Operationalizing Patient Experience Data as a System Feedback Asset in Learning Health Systems: Learnings From Model of Care Redesign at Trillium Health Partners

ABSTRACT

Introduction

Learning Health Systems (LHS) offer a pathway to continuously improve care by integrating data and lived experience, yet few practical examples demonstrate how LHS can be operationalized in hospital settings in ways that embed the voices of patients and families. This manuscript describes how patient experience (PX) data are integrated into the design of a new model of care (MoC) that responds to both system demands and the realities of patients and families.

Methods

Researchers partnered with clinical leaders to integrate patient and family voices into the design of a new MoC using PX survey data, interviews with patients and families admitted to medicine units, and naturalistic walk‐throughs of each medicine unit. Data collection explored care priorities, interactions with staff, physical environments, and reflections on positive and challenging aspects of care. Analysis followed a three‐step iterative process: unit‐level synthesis, cross‐unit thematic comparison, and data integration.

Results

In total, 390 patients participated in the PX survey and 19 patients/families participated in interviews. Patients situated their experiences within a health care system under strain, highlighting both strengths and gaps. Many expressed appreciation for nurses' compassion, but emotional support for fears, anxieties, and worries was inconsistent, particularly when staff were overstretched or transient. Long wait times for transport, procedures, or basic care needs were reported as sources of anxiety and diminished dignity. Naturalistic walk‐throughs documented how crowding, noise, lighting, and wayfinding shaped patients' comfort and privacy.

Conclusions

This work provides a practical example of how an LHS approach can be operationalized by embedding the voices of patients and families into care redesign. Findings from the PX survey and interviews are directly shaping a new MoC at Trillium Health Partners (THP) that is evidence‐informed and co‐designed with patients and families, providing a practical example of how PX data can drive continuous, system‐level learning.

1. Introduction

Health systems both in Canada and internationally face mounting pressure to provide safe, equitable, and person‐centered care amid rising patient complexity, workforce shortages, and constrained resources [1, 2, 3]. Pressures to engage in continuous improvement are especially visible in hospital medicine units, spaces where older adults, people living with frailty, and patients with multiple chronic or cognitive conditions require coordinated, multidisciplinary attention. High demand can pull teams toward task‐oriented routines (e.g., strict medication rounds and vital signs) that contribute to delays in responding to emerging patient needs (e.g., toileting, warmth, and pain relief), varied emotional support, and fragmented communication with patients and families. Such conditions, along with factors such as noise, crowding, and frequent staff turnover, can influence how patients perceive the quality, safety, and person‐centredness of their care [4, 5].

In response, systems are increasingly adopting Learning Health System (LHS) approaches that integrate routinely collected organizational data with PX data to enable continuous improvement [6, 7]. PX measures of communication, responsiveness, and perceived dignity provide actionable signals about how care is delivered and can guide unit‐level redesign [8]. Consistent with the LHS framework, PX data serve as fuel for continuous cycles of evidence generation, co‐design, and implementation [9]. Yet practical examples of how these principles can be operationalized in hospital medicine contexts, and how patient, caregiver, and family voices can be integrated into that learning cycle, remain scarce [4, 8, 10]. Moreover, PX data are often limited to survey‐based approaches, which can constrain the depth and contextual richness of experiences captured [11, 12].

Trillium Health Partners (THP), one of Canada's largest and most diverse hospital systems, offers an opportune context for applying LHS principles. Facing rising patient volumes, workforce shortages, and increasing case complexity, THP is undertaking a major infrastructure and service expansion, which is expected to be completed by 2033. Central to this effort is the redesign of the model of care (MoC) across 12 medicine units to align workforce capacity, patient needs, and quality of care. Guided by an LHS approach, this redesign integrates real‐time evidence from multiple domains (i.e., clinical outcomes, staffing data, safety metrics, and PX insights) to inform decisions about how care is delivered. However, the MoC redesign is ongoing, and this paper focuses specifically on the PX component of the broader project. The aim of this component was to generate insights to inform the redesign of MoC and to prioritize design considerations based on PX data, including how patients and families experience their care environments, interactions, and unmet needs.

1.1. Research Objectives

The objective of this study was to describe how an LHS approach can be operationalized in a hospital by integrating PX data into the redesign of MoC. Specifically, we sought to: (1) capture patient, caregiver, and family experiences across medicine units with varying PX performance; (2) identify strengths, hurdles, opportunities, and pressures from the perspective of those receiving care; and (3) generate actionable insights to inform the redesign of MoC at THP.

