Video Abstract
Patient and Family Centered I-PASS (PFC I-PASS) emphasizes family and nurse engagement, health literacy, and structured communication on family-centered rounds organized around the I-PASS framework (Illness severity-Patient summary-Action items-Situational awareness-Synthesis by receiver). We assessed adherence, safety, and experience after implementing PFC I-PASS using a novel “Mentor-Trio” implementation approach with multidisciplinary parent-nurse-physician teams coaching sites.
Hybrid Type II effectiveness-implementation study from 2/29/19-3/13/22 with ≥3 months of baseline and 12 months of postimplementation data collection/site across 21 US community and tertiary pediatric teaching hospitals. We conducted rounds observations and surveyed nurses, physicians, and Arabic/Chinese/English/Spanish-speaking patients/parents.
We conducted 4557 rounds observations and received 2285 patient/family, 1240 resident, 819 nurse, and 378 attending surveys. Adherence to all I-PASS components, bedside rounding, written rounds summaries, family and nurse engagement, and plain language improved post-implementation (13.0%–60.8% absolute increase by item), all P < .05. Except for written summary, improvements sustained 12 months post-implementation. Resident-reported harms/1000-resident-days were unchanged overall but decreased in larger hospitals (116.9 to 86.3 to 72.3 pre versus early- versus late-implementation, P = .006), hospitals with greater nurse engagement on rounds (110.6 to 73.3 to 65.3, P < .001), and greater adherence to I-PASS structure (95.3 to 73.6 to 72.3, P < .05). Twelve of 12 measures of staff safety climate improved (eg, “excellent”/“very good” safety grade improved from 80.4% to 86.3% to 88.0%), all P < .05. Patient/family experience and teaching were unchanged.
Hospitals successfully used Mentor-Trios to implement PFC I-PASS. Family/nurse engagement, safety climate, and harms improved in larger hospitals and hospitals with better nurse engagement and intervention adherence. Patient/family experience and teaching were not affected.
What’s Known on This Subject:
Family-centered rounds improve patient safety and hospital experience. However, family-centered rounds buy-in, adherence, and implementation strategies vary. Mentored-Implementation can successfully coach sites in implementing hospital-based interventions but is typically physician-centric and does not include patients/families and nurses.
What This Study Adds:
We successfully implemented a family-centered rounds intervention across 21 hospitals using novel Mentor-Trios involving parent-nurse-physician triads coaching sites. Safety climate and rounds adherence improved post-implementation; resident-reported harms decreased in larger hospitals and those with greater intervention adherence and nurse engagement.
Communication failures frequently cause harmful errors.1 High-reliability structured provider communication improves patient safety.2 However, communication with patients and families is understudied. Patient and family-centered rounds (PFCR) improve provider-patient and family communication3 –6 and are a best practice per the Institute for Patient and Family-Centered Care3 and the American Academy of Pediatrics.7,8 Although common in pediatrics, PFCR have not fully achieved their potential to improve communication because they are often implemented inconsistently, more physician- than family-centered, replete with medical jargon, and inconsistently engage patients and families and nurses.4,5,9 –12 Support for PFCR also varies among residents, attending physicians, and nurses.13
We previously coproduced with parents, nurses, and physicians a PFCR intervention, Patient and Family Centered I-PASS (PFC I-PASS).14,15 PFC I-PASS emphasized family and nurse engagement, health literacy, and high-reliability structured bidirectional written and verbal communication organized around the I-PASS (Illness severity-Patient summary-Action items-Situational awareness-Synthesis by receiver) framework. After implementing PFC I-PASS in 7 centers, harmful medical errors dropped 38% and hospital experience improved; rounds duration and teaching remained unchanged.14
To inform widespread dissemination, we sought to test PFC I-PASS implementation and real-world effectiveness in a broader group of US community and tertiary pediatric teaching hospitals. Recognizing that hospital-based quality improvement (QI) interventions are often physician-centric or nurse-centric,16 we implemented PFC I-PASS using a novel interdisciplinary implementation strategy involving coaching by parent, nurse, and physician mentors (Mentor-Trios). We hypothesized that intervention adherence and sustainment, safety climate, harmful errors, and hospital experience would improve post-implementation.
