Patients

We recruited a convenience sample of caregivers whose children underwent recent surgical CHD repair. The caregiver patients in this study were defined as either the parents or legal guardians of children discharged after surgical repair of their CHD. Only caregivers who spoke English or Spanish, who had children with two-ventricle physiology, and whose child had a recent surgical procedure defined as within 30 days of discharge were included. We excluded caregivers of children with single ventricle physiology or children discharged >30 days after surgery because those patients are more medically complex and merit separate study.

The clinicians who participated were a convenience sample of healthcare professionals that provide inpatient care to patients with the above-mentioned criteria. These clinicians included physicians (faculty members of the cardiac care unit), registered nurses and advanced practice registered nurses. Clinicians were required to have at least 6 months’ experience to ensure they had a clear understanding of the discharge process.

Patient recruitment

The primary investigator contacted caregivers via telephone 5–7 business days post-discharge and prior to their first follow-up outpatient surgical visit. The primary investigator was not involved in the care of the children of any enrolled caregivers. The first caregiver listed in the electronic medical record was contacted. A second caregiver was contacted if the first did not answer. Only one caregiver per patient was interviewed. Caregivers provided informed consent prior to the interview.

The primary investigator recruited clinicians based on their experience and active involvement with the patient discharge process, approached them during working hours and obtained consent for all patients.

Interviews

The primary research team and experts in the perioperative care of infants and children undergoing congenital heart surgery, including cardiac care unit doctors and surgical advanced practice registered nurses, defined important aspects of care that caregivers should know prior to discharge. The aspects of care identified were: who to call in an emergency, comfort with pain management, medication administration, incision care and comfort with scheduling follow-up appointments. The research team then developed an interview guide (Appendix A) that addressed the following domains:

1. (1) sense of caregiver preparedness and/or discharge readiness,

2. (2) how the discharge process occurred,

3. (3) general gaps in preparedness,

4. (4) caregiver knowledge of the pre-identified important aspects of care that caregivers should understand for discharge,

5. (5) suggestions for improvements in the discharge process,

6. (6) input on a mobile application for discharge teaching,

7. (7) input on other technology-based support systems for discharge teaching.

Telephonic interviews were approximately 15–20 min. All caregivers received a $10 Visa gift card after completing the interview. Interviews were conducted until no new themes or domains emerged as determined by the interviewer.

The primary investigator conducted in-person semi-structured interviews with clinicians using a guide (Appendix B) developed similarly to the caregiver interview guide by the primary research team with input from the institution’s cardiac care unit leadership (physicians including the medical director, advanced practice registered nurses, and registered nurses). These interviews were conducted in a private room without distractions. The interview guide contained questions asking about what clinicians feel caregivers should know at the time of discharge, if caregivers are prepared to go home the day of discharge, what kind of learning tools are used, and specific information caregivers receive from the above-mentioned important aspects of care that parents should understand prior to going home.

Transcription of interviews

All interviews were audio-recorded and transcribed verbatim using a transcription service. Transcripts were de-identified prior to analysis.

Interviews with Spanish-speaking caregivers were conducted with a certified Spanish language interpreter via telephone. Only the English portions of the interviews were transcribed. The principal investigator (who is proficient in Spanish) reviewed the audio recordings to ensure the accuracy of the English translation by the interpreters during the interview.

Data collection

From the patient’s electronic medical record, we obtained: age, Society of Thoracic Surgeons–European Association for Cardio-Thoracic surgery mortality category; a measure of risk of mortality depending on the type of cardiac surgery, with higher scores indicating a higher risk of mortality, number of previous cardiac surgeries (excluding cardiac catheterisations and minor procedures), presence of a genetic syndrome, day of the week discharge occurred, medications prescribed at discharged and hospital length of stay. ^{Reference Cavalcanti, Sa and Santos14} We also recorded information about post-operative complications including clinically relevant bleeding (defined as haemoglobin drop of 20 g/L or more or requiring two or more separate transfusions of packed red blood cells, or bleeding that required medical or surgical intervention, not including menstrual bleeding), neurological complications and arrhythmias. These additional data was collected since such events could impact the amount of information conveyed to caregivers at discharge and could impact caregiver cognitive load. ^{Reference White, Fredericks, Mannarino, Janofsky and Faustino15}

Information collected about the caregivers included their relationship to the patient, prenatal knowledge about their child’s CHD diagnosis, relationship status (single or married couple), living distance from the hospital, insurance information (i.e., private versus public) and preferred language.

Health literacy data were recorded to examine the correlation of health literacy and knowledge gaps after discharge. ^{Reference Sanders, Federico, Klass, Abrams and Dreyer16} Health literacy was determined by using a three-item hospital-registered, nurse-administered admission screening tool validated in adults consisting of questions about difficulty performing reading tasks, understanding written information and highest level of education. ^{Reference Cawthon, Mion, Willens, Roumie and Kripalani17,Reference Chew, Griffin and Partin18} We transformed responses into levels of health literacy as previously described by Chew et al. ^{Reference Chew, Griffin and Partin18}

We inputted all patient and caregiver data into the Research Electronic Data Capture data management system. ^{Reference Harris, Taylor, Thielke, Payne, Gonzalez and Conde19}

Data analysis

We used a summative content analysis approach. A summative content analysis approach, as described by Hsieh and Shannon, “…involves counting and comparisons, usually of keywords or content, followed by the interpretation of the underlying context”. ^{Reference Hsieh and Shannon20} Two members of the research team used the above-mentioned domains (i.e., sense of caregiver preparedness and/or discharge readiness, how the discharge process occurred, general gaps in preparedness, caregiver knowledge with the pre-identified important aspects of care, suggestions for improvements and input on mobile application for discharge teaching) as initial coding categories. ^{Reference Hsieh and Shannon20,Reference Atkinson and Abu el Haj21}

After reviewing the transcripts from the clinicians and the caregivers, the categories and definitions were identified and redefined as necessary. Any potential subcategories were developed in a subsequent review of the transcripts. Any other portion of the transcript that was not initially categorised with the initial coding schema but were agreed upon to be significant were given new codes, allowing new concepts to develop. The same two members of the research team (i.e., the reviewers) used Dedoose© coding software only for the coding on the transcripts. ²² The reviewers also determined if the answers to questions targeting caregiver comprehension of the specific areas demonstrated sufficient knowledge of each area and coded as such.

The reviewers met regularly to discuss codes, ensure reliability of the coding data by going over each transcript and discussing why each code was given and resolve any coding discrepancies from the transcripts. The coding schema was updated as necessary until a final coding dictionary was developed. After this initial process, the reviewers then re-coded the data with the revised codes in the final coding dictionary.

Descriptive statistics were used to analyse demographic data using Excel.