Study design and sample
A retrospective chart review was conducted using the data of patients discharged with a diagnosis of HF (International Classification of Disease-10 codes 150.1–150.4, 150.9) from an academic medical center in Korea between January 2014 and March 2017. Eligible patients were admitted due to HF exacerbation and were discharged alive. We initially screened 818 records and excluded 516 because their primary reason for the index hospitalization was not related to HF or because they died during the index hospitalization. Of the 302 records retained, 101 were excluded as they did not include data on variables of interest (e.g., living status, post-discharge clinic appointments). As a result, 201 records were included in the study.
Data collection
Electronic medical records were reviewed to abstract the following information from the index hospitalization: demographic and clinical information and information related to the discharge process. To identify patients’ prognosis, their electronic medical records were reviewed for a median of 699 days (interquartile range: 441–789) after the index hospitalization. A data collection protocol was developed to consistently abstract the data. Any ambiguity among the research nurses was discussed by a principal investigator.
Information related to the discharge process. The discharge process factors included in this study were post-discharge clinic appointments, patient education for HF self-care, and family participation in patient education. According to a scientific statement from the American Heart Association (AHA), systematic implementation of the principles of transitional care programs to prevent rehospitalizations was recommended for at-risk patients with HF [8]. The suggested components of the transition of care include medication reconciliation, early post-discharge contact and communication with patients and/or clinicians, patient education related to self-care, interdisciplinary care coordination, and communication of patient health information to patients and their post-discharge healthcare provider [8]. However, the systematic approach for the transition of care (e.g., care coordination) had not been implemented in Korea, including in the facility where this study was conducted. At the clinic, patients saw the same clinicians who evaluated them during the index hospitalization, and medication reconciliation was performed before discharge for all patients in Korea. Of the suggested components listed in the AHA scientific statement, information about scheduling post-discharge clinic appointments and patient education on self-care was available in the medical records. The days of the patient’s first clinic visit after discharge were computed. The educational contents about HF before discharge were also abstracted from the medical records and categorized as follows: pathophysiology, medication, low sodium diet, symptom management, and exercise.
The significant role of caregivers in the self-care of patients with HF is well-known [9, 10], and the American Association of Heart Failure Nurses’ position paper clearly states the importance of family involvement in educating patients [11]. Therefore, information about the participation of family caregivers in patient education is also included as one of the discharge process factors. Patients with medical records indicating the participation of family members during patient education were considered to have the participation of family caregivers in their education.
HF event-free survival. The outcome of this study was defined as the composite end point of time to the first HF-related event, which was composed of HF-related death, rehospitalization, or ED visit due to HF exacerbation over 2 years after the index hospitalization. This information was abstracted from patients’ electronic medical records after the index hospitalization was reviewed for up to 2 years.
Demographic and clinical information. For the demographic information, age, sex, education level, and living arrangement (living alone vs. living with others) were collected. For the clinical information, data on left ventricular ejection fraction, etiology of HF, changes in weight between admission and discharge, the number and names of the discharge medications, and comorbid conditions were collected. Comorbidity burden was measured using the Charlson comorbidity index (CCI) [12].
Data analysis
To compare the patients’ demographic and clinical information and information related to the discharge process between patients with and without HF-related events, an independent t-test or chi-square test was conducted, as appropriate.
Multivariable Cox proportional hazards regression was used to examine whether HF event-free survival was predicted by discharge process factors (i.e., days of post-discharge clinic appointment, the number of educational contents about HF before discharge in which the patients received, and the participation of family members during patient education). Covariates included in the multivariable Cox proportional hazards regression model were age, education level, living arrangement, comorbidity burden measured with the CCI, and changes in weight between admission and discharge. All statistical analyses were performed using SPSS, version 20.0, software (IBM SPSS). Statistical significance was set at p < 0.05.