Methods
Design
We conducted a prospective hybrid-type-II-effectiveness-implementation17 study on 21 US teaching hospitals’ pediatric inpatient units from February 29, 2019 to March 13, 2022. Hybrid-type II-effectiveness-implementation studies equally focus on intervention effectiveness and implementation.17 We collected staggered data across sites (≥3 months pre-implementation, then 12-months intensive Mentored-Implementation data collection/site). Boston Children’s Hospital’s Institutional Review Board provided approval.
Sites
Thirty sites applied to participate, providing information about hospital, residency program, study team, PFCRs, and I-PASS handoffs experience. A multidisciplinary parent-nurse-physician team selected 21 sites (Supplemental Fig 5) based on geographic diversity, hospital type (tertiary or community), residency size, and PFCR experience.
Sites included 14 nested children’s hospitals, 4 freestanding children’s hospitals, and 3 general hospitals. Twelve were tertiary care teaching hospitals, and 9 community teaching hospitals. Residency size ranged from 16 to 150 residents. Rounding team structure varied (eg, interns and third-year-residents versus interns and second-year-residents). Twenty hospitals implemented on general pediatric (nonintensive care) units, and 1 on pediatric neurology. Sites variably practiced PFCRs (<5 years [n = 5], 5–10 years [n = 9], >10 years [n = 7]); none previously used I-PASS for PFCR.
Intervention
Our intervention, PFC I-PASS,14,18 is a structured verbal and written rounds communication framework organized around I-PASS structure, supported by communication, teamwork, and health literacy best practices to ensure a shared mental model (Fig 1).14 We adapted PFC I-PASS for larger-scale implementation by streamlining resident training and incorporating novel educational techniques (eg, flipped-classroom, virtual learning).15,19
PFC I-PASS intervention. Key intervention supporting activities include education and training activities for medical students, nurses, residents, and attending physicians using a flipped classroom and interactive simulations, as well as implementation and sustainment activities that ensure practice change, including observations and feedback. The PFC I-PASS rounding process included several required elements, including prerounds planning, accounting for interpretation needs, taking infection-control precautions, and ensuring all key team members were present. During rounds, providers introduced team members, invited patient or family to share first, nurse to share second, and then resident or advanced care provider to share next. Health literacy best principles (eg, avoiding medical jargon, using simple vocabulary), active listening, and nonverbal cues were incorporated into rounds. Rounding communication was structured through the I-PASS format, with Illness severity provided by the patient or family, Patient summary provided by the nurse and resident or advanced care provider, Action items and Situational awareness provided by the resident or advanced care provider, and Synthesis provided by the patient or family with input from nursing. Providers then completed a written summary of rounds (Rounds Report). The Mentored Implementation model, pioneered by the Society of Hospital Medicine, involves dedicated mentorship of each site by a mentor with expertise in the intervention. In our study, we modified Mentored Implementation to include a novel “Mentor-Trio” approach where each site was paired with a Mentor-Trio consisting of a parent, nurse, and physician mentor. Mentor-Trios met with Mentee-trios (parent, nurse, and physician champions) at each of the 21 sites monthly during the duration of the project to review stakeholder engagement, assess intervention adherence through review of monthly run charts, brainstorm potential plan-do-study-act cycles, and discuss intervention challenges and successes.
PFC I-PASS intervention. Key intervention supporting activities include education and training activities for medical students, nurses, residents, and attending physicians using a flipped classroom and interactive simulations, as well as implementation and sustainment activities that ensure practice change, including observations and feedback. The PFC I-PASS rounding process included several required elements, including prerounds planning, accounting for interpretation needs, taking infection-control precautions, and ensuring all key team members were present. During rounds, providers introduced team members, invited patient or family to share first, nurse to share second, and then resident or advanced care provider to share next. Health literacy best principles (eg, avoiding medical jargon, using simple vocabulary), active listening, and nonverbal cues were incorporated into rounds. Rounding communication was structured through the I-PASS format, with Illness severity provided by the patient or family, Patient summary provided by the nurse and resident or advanced care provider, Action items and Situational awareness provided by the resident or advanced care provider, and Synthesis provided by the patient or family with input from nursing. Providers then completed a written summary of rounds (Rounds Report). The Mentored Implementation model, pioneered by the Society of Hospital Medicine, involves dedicated mentorship of each site by a mentor with expertise in the intervention. In our study, we modified Mentored Implementation to include a novel “Mentor-Trio” approach where each site was paired with a Mentor-Trio consisting of a parent, nurse, and physician mentor. Mentor-Trios met with Mentee-trios (parent, nurse, and physician champions) at each of the 21 sites monthly during the duration of the project to review stakeholder engagement, assess intervention adherence through review of monthly run charts, brainstorm potential plan-do-study-act cycles, and discuss intervention challenges and successes.
