Access to online patient portals among individuals with depression and anxiety

 Data sample

We analyzed data from the fifth iteration of the National Cancer Institute's Health Information National Trends Survey (HINTS 5). HINTS is a nationally representative survey of the US adult population conducted every 1–3 years since 2003 to collect data related to health information, health behaviors, ScienceDirect's AI-generated Topic Pages" class="topic-link" style="margin: 0px; padding: 0px; text-decoration-line: underline; text-decoration-thickness: 1px; text-decoration-color: rgb(46, 46, 46); color: rgb(46, 46, 46); word-break: break-word; text-underline-offset: 1px;">cancer risk factors, and communication technology. The target sample for the HINTS was non-institutionalized, civilian adults (aged ≥18 years) living in the U.S.

Data used in this study comprised respondents of HINTS 5, Cycle 1 (2017), HINTS 5, Cycle 2 (2018); HINTS 5, Cycle 3 (2019); and HINTS 5, Cycle 4 (2020). Data collection periods data were from January 2017 through May 2017 (HINTS 5, cycle 1), January through May 2018 (HINTS 5, cycle 2), January through April 2019 (HINTS 5, cycle 3), and February through June 2020 (HINTS 5, cycle 4). HINTS 5 cycles 1, 2, and 4 were administered as a single-mode mailed survey. However, HINTS 5 cycle 3 differed from the other iterations by using a mixed-mode—both self-administered mail and web survey to collect participant responses. A total of 13,360, 14,586, 23,430, and 15,347 households received the questionnaires in cycles 1, 2, 3, and 4 respectively, and the response rates were 32.4% for cycle 1, 32.9% for cycle 2, 30.3% for cycle 3, and 36.7% for cycle 4.

HINTS 5, Cycle 1 comprised of 3285 respondents; HINTS 5, cycle 2, had a total of 3504 respondents; HINTS 5, cycle 3, had a total of 5438 respondents; and HINTS 5, cycle 4, had a total of 3865 respondents. All iterations of the HINTS 5 employed a two-stage, stratified random sampling technique. The ScienceDirect's AI-generated Topic Pages" class="topic-link" style="margin: 0px; padding: 0px; text-decoration-line: underline; text-decoration-thickness: 1px; text-decoration-color: rgb(46, 46, 46); color: rgb(46, 46, 46); word-break: break-word; text-underline-offset: 1px;">first stage involves the selection of non-vacant residential addresses obtained from the Marketing Systems Group (MSG). In the ScienceDirect's AI-generated Topic Pages" class="topic-link" style="margin: 0px; padding: 0px; text-decoration-line: underline; text-decoration-thickness: 1px; text-decoration-color: rgb(46, 46, 46); color: rgb(46, 46, 46); word-break: break-word; text-underline-offset: 1px;">second stage, an adult from each household was selected for participation in the survey using the ‘Next Birthday’ method. For HINTS 5, cycle 1, the database of residential addresses was grouped into three strata including addresses in areas with high concentrations of people from racial or ethnic minority groups; addresses in areas with low concentrations of people from racial or ethnic minority groups; and addresses located in counties comprising Central Appalachia, regardless of racial or ethnic minority status. In HINTS 5, cycles 2, 3, and 4 two strata were used and included (1) addresses in areas with a high concentration of people from racial or ethnic minority groups, and (2) addresses in areas with a low concentration of people from racial or ethnic minority groups.

For this study, depression and/or anxiety status was determined from responses to the survey item on depression/anxiety diagnosis, which was worded as: “Has a doctor or other health professional ever told you that you had depression or anxiety disorder (yes/no)?” Participants who responded “Yes” were classified as having depression and anxiety. After combining all four HINTS iterations, a total of 3585 participants self-reported a history of depression and/or anxiety disorder. Of these, 391 participants with missing data for portal usage were excluded. Thus, the final eligible sample consisted of 3194 respondents.

 Exposure variable

The main variable of interest was portal use within the past 12 months preceding the survey. This was determined by asking whether participants have accessed patient portals of ScienceDirect's AI-generated Topic Pages" class="topic-link" style="margin: 0px; padding: 0px; text-decoration-line: underline; text-decoration-thickness: 1px; text-decoration-color: rgb(46, 46, 46); color: rgb(46, 46, 46); word-break: break-word; text-underline-offset: 1px;">EHR in the past 12 months. Participants who responded using patient portals one or more times were categorized as portal users and those reporting no use of the portal in the preceding 12 months were characterized as non-users of patient portals.

Use of patient portals

To better understand the patterns and variations in how people with depression and/or anxiety use their medical records (patient portals), which may give us insight into familiarity with different uses of the portal, we evaluated participants' reported uses of their patient portals. Measures indicating reasons for using the patient portals included viewing test results, medication refills, securely messaging health care providers or staff, adding health information such as side effects to share with a provider, downloading health information to a personal device such as a phone, and using the information to decide on treatment. Additionally, participants perceived usefulness of patient portals was ascertained from one survey item asking if they found patient portals useful. Responses to all these questions were (yes vs no), and they were dichotomized.

 Barriers to portal use

We also sought to assess the reasons for the non-use of patient portals among individuals with depression and or anxiety who reported not using patient portals in the preceding 12 months. HINTS 5 asked nonusers whether any of seven reasons why they had not accessed an online medical record in the past twelve months: no way to access the website (that is, no internet), not having such a record, preferring to speaking directly to a provider, having concerns about the privacy or security of the website, having difficulty logging into their portals, having limited experience with computers, and having no need to access the record. Participants were allowed to select more than one reason. Each of these reasons was categorized as a binary variable (Yes vs No)

Participant characteristics

Informed by previous studies, as well as the Andersen Model of health services utilization, the specific sociodemographic variables and health-related variables included in the present study were age, gender, race, educational level, household income, insurance status, comorbidities, and rural-urban residence. Gender was categorized as male or female. Age was categorized as 18–34 ​y, 35 to 49 ​y, 50 to 64 ​y, and 65 ​y or older. The race was categorized as white, Black or African American, Hispanic and others. Level of education was categorized as a high school graduate or less, some college, and college graduate or postgraduate. Rural or urban residence was defined using the US Department of Agriculture's (USDA) 2003 Rural-Urban Continuum Codes. Codes 1 to 3 were classified as urban, while codes 4 to 9 were classified as rural. Household annual income was categorized as lowest to low income (less than USD 20,000 to 34,999 USD) and middle to high income (USD 35, 000 to USD 75,000 or more). Respondents were classified as having a comorbidity if they had one or more of the following conditions: diabetes mellitus, hypertension, heart disease, or ScienceDirect's AI-generated Topic Pages" class="topic-link" style="margin: 0px; padding: 0px; text-decoration-line: underline; text-decoration-thickness: 1px; text-decoration-color: rgb(46, 46, 46); color: rgb(46, 46, 46); word-break: break-word; text-underline-offset: 1px;">lung disease.