A hybrid type II design combined a quasi-experimental pre-post study with a multi-methods implementation evaluation. The article is structured in accordance with the Standards for Reporting Implementation Studies (StaRI) Statement (8). Figure 1 provides an overview of the study method, including both the intervention components and the implementation process.

The study was conducted in a 40-bed medical unit of a regional hospital in French-speaking Switzerland. The control phase was conducted from June to October 2024, followed by the intervention phase from November 2024 to March 2025. The target population for the intervention consisted of older inpatients, meeting the following inclusion criteria: aged ≥50 years, discharged to home planned at the beginning of the hospitalization, hospitalized for more than 48 h, able to read and speak French, and capable of providing written informed consent. We used a ≥50-year age criterion because multimorbidity is increasingly common from mid-life onward in Switzerland, and a higher age cutoff would have substantially reduced recruitment, as many adults over 65 are typically discharged to rehabilitation rather than directly home. A minimum 48-hour stay was required to ensure sufficient exposure to ward care and, in the intervention arm, time to initiate the PODS. Patients were excluded if, according to the healthcare team, they were unable to follow the study procedures due to cognitive impairment or language barriers.

The implementation component of the study targeted the unit's healthcare professionals who were expected to help patients complete the PODS, which included nurses, physicians, liaison nurses, dietitian, occupational therapists and physiotherapists.

A total sample size of 130 participants (65 in the control group and 65 in the intervention group) was determined to detect a mean difference of 10 points on the primary outcome measure (assumed SD = 18, β = 0.80, α = 0.05), accounting for a 20% attrition rate. For both groups, newly admitted patients were routinely screened for eligibility by the on-site project lead, nurse managers, or clinical nurse specialists, who informed the bedside nurses accordingly and the research collaborator for information and consent.

A working group composed of professionals from the unit was established to support the implementation process and served as a key source of data. The group was formed by the on-site project lead—appointed by the nursing directorate in collaboration with the unit nurse manager—together with the nursing directorate. Professionals from each discipline were identified and approached by the project lead to join the group: two physicians, three clinical nurse specialists, two staff nurses, one healthcare assistant, one physiotherapist, one dietitian, one liaison nurse, and the on-site project lead herself, who was an assistant nurse manager. Over the course of the implementation, additional members were invited as needed to ensure continued representation across all disciplines, and all team members were invited to complete the questionnaire for assessing factors influencing the implementation of complex interventions and study surveys on the feasibility and fidelity of the intervention. For qualitative data collection related to implementation (focus groups, interviews), participants were identified by the nurse manager and selected based on their availability on the chosen dates, while ensuring maximum possible representation of the different professional disciplines.

The intervention is reported in accordance with the Template for Intervention Description and Replication (TIDieR) checklist (39). It took place in an acute care unit of a regional hospital in Switzerland and was designed to structure discharge teaching, strengthen patient understanding, and involve caregivers. The Patient-Oriented Discharge Summary (PODS) constituted the core component of the intervention, complemented by the systematic use of teach-back and the involvement of caregivers. To ensure consistent use of the PODS, the healthcare team was provided with various resources, including posters summarizing the PODS content and standardized workflow, a training video developed by the working group, examples of completed PODS, refresher discussions during team meetings and visual reminders in nurses office and health electronic record. To support consistent implementation, the nurse manager reminded the team each morning about which patients had an active PODS.

The PODS sections were based on the official French translation provided in the PODS Toolkit (https://pods-toolkit.uhnopenlab.ca). We used this version as a starting point and conducted a cultural adaptation for our local context, further refining the wording and the layout in collaboration with the senior patient partner and the unit's healthcare team (Supplementary Material S1). The senior patient partner was actively involved in the research project as a member of the research team. She contributes on an ad hoc basis, particularly when a patient perspective is needed to inform specific components of the intervention or study procedures. The PODS contained six sections of information to be completed: (1) reason for hospitalization (2) health problems; (3) warning signs and symptoms to watch for, (4) changes in daily activities and health behaviors, (5) upcoming follow-up appointments, and (6) contact information. The bedside nurse was responsible for providing the PODS to the participants and explaining the purpose of the document. The PODS was kept in the participants' room. The intervention was delivered throughout the patient's hospital stay rather than in a fixed number of sessions. The intervention was delivered face-to-face at the bedside by multiple healthcare professionals, each responsible for specific sections of the PODS based on their clinical role. The bedside nurse responsible for admission introduced the PODS and explained its purpose. Throughout the hospital stay, nurses progressively completed the PODS during routine care encounters and based on daily medical rounds, particularly for the sections related to health problems and symptoms. Physiotherapists added information related to mobility, physical activity, and functional recommendations when physiotherapy sessions occurred. Dietitians contributed to the section on lifestyle and dietary behaviours when clinically indicated. Near the end of the stay, liaison nurses completed or verified the contact information and community follow-up arrangements. Physicians contributed content for warning signs and symptoms requiring medical attention, and nurses added details regarding follow-up appointments. On the day of discharge, the bedside nurse was responsible for reviewing and verifying the information recorded in the PODS.