Our implementation strategy (Mentored Implementation), pioneered by the Society of Hospital Medicine20 in multiple domains,21 –23 involves content experts (typically physicians) intensively coaching hospitals to implement large-scale QI projects. Informed by principles of coproduction24 with families and nurses, we refined Mentored Implementation using a novel Mentor-Trio approach, where triads of parent-nurse-physician mentors drawn from 7 parents, 6 nurses, and 16 physicians coached sites in implementing PFC I-PASS. Most Mentors previously participated in our 7-center PFC I-PASS Study.14 Experienced senior physician-mentors further guided Mentor-Trios. Mentors participated in monthly all-mentor calls for mentor training and support on PFC I-PASS and received pretemplated slide decks with standardized, tailored prompts for monthly site calls. We paired central Mentor-Trios with local “Mentee-Trios” (parent, nurse, and physician project champions) leading site-level implementation. Mentor-Trios met with sites monthly via teleconference to review stakeholder engagement, intervention adherence (run charts), plan-do-study-act cycles, and intervention challenges and strategies.
When COVID-19 began, sites varied in implementation phase: pre-implementation (n = 1), early-implementation (months 1–6) (n = 13), late-implementation (months 7–12)(n = 7). We made several ensuing rounds adaptions (eg, incorporating telemedicine, Supplemental Fig 6) and paused data collection for 3 to 6 months across sites (March 2020–September 2020).
Data Collection and Measurement
To measure intervention adherence, site providers and, rarely, site parents observed rounds in real-time for ≥3 patients per week per site using a previously developed rounds observation tool.2,5,14 During the pandemic, some observations occurred virtually. Observers underwent video- and simulation-based training and rated key rounding aspects, including format (ie, rounding at bedside) (closed-ended); I-PASS component adherence (dichotomous); communication quality (4-point-Likert-extent-scale); nurse presence (4-point-categorical-scale), family and nurse engagement (5-point-behaviorally-anchored-scale); and teaching (dichotomous).
To measure safety climate and resident-reported harms, we surveyed resident-physicians at end-of-rotation (approximately every 2–4 weeks) and nurses and attending physicians at staff meetings. Staff provided completion-based-consent, completing surveys on paper or REDCap. For safety climate, we assessed select Agency for Healthcare Research and Quality Survey on Patient Safety Culture25 5-point-Likert-scale items, including working as a team (strongly disagree to strongly agree), staff freely speaking up (never to always), and safety grade (failing to excellent). Participants also reported minor or major harms (adverse events) caused by communication failures using items modified from prior studies.14,21 Minor harms had limited clinical consequence (ie, requiring more frequent monitoring or transient discomfort, without prolonging hospitalization, significant organ dysfunction, or worsening clinical condition). Major harms had significant clinical consequence (ie, deteriorating clinical status, organ dysfunction, prolonged hospitalization, disability beyond discharge, death). We refined staff surveys using cognitive interviewing and piloting.
For patient and family experience, we surveyed 2 to 4 randomly selected Arabic, Chinese, English, or Spanish-speaking inpatients (≥13-years-old) or parents and caregivers (“families”) weekly per site about safety climate, experience, and demographics. Families provided verbal consent with interpreters when necessary; patients ≥13-years-old provided assent. Patients and families self-completed paper (all 4 languages) or REDCap (English) surveys. Surveys included select 5-point Likert-scale Children’s Hospital Safety Climate Questionnaire26 items about speaking up, questioning decisions, and fear of asking questions (strongly disagree to strongly agree). They also included hospital experience items modified from prior research relating to how well doctors and nurses addressed concerns and made them feel part of the team (not at all to extremely).14 They also completed Universal Health Literacy Precautions Toolkit items for language proficiency,27 the 3-item screener for health literacy,28 and self-reported demographics (including race and ethnicity). Surveys had a sixth grade Flesch-Kincaid reading level and underwent piloting and cognitive interviewing and professional translation into Arabic, simplified Chinese, and Spanish.
