Human-centered Design of Patient Interviews: Capturing the Experiences of Patients Recovering From Emergency Abdominal Surgery

Abstract

Objectives:

Recognizing the challenges of at-home recovery following a truncated presurgical pathway for emergency laparotomy (EL) patients, we sought to identify barriers and facilitators to optimal recovery. This study aimed to develop a human-centered interview guide to capture the experiences of patients recovering at home after EL.

Methods:

We employed an iterative human-centered design (HCD) approach to interview guide development, structured across 3 cycles. Each cycle refined the interview guide based on mock interviews and feedback from an interdisciplinary team. In cycle 1, we focused on creating a patient-centered, understandable guide. In cycle 2, we introduced a preinterview survey to tailor and shorten the interview process. In cycle 3, we automatically integrated patient responses from the survey into the interview guide, streamlining the workflow for interviewers.

Results:

The HCD process yielded a comprehensive, efficient interview guide responsive to both patient and interviewer needs. The integration of a preinterview survey reduced the cognitive load for patients and minimized interviewer preparation time, facilitating in-depth patient discussions on EL recovery experiences.

Conclusions:

This study underscores the value of HCD in research measurement design and tool development. The finalized guide enhances patient-centered data collection, reduces interviewer errors, and supports meaningful insights into EL recovery. This reusable protocol may benefit other researchers working on similar patient safety studies.

Emergency general surgery (EGS) has a substantial disease burden, and there is a critical need for improved outcomes. EGS accounts for 11% of US hospital admissions (3 million), with high risks, patient acuity, complication rates, and costs.^1–3 The subset of emergency laparotomy (EL) patients suffer the worst EGS outcomes.¹ Our Re-engineering Surgical Recovery and Transitions Using Technology Patient Safety Learning Laboratory (REST-PSLL), supported by Agency for Healthcare Research and Quality, focuses on developing innovations to support EL patients, particularly at home. This focus on at-home recovery is critical. Research shows that for many patients, the ability to recover at home is a vital aspect of postoperative quality of life.⁴ Understanding and supporting this phase of the patient journey is therefore a key priority.

Compared with nonemergent patients, the presurgical pathway for EL patients is truncated (Fig. 1). For example, patient education and caregiver planning often cannot be done ahead of the surgery. In a Delphi study involving EL patients and caregivers, patients expressed that this truncated presurgical time limited their opportunity to receive information, noting that EL patients “valued further information post-procedure, almost as a retrospective consent discussion. This may specifically relate to the rapid decision-making process due to time limitations in the emergency setting.”⁴ In interviews with EL patients, Park et al⁵ found that these patients expressed concerns related to fear and uncertainty, inpatient experience, information needs, acute postoperative concerns, and longer-term physical and emotional outcomes that persisted years after surgery. Our current understanding of barriers to and facilitators of optimal EL recovery remains incomplete, and this understudied phase of care is, therefore, ripe for innovative solutions. REST-PSLL aims to iteratively design, develop, implement, and evaluate innovative solutions for EL recovery, particularly in the home setting. By home, we mean any non-acute location a patient might go to after surgery (e.g., their own home; a residential care facility; a friend’s house).

FIGURE 1.

Surgical pathway for EL patients, and focus of the REST-PSLL project.

There is now a large body of research demonstrating the value of using human-centered or user-centered design approaches when creating digital health technologies with patients or consumers.^6,7 Our team is taking this approach to design solutions for EL patients. We are interviewing EL patients approximately 1 to 2 months after their surgery to understand their experiences during recovery from EL and to identify both barriers and facilitators to at-home recovery. Given that our interviews take place at a time when patients may still be recovering, we need to ensure that these interviews are thoughtfully designed and minimize the burden for patients. We, therefore, took an iterative, human-centered design (HCD) approach to the design of our interview guide. Figure 2 shows how the HCD process of the interview guide fits into the HCD process of EL interventions. The purpose of this manuscript is to describe our HCD approach to iteratively develop an interview guide that will comprehensively capture the experiences of EL patients, which can serve as a model for others who are engaging patients at vulnerable times.

FIGURE 2.

HCD of the interview guide and interventions.

METHODS

This study was approved by the University of Minnesota’s Institutional Review Board (IRB). All individuals who participated in the mock interviews had clinical backgrounds and personal histories as surgical patients to provide realistic feedback on the interview guide’s content and flow.