The frequency and duration of these interactions varied depending on patients' needs, resulting in a variable “dose” across participants. The intervention was intentionally designed to be highly personalized, with all PODS sections completed using patient-specific information. Using a blank template was intentionally employed to encourage active patient participation. By progressively filling out the PODS during the hospitalization, it functioned not only as an information record but also as a pedagogical tool: patients engaged in recalling, reformulating, and co-producing key information, which supported memory retention and ownership of their discharge plan. The teach-back technique was systematically integrated into this process: when reviewing PODS content, healthcare providers were also reminded to use the teach-back principles. This allowed misunderstandings to be identified and corrected immediately. Whenever possible, caregivers were identified early during admission and invited to participate in the process. Their role in supporting both patient comprehension and post-discharge care was emphasized during staff training.

No substantial modifications were made to the intervention during the study period.

The implementation followed the EPIS framework (Exploration, Preparation, Implementation, Sustainment) (9). The preparation phase took place from February 2024 to October 2024, while the implementation phase was conducted from November 2024 to March 2025. An on-site project lead and an interdisciplinary working group composed of staff members were designated to guide the implementation process. Together with the research team, this working group also contributed to identifying and monitoring implementation determinants and selecting the implementation strategies that would address the determinants.

The preparation phase involved maintaining the usual discharge practices while recruiting and collecting data from participants in the control group. During ths phase, a contextual analysis of the unit was conducted using the Consolidated Framework for Implementation Research 2 (CFIR2) with the interdisciplinary working group in a collective format to identify potential barriers and facilitators (10). To facilitate the identification of barriers and facilitators, the study investigator used the CFIR Card Game, a structured tool that presents CFIR constructs as individual cards (11). Healthcare professionals from the working group discussed how these constructs applied to their setting, encouraging collaborative reflection and dialogue between the clinical and research teams. To confirm or challenge the implementation determinants identified through the card game session, a first semi-structured focus group was conducted in the unit by the investigator and the research collaborator with other staff members, including nurses, liaison nurses, physiotherapists, physicians (Supplementary Material S2). All members of the unit team were also invited to complete the Normalization Measure Development (NoMAD) questionnaire, a validated tool for assessing factors influencing the implementation of complex interventions from the perspective of healthcare professionals (12, 13).

Based on the findings from the contextual analysis, the implementation strategies guiding the project were identified and then mapped to the Expert Recommendations for Implementing Change (ERIC) taxonomy (14). Regular workshops held with the interprofessional team as part of the project management approach offered a structured space to collaboratively identify implementation strategies. Some were primarily initiated and led by the research team, while others were driven by hospital leadership and unit staff. To validate and refine the strategies developed by the working group, the investigator and the research collaborator held a second focus group with unit staff members to confirm alignment with their needs and incorporate complementary perspectives (Supplementary Material S2). Several strategies were designed and deployed prior to implementation and as the implementation unfolded, additional strategies were developed or adjusted in response to emerging barriers and facilitators.

At the end of the preparation phase, the unit team received formal training in using the PODS. During the implementation phase, all patients aged 50 years and older who were admitted to the unit, planned for discharge home, and whom the on-site project lead and nurse manager considered appropriate, received the PODS within 24 h of admission. Implementation determinants and outcomes were continuously monitored and shared with the working group. Monthly follow-up meetings provided an ongoing feedback loop and enabled the refinement of the implementation strategies as needed.

The implementation phase lasted 5 months. At the end of this phase, four interviews were conducted by the investigator and the research collaborator to identify persistent obstacles to PODS implementation. These interviews were conducted with a physiotherapist, two nurses and a nurse clinician (Supplementary Material S2).

