Online Peer Influence on IUD Behaviors and Attitudes

Study Design:

University of Pennsylvania researchers will conduct a randomized controlled trial to examine whether there is an association between social exposure to women who use intrauterine contraception (IUC) as their method of birth control and change in IUC attitudes and behavior among women who have never used IUC. The intervention used to test this association will be an online social community, designed by the researchers, in which women will be invited to share their personal experiences with birth control methods they have used. Non-IUC user participants will be randomized to either the intervention or control conditions. The intervention condition will consist of small online social groups in which non-IUC user participants will interact with IUC user participants. The control condition will consist of online groups identical to intervention groups, except that no IUC user participants will be in these groups. Participants will interact within small groups for twelve days each. Non-IUC user participants will complete surveys before and after the twelve-day intervention that measure IUC attitudes and behavior. UCSF researchers will conduct analysis on deidentified data obtained from this study.

Hypothesis:

It is hypothesized that among women who have never used IUC, there is an association between social exposure to IUC users and change in attitudes and behavior related to IUC use. It is also hypothesized that exposure to IUC users is associated with positive change in knowledge of IUC use and efficacy and positive change in IUC-related information-seeking.

Aims:

The study has three aims:

Aim 1: To determine whether exposure to IUC users in an online community is associated with change in women's attitudes and behavior related to IUC use.

Aim 2: To determine whether an association between exposure to IUC users and change in women's IUC attitudes and behavior varies by women's race/ethnicity.

Aim 3: To observe the content of women's interactions about IUC and other birth control methods in an online social community.

Data analysis:

In order to address Aim 1, we will first conduct univariate analysis and obtain descriptive statistics (including mean, median, mode, variance, and range) on demographic characteristics and IUC attitudes and behaviors among participants in the intervention and control groups. In bivariate analysis, we will use chi-square tests to examine the relationship between study condition and IUC-related behaviors, as measured on the post-survey. We will use t-tests to examine relationships between condition and post-survey IUC attitudes, and matched t-tests to examine the relationships between conditions and changes in IUC attitudes. If data on attitudes is not normally distributed, we will use the appropriate non-parametric tests instead of t-tests.

We will conduct multivariate analyses for those outcomes associated with intervention condition in bivariate tests. Multivariate analysis for Aim 1 will include mixed-effects regression to examine associations at both the individual and cluster levels. We will use multiple linear regression for continuous outcomes and multiple logistic regression for binary outcomes. Regression models will test associations between intervention and control conditions and outcome IUC attitude and behavior variables. We will obtain regression coefficients from linear regression models and odds ratios from logistic regression models in order to describe the associations between study arms and outcomes.

In order to address Aim 2, we will examine bivariate relationships between the racial/ethnic categories of participants and outcome attitude and behavior variables, using chi-square tests for binary outcomes and t-tests or appropriate non-parametric tests for continuous outcomes. We will use multiple linear and/or logistic regression to test possible associations between race/ethnicity and outcome variables, controlling for other background variables. Resulting regression coefficients and odds ratios will describe whether the effect of the intervention was greater for women in some racial/ethnic groups as compared to others.

In order to address Aim 3, we will perform qualitative analysis of conversations that take place among group members in the online social groups, using grounded theory methodology. We will examine qualitative data from both the intervention and control groups.

Privacy:

Recruitment and participation in the study will occur online. Participants will only interact with the study from remote locations, so they will not be seen entering or leaving any location affiliated with the study, or interacting with any study personnel. Once registered in the study, participants will only be contacted via secure email that does not indicate the nature of the study in the subject line. The study website will be password-protected to prohibit access to Internet users not participating in the study. Participants' screen names on the site will not contain their real names, and they will not be identifiable to other study participants.

There is minimal risk of privacy loss. Study participants may not want others to know about their interest in or use of birth control, and may have feelings of embarrassment or shame in the event of a loss of study privacy. Additionally, birth control use may be a sensitive topic in the personal relationships of some study participants, and a loss of privacy could lead to strain or conflict in these personal relationships.

Confidentiality:

Data collected by researchers at the University of Pennsylvania will be stored on a secure network at their site. Only University of Pennsylvania study staff will have access to data on this network. University of PEnnsylvania staff will remove all identifiers from the data and export the data to researchers at UCSF via secure email. Deidentified data will be stored at on a secure UCSF server. Only the study staff will have access to the data, and we will not have access to the key.