2. Methods

2.1. Approach

This mixed‐methods study was designed to integrate real‐time, experience‐based evidence into care redesign and decision‐making processes following the Institute for Better Health's LHS framework [9]. The goal was to generate timely, actionable insights to inform the redesign of MoC by identifying strengths, challenges, and opportunities from the perspectives of those receiving care. We conducted an exploratory, qualitative study with a rapid data collection design that integrated three complementary data sources:

1. PX survey data.

2. Rapid qualitative interviews with patients and families.

3. Naturalistic walk‐throughs of medicine units.

2.2. Setting and Unit Selection

THP operates 12 medicine units across two hospital sites, serving one of Canada's most demographically and clinically diverse populations, as over 60% of the population are immigrants [13]. Based on PX survey scores recorded in December 2024, 4 units were purposively selected to represent variation in PX performance and operational context. One high‐scoring and one low‐scoring unit were chosen at each of the two hospital sites, for a total of 4 units, provided each unit had at least 60 survey respondents. High‐performing units were defined as those with the highest overall PX score (on a 0–10 rate your overall experience survey question), and low‐performing units were defined as those with the lowest overall PX score among all 12 medicine units at THP. The sample size threshold was determined pragmatically to ensure sufficient data representation and reliability while reflecting the number of responses available across all medicine units at the time of selection. Although unit selection used a single month of PX data, December 2024 scores were reviewed against preceding months to confirm that the relative ranking of high‐ and low‐performing units was relatively consistent across months and supported by an adequate number of responses, reducing the likelihood that unit selection reflected short‐term fluctuation. This sampling strategy enabled the comparison of unit‐level experiences within the same organizational environment across the two sites and between high‐ and low‐performing units. All data were collected from the four selected medicine units.

2.3. PX Survey

The PX survey captured both quantitative and qualitative dimensions of care and assessed overall PX. PX survey questions asked patients to rate aspects of their hospital stay using the following questions on a scale of 1–4 (1 = never, 2 = sometimes, 3 = usually, 4 = always):

• Question 1 [communication]: Do you feel that there was good communication about your care between doctors, nurses, and hospital staff?

• Question 2 [information]: During this hospital stay, did you get all the information you needed about your condition and treatment?

• Question 3 [attentiveness]: Were you able to get a member of hospital staff to help you when you needed attention?

• Question 4 [emotional support]: Did you get the support you needed to help you with anxieties, fears, or worries you had during this hospital stay?

• Question 5 [engagement]: Were you involved as much as you wanted to be in decisions about your care and treatment?

• Question 6 [discharge clarity]: Before you left the hospital, did you have a clear understanding of all of your prescribed medications, including those you were taking before your hospital stay?

• Question 7 [discharge information]: Did you receive enough information from hospital staff about what to do if you were worried about your condition or treatment after you left the hospital?

The PX survey items are drawn from the Ontario Hospital Association (OHA)‐coordinated provincial PX measurement program, which provides common, validated PX instruments for use across Ontario hospitals [14]. Since 2014, these instruments have been selected and refreshed through OHA‐led review committees comprised of representatives from hospitals, physicians, patients, Health Quality Ontario (now part of Ontario Health), the Ministry of Health and Long‐Term Care, and the former Local Health Integration Networks, with input from national and international subject‐matter experts [14]. The patient voice was embedded in instrument development through a digital field test and a formal cognitive test led by a third‐party qualitative researcher, and a THP Patient and Family Partner reviewed the survey.

At THP, the survey is self‐administered in English and distributed to admitted patients by email (approximately every 30 days) and via posted QR codes. The PX survey is administered routinely to all admitted patients as part of standard hospital operations, rather than as an instrument specifically designed and implemented for this study. Because it was distributed automatically to all patients who received care at THP, the total number of patients invited is not captured, and a formal response rate cannot be calculated; only the number of completed responses is known. In addition to the fixed‐response rating items, the survey included optional open‐ended comment fields, which provided the qualitative dimension of the survey data.

2.4. Rapid Qualitative Interviews With Patients and Families

A team of three PhD‐trained qualitative researchers conducted unit visits between April 16 and May 1, 2025. Semi‐structured, rapid qualitative interviews [15, 16] were guided by an interview protocol co‐developed with the MoC team through discussions, available in full in the Supporting Information. Rapid qualitative approaches compress conventional qualitative timelines through team‐based data collection, structured documentation, and concurrent data collection and analysis, trading prolonged immersion for credible, decision‐relevant findings on operational timescales [15, 16]. This approach was selected because it matched the rapid‐cycle tempo of the LHS approach and the decision timeline of the MoC redesign. Topics that were discussed with patients during interviews included care priorities, perceptions of timeliness and responsiveness, relationships with staff, the influence of the physical environment, and suggestions for improvement. Researchers introduced themselves as members of the hospital's research institute and clarified their independent role from the care team, explained that participating in the interview would not affect the patient's care, and obtained consent to audio‐record the interview. Conversations ranged from 5 min to 1 h (median: 19:29) in duration according to patient preferences, and were audio‐recorded using Otter.ai, reflecting the amount of detail patients wanted to provide at the time of the conversation [17]. Each participant received a $10 gift card for their time and contribution. Demographic and social identity (e.g., age, sex, gender, race/ethnicity, and primary diagnosis) data were captured.

Eligible participants were patients aged 18 or older admitted to one of the four selected units during the study period (April to May 2025). We aimed to recruit 4–6 patients (and families, if available) from each unit. In consultation with charge nurses, researchers identified patients who had the cognitive ability to participate, sufficient English proficiency, clinical stability, and no additional infection‐control precautions precluding in‐room visits.