Standard statistical analyses were used to describe the data, that is, mean and standard deviation or median and interquartile range for continuous variables and absolute and relative frequencies for categorical data. Given the nature of the CTM-15 outcomes, linear regression was used to analyze the effect of the intervention on both the CTM-15 total score and the individual item-level scores. Although item-level responses are ordinal, we prioritized linear models for the primary analysis because their coefficients provide an intuitive and easily interpretable effect estimate. As a sensitivity analysis, we re-estimated all item-level models using ordered logistic regression, thereby altering the analytical assumption to reflect the ordinal scale of the data. This allowed us to assess whether our findings were robust to a modelling approach more strictly aligned with the data's measurement properties. The analyses were adjusted for a predefined (based on the literature) set of potential confounders, including educational background, living arrangement (cohabitation), recent hospitalizations, home care utilization, and functional status. Missing data were not imputed; therefore, analyses were conducted on an available-case population (i.e., complete-case analyses). Data were evaluated for outliers using standard descriptive statistics (percentiles) and visual inspection (histograms, scatterplots, and boxplots). No outliers were found. Differences between the two groups regarding post-discharge difficulties (ELADEB card-sorting results) were examined descriptively. We compared the frequency with which each domain was classified as a problem across groups and explored patterns in the distribution of reported difficulties. All data preparation and statistical analyses were performed using Stata Version 19 (22).

Qualitative data were drawn from the initial working group session, during which the CFIR card game was used, from two focus groups conducted with interprofessional team members, and from four interviews at the end of the implementation phase. These recordings were transcribed verbatim. A rapid directed content analysis was conducted, guided primarily by the CFIR 2.0 codebook (40, 41). The CFIR 2.0 provided the primary deductive structure for coding, and inductive codes were created when data did not fit the existing constructs. All transcripts were imported into MAXQDA software (23). Two researchers conducted the analysis: the principal investigator, a doctor in nursing science with extensive expertise in implementation research and in applying the CFIR framework, and a research collaborator, a specialized nurse clinician trained in qualitative methods. Prior to full analysis, they jointly coded a sample of transcripts to ensure consistency. They then coded the remaining transcripts independently, meeting regularly to compare interpretations, resolve discrepancies, and refine the coding scheme. Because each CFIR construct can function either as a barrier or a facilitator, we created additional codes to explicitly distinguish these two possibilities during analysis. This required revisiting earlier transcripts to classify each coded segment accordingly and to ensure that the distinction was analytically meaningful within the content analysis.

The study was approved by the Cantonal Ethics Committee on Human Research of Vaud (2023-02048) in accordance with Swiss regulations and institutional guidelines for research involving human participants. For all recorded sessions— working group meetings, focus groups, or post-implementation interviews with healthcare professionals—written informed consent was obtained prior to participation.

Of the 217 eligible patients invited to participate, 87 declined to participate, most often because they felt too occupied or overwhelmed by health problems, or because they anticipated minimal difficulties following discharge. A total of 66 participants were included in the control group and 64 in the intervention group (Supplementary Material S3). Among these, 11 participants were excluded from the control group (lost to follow-up, n = 5; change of discharge destination, n = 5; hospital readmission, n = 1) and eight from the intervention group (lost to follow-up, n = 5; change of discharge destination, n = 2; transferred to another department, n = 1). Data from 55 to 56 participants were included in the final analysis. Table 1 summarizes the participants' demographic and clinical characteristics.

Unlike the pre-filled versions tested elsewhere (26–28), a distinctive feature of our study was the use of a collaborative blank version of the PODS, completed progressively with patients. This format fostered individualized teaching and patient engagement, but it also contributed to variability in completion, potentially reducing the tool's usefulness once patients returned home. This design choice is key to interpreting our findings: PODS proved effective in structuring discharge communication and fostering a sense of preparedness but was less effective in providing concrete post-discharge guidance. The blank format encouraged personalization but sections requiring detailed instructions—such as warning signs or follow-up actions—were often incomplete. Similar to our study, Hahn-Goldberg et al. (24) and Okrainec et al. (25) tested a blank collaboratively completed version of the PODS. However, unlike our findings, the results did not show statistically significant improvements in patients' understanding of medications or in their knowledge of what to do if problems arose after discharge. These findings contrast with studies evaluating pre-filled or standardized PODS versions, which tend to ensure completeness and usability at home, but may offer fewer opportunities for co-production during hospitalization. For example, Black et al. (26) reported that 88% of patients found the PODS at least partly helpful, 79% felt adequately prepared for discharge, and 98% of patients and families believed that the PODS added value to the discharge experience. Similarly, Schofield et al. (27) tested a heart failure–specific adaptation of the PODS, which included pre-filled standard information. This version was associated with higher rates of recall of written information, more frequent use of the materials as visual reminders, and improved adherence to dietary and physical activity recommendations after 30 days (27). However, in that study, patients often did not consult the materials during the initial days post-discharge, and individual teaching provided by ward staff was identified as the most influential factor in behavior change (27). An adaptation of the PODS for intensive care also showed positive results, particularly in terms of patients' understanding the medical condition (28). Additionally, 80% reported having a conversation with the ICU team about the transition and the next steps (28). Similarly, a comparative study in Ontario found that hospitals with full PODS implementation had higher odds of patients reporting discussions with staff about the needed help and receiving written information on post-discharge symptoms (25). Future adaptations could therefore combine standardized pre-filled content for core instructions with individualized sections completed collaboratively, ensuring that both informational accuracy and patient engagement are maintained.