Outcomes
Our primary implementation outcome was adherence to all 5 I-PASS components on rounds (dichotomous). Our primary clinical-effectiveness outcome was overall resident-reported harm (major+minor) rates.21 Secondary outcomes included top-2-box-rated safety climate and patient and family experience scores. (“top-2-box” = top-2 Likert-scale options, eg, strongly agree and agree). We defined implementation, clinical-effectiveness, service, and experience outcomes using implementation research standards.29 Process outcomes included family engagement (top-2-box), nurse engagement (top-2-box), plain language (top-2-box), and teaching (dichotomous).
Analyses
We compared respondent sociodemographic characteristics by intervention time-period (pre vs early- vs late-implementation) using χ-square tests or Fisher’s exact tests. We estimated adherence rates to key intervention elements for each implementation month (normalized to intervention go-live, not calendar-month) using mixed-effects logistic regression with random intercepts controlling for site-level clustering. These included individual I-PASS components (eg, illness severity) and composite outcomes (adherence to all I-PASS components). Interrupted-time-series30 tested for intervention effect on linear trend over time of estimated monthly adherence rates.
We used mixed-effects Poisson regression to estimate resident-reported harm rates, testing for changes over time via tests for linear trend over time-periods.
For safety climate and experience, we dichotomized the sample into participants with and without top-2-box scores (4–5 vs 1–3 of 5). We examined bivariate associations between having top-2-box scores and intervention time-period using mixed-effects logistic regression and tested for trends over time. We adjusted top-2-box staff safety climate scores for characteristics varying across time-periods.
We explored whether the intervention’s effect on key outcomes in the late-implementation period differed by hospital size (number of pediatric staffed beds), family engagement, nurse engagement, or adherence to all I-PASS components. We dichotomized hospital size, family engagement, nurse engagement, and adherence by median hospital values for each characteristic. To test for effect-modification of hospital characteristics, we evaluated differences in trends in outcomes over time by hospital characteristic by including an interaction-term between the characteristic and time-period in regression models. Statistical significance was achieved with a 2-sided P value <.05. We used REDCap31 for data collection and management and SAS v9.4 (SAS Institute) for analyses.
Results
Participant Characteristics
We conducted 4457 rounds observations, and received surveys from 2285 patients and families, 1240 resident-physicians, 819 nurses, and 378 attending-physicians. Survey response rates were 72.8%, 50.6%, 33.1%, and 41.7%, respectively (53.0% overall). Most patient and family participants were parents (89.2%), female (82.6%), 18 to 34 years-old (48.1%), had some college (64.4%), and an income <$50000 per year (45.2%). The most common race and ethnicity was white non-Hispanic (49.1%), followed by Hispanic (23.0%), and Black non-Hispanic (14.0%); 13.8% of respondents had limited English proficiency and 24.9% had limited health literacy. Patient and family characteristics did not differ significantly between implementation periods. Staff were predominantly resident-physicians (50.9%), female (75.4%), 18 to 34 years-old (69.2%), white non-Hispanic (65.0%), and in practice 1 to 5 years (40.2%). Staff role, gender, age, and years in practice significantly varied by implementation-period (Table 1-4).