2.5. Naturalistic Walk‐Throughs

During each unit visit, researchers also conducted naturalistic walk‐throughs to capture the environmental and workflow contexts. For the walk‐throughs, researchers began at the units' entrances and slowly explored each unit, noting enablers and barriers to navigation. Fieldnotes and audio reflections documented aspects such as layout, crowding, accessibility, lighting, noise, signage, privacy, and movement patterns. These observations provided contextual depth to patient accounts and enabled triangulation between what patients described and what was observed. Walk‐throughs were conducted by the same researchers immediately before or after patient conversations on each unit and were documented through voice recordings via Otter.ai. This observational approach, consistent with naturalistic and environmental walk‐through methods used to examine how healthcare spaces shape PX [4, 5], situated patient accounts within the physical and operational characteristics of each unit, including wayfinding from the main atrium, corridor width and clutter, equipment storage around the nurses' station, room crowding, and night‐time noise.

2.6. Data Analysis

All recordings from Otter.ai were automatically converted to transcripts using the same application. Recordings were independently reviewed and transcripts corrected by the research team. Findings were discussed during team synthesis meetings that featured open conversation discussing unique or surprising findings and common insights across units. The analysis followed a three‐step process emphasizing rapid learning and translation:

1. Unit‐level synthesis: For the quantitative portion, the dashboard summary was reviewed for each unit that provides proportion of responses for each PX question. The dashboard data was presented as frequency counts of each PX question. The overall PX Score for each unit is the arithmetic mean of patients' responses to the single 0–10 “rate your overall experience” item, calculated as the sum of each rating multiplied by its number of responses divided by the total number of responses (e.g., ratings of 10 from three respondents, 9 from four, 7 from one, 1 from one, and 0 from one yield (30 + 36 + 7 + 1 + 0)/10 = 7.4 out of 10). Because this score is derived solely from the 0–10 overall experience item, which does not include “Not Applicable” or “I did not need attention” response options, these categories do not affect the PX Score; on the individual PX questions where they appear, such responses were retained as distinct response categories. For the qualitative portion, each researcher who collected data reviewed their transcripts and notes to summarize salient observations, environmental features, and PX interview findings for each unit into a single unit summary.

2. Cross‐unit comparison: The lead researcher (UM) reviewed all summaries and identified distinct themes relevant across the 4 units to highlight similarities and differences and an integrative understanding of issues, strengths, and opportunities.

3. Data integration: The lead researcher (UM) triangulated survey data with qualitative findings in two ways: convergence checking, which compared qualitative themes with corresponding PX survey domains to assess alignment or divergence; and contextual triangulation, which used fieldnotes from walk‐throughs to situate interview accounts within observable environmental or workflow conditions. Integrated insights were shared with the broader MoC team. The qualitative findings below are linked to PX survey questions to clarify connections, except for the physical environment, which is not captured in the PX survey.

To translate findings, the team developed a SHOP table (Strengths, Hurdles, Opportunities, Pressures), adapted from the SWOT framework [18]. Patients and families naturally identified strengths, hurdles, and pressures through surveys or interviews, while opportunities were co‐generated through analytic reflection during research team meetings.

2.7. Ethics Statement

This study received an exemption from the THP Research Ethics Board because it qualified as an evaluation of a local quality improvement initiative. Ethical and data storage principles were still followed, including obtaining verbal informed consent and ensuring patient confidentiality and privacy.

3. Results

A total of 390 surveys from patients were included in the PX survey analysis. Of the 390 survey respondents, 373 reported demographic characteristics: the mean age was 66.3 years, and 54% were men and 44% were women. Compared with 2021 Census estimates for Mississauga [19], the respondent sample under‐represented several racialized groups: South Asian respondents comprised 32.3% of racialized respondents versus 41.0% in the Census, and Chinese respondents 9.2% versus 11.8%, with Arab patients also under‐represented, whereas Black respondents were close to parity (12.5% vs. 11.1%). A total of 19 patients and families participated in interviews across the four medicine units. Of these, 17 were with patients alone and two were dyad interviews that included a family caregiver alongside the patient. Because of two caregivers, insights were analyzed within each dyad rather than as a separate group. The median age of interview participants was 74 years (range: 39–91); just over half were women (52.6%). The sample reflected ethnic and linguistic diversity, comprising participants who identified as South Asian, Caribbean, Vietnamese, and White‐European. Patient conditions ranged from acute exacerbations (e.g., COPD, asthma, infections) to chronic disease complications (e.g., dialysis, sickle cell crisis).

We first present the dashboard summary of all PX questions for each medicine unit in Table 1. We then map qualitative findings from the rapid qualitative interviews in each unit.

TABLE 1.

Top‐box results (%) for each PX question by unit.

	Unit 1 (n = 121)	Unit 2 (n = 79)	Unit 3 (n = 76)	Unit 4 (n = 114)
Average Overall PX Score (0–10)	7.3	8.0	7.5	8.5
Q1 Communication	58	63	62	70
Q2 Information	50	65	47	62
Q3 Attentiveness	67	70	75	76
Q4 Emotional support	51	60	55	61
Q5 Engagement	46	58	46	61
Q6 Discharge clarity	67	71	70	74
Q7 Discharge information	55	54	55	62

Note: Top‐box reflects the percentage of respondents selecting the most positive response option (“Always,” “Yes, always,” or “Completely”). Full response distributions available in the Supporting Information.

