Interactive Web-based Sexual Health Literacy Program and Safe Sex Practice in University Students

Subject recruitment procedures:

Recruitment of participants will be conducted by using bulk electronic mails, promotional brochures, posters and campus booths. An invitation card with a QR code of the web-based intervention will be used for inviting female university students to enrol the study. Online enrolment will be done for screening participants for eligibility. From the investigators' past experience in recruiting university students, the investigators shall be able to reach over 30,000 female university students in five universities.

Participants will be asked to complete the web-based baseline survey at enrolment. The web-based questionnaire will be set up under an online platform. Participants may use their mobile devices including laptop computers and smartphones to complete the questionnaire. Participation is entirely voluntary. Written informed consent will be obtained via the web-based platform before completing the baseline questionnaire.

After completion of the questionnaire, the recruited students will then be randomized to either intervention group or control group, according to a list prepared by blocked randomization of a size decided by a randomizer. The block size and order of allocation will be kept securely in the randomizer to avoid selection bias. Allocation concealment will be done at the online platform according to the participants' enrolment sequence. The participants will then be guided to a web-page according to their group allocation.

At 3-month and 6-month after the entry of the study, participants will receive a link through their e-mail to complete a follow-up questionnaire (T2). Three months later, another e-mail link will be sent to participants for completing the last follow-up questionnaire (T3). If no response from a participant, a follow-up telephone text-message and/or telephone call will be sent after 1 week as a reminder.

Sample size assessment:

Female university students (n = 500), across disciplines and year of study, will be recruited from five universities with dormitory or residential halls in Hong Kong. The sample size calculation is based on the primary comparison of behavioural change in consistency of condom use. A previous study showed the mean percentage of condom use was 67.8% (SD 39.3%) in people under an ICBI and 23.5% (SD 35.3%) in those under usual care, corresponding to a moderate to large Cohen effect size of 0.6.(20) To detect a conservative small Cohen effect size of 0.3 with 80% power and maximum 5% false positive error rate by a two-sided two-sample t-test, the investigators need a total of 352 (176 per group) female university students. Assuming 30% attrition based on a previous study using web-based intervention (21), the investigators plan to recruit a total of 500 female university students.

Quality assurance plan:

The web-based intervention is password-protected and "Pseudo-Codes" will be used to ensure anonymity and confidentiality to minimize social desirability of the responses. By asking the participant to create his/her own "pseudo-code", a code that can be accurately retyped each time the participant completes pre-test and post-test. There will otherwise no other vulnerable subjects involve in the study.

Data check and analysis method:

To assess the efficacy of ICBI in increasing consistency of condom use, a linear mixed effects model will be adopted with intervention group and baseline consistency of condom use as the covariates. The group by time interaction shall be assessed for changes of ICBI effects over time. When the interaction effect is insignificant, it will be removed from the model and the overall ICBI will be estimated. Otherwise, liner contrasts will be used to assess the ICBI effect at each time epoch. In case of baseline imbalance, additional analysis with adjustment of variables that show imbalance at baseline will be performed. If the data will not normally distributed, the corresponding analyses methods, such as zero/one inflated beta (ZOIB) regression model, will be applied based on the data characteristics.

The intention-to-treat principle will be adopted and all study subjects will be included in the analysis. Missing values at 3- and 6-month follow-up will be replaced by the last observed value and multiple imputation. Missing values at baseline will be replaced by values from 3-month or 6-month follow-up. If there will be no value obtained at any of the 3 time points, the participant will be excluded from the analysis. For other outcomes, the t test will be applied for continuous data and the chi-square test will be applied for categorical data. The analysis will be performed by using the R software (version 3.6.1; R Core Team). Each estimated effect will be accompanied by a 95% confidence interval and 5% level of significance. Questions will be downloaded from the inquiry system. Content analysis will be performed to categorise the collected responses.

Data collection and management:

The investigators and the research assistants of the research team are responsible for data collection and they will be permitted to access to source data and study record. Daily logs will be recorded by the research assistants to monitor the study progress including the number of respondents approached, interviews completed or refused and incomplete interviews. All data entered in the database will be verified and cleaned. After completion of data entry, computer logic checks will be run for consistency of related code. Necessary corrections will be made to the database.