003), some couples may need both separate and joint intervention sessions to meet their needs. Perhaps less stable couples may need joint sessions that address psychosocial and relationship factors whereas more stable couples could be well-served by individually targeted interventions or shorter interventions for couples.
Finally, intervention research should address questions about recruitment and study design. Studies should identify alternative ways to recruit and retain couples. El-Bassel et al. (2003) relied more heavily on primary care facilities and appeared to have recruited different types of couples than we did. Thus, future studies might consider a different mix of recruitment settings. Study designs may need to be more complex and measure a larger number of psychosocial variables (e.g., self-efficacy for all types of contraceptives). Designs might need to include an individually targeted standard of care arm as well as arms for individuals and couples that address individual psychosocial and/or relationship factors. Investments in such complex trials requires a more complete understanding of the role of individual psychosocial and relationship factors for different types of couples.
Acknowledgments
Members of the PARTNERS Project include S. Marie Harvey, Principal Investigator for the Los Angeles (CA) site; Heather C. Huszti, Principal Investigator for the Oklahoma City (OK) site; and Christine Galavotti, Katina A. Pappas-DeLuca, and Joan Marie Kraft, CDC Project Officers. The authors thank Stephen G. West, Mary Gerend, and Aaron Taylor for their statistical consultation. The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the Centers for Disease Control and Prevention.
Appendix I
Items Used to Create Measures
Psychosocial and relationship factorsReferences
- . A test of major assumptions about behavior change: A comprehensive look at the effects of passive and active HIV-prevention interventions since the beginning of the epidemic. Psychological Bulletin. 2005;131:865–897
- . Couples and reproductive health: A review of couple studies. Studies in Family Planning. 1996;27:291–306
- . Reproductive health care: Services oriented to couples. International Journal of Gynecology and Obstetrics. 1998;61:275–281
- . Women at risk of HIV/STD: The importance of male partners as barriers to condom use. AIDS and Behavior. 1998;2:75–85
- . HIV/AIDS Surveillance Report. Atlanta: US Department of Health and Human Services, Centers for Disease Control and Prevention; 1999;[11, no. 2, pp. 1–44]
- . Sexually Transmitted Disease Surveillance, 2002. Atlanta: US Department of Health and Human Services, Centers for Disease Control and Prevention; 2003;
- . HIV/AIDS Semi-Annual Surveillance Survey, 2003. 2003;Available from http://lapublichealth.org. Accessed MONTH DD, YYYY.
- The efficacy of a relationship-based HIV/STD prevention program for heterosexual couples. American Journal of Public Health. 2003;93(6):963–969
- . Disparities in rates of unintended pregnancy in the United States, 1994 and 2001. Perspectives on Sexual and Reproductive Health. 2006;38(2):90–96
- . The role of theory in HIV prevention. AIDS Care. 2000;12(3):273–278
- . Men’s perceptions of their roles and responsibilities regarding sex, contraception and childbearing. Family Planning Perspectives. 1996;28(5):221–226
- . Couple dynamics in sexual and reproductive decision-making among Mexican immigrants. In: Severy LJ, Miller W editor. Advances in population: Psychosocial perspectives. Vol. 3:London: Jessica Kingsley Publishers; 1999;p. 251–279
- A conceptual model of women’s condom use intentions: Integrating intrapersonal and relationship factors. AIDS Care. 2006;18:698–709
- . Relationship power, sexual decision-making and condom use among women at risk for HIV/STDs. Women & Health. 2002;36(4):69–84
- A randomized study of a pregnancy and disease prevention intervention for Hispanic couples. Perspectives on Sexual and Reproductive Health. 2004;36(4):162–169
- . The essentials of contraceptive technology. Baltimore: Population Information Program, Center for Communication Programs, The Johns Hopkins School of Public Health; 1997;
- . Close relationships and elevated HIV risk behavior: Evidence and possible underlying psychological processes. Review of General Psychology. 1997;1(1):72–107
- Nova Research Company. (2003). Questionnaire Development System (QDS) Manual (Version 2.1). Bethesda, MD. Available from www.novaresearch.com. Accessed October 12, 2006.
- . Measuring sexual relationship power in HIV/STD research. Sex Roles. 2000;42:637–660
- . Psychosocial correlates of heterosexual condom use: A meta-analysis. Psychological Bulletin. 1999;125(1):90–132
- . Relationship dynamics, ethnicity and condom use among low-income women. Family Planning Perspectives. 2000;32:82–88101
- . County and city data book: A statistical abstract supplement 2000. Washington, DC: US Government Printing Office; 2001;
- . Gender-related correlates and predictors of consistent condom use among young adult African-American women: A prospective analysis. International Journal of STD & AIDS. 1998;9:139–145
- . Examining communication and assertiveness as predictors of condom use: Implications for HIV prevention. AIDS Education and Prevention. 2000;12(6):492–504
Joan Marie Kraft, PhD, is a Behavioral Scientist in the Women’s Health and Fertility Branch in the Division of Reproductive Health, U.S. Centers for Disease Control and Prevention. Her research focuses on developing and evaluating interventions to reduce unintended pregnancy, HIV and STDs.
S. Marie Harvey, DrPH, MPH, is Chair and Professor of the Department of Public Health, Oregon State University. She brings a social and behavioral science perspective to research on the factors that influence women’s sexual and reproductive health. Dr. Harvey conducts research on the prevention of unintended pregnancy and HIV/STDs among high-risk women, men and couples and the influence of relationship and contextual factors on sexual risk-taking. She uses research findings to inform policies and practices that will improve the public’s health.
Sheryl Thorburn, PhD, MPH, is an Associate Professor in the Department of Public Health, Oregon State University. Dr. Thorburn’s research interests include social and cultural influences on sexual and reproductive health, including HIV prevention behavior; acceptability of reproductive technologies and HIV prevention methods; and discrimination in health care, related attitudes and beliefs, and their effects on the health-related behaviors and outcomes of disadvantaged groups.
Jillian T. Henderson, PhD, MPH, is an Assistant Professor in the Department of Obstetrics, Gynecology and Reproductive Sciences at the University of California, San Francisco. Her research focuses on contraception and access to reproductive health care.
Sam F. Posner, PhD, joined the Centers for Disease Control and Prevention in 1998. He is the Associate Director for Science for the Division of Reproductive Health (DRH). As the Senior Scientist for DRH he reviews all research and programmatic activities in the Division. His research focuses on prevention of HIV, STD, and unintended pregnancy in domestic and international settings, with a focus on acceptability of barrier methods of protection for women at risk of HIV/STD and unintended pregnancy.
Christine Galavotti, PhD, is Chief of the Applied Sciences Branch in the Division of Reproductive Health, U.S. Centers for Disease Control and Prevention (CDC). She joined CDC in 1988. She focuses on designing and evaluating behavioral interventions to prevent unintended pregnancy, HIV and STDs among persons at risk.
Conducted as part of the PARTNERS Project, which was supported by cooperative agreements #U30/CCU 915062-1-0 and #U30/CCU 615166-1-0 with the U.S. Centers for Disease Control and Prevention (CDC).
PII: S1049-3867(06)00116-2
doi:10.1016/j.whi.2006.10.006
© 2007 Jacobs Institute of Women’s Health. Published by Elsevier Inc. All rights reserved.