3.1. PX Survey

Table 1 presents the dashboard summary of all PX questions for each of the four medicine units. Overall, analysis of patients who responded to each PX question indicates that units with lower PX scores tend to have more negative extreme responses (e.g., “definitely no,” “never”). This suggests that quality improvement efforts should focus on understanding which experiences contribute to the more extreme responses and then address these concerns to achieve the largest increases in overall PX scores.

3.2. Synthesis of Findings From Rapid Qualitative Interviews

Across all 4 units, patients and their families expressed the following common themes about their experience in the unit. Each quote below is associated with the unit and participant number.

3.2.1. Communication Gaps and Uncertainty (Q1: Communication)

Patients consistently described fragmented, delayed, or ambiguous communication regarding tests, procedures, care plans, and discharge that resulted in a sustained sense of uncertainty that shaped their hospital experience. Two distinct forms of uncertainty emerged within this theme: diagnostic uncertainty, reflecting gaps in clinical communication and patients' ability to make sense of what was happening to their bodies, and care process uncertainty, reflecting ambiguity about scheduling, transport, discharge timelines, and whom to contact when plans changed. Information was often delivered in pieces, lacked follow‐up, or was disconnected from patients' embodied experiences that left them on their own to make sense of what was happening to their bodies and what would happen next with their care: “They might send me home today, but no one can tell me why the leg still burns” (3‐1).

Patients struggled to reconcile functional decline or persistent symptoms with the absence of clear explanations, timelines, or diagnostic closure. A patient (3) described repeated admissions since October, contrasting her prior independence (“I drive, I garden”) with the shock of sudden immobility (“then suddenly I have no legs”). Another patient (3‐1) went from gratitude at the start of her hospital experience for having access to care to bewilderment because of unresolved symptoms. Beyond this diagnostic uncertainty, patients also described care process uncertainty. Another patient described being moved and delayed without explanation, and being given approximate timeframes that changed rapidly without warning—describing being told that a procedure would occur “around 1 or 1:30 pm,” yet not knowing what the procedure involved or whom to contact if plans changed.

3.2.2. Compassionate Staff, Tempered by System Strain (Q4: Emotional Support)

Across all units, patients praised staff for their kindness, attentiveness, and empathy. Emotional support was most often experienced through small, relational gestures that conveyed presence, dignity, and care. Patients frequently used affirming language, such as “courteous,” “helpful,” and “very thoughtful,” to describe their interactions with staff. Moments of compassionate care were often subtle but deeply meaningful. One patient (1–5) described how a nurse pre‐emptively brought pillows and blankets to two patients positioned in corridor beds before being asked, interpreting this as a sign of thoughtfulness and anticipation of need. Another patient (1–6) recalled how their nurse repeatedly replenished hot water throughout the day, a small act that conveyed attentiveness.

At the same time, patients were acutely aware that emotional support was constrained by workload pressures and staffing shortages:

“I've been through a lot with a lot of nurses… I know at the beginning when I was here I wasn't really getting the right ones… They would take a really long time, which I do understand… I've been through a couple situations that really upset me. But I would say the last two weeks, it's been really, really good help. The nurses come fast” (4‐3).

Many patients explicitly framed these experiences as consequences of system strain. As one patient (2–5) put it, “If a nurse is overwhelmed, she can't be as kind.” Emotional availability, in this sense, was understood as contingent on time, competing demands, and the sheer volume of work. Patients frequently attributed inconsistent emotional support not to individual staff but to broader systemic pressures. Patients repeatedly emphasized that “the problem is not the people that work here. It's the system” (2–4). Another patient echoed this sentiment, noting that staff were overworked. Even when patients described having to become angry or assertive to receive care, they framed this as a response to structural constraints rather than individual neglect: “You can't blame them…they put up with a lot too…it's just the system” (2–4).

3.2.3. Waiting as a Silent Experience (Q3: Attentiveness)

Across units, patients described prolonged waiting for imaging, transport, call bell responses, repositioning, or discharge, often with silence because there were no explanations or updates. These delays were experienced as interruptions to care marked by discomfort, vulnerability, and uncertainty about what would happen next.

For patients who required assistance with basic bodily needs, waiting was particularly distressing:

“I know how busy they are… one time I waited two hours [for help to go to the washroom]… it was very upsetting of course through a lot of stomach issues. Yeah, it was really a bad experience” (4‐3).

Despite these experiences, this patient hesitated to ask for help or delayed using the call bell, out of empathy for overburdened staff:

“I see myself pressing the call button only when I really, really, really [need something]… I feel bad making them work more” (4‐3).

