Background: Patients’ confidence managing their health after discharge is essential to effective transitions of care. After Visit Summaries (AVSs) are a standard way of communicating discharge instructions. Past studies proposed solutions to improve discharge instructions and information retention, including simplified information pages, structured discharge letters, and graphic-based discharge information (DeSai et al., Lin et al., Hill et al.). Other studies focused on reading level as the AMA and NIH recommend 6th-7th grade reading levels for patient understanding. One achieved success through templates, provider reminders/tips, and feedback programs (Unaka et al.), while another prospectively edited instructions to reach a desired reading level and found statistically significant improvements in post-discharge phone calls and readmissions (Choudhry et al.). To inform local improvement initiatives focused on patients’ confidence managing their health after discharge, we determined the reading levels of and variability of content within discharge instructions.
Methods: We extracted data from a random selection of discharges sourced from a two month sample of discharged patients cared for by Hospital Medicine providers at a large academic medical center. No service lines or patient populations were excluded. We created a data extraction tool to identify variables from each discharge including discharging provider credentials, patient demographics, binary checks of whether various details were included, and reading level (determined by Flesch Kincaid score). We performed descriptive statistics on all variables. We tested significant differences in outcomes across some independent variables including provider credentials and discharging team.
Results: Two hundred and thirty discharges were reviewed. 19.6% of AVSs lacked provider written discharge instructions. Of the remaining AVSs (188) that had written instructions, 6.9% had Flesch Kincaid scores reflecting a 7th grade reading level or lower. The average score of English instructions was 53.4 (19 – 87), or a 10th grade reading level. In terms of content, diagnoses (95%) and medications (68%) were the most common variables included, whereas diet (5.8%), home monitoring (6.9%), and patient restrictions (6.9%) were the least included. Return precautions were included 17.5% of the time. Of the 13 instructions written for Spanish speaking patients, only 5 were translated into written Spanish. Of other languages, only 2/7 were written in the patient’s preferred language.
Conclusions: There is sizable variability in AVSs at our institution. The AMA and NIH recommend 6th-7th grade reading level in medical writing targeted towards patients, so our rate of 6.9% of meeting a conservative 7th grade level is not acceptable. It was also shown that a sizable proportion of patients were discharged without any instructions. In the future, electronic medical record systems should be leveraged to make certain fields and content mandatory to decrease this variability. A potential solution could use templates. Furthermore, for non-English speaking patients, only 36.8% of instructions were written in the patients’ preferred language. Hopefully standardizing how our institution writes instructions and broadening the languages available will allow patients to feel more empowered in managing their own health, particularly once they leave the hospital.