Patient and Family Characteristics (n = 2285)
| Characteristic | Overall, N (2285) (%) | Pre-implementation, N (619) (%) | Early-implementation, N (866) (%) | Late-implementation, N (800) (%) |
|---|---|---|---|---|
| Relationship to patient (n = 2256) | ||||
| Patient | 155 (6.9) | 46 (7.5) | 61 (7.1) | 48 (6.1) |
| Parent | 2012 (89.2) | 541 (88.3) | 765 (89.6) | 706 (89.5) |
| Grandparent | 58 (2.6) | 19 (3.1) | 16 (1.9) | 23 (2.9) |
| Othera | 31 (1.4) | 7 (1.1) | 12 (1.4) | 12 (1.5) |
| Age, y (n = 2248) | ||||
| <18 | 100 (4.4) | 33 (5.4) | 38 (4.5) | 29 (3.7) |
| 18–34 | 1081 (48.1) | 299 (49.0) | 422 (49.6) | 360 (45.7) |
| 35–54 | 1000 (44.5) | 261 (42.8) | 368 (43.2) | 371 (47.1) |
| ≥55 | 67 (3.0) | 17 (2.8) | 23 (2.7) | 27 (3.4) |
| Gender (n = 2230) | ||||
| Female | 1843 (82.6) | 501 (82.3) | 693 (82.2) | 649 (83.4) |
| Male | 387 (17.4) | 108 (17.7) | 150 (17.8) | 129 (16.6) |
| Race and ethnicity (n = 2238) | ||||
| Asian, non-Hispanic | 132 (5.9) | 35 (5.8) | 48 (5.7) | 49 (6.2) |
| Black, non-Hispanic | 314 (14.0) | 81 (13.4) | 110 (13.0) | 123 (15.7) |
| Hispanic or Latino | 515 (23.0) | 142 (23.5) | 195 (23.0) | 178 (22.7) |
| White, non-Hispanic | 1099 (49.1) | 300 (49.7) | 412 (48.5) | 387 (49.3) |
| Other, non-Hispanicb | 178 (8.0) | 46 (7.6) | 84 (9.9) | 48 (6.1) |
| Language proficiency (n = 2253) | ||||
| Limited English proficiency (LEP)c | 312 (13.8) | 83 (13.7) | 121 (14.1) | 108 (13.7) |
| English proficient | 1941 (86.2) | 524 (86.3) | 736 (85.9) | 681 (86.3) |
| Language preferred by patients or families with LEP (n = 312) | ||||
| Arabic | 2 (0.6) | 1 (1.2) | 1 (0.8) | 0 (0) |
| Chinese | 4 (1.3) | 1 (1.2) | 1 (0.8) | 2 (1.9) |
| English | 159 (51.0) | 48 (57.8) | 61 (50.4) | 50 (46.3) |
| Portuguese | 1 (0.3) | 1 (1.2) | 0 (0) | 0 (0) |
| Spanish | 121 (38.8) | 29 (34.9) | 45 (37.2) | 47 (43.5) |
| Otherd | 14 (4.5) | 3 (3.6) | 8 (6.6) | 3 (2.8) |
| Health literacy (n = 2247) | ||||
| Limited health literacye | 560 (24.9) | 149 (24.6) | 231 (27.1) | 180 (22.8) |
| Adequate health literacy | 1687 (75.1) | 456 (75.4) | 623 (73.0) | 608 (77.2) |
| Education (n = 2232) | ||||
| No college | 794 (35.6) | 220 (36.5) | 308 (36.5) | 266 (33.8) |
| ≥Some college | 1438 (64.4) | 383 (63.5) | 535 (63.5) | 520 (66.2) |
| Annual household income (n = 2285) | ||||
| <$14 999 | 285 (12.5) | 71 (11.5) | 112 (12.9) | 102 (12.8) |
| $15 000–$49 999 | 747 (32.7) | 225 (36.4) | 276 (31.9) | 246 (30.8) |
| $50 000–$99 999 | 540 (23.6) | 138 (22.3) | 200 (23.1) | 202 (25.3) |
| ≥$100 00 | 479 (21.0) | 119 (19.2) | 196 (22.6) | 164 (20.5) |
| Missing | 234 (10.2) | 66 (10.7) | 82 (9.5) | 86 (10.8) |
| Characteristic | Overall, N (2285) (%) | Pre-implementation, N (619) (%) | Early-implementation, N (866) (%) | Late-implementation, N (800) (%) |
|---|---|---|---|---|
| Relationship to patient (n = 2256) | ||||
| Patient | 155 (6.9) | 46 (7.5) | 61 (7.1) | 48 (6.1) |
| Parent | 2012 (89.2) | 541 (88.3) | 765 (89.6) | 706 (89.5) |
| Grandparent | 58 (2.6) | 19 (3.1) | 16 (1.9) | 23 (2.9) |