3.2.4. Physical Environment (Not Captured in PX Survey)

The unit's physical environment directly shaped patients' experiences of care, safety, and comfort. Patients noted positive features such as access to daylight and windows; however, these were frequently offset by spatial constraints, cluttered corridors, hallway beds, and compromised privacy. Environmental stressors—particularly noise, lighting, and navigational complexity—were described as adding strain to already vulnerable moments of hospitalization. These accounts aligned closely with the research team's naturalistic walk‐throughs, which independently documented narrow and cluttered corridors; equipment and pharmacy carts compressing the space around the nurses' station; variable night‐time noise; and wayfinding challenges in reaching units located several floors from the main atrium. Triangulating patient narratives with these direct observations strengthened confidence that environmental strain was a structural feature of the units, and these walk‐through observations are reflected alongside patient accounts in Table 2.

TABLE 2.

Strengths, hurdles, opportunities and pressures.

Strengths

Hurdles

Compassionate, Attentive Staff: A universal strength across all units was the presence of kind, compassionate, and attentive staff. Patients consistently described staff as thoughtful and emotionally present. In unit 3, staff were celebrated with praise. Similarly, in units 2 and 4 (units with higher PX scores), patients reported staff going beyond standard duties through proactive gestures such as providing blankets or emotional support in moments of distress. In unit 1, patients acknowledged a noticeable joy among staff, and patients viewed some staff as strong advocates for person‐centered care. However, not all staff were compassionate and attentive, and this is discussed more in hurdles and pressures. Responsiveness to Basic Physical and Emotional Needs: Staff across units frequently anticipated and responded to patients' needs without being asked, such as offering pillows, hot water, or simply showing emotional presence through physical touch or gentle words. This kind of proactive attentiveness was explicitly noted in units 2 and 4, which had overall higher PX scores. Physical Environment: Natural light and daylight exposure were positive aspects of all units, contributing to a therapeutic and uplifting atmosphere. Private rooms, clean washrooms, and physical infrastructure that supported patient comfort and privacy were strengths of units 2 and 4, which had overall higher PX scores. Supportive Communication and Understanding: The presence of bilingual staff in unit 3 was highlighted as a key asset, fostering a sense of cultural responsiveness and safety, especially for non‐English‐speaking patients.

Prolonged Wait Times and Delayed Responses: Patients in each unit reported frustration with delays while waiting for imaging, porters, medication, or staff responses to call buttons. In units 1 and 2, delays were broad and touched many different phases of care; in unit 3, call bell responsiveness was specifically problematic for patients. These delays not only impaired comfort and timely care but also contributed to anxiety and a sense of being overlooked. Communication Gaps and Uncertainty: A critical and recurring issue was confusion or uncertainty due to inconsistent or absent communication about care plans, treatment status, discharge timelines and supports. Patients in units 2, 3 and 4 described profound uncertainty, often prolonged, that generated stress and emotional fatigue. Poorly located information boards in rooms (unit 4), inattentive staff (unit 1), or fragmented communication across teams created an emotional fog in which patients were unsure of what was happening and why. Challenges with Physical Environments: Physical layout and clutter posed serious hurdles for some patients even though it was a strength for other patients. In units 2, 3 and 4, cluttered corridors obstructed mobility and undermined safety, especially for patients in recovery or with limited mobility. Crowded or shared rooms, hallway beds with minimal privacy, and inadequate family seating (unit 1 and 4) reduced patients' sense of dignity, privacy, and human connection. Emotional and Psychological Vulnerability: In both units 3 and 4, patients shared limited access to structured emotional and mental health supports, particularly for elderly, immigrant, or socially isolated individuals. In unit 4, one patient reported heightened emotional vulnerability based on previous personal experiences. While staff offered compassion, this was often ad hoc without the offer of more formal mental health supports and therefore left gaps in their care. Discharge Process is Stressful: In units 3 and 4, the discharge process emerged as a major hurdle because it was unclear, delayed, and lacking personalization, particularly for those with complex social circumstances (e.g., unhoused patients). Unclear transitions increased fear and distress, particularly among patients with fewer resources or those unfamiliar with navigating social systems after being discharged from the hospital. Cultural and Personal Needs Sometimes Overlooked: In unit 3, some patients noted the absence of simple but culturally meaningful comforts, like access to hot water, detracted from patients' sense of comfort and recognition. These comforts had large symbolic weight for some patients, not having them signaled a lack of attention to their cultural expectations in care.

