Using purposeful sampling for this exploratory, descriptive study, the investigators incorporated methodological pluralism to interview patients and observe any artifacts they used to track care issues and their care questions. The Institutional Review Boards and hospital leaders granted approval for the study, and the team worked with unit managers during data collection procedures.
The study was conducted at a large cancer center in the western United States; its 25-bed inpatient general surgery unit was the data collection site. This unit admits approximately 460 patients per month for ear-nose-throat, thyroid, colorectal, prostate, gastrointestinal, Whipple, and cardiothoracic surgical procedures.
Participant Selection Criteria
Purposeful sampling with maximum variation was employed to select 20 patients across (1) demographic characteristics, (2) levels of care complexity (comorbidities, no comorbidities), (3) types of surgical procedures and (4) recovery course (routine, non-routine). The authors collaborated with the nurse manager and charge nurses on the patient care unit to apply selection criteria to potential study participants. Patient inclusion criteria were: adult (age 18–85), ability to converse in English, willingness to participate, and cognitive capabilities and medical condition conducive to participation.
The team used a multimethod approach for data collection: naturalistic observations, interviews, field notes, and artifact (tool) capture. Patients were asked to complete demographic forms with the following data: age, educational level, number of times as an inpatient at the facility, surgical procedure, and number of days since their day of surgery.
After obtaining informed consent from participants, two of the authors (ns and lh) conducted and audio-recorded patient interviews in the patients' rooms, typically at the bedside. A scripted interview guide, available from the authors, was developed to standardize the questions and expected probes. Three major topics were explored: preferences about participating in handoffs/perceived barriers, information requirements and any tools inpatients use now such as paper, computerized tools, or whiteboards in the room. The interview questions were crafted by integrating material from a literature review and from a discussion with the cancer center's director of patient and family support. The questions were reviewed by the team, piloted, and honed.
For patients who indicated they used tools, the team audio-recorded participants using "think aloud" methods to describe tool usage, organization, content and utility.
Additionally, the authors recorded field notes about the context of interviews and any tools in patients' rooms such as communication whiteboards. The researchers de-identified paper tools and white boards and then photographed them.
Interviews were transcribed by a professional transcription company and verified for accuracy against audio recordings by the research team. They were uploaded with artifact photos and field notes to a secure computing environment.
The authors analyzed the data using conventional content analysis and the tool Atlas tiTM version 7.0.85. Qualitative content analysis was the most suitable for interview analyses because it is the least interpretive approach and it supports counting responses as appropriate. Data analysis followed the three steps recommended by Bernhard and Ryan: (1) first cycle provisional coding to establish definitions and boundaries among codes, (2) second cycle coding of all of the transcripts, and (3) third cycle consolidation of categories to examine the data for themes. Coding was recursive with the analysis cycling between first and second cycle coding. The team achieved consensus on the codes and their application. After second cycle coding, the authors individually reassessed all transcripts for coding consistency. The team then reviewed discrepancies to achieve consensus on final coding.
Themes were inductively derived for third cycle coding. Categories were combined after team discussion, eg, patient information needs and family information needs. To maintain rigor throughout the analysis, features of trustworthiness were given careful attention: confirmability to maintain neutrality and remain true to the participants' views; auditability as it pertains to the process of inquiry, findings, interpretations, and recommendations; credibility via member checks.