Our approach was grounded in Systems Engineering Initiative for Patient Safety (SEIPS) 3.0 framework (Fig. 1), which, in a systematic review across 16 original studies, was found to be an effective strategy for studying and improving health-related work and to generate useful products.^8,9 By using the SEIPS 3.0 approach to guide our data collection, we were better able to capture a holistic understanding of the recovery process over time and the factors that may help or hinder at-home recovery.^8,9 We also built on the work of Acher et al, who used SEIPS to understand patient and provider perspectives on contributors to surgical readmissions, which included exploratory questions related to incision care, drain care, ostomy care, medication, and nutrition.¹⁰ We created a prototype interview guide in document format based on the modified SEIPS 3.0 framework^8,9 and interview guide by Acher et al,¹⁰ with input from our interdisciplinary team composed of surgeons, nurses, and a human factors expert. Our iterative HCD approach to interview guide design involved 3 design cycles. Cycle 1 focused on the needs of both the patient/interviewee and the interviewer, cycle 2 focused primarily on the needs of the patient/interviewee, and cycle 3 focused primarily on the needs of the interviewer.

Our first design cycle focused on redesigning the content and flow of the prototype interview guide. We conducted 6 mock interviews. One research team member acted as the interviewer and a research team member with a history as a surgical patient acted as the interviewee. Our second design cycle focused on shortening the interview by moving questions to an asynchronous questionnaire completed ahead of the interview. We tested this approach through 2 mock survey completions and interviews. One research team member acted as the interviewer and a nonresearch team member with history as a surgical patient acted as the interviewee. Our third design cycle focused on streamlining the preparation work for the interviewer. We tested this approach with 3 final mock interviews. One research team member acted as the interviewer and another research team member acted as the interviewee. We iteratively revised the guide after each mock interview.

The final versions of the preinterview questionnaire and the interview guide are available in the Online Supplement (Supplemental Digital Content 1, http://links.lww.com/JPS/A764).

RESULTS

Cycle 1

For cycle 1, the primary design challenge was to translate a clinical and systems-engineering framework (SEIPS) into a conversation that was logical and intuitive for a patient recalling a complex and often traumatic experience. In particular, the mock interviews identified the significant role that in-hospital processes play in at-home recovery for EL patients. For example, mock interviews revealed that education on identification of complications happens primarily post surgery for EL patients. The team hypothesized that a guide structured chronologically to mirror the patient’s actual journey—from inpatient stay to discharge and recovery at home—would be more effective than a guide organized by abstract theoretical SEIPS domains. We added a series of questions to understand what happened during the patient’s time in the hospital and reordered the guide to support logical reflection on these phases: post-surgery inpatient; day of discharge; time at home; and current experience. This patient-centered flow was intended to reduce the cognitive burden of recall and allow for a more natural narrative. We also revised terminology to ensure it was patient-centric, understandable, and inclusive (e.g., using the phrase “place after discharge” instead of “home”), to ensure the guide was accessible and welcoming to all participants, regardless of their living situation. Finally, we made protocol changes to help patients understand the context for the questions, specifically sharing our screen during the interview to show relevant visuals (e.g., a drain). At the end of this cycle, we had an understandable and comprehensive interview guide.

Cycle 2

The interview guide at the end of design cycle 1, while comprehensive, was long. On examination of the guide, we identified that we could ask many of the questions ahead of time using an asynchronous questionnaire. The research team weighed the trade-off between adding a preinterview task for patients and the potential benefit of a shorter, more focused interview. We hypothesized that reducing the cognitive load during the synchronous interview would be a net positive for patients who were still in their recovery period. The preinterview survey was designed to be completed asynchronously, allowing patients to fill it out at their own pace. This approach also aimed to empower patients by giving them an opportunity to see the topics in advance. We extracted these questions and implemented them as a REDCap questionnaire for participants to complete before their interview. The REDCap questionnaire implemented branching logic to tailor the questions based on patients’ answers. One example of this branching logic was:

Do you have an ostomy bag as a result of this surgery? (Yes/No)

If Yes: Before leaving the hospital, did you receive information about how to care for your ostomy bag? This information could have involved conversations with care team members and/or handwritten, printed, audio, or electronic materials. (Yes/No/Unsure)

If Yes: How did you receive the information about how to care for your ostomy bag? (Select all that apply) from “conversation with a care team member; handwritten notes; printed materials; text on a screen or device; audio and/or video on a screen or device; other.”

Figure 3 shows how these questions were displayed to the patient.

FIGURE 3.

Example questions from the preinterview questionnaire.

The interviewer could then review a participant’s answers before the interview and tailor the interview questions based on their responses. For example, if a patient responded yes to the questions in Figure 3, we would ask during the interview to describe what was helpful and/or not helpful about how they received the information about ostomy care.

We were also able to synthesize sections of the interview guide based on these responses. For example, if a participant answered yes to multiple questions in the survey about wound, drain, and ostomy care, when the interviewer asked the question about what was helpful and/or not helpful about how they received the information, the interviewer could say “Feel free to describe each separately or together.” At the end of this cycle, we had a presurgical REDCap questionnaire and much shorter interview guide informed by the participant’s responses to the presurgical REDCap questionnaire.

