HPV Vaccination: An Investigation of Physician Reminders and Recommendation Scripts

Study Rationale:

Human papillomavirus (HPV) vaccine uptake among females and males remains unacceptably low. Health care provider recommendation of HPV vaccine is a major driver of vaccine acceptance. Conversely, the failure of health care providers to recommend HPV vaccine has repeatedly been identified as the primary factor in non-vaccination. Among many recommendations for increasing HPV vaccination rates are several that specifically target health care providers. Two approaches are implementation of electronic health records-based decision support systems and interventions designed to help physicians recommend HPV vaccine in a routine, matter-of-fact manner that addresses uneasiness in discussing sexual activity with parents of adolescents.

A randomized, 3-arm interventional study is proposed.

The 3 arms (randomized at the level of health care provider) will be:

1. usual practice;
2. automated reminders to recommend HPV vaccine for eligible male and female adolescents during clinic visits;
3. automated reminders PLUS a suggested recommendation script.

Here is a draft of the recommendation script that the study team is proposing for the third arm of the study: "Three vaccines are recommended for <first name>, meningococcal to prevent meningitis, HPV to prevent cancer, and Tdap to prevent tetanus. All three are recommended at this age". The script draft was based on the principles of Diffusion Theory and on theories of learning and memory.

Diffusion Theory predicts that adoption of an innovation (in this case the recommendation script) is predicted by three central factors:

1. The innovation should be seen as superior to the approach it is intended to replace (ADVANTAGE);
2. The innovation should be viewed as consistent with pre-existing beliefs and experiences (COMPATIBILITY); and
3. The innovation should be seen as easy to understand and implement (COMPLEXITY).

It is believed that the proposed script meets the ADVANTAGE criterion in that it provides a simple, straightforward way of introducing the HPV vaccine into conversations with adolescents and their parents, an issue with which some physicians struggle. The script meets the COMPATIBILITY criterion in that it makes the HPV vaccine recommendation consistent with the way other vaccines are typically recommended in conversations with adolescents and their parents. Finally, the proposed script meets the COMPLEXITY criterion in that it is simple and will likely be time-saving in most cases. Furthermore, the existing electronic health record (EHR) infrastructure at the target clinics will support easy introduction of the recommendation and script into clinical workflow.

Research in learning and memory has shown that information presented first or last in a sequence has the greatest salience and is therefore most likely to be remembered. In our suggested script, HPV vaccine was purposely in the middle, between Tdap and meningococcal vaccines to decrease the salience of HPV vaccination, thereby helping physicians to treat the three vaccines in a consistent manner. When HPV vaccine is mentioned last, there may be a tendency to hesitate or to treat it as separate from the other recommended vaccines.

Approximately 30 pediatric health care providers will be randomized across the 3 arms in equal numbers. The primary outcome of interest, HPV vaccination, will be evaluated as a patient outcome nested within provider. Physician acceptance of the recommendations, comparing changes in vaccination rates with qualitative feedback from providers will be further assessed. To gather qualitative feedback, 3-4 pediatric health care providers in each clinic will be interviewed. Participants will be purposely selected based on provider demographics, including gender, years in practice, and role within the clinic organizational structure. The interviews will be guided by the Consolidated Framework for Implementation Research (CFIR), a theory and evidence-based framework that represents the accumulated results of over 50 years of research on implementation and diffusion. Mixed methods will enable the team to examine not only impact on vaccination rates but also the characteristics of the reminder (e.g., usability, support of workflow), its implementation, and the context of setting in which it was implemented (e.g., clinical structure).

Aims and Objectives Aim 1. To evaluate the effects of physician-targeted automated HPV vaccination reminders alone and reminders plus a recommended script on HPV vaccination rates among 9-12 year old male and female patients.

Hypothesis. It is expected that a linear increase in HPV vaccination rates across the usual practice control group and the two intervention conditions, such that the reminder + script group will have the highest first dose HPV vaccination rates, followed by the reminder-only group will be noted.

Secondary objective: To evaluate rates of return for second dose. Exploratory objective: To evaluate experiences of pediatric health care providers with the reminder and script interventions.