Our findings emphasize the relational value of the PODS. Patients reported limited use of the tool after discharge but consistently highlighted the usefulness of the conversations it prompted with healthcare professionals. These findings suggest that the PODS is effective less as a written document than as a structure that supports communication, shared understanding and active participation. Consequently, its value emerges mainly during hospitalization, where it acts as a relational and cognitive scaffold rather than a static information tool. This raises broader questions about the role of discharge tools: their main contribution may lie in structuring and enhancing teaching and dialogue during hospitalization, rather than to sustaining self-management behaviors once patients are at home. A literature review by Okrainec et al. (29) supports this interpretation, showing that patient-centered discharge tools improve comprehension but do not consistently translate into better adherence or outcomes (29). Similarly, previous studies have not demonstrated that PODS reduces unplanned healthcare utilization or hospital readmissions (24, 30).

This relational function reflects the core purpose of the PODS—a tool designed to guide discharge teaching and synthetize key information for the post-hospital period. Healthcare professionals repeatedly emphasized this pedagogical role, describing the PODS as “a structure that helps us not forget the essentials” and “a support for the conversation rather than a document to fill”. In surveys, 70% of respondents recognized its value for discharge teaching, and 85% rated it as helpful for patients and caregivers. Together, these findings confirm that the PODS primarily serves to organize and clarify information during discharge preparation. However, patients still face numerous practical challenges beyond what the tool covers, such as transportation difficulties or limitations in daily activities. These domains fall outside the scope of the PODS, which focuses mainly on medical and self-care information. Consequently, while it enhances comprehension of clinical instructions, it does not address broader determinants of a successful transition—such as mobility, social support, or community resources. This may help explain why improved perceptions of care transition quality did not translate into fewer post-discharge problems: many of these difficulties stem from contextual and social barriers rather than from informational gaps.

Our findings suggest that this tool alone may not be sufficient to address these issues. The use of blank PODS forms completed collaboratively with patients, rather than pre-filled versions with standardized content, may have limited the consistency and completeness of the information recorded. Some participants likely received the PODS with minimal content, which may not have provided adequate guidance for managing symptoms and other post-discharge challenges. The predominance of physical health problems among reported difficulties further highlights the complexity of post-hospital recovery, which extends beyond information provision. Effective transitional support therefore requires multifaceted strategies that combine discharge teaching with ongoing clinical, social and community-based support to better address patients' diverse needs once home.

In addition to providing evidence on the effects of PODS for patients, this study highlights how implementation determinants shape clinical outcomes. Workload constraints, limited workflow compatibility, and variable interprofessional engagement influenced PODS use. For example, incomplete sections of the tool reflected both time pressure and inconsistent ownership across professional groups, limiting the guidance that patients received at home. Fidelity analysis confirmed this: while general sections, such as the reason for hospitalization, were usually completed, more action-oriented components (e.g., what to do, warning signs, or contact details) were often omitted. As a result, the information received by PODS patients was personalized and clear, but not always comprehensive enough to support symptom management after discharge. This helps explain why patients felt better prepared in the hospital but still reported similar post-discharge problems.

Concerns about workload were common at the start of implementation but decreased as staff gained experience, echoing findings by Black et al. and Hahn-Goldberg et al. (24, 26). They also both reported that greater experience with PODS increased professionals' confidence that patients were prepared for discharge. We did not assess this quantitatively, but interviews in our study revealed that while some staff perceived PODS as improving patient readiness, others doubted whether patients would use it at home. Thus, the impact of PODS depended not only on the intervention itself but also on its appropriation by the staff. Compared to pre-filled formats, which guarantee a minimum information set, blank forms may be more sensitive to workload and time constraints, but they also better support personalized, discussion-based teaching in our context (38).

