An Intervention Promoting HPV Vaccination in Safety-net Clinics

The primary research goal is to implement and evaluate evidence-based intervention strategies targeted to adolescent girls and their parents that stimulate patient demand for the human papillomavirus (HPV) vaccine and compliment provider-directed strategies already employed by Parkland clinics.

The project has three specific aims:

• Aim 1. Determine if a mailed packet with an invitation letter and educational material (HPV-specific or general brochure about all recommended adolescent vaccines) delivered prior to clinic visit, increases 1 dose coverage of the HPV vaccine series.
• Aim 2. Determine if a telephone recall delivered 2 weeks after the clinic visit, increases 1 dose coverage among parent-patient dyads who decline the vaccine during the clinic visit.
• Aim 3. Determine if telephone reminders to those overdue for Dose 2 or 3 of the HPV vaccine series, increase 3 dose coverage of the series.

Phase 1 - Cognitive interviews: To ensure the usability and cultural appropriateness of all project materials, investigators will conduct 40 in-person, cognitive interviews with parents of female patients ages 11-18 attending Parkland clinics.

Phase 2 - Randomized controlled trial. Investigators will select and invite 1200 families whose daughter has an upcoming appointment at one of four Parkland clinic sites. Eligible patients will be randomized to either the Intervention group or the General Adolescent Vaccine Brochure group. The Intervention group participants will be mailed a packet containing an invitation letter to join the study and HPV vaccine-specific educational materials; General Adolescent Vaccine Brochure group participants will be mailed a packet with educational materials on all four of the recommended adolescent vaccines (HPV, meningococcal, TDAP, and influenza). A few days after the packet has been sent, a research assistant will call participants to verify that the letter was received and assess interest in completing a pre-clinic visit survey.

Participants who agree to fill out the in-clinic survey will be asked to come to their clinic appointment 40 minutes early to obtain verbal informed consent from the parent and verbal assent from the adolescent female patient. Then, the parent will complete a 20-30 minute survey about the HPV vaccine and interactions with their daughter's provider. The daughter will complete a short survey that asks about use of social media and health concerns.

Intervention group participants who complete the in-clinic survey are also eligible to receive telephone recalls and reminders. If the Intervention group parent refused the vaccine during the initial clinic visit, a telephone recall will be delivered after 2 weeks offering to schedule an immunization-only appointment. If the Intervention group participant misses the 2nd and/or 3rd doses, telephone reminders will be delivered. For participants who complete the in-clinic survey and are randomized to the General Adolescent Vaccine Brochure group, parents who refused will be contacted after 2 weeks to assess attitudes and intentions toward the HPV vaccine.

Participants in both the Intervention and General Adolescent Vaccine Brochure group will have their medical record reviewed to assess HPV vaccine outcomes, receipt of the other three recommended adolescent vaccines, and contact with the clinic during the 12 months following randomization.

Ancillary Analyses

The investigators will perform stratified analyses and multivariate logistic regression modeling to examine whether certain subgroups were more likely initiate and complete the series (e.g, vaccine uptake might vary by characteristics such as preferred language, race/ethnicity, patient age, or parent educational status). The Investigators will use similar methods to also examine our data for confounding, for example, if initiation differs by clinic site. If no interactions are found (e.g., outcome estimates are similar for different subgroups), the investigators will report summary chi-square tests that only adjust for potential confounders.

The investigators will also use the dates that patients received the 1st, 2nd, and 3rd HPV doses to run a time-to-recurrent event analysis. The investigators will model the completion of the 3-dose series as recurrent event data, 72; 73 with event times from intervention to Dose 1, Dose 1 to 2, and Dose 2 to 3. A proportional means model will be constructed with the outcome being the recurrent events, and the variables of interest including intervention group assignment, sociodemographic, clinic site, knowledge, and attitudes. A significant intervention indicator effect will conclude that the intervention generally shortens the time between initiation and receipt of all doses.

Secondary Analyses

The investigators will conduct the following secondary analyses to evaluate whether parents who receive the intervention have greater HPV vaccine knowledge, information seeking, and PAPM stage progression compared to parents receiving usual care. Among those who consented for additional contact and completed pre-clinic and follow-up surveys, the investigators will explore associations among intervention assignment and intermediate factors including parents' HPV vaccine knowledge, attitudes, perceived risk, information-seeking, and HPV vaccine decisional stage. Descriptive statistics will be calculated and outliers identified. A correlation matrix will be run to check for multicollinearity (strong correlations between intermediate factors). The investigators will run univariate and multivariate logistic regression models to determine whether intervention assignment is associated with greater knowledge of the HPV vaccine, more positive attitudes toward and information seeking about the HPV vaccine, and PAPM stage progression.

To explore generalizability to all safety-net patients in the clinics, the investigators will compute t tests or chi-square tests and examine whether patients who 1) decline to participate or 2) do not consent for reminder and recall intervention components differ from those randomized with respect to variables available in the EMR (patient demographics, health care utilization patterns, clinic visit characteristics).