Cycle 3

The interview guide at the end of design cycle 2, while a significant improvement for the patient interviewee, required the interviewer to manually review the REDCap questionnaire responses and tailor the interview guide. For cycle 3, the design focus shifted to the interviewer’s experience, as the previous cycle had introduced a new workflow burden. The design tension was that the manual process of reviewing survey responses to customize each interview was both time-consuming and prone to human error, such as overlooking a patient’s answer and failing to ask a relevant follow-up question. The team reasoned that automating this process by piping survey data directly into the interviewer’s guide would significantly improve efficiency and consistency. By reducing the interviewer’s cognitive load related to preparation and in-the-moment synthesis, the automation would free them to focus more on active listening and building rapport with the patient, ultimately leading to a higher quality interview and richer data. We therefore implemented the interview guide as REDCap questionnaire, where participant responses to the preinterview questionnaire were automatically piped into the interview guide in relevant locations. We also added free-text blocks at relevant locations where the interviewer could type notes during the interview. Figure 4 shows an example of one section of the interview guide. The black boxes show examples of information automatically piped in from the preinterview questionnaire. This streamlines workflow for the interviewer, eliminating the need for any tailoring work before the interview. At the end of this cycle, we had a presurgical RedCap questionnaire, shorter interview guide informed by the participant’s responses to the presurgical RedCap questionnaire, and a RedCap interview guide with the participant’s responses to the presurgical RedCap questionnaire piped into the RedCap interview guide.

FIGURE 4.

Interview guide with automatic piping of responses.

Following the completion of these design cycles, the finalized interview guide and preinterview questionnaire were deployed in our ongoing study with EL patients. To date, we have conducted interviews with 13 EL patients to-date. Typical time for patients to complete the preinterview survey has ranged from approximately 5 to 25 minutes, and the typical duration of the interview has ranged from approximately 45 to 60 minutes. The interview flow has been smooth, and the automated data piping has functioned as designed, saving an estimated 20 minutes of preparation time per interview. Patients have completed the preinterview survey without issue and have not expressed concerns about the task. The focused nature of the interviews has allowed for rich, in-depth discussions. For instance, by having baseline information on ostomy care education, interviewers were able to probe deeper into the quality and effectiveness of that education rather than simply confirming its existence.

DISCUSSION

Traditionally, HCD is applied to design patient-centered innovations; here, however, we applied it to the research methods that inform those solutions. By employing this HCD approach to our interview guide design, we successfully created a streamlined interview guide that is both responsive to the patients’ and interviewer needs. The iterative process required repeated mock interviews and refinements to balance comprehensive data collection with end-user needs. By engaging multiple perspectives and iteratively refining our guide, we enhanced the experience for both the patient interviewee and the study team interviewer, thus underscoring the value of HCD for developing effective, user-centered research tools.

The integration of a preinterview questionnaire and automated branching logic allowed for customization without extensive manual adjustment, saving the interviewer time. By designing a user-friendly, context-aware interview guide, we also reduced the risk of interviewer errors during preinterview preparations and during the interview itself. These innovations made the interview shorter and more focused, allowing us to engage in more meaningful discussions with patients about their recovery experiences. As a result, the interview process became more patient-centered, allowing us to gather in-depth insights into factors that facilitate or hinder EL at-home recovery. In addition, this approach allowed us to more easily gather structured responses that would have required manual extraction from the qualitative interview transcripts.

Our HCD approach also yielded a reusable and adaptable protocol that can benefit other research teams who have similar research goals. This process reinforces the notion that investment in HCD for research tools, while time-intensive, can ultimately yield substantial dividends by reducing errors, improving user satisfaction, and enhancing the quality of data obtained.

Our approach does have limitations. The iterative design process was time- and resource-intensive, which may limit the feasibility of having smaller research teams or those with constrained timelines utilize this development process. Furthermore, while this manuscript describes the development process of our research tool rather than the clinical results obtained from it, we recognize that the value of this guide will ultimately be determined by the quality of insights it helps generate in our ongoing primary study. Our preliminary results from the 13 EL interviews to date show it is indeed generating novel and important insights. In addition, while we designed the interview guide to be flexible, the reliance on the preinterview REDCap questionnaire and automated piping may introduce challenges if patients have difficulty navigating digital surveys or if technological issues arise, potentially impacting the effectiveness of the interview process. That said, this concern has not played out in our data collection with the 13 EL patients interviewed to date.

CONCLUSIONS

This study demonstrates the value of HCD in developing research tools that effectively capture complex patient experiences, in this case, EL recovery. Our iterative HCD approach led to a flexible, patient-centered interview guide that streamlines the interview process, reduces cognitive load for patients and interviewers, and enables richer data collection by focusing on both patient and interviewer needs. The integration of a preinterview questionnaire with automated data piping enhances efficiency and reduces errors, allowing for more focused discussions on patients’ recovery experiences. This work highlights the importance of thoughtfully designed research instruments, which capture patient insights that can lead to innovative solutions.

Supplementary Material

pts-22-127-s001.pdf ^{(191.8KB, pdf)}