Another key factor was uneven engagement across professional groups. Although designed as an interprofessional tool, PODS was mostly championed and completed by nurses, while other professionals engaged inconsistently due to workflow barriers or limited perceived value. This nursing-led appropriation limited the PODS capacity to integrate broader interprofessional contributions that might have strengthened continuity of care. Similar challenges have been reported in previous studies, in which sustained engagement across disciplines proved difficult. Black et al. (26) showed that visible, consistent engagement from unit leaders and clinical champions were critical for normalizing PODS use by reinforcing it during daily rounds, team huddles, and staff onboarding. In our setting, the working group noticed that when the nurse manager and clinical nurses actively reminded staff and monitored EHR entries, penetration rates increased, and motivation was sustained. Conversely, disruptions, such as the influenza outbreak, coincided with drops in completion rates and staff engagement. This suggests that leadership and continuous reinforcement are not only facilitators of implementation but also determinants of the consistency and quality of patient outcomes. Additional strategies might also have supported broader interprofessional engagement, such as regularly clarifying role expectations for each profession, offering periodic profession-specific feedback, highlighting interprofessional success stories, securing stronger visible endorsement from hospital leadership, explicitly linking the project to other institutional goals, and identifying influential clinicians as informal ambassadors. These options were beyond the scope of the current project but could be explored in future implementation.

Patient feedback can help interpret these patterns. Patients valued the discussions prompted by PODS, even when they did not use the document itself after discharge. This highlights that PODS' primary contribution may lie in its relational function—structuring discharge conversations and supporting teaching—rather than in its role as a written discharge summary. Where the PODS was used interactively—through a supportive dialogue between healthcare professionals and the patient—patients may have felt more prepared. Conversely, when completion was fragmented or rushed, the relational component of the interaction was reduced, which may have weakened its impact once the patient returned home. However, the blank format of the tool requires patients to actively document their information. Feedback from both patients and professionals indicated that some were not interested in or did not see the purpose of this exercise, raising broader questions about expectations regarding patient engagement in discharge preparation. In the acute hospital context, fatigue, the impact of illness, and—for older adults—the additional effects of aging may limit patients' readiness to engage with such a tool. This cultural aspect may also differ across contexts: in Switzerland, encouraging patients to co-produce their discharge summary represents a relatively new expectation, whereas in Canada PODS has been more widely sustained in some regions (24, 25). Integrating PODS into electronic health records could make the form easier for staff to complete. Some parts could be automatically filled in with standard information, while leaving space for healthcare professionals to add details specific to each patient. However, this should not replace the face-to-face explanations and discussions at discharge, which patients find particularly helpful.

The use of a mixed-methods approach, combining surveys, focus groups, interviews, and field observations, provided a comprehensive understanding of the PODS implementation process in the study. The implementation was guided by the CFIR, which was used at baseline for the contextual analysis and throughout the monitoring phase. This structured approach helped identify key determinants, guided the selection and adaptation of strategies, and ensured systematic analysis. Using the CFIR also strengthens the transferability of our findings by enabling comparison with other international studies applying the CFIR, as it is the most widely used determinant framework. The study was conducted only in a single unit of one hospital, but testing the implementation under real-world acute care conditions, with healthcare team members, and within a setting characterized by organizational constraints strengthens the transferability of the results to other hospital environments. Several methodological limitations should be acknowledged. First, the study used a pre–post design with two groups but without randomization, which limits causal inference about the effects of the PODS on patient outcomes. In our study, fidelity was indirectly assessed through completion rates, which provided only partial insight. The absence of a clear trend in post-discharge problems may be explained by marked variability in PODS completion, with some forms fully completed, others partially completed, and others left largely blank. Completing the CTM-15 and assessing post-discharge problems 5–10 days after discharge may be subject to recall bias, as patients might have forgotten some of the difficulties faced in the first days at home or no longer perceived them as problems—either because they had adapted or adjusted their expectations. Finally, not all professional groups involved in discharge processes were represented in the qualitative components of the study. In particular, some professions—including physicians—did not participate in the post-implementation interviews nor in all working-group workshops, mainly due to workload and scheduling constraints. This limits the diversity of perspectives captured regarding the implementation, especially given the central role of certain professional groups in producing discharge summaries.