Opportunities	Pressures
Strengthen Communication and Reduce Patient Uncertainty Bedside Journey Boards: Use digital tablets or laminated cards at bedside showing pending tests, estimated times, care team names, and real‐time updates. Supports continuity and reduces anxiety about unknowns. Bedside Updates and Structured Multidisciplinary Rounds: Implement regular, predictable updates to patients from nurses and allied health to clarify care plans and discharge expectations. Daily “What to Expect Today” Huddles: Delivered in the patient's preferred language to enhance comprehension and reduce emotional distress from uncertainty. Access to Medical Records and Test Results: Create channels for patients to view their results and records in real time to reduce confusion and build trust. Improve Response Times and Compassionate Care 15‐Minute Response Pledge with Rapid Assistance Role: Introduce aide/volunteer support for high‐demand times to meet toileting and urgent needs. Proactive Rounding for Emotional Support: Protected time each day for staff to check in with patients emotionally, especially during non‐peak hours. Rapid‐Response Protocols for Acute Discomfort or Risk: Establish clear thresholds for timely interventions that prioritize patient safety and emotional wellbeing. Deploy volunteers to support non‐clinical basic needs: Utilize trained volunteers to assist with tasks such as providing water, blankets, or emotional support during long waits, alleviating pressure on clinical staff and ensuring patients feel seen and supported during times of high demand. Reduce Physical and Environmental Strain Regular Hallway Clean‐ups: Maintain clear corridors by storing unused equipment in dedicated areas while ensuring staff access. Noise Mitigation and Sleep Support Bundles: Provide earplugs, soft‐close bins, cognitive distraction activities for restless patients, and dim corridor lighting during night hours. Quiet‐at‐Night Pact: Agreement signed by patients and staff to support a restful environment through behavior and infrastructure (e.g., silent hinges, dim lighting).	Workforce and Staffing Shortages: A central pressure across all units was understaffing and excessive workload, particularly for nurses. In units 3 and 4, patients explicitly linked this to delays in call bell response and diminished emotional presence. In units 2 and 4, these staffing issues were described by patients as visibly impacting the quality and consistency of compassionate care, and they suggested that even well‐intentioned staff are sometimes pushed beyond their limits, leading to hurried interactions and missed opportunities for empathetic, compassionate and person‐centered care. Patients Absorbing Systemic Pressures: A powerful and recurring insight is how patients internalize systemic pressures by often suppressing their own needs out of concern for overburdened staff. In units 1 and 4, patients delayed or withheld care requests, even when it compromised their wellness. Social Vulnerabilities and Structural Inequities: units 1 and 4 brought forward patients' socio‐material vulnerabilities (i.e., homelessness, immigration status, isolation), which can compound medical challenges and create complex care needs. These pressures are felt by patients and also absorbed by staff. Caregiver and Family Burdens: In units 1 and 4, caregiver burden emerged as a critical pressure point. Families were expected to “fill the gaps,” often without formal support, training, or respite. Patients in unit 1 described feeling pressured to have caregivers supplement their care, while unit 4 patients detailed the emotional and logistical strain on families, especially those caring for individuals with special needs or limited mobility. Loss of Humanity in the System: Unit 1 patients introduced a particularly reflective pressure—the erosion of trust and belief in the system's humanity. Patients referenced past traumas and recurrent disappointments that colored their current stay, creating a sense of moral injury, not just due to what happened, but because of what the system seemed to have become for them and staff.
Address Social and Structural Vulnerabilities Transition Navigators for Complex Discharges: Assign a navigator at admission for patients who are unhoused, uninsured, or lacking social supports. “Going Home Together” Conferences: Collaborative planning with patients, families, and care providers to coordinate post‐discharge care. Caregiver Support Initiatives: Offer structured help and education for families managing care at home. Improve Privacy and Dignity Repair and Provide Movable Privacy Partitions: Ensure temporary beds have sufficient visual and acoustic privacy to restore dignity and comfort. Culturally Responsive and Humanizing Practices Cultural Comfort Menu & Checklist: Includes hot water, multilingual signage, and space for personal or spiritual objects. Empower Patient Participation and Autonomy Mobility and Self‐Care Protocols: Educate patients on what they can do independently in the unit (e.g., walking around). Expanded Patient Engagement Materials: Provide frequent updates, contests, and interactive education to boost morale and engagement. Meaningful Staff Support and Wellbeing Staffing Acuity Index: Auto‐flag staffing shortages and deploy relief staff proactively to prevent burnout and neglect. Interpersonal Skills Training & Burnout Prevention: Provide training on communication and empathy for units in constant high‐stress clinical environments. Emphasize the development of interpersonal competencies, especially for transient staff (e.g., PSWs coming in and out of units for a specific purpose), that support effective collaboration among staff and foster compassionate interactions with patients and families. Burnout Prevention: Explore strategies to mitigate burnout by addressing the challenges that come with staff turnover and fragmentation, particularly among transient staff such as PSWs. Prioritize efforts to better integrate transient staff into unit workflows and interactions to ensure continuity of care and more cohesive interactions with patients and families. Patient and Staff Advocacy Forum: Create a hospital‐supported space where patients, families and frontline staff collaboratively identify systemic issues and co‐author letters or statements to policymakers, demonstrating institutional solidarity and commitment to broader health system change. Integrate Mental Health Integration of Mental Health Professionals: Include social workers or psychologists in rounds and assessments for patients showing distress. Emotional Distress Screening at Admission: Identify high‐risk patients early and offer structured psychological support.

Patients described hallways crowded with chairs, carts, medical equipment, and tables, which created barriers to safe movement and heightened anxiety, especially for those with mobility challenges. One patient undergoing daily rehabilitation described how navigating these spaces required constant vigilance:

“I'm doing therapy everyday with the team here…I do bump into it [the chairs and other equipment and tables lining the hallways]…I have to make sure I focus on my footsteps” (4‐3).