Materials and Methods:

The Child Health Improvement through Computer Automation (CHICA) system, described in detail above, has already been developed and has been in operation since November of 2004. The functionality is well documented. New components of the system will have to be developed for this study. The first of these will be the activation of CHICA's ability to randomize its alerts by provider. The second will be the development of the Arden Syntax rules (Medical Logic Modules) to query the data from the immunization registry and generate appropriate prompts and reminders to the clinician. Once written, these rules must be encoded in CHICA and extensively tested with clinical data before they are deployed. Third, the study team will develop a mechanism by which CHICA can query the immunization registry at some time point after the clinic visit to determine whether an HPV vaccine was administered.

Include potential difficulties and limitations of the available methods and strategies for overcoming those methods. Randomization by physician depends on having accurate data from the clinic's registration system about which physician the patient will see. The study team has used this information in the past without significant problems. However, if it is noted that the data are missing or unreliable in a significant number of cases, the study team can ask for the information to be entered into the system by clinic personnel prior to the visit.

Programming of Arden rules is part and parcel of operating the CHICA system. So the study team does not anticipate problems with developing and testing these. The study team closely monitor the display and response to CHICA reminders. If there is a problem, the study team will be able to detect and correct it quickly.

The ability to query the CHIRP immunization registry by Health Level Seven International (HL7) messaging has recently been developed. The responses from CHIRP are not perfect, but are well over 90% reliable. The study team will build our system for querying CHIRP for HPV data to track and reissue failed queries in order to maximize the available data.

The location of the study sites. As previously noted, the study will be conducted in the five clinics in the Eskenazi Health System that currently use CHICA. They are: Eskenazi Outpatient Care Center, Blackburn, Pecar, Eskenazi 38th St., and Forest Manor.

Outcome measurement. The outcome of primary interest, HPV vaccine uptake, will be recorded by having the CHICA system re-query the CHIRP system after each visit. The study team will record whether HPV vaccine was administered. The CHICA system will be programmed to track patients who are enrolled in the study. Between 7 and 30 days after the clinic visit, CHICA will send an HL7 query for vaccination (VXQ) message for each patient to the CHIRP immunization registry of the Indiana State Department of Health. In response to this query message, CHIRP will send an HL7 vaccination record response (VXR) message with the patient's immunization record. Because all of the CHICA clinics report their immunization data to CHIRP, this will be a reliable way to assess whether the patient received an HPV vaccine.

Predictors of HPV vaccine uptake. The study team will use the re-query of CHIRP to determine which other vaccines (e.g., meningococcal; Tdap; influenza) were administered. Other variables of interest that will be recorded will be provider seen, patient age, and patient gender.

The 5 CHICA clinic study sites are located throughout metropolitan Indianapolis. These clinics serve largely low income (70% Medicaid) and minority (33% Hispanic, 49% African American) children. CHICA has provided data to over 37,000 children since first implemented in 2004. As indicated in Part 1 (Introduction), these CHICA clinics have been used successfully for multiple pediatric health service intervention studies. The real-world implementation being used in this study is a great strength. It will enable us to evaluate how a reminder system and recommended script influence HPV vaccination practices in busy pediatric clinics. An inherent limitation with this kind of study is that the study team is limited in the kind of data that can be collected. The study team cannot, for instance, collect attitude/knowledge information from parents. However, it is believed that the strengths of this research methodology far outweigh the limitations.

Qualitative interviews. Interviews will be digitally recorded and transcribed. Interviewers will be asked to record field notes related to the interview to capture information not included in the transcripts (e.g., tone of voice). Digital recordings will be transferred to a password-protected computer, in a locked office. Recordings and transcripts will be saved for a 3-year period from the end of the study. At that point in time, digital recordings and transcripts will be permanently erased. Content analysis will enable investigators to identify patterns in respondents' answers. The study team will specifically look for barriers to implementation and use of the recommendations and scripts as well as contextual aspects that facilitated routine use of the recommendations and scripts to improve adherence to recommended HPV vaccination guidelines.