Association of Patients Reading Clinical Notes With Perception of Medication Adherence Among Persons With Serious Mental Illness

This survey study examines how patients with a mental illness diagnosis who read at least 1 clinical note in the last year perceived its association with their medication adherence.


Introduction
Starting April 5, 2021, new US federal rules will mandate that all clinicians (including physicians, physician assistants, and nurse practitioners) must share clinical notes with patients via online health portals with few exceptions. 1 Surveys show that clinicians worry that patients with mental health diagnoses will become anxious, confused, or upset after reading their visit notes. 2 In this study, we examined how patients with a mental illness diagnosis who read at least 1 clinical note in the last 12 months perceived how reading the note affected their adherence to prescribed medication.

Methods
We reanalyzed data from a web-based survey of patient experiences with access to their outpatient visit notes. 3 Data were collected in 2017 from adult patients in 3 diverse health systems in the US, who had accessed at least 1 ambulatory note in the previous 12 months. All participants gave written informed consent. The institutional review boards at Beth Israel Deaconess Medical Center in Boston, Massachusetts, Geisinger Health System in Pennsylvania, and the University of Washington Medicine in Seattle approved the survey and study protocol at their respective sites. The American Association for Public Opinion Research (AAPOR) reporting guideline was used in this survey study.
Of the 136 815 patients who received a survey invitation, 29 656 responded for a response rate of 22% using the AAPOR guidelines. 4 We excluded all participants who did not report taking or being prescribed medication in the 12 months before the survey. While there is no consensus or federal definition of serious mental illness (SMI) in the US, the term is often used to describe major depression, bipolar, and schizophrenia-related disorders. 5 Using the International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10), we classified each respondent as having (1) SMI (major depressive disorder, schizophrenia, schizoaffective disorder, or bipolar-related disorder [ICD-10 codes F20.0-F29 and 31.0-F33.9]), (2) other mental illness (ICD-10 codes F00-F99, except for those listed earlier), or (3) no mental illness. We used a 2-sided t test, χ 2 test, and Fisher exact test for statistical analyses, and significance was set at P < .001. All analyses were completed using SAS software version 9.4 (SAS Institute) and performed between August 2020 and January 2021.

Results
In the sample of 29 656 respondents, 23 576 (79%) had read at least 1 note, and among them, 19 411 (82%) were taking or had been prescribed a medication in the past 12 months. Diagnostic codes were missing for 31 participants, leaving an analytic sample of 19 380; among the respondents 12 112 (62%) were women and 14 775 (76%) were aged 45 years or older.  a This question was only asked at 2 of 3 health systems (Beth Israel Deaconess Medical Center, Boston, and Geisinger Health System, Pennsylvania); therefore, the proportion was calculated using the denominator from only these health systems. All other questions were asked at all 3 health systems, which included the University of Washington Medicine, Seattle.
b P value for this figure was calculated with Fisher exact test because of the reduced sample size for cell frequencies, which were not indicated in Table 2 of less than 5. Two of the patients with severe mental health conditions responded that they were less likely to take medications as prescribed, and this category was not included in the