Noise and light at night were additional sources of distress. Some patients reported difficulty sleeping due to sounds from roommates or staff activity, with one patient noting an inability to rest. These disruptions undermined recovery and contributed to fatigue and irritability.

Navigation within the hospital was also described as challenging, particularly for patients and families unfamiliar with the layout. One unit was a cul‐de‐sac within a labyrinthine corridor, requiring passage through multiple units to reach it.

We provide a synthesis of qualitative findings across all units in a narrative format using the SHOP framework shown in Table 2. This synthesis was generated through a combination of suggestions from patients/families we interviewed, the PX survey, and discussions with research team members and the broader MoC project team. Some opportunities mentioned in the table below are directly from patients, while others were interpreted through patient discussions of strengths and hurdles.

4. Discussion

4.1. Study Summary

This study gathered data from patient, caregiver and family interviews, unit observations, and PX surveys to explore care experiences in four medicine units at Trillium Health Partners. Across all units, patients described a system under strain marked by long waits, inconsistent communication from staff, and unit layouts that sometimes compromised privacy and comfort, all against a backdrop of compassion and empathy for the staff who were overburdened. In the following sections, we reflect on learnings from data gathered for medicine unit MoC design and we argue that PX provides details on how we can create a better experience for patients and families.

4.2. From Quick Fixes to System Learning: Positioning PX Data as a System Feedback Asset

Most strengths, hurdles, opportunities and pressures referenced specific improvements that individual staff or teams can make targeted at one or two PX questions. This emphasis is consistent with the direction of several research studies and interventions in the healthcare improvement literature, which focus on what providers can do differently (e.g., training, scripts, and protocols aimed at changing communication and responsiveness). These efforts have produced valuable gains in certain contexts: communication training has improved patient satisfaction and clinical outcomes [20, 21]; hourly and intentional rounding has enhanced perceptions of nursing responsiveness [22, 23]; and standardized communication frameworks such as SBAR (Situation, Background, Assessment, Recommendation) have strengthened patient hand‐offs [24, 25]. Yet, evidence also cautions that such strategies can emphasize transactional (i.e., focus on diagnosis and treatment) rather than person‐centered care (i.e., focus on the person's goals, care and life circumstances). When systems are under strain [26], their implementation often encounters barriers related to workload and sustainability [27]. Under such conditions, clinicians may prioritize completing required behaviors encouraged by interventions over engaging with patients and their families because they are auditable, time‐bound and aligned with performance metrics. As a result, interventions can inadvertently narrow clinical encounters to brief, standardized exchanges that produce transactional forms of care when patients may need emotional support.

From a systems thinking perspective, these dynamics reflect what Forrester [28] and Senge [29] described as “fixes that fail” or single‐loop learning [30]: interventions that address symptoms without tackling root causes. When solutions focus on individual behavior, they can shift the burden onto clinicians rather than addressing systemic drivers such as understaffing, fragmented communication, and resource constraints [31]. Research confirms this pattern: provider‐focused interventions often target visible symptoms rather than structural feedback loops [32], mask design flaws through individual workarounds [33, 34], and rely too heavily on behavioral strategies while systemic structures go unaddressed [35]. As a result, the system achieves short‐term gains while perpetuating the pressures that erode sustainable, relational care.

PX provides a potential pathway out of this cycle by shifting attention from narrow performance indicators to patients' lived and living experience of care. Where quick fixes focus on surface‐level symptoms, patient stories can help trace these issues back to their systemic roots, a process referred to as double‐loop learning [30]. There is now a substantial body of literature demonstrating that meaningfully engaging patients, families, and communities—and co‐designing with them—is crucial not only as a form of engagement but as a catalyst for collective inquiry into the root causes of service failures and ensuring that interventions address systemic drivers rather than serving as short‐term fixes [36, 37, 38, 39].

Our findings illustrated this point: patients did not simply suggest the need for improved communication or faster response time, but described communication gaps, prolonged waiting, emotional vulnerability, and environmental barriers as potential downstream effects of understaffing, fragmented processes, and uncertainty around discharge. Distinguishing diagnostic uncertainty from care process uncertainty also points to different redesign responses, with structured diagnostic communication and shared decision‐making for the former, and reliable operational status updates for the latter. By capturing how patients experience the downstream effects of organizational strain, PX data can reveal where system design, resource allocation, and culture misalign with the realities of patients and staff. For example, experiences of diagnostic limbo, delayed call bell responses, or patients withholding requests for help out of concern for overburdened nurses point to structural constraints rather than isolated individual shortcomings, a pattern that we discerned through team‐based analysis in which PX data highlighted what mattered to patients and subsequent interpretive work unpacked why these experiences were occurring. In this sense, PX functions as a feedback signal that surfaces what matters most to patients, prompting further examination of underlying systemic conditions. However, when reduced to satisfaction scores or solely structured PX surveys, PX risks reproducing the very quick‐fix dynamics that it seeks to overcome by focusing solely on indicators without considering the systemic conditions that produce and reinforce them.

