Novel Interventions and Assessments Using Patient Portals in Adolescent Research: Confidential Survey Study

While adolescents can receive confidential health care without parental or guardian notification, they are rarely asked about their experiences and opinions regarding their care because participation in research often requires parental consent. Anonymous research with adolescents via confidential patient portals may ameliorate this research gap.

Because use of a confidential online adolescent patient portal is high at our academic institution, we hypothesized that adolescents would also respond to survey-based research via the portal, especially if asked anonymously and without parental consent. We used a clinical scenario of needing to better understanding adolescent and young adults' views about their health and health care, including information on a long-acting reversible contraceptive (LARC) to test if and how they will use a portal for research.

Upon receiving Institutional Review Board approval, we sent 2 portal-based surveys about confidential services to 2 groups of females, ages 14 to 25 years, who had attended an adolescent clinic in the past 3 years. This clinic mostly serves Medicaid recipients (80%) and is racially and ethnically diverse with half of patients identifying as African American and roughly 10% Hispanic. The control group was a random sample of female patients who never received a LARC (n=150) and the intervention group included all female adolescents who had received a LARC from the same clinic (n=107). This second sample was manually cross-checked to confirm they had an office visit for this reason. Consenting for themselves, the control group received an email through the patient portal with a link and a request to perform an assessment. The survey for the control group included items assessing health literacy and health communication preferences. The survey for the intervention group included health literacy items as well as items to assess their opinions and perceptions regarding LARCs. We tracked click-through rates and opened messages; each participant received 4 reminders.

While only 3 participants fully completed either survey, email read rates (29/107 [27.1%] of LARC recipients and 39/150 [26.0%] of controls) were encouraging. Additionally, of those who opened the messages, almost twice as many of the LARC recipients (10/107 [9.3%]) read through the entire survey, while less than half read the entire survey as compared to those who received the survey asking about health literacy and health care preferences (6/150 [4.0%]).

The methodology of using adolescent portals for online surveys provides a new avenue for research even though the study did not yield sufficient participation to understand these adolescents' preferences. Future studies need to test if a different survey topic would engage adolescents or if other methods like text-based reminders would improve participation.