| Othera | 31 (1.4) | 7 (1.1) | 12 (1.4) | 12 (1.5) |
| Age, y (n = 2248) | ||||
| <18 | 100 (4.4) | 33 (5.4) | 38 (4.5) | 29 (3.7) |
| 18–34 | 1081 (48.1) | 299 (49.0) | 422 (49.6) | 360 (45.7) |
| 35–54 | 1000 (44.5) | 261 (42.8) | 368 (43.2) | 371 (47.1) |
| ≥55 | 67 (3.0) | 17 (2.8) | 23 (2.7) | 27 (3.4) |
| Gender (n = 2230) | ||||
| Female | 1843 (82.6) | 501 (82.3) | 693 (82.2) | 649 (83.4) |
| Male | 387 (17.4) | 108 (17.7) | 150 (17.8) | 129 (16.6) |
| Race and ethnicity (n = 2238) | ||||
| Asian, non-Hispanic | 132 (5.9) | 35 (5.8) | 48 (5.7) | 49 (6.2) |
| Black, non-Hispanic | 314 (14.0) | 81 (13.4) | 110 (13.0) | 123 (15.7) |
| Hispanic or Latino | 515 (23.0) | 142 (23.5) | 195 (23.0) | 178 (22.7) |
| White, non-Hispanic | 1099 (49.1) | 300 (49.7) | 412 (48.5) | 387 (49.3) |
| Other, non-Hispanicb | 178 (8.0) | 46 (7.6) | 84 (9.9) | 48 (6.1) |
| Language proficiency (n = 2253) | ||||
| Limited English proficiency (LEP)c | 312 (13.8) | 83 (13.7) | 121 (14.1) | 108 (13.7) |
| English proficient | 1941 (86.2) | 524 (86.3) | 736 (85.9) | 681 (86.3) |
| Language preferred by patients or families with LEP (n = 312) | ||||
| Arabic | 2 (0.6) | 1 (1.2) | 1 (0.8) | 0 (0) |
| Chinese | 4 (1.3) | 1 (1.2) | 1 (0.8) | 2 (1.9) |
| English | 159 (51.0) | 48 (57.8) | 61 (50.4) | 50 (46.3) |
| Portuguese | 1 (0.3) | 1 (1.2) | 0 (0) | 0 (0) |
| Spanish | 121 (38.8) | 29 (34.9) | 45 (37.2) | 47 (43.5) |
| Otherd | 14 (4.5) | 3 (3.6) | 8 (6.6) | 3 (2.8) |
| Health literacy (n = 2247) | ||||
| Limited health literacye | 560 (24.9) | 149 (24.6) | 231 (27.1) | 180 (22.8) |
| Adequate health literacy | 1687 (75.1) | 456 (75.4) | 623 (73.0) | 608 (77.2) |
| Education (n = 2232) | ||||
| No college | 794 (35.6) | 220 (36.5) | 308 (36.5) | 266 (33.8) |
| ≥Some college | 1438 (64.4) | 383 (63.5) | 535 (63.5) | 520 (66.2) |
| Annual household income (n = 2285) | ||||
| <$14 999 | 285 (12.5) | 71 (11.5) | 112 (12.9) | 102 (12.8) |
| $15 000–$49 999 | 747 (32.7) | 225 (36.4) | 276 (31.9) | 246 (30.8) |
| $50 000–$99 999 | 540 (23.6) | 138 (22.3) | 200 (23.1) | 202 (25.3) |
| ≥$100 00 | 479 (21.0) | 119 (19.2) | 196 (22.6) | 164 (20.5) |
| Missing | 234 (10.2) | 66 (10.7) | 82 (9.5) | 86 (10.8) |
Patient and parent characteristics derived from self-reported survey data. Counts may not add up to column totals because of missing data (<3% missing except for annual household income). Patient and parent characteristics derived from self-reported survey data.
Other relationship category includes 31 respondents who provided the following write-in categories: l/g(n = 1); foster parent (n = 4); primo (n = 1); sister/primary caregiver (n = 1); cousin (n = 2); aunt, legal guardian (n = 1); aunt (n = 2); ant [sic] (n = 1); guardian (n = 3); uncle (n = 1); foster mom (n = 2); foster parent (n = 1); brother (n = 1); sister (n = 4); hermana (n = 1); girlfriend (n = 1); sister of patient (n = 1), guardian/caregiver (n = 1); 2 who did not further specify the “other” relationship.