PX data must therefore be approached as a generative, relational process, situating patient and family expectations within the broader context of a system under strain and fostering collaborative dialog among patients, families, clinicians, and leaders. This requires learning from providers and staff, whose experiential knowledge connects what patients report to the organizational and workflow conditions that shape those experiences.

4.3. Operationalizing PX Data as Fuel for LHS

Reframing PX data as a system feedback asset is crucial for organizations seeking to operate as an LHS [9]. When PX data are interpreted primarily as feedback on staff performance, they risk perpetuating the “quick fix” cycle. To break this cycle, PX must be positioned as fuel for the LHS engine: a routine input alongside clinical, quality, and operational data that powers iterative learning and adaptation.

To move from quick fixes, we propose three layers of organizational change based on how PX is leveraged and implemented:

• Structural (explicit) changes: Embed PX into enterprise dashboards and board reviews as early‐warning data on system design and resource stress, directly linking PX evidence to decisions about staffing models, care environments, and policy priorities.

• Relational (semi‐explicit) changes: Establish joint patient–staff learning forums and embed co‐design throughout improvement processes, enabling participants to interpret PX data together and translate insights into feasible actions.

• Transformative (implicit) changes: Reframe the organizational narrative from “PX as evaluation of staff” to “PX as a system feedback asset,” using PX data to guide implementation and rapid‐cycle evaluation consistent with LHS gears 4 and 5 [12].

Operationalizing PX within an LHS required moving from data collection to data connection by linking patient insights directly to operational learning, governance, and redesign. This project surfaced several practical lessons for translating PX from bedside stories into a system asset:

PX survey data alone is not enough to capture the nuanced differences between patients and families. Structured PX surveys provided a useful high‐level signal but were limited in their ability to explain how experiences differed across patients, caregivers, and families. Regardless of whether a medicine unit had a high or low overall PX score, patients across all units reported both strengths and challenges; these nuances emerged more fully through open‐ended survey comments and interviews that elicited context, meaning, and diverse perspectives.

1. PX survey data is not enough to capture the nuanced differences between patients and families. Structured PX surveys provided a useful high‐level signal but were limited in their ability to explain how experiences differed across patients, caregivers, and families. Within the same units, patients, caregivers, and families often reported similar benefits alongside markedly different challenges. Moreover, regardless of whether the medicine unit had a high or low overall PX score, patients across all units reported both strengths and challenges. We were able to identify these nuances more completely through open‐ended data on PX survey and interviews with patients, caregivers, and families that elicited context, meaning, and diverse perspectives.

2. Broaden recruitment to widen voices. Recruitment through nurses supported safety and flow, but likely filtered out patients facing language, cognitive, or social barriers who may experience disproportionate strain. Future cycles should include direct researcher recruitment, on‐site interpretation, and community partner outreach to strengthen equity and reduce selection bias [12].

3. Rapid methods. Rapid methods yielded timely guidance for MoC redesign but limited longitudinal depth. Building repeat touchpoints into the LHS cycle would enable tracking of how PX shifts with system changes.

4. Integrate patient, family, and caregiver engagement as a core learning mechanism. This phase did not include deeper co‐design due to limited scope, but involving patients and caregivers through advisory groups or experience‐based co‐design could generate richer insights [40]. Future LHS phases should incorporate structured engagement to support participatory governance.

5. Translate findings with an intuitive heuristic. Adapting SWOT into SHOP (Strengths, Hurdles, Opportunities, Pressures) produced outputs that were both respectful of patient narratives and usable in redesign.

4.4. Strengths and Limitations

This study has several strengths and limitations. Its strengths include the integration of multiple complementary data sources (PX survey data, rapid qualitative interviews, and naturalistic walk‐throughs) within a real‐world LHS initiative, enabling triangulation of patient experiences with insights from the survey and naturalistic walk‐through. Several limitations should also be noted. First, the study was exploratory and based on a small sample of interviews collected through a rapid design so the findings are intended to be hypothesis‐generating. Second, interview participants were identified in consultation with unit charge nurses, which supported patient safety and flow but likely filtered out patients facing language, cognitive, or social barriers, who may experience the most acute effects of system strain. Third, because of the rapid nature of the work, patients and families were not involved in designing the interview guide or in selecting the environmental features assessed during walk‐throughs. Finally, unit selection drew on a single month of PX data, although rankings were checked against preceding months to reduce the risk of regression to the mean. Future LHS cycles should broaden recruitment, incorporate interpretation and direct researcher outreach, and embed patients and families in study design and in the identification of improvement priorities. Finally, the comparison of survey respondents with Census estimates reported above indicates some response bias in routinely collected PX data, with several racialized groups (notably South Asian, Chinese, and Arab patients) under‐represented relative to the surrounding population; this suggests that some of the most structurally marginalized patients may be under‐represented in PX survey data. Quantifying Indigenous identity is a further limitation, as Indigenous patients may not disclose their identity in hospital settings.

Funding

This work was supported by Trillium Health Partners.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Data S1: Supporting Information.

Table 1: Dashboard Summary of Each Unit Across PX questions accessed on June 13, 2025.

Data Availability Statement

The data that supports the findings of this study are available in the supplementary material of this article.