The “Other, non-Hispanic” category for race and ethnicity includes 69 respondents who selected more than 1 race: white and other (n = 6); Native Hawaiian/other Pacific Islander and other (n = 1); Native Hawaiian/other Pacific Islander and white (n = 5); Native Hawaiian/other Pacific Islander and white and other (n = 1); Black and other (n = 2); Black and white (n = 17); Black and white and other (n = 1); Black and Native Hawaiian/other Pacific Islander (n = 1); Asian and other (n = 2); Asian and white (n = 4); Asian and Native Hawaiian/other Pacific Islander (n = 6); Asian and Native Hawaiian/other Pacific Islander and other (n = 1); Asian and Native Hawaiian/other Pacific Islander and white (n = 3); Asian and Black (n = 1); American Indian or Alaskan Native and white (n = 13); American Indian or Alaskan Native and Native Hawaiian/other Pacific Islander (n = 1); American Indian or Alaskan Native and Black (n = 1); American Indian or Alaskan Native and Black and white (n = 2); American Indian or Alaskan Native and Asian and Black (n = 1). Forty four respondents also provided the following write-in categories: Middle Eastern (n = 4); from Middle East (n = 1); mixed (Indian/Black Jamaican) (n = 1); Arabian (n = 1); Arabic (n = 2); Samoan (n = 2); French Indian/German (n = 1); mix Hawaiian (?)[sic] (n = 1); India (n = 1); Ukrainian (n = 1), human (n = 2); multiracial (n = 2); Native Hawaiian (n = 1); Filipino (n = 2); Aboriginal American (n = 1), Pakistani (n = 1); Indian (n = 2); mixed (n = 1); Micronesian (n = 1); Latino (n = 1); Latina (n = 1); Hispanic (n = 2); Armenian (n = 1); Native (n = 1), Puerto Rican (n = 1); Jewish (n = 1); Phillipino [sic] (n = 1); everything – mix plate (n = 1); Moroccan (n = 1); Somali (n = 2); North African/Middle Eastern (n = 1); Samoans (n = 1); Indian, Black, and white (n = 1). The remainder of respondents selecting “other” did not further specify their race.
Limited English proficiency was defined as responding anything other than “very well” to the question “how well do you speak English?” Response items ranged from very well, well, not well, to not at all. This item was taken from the Guide to Implementing the Health Literacy Universal Precautions Toolkit.27
Other languages preferred by patients and families with LEP includes 14 respondents who provided the following write-in categories: American sign language (n = 1); Bengali (n = 1); Haitian Creole (n = 1); Hindi (n = 1); Polish (n = 1); Russian (n = 2); Somali (n = 4); Urdu (n = 1); and 2 who did not further specify the “other” language.
Limited health literacy was defined as selecting a non-top-box answer for 1 or more of the following 3 questions: (1) “How often do you have someone (like a family member, friend, hospital/clinic worker, or caregiver) help you read hospital materials?” (non-top-box responses: all, most, some, or a little of the time); (2) “How often do you have problems learning about your medical condition because of difficulty understanding written information?” (non-top-box responses: all, most, some, or a little of the time); and (3) “How confident are you filling out forms by yourself?” (non-top-box responses: quite a bit, somewhat, a little bit, or not at all). This item was taken from the Chews 3-item screener.28
Resident Characteristics (n = 1240)
| Characteristic | Overall, N (1240) (%) | Pre-implementation, N (363) (%) | Early-implementation, N (457) (%) | Late-implementation, N (420) (%) | P |
|---|---|---|---|---|---|
| Age, y (n = 1221) | .22 | ||||
| 18–34 | 1155 (94.6) | 336 (94.9) | 425 (94.2) | 394 (94.7) | |
| 35–44 | 58 (4.8) | 13 (3.7) | 24 (5.3) | 21 (5.1) | |
| ≥45 | 8 (0.7) | 5 (1.4) | 2 (0.4) | 1 (0.2) | |
| Gender (n = 1225) | .09 | ||||
| Female | 826 (67.4) | 240 (67.2) | 307 (67.6) | 279 (67.4) | |
| Male | 352 (28.7) | 109 (30.5) | 127 (28.0) | 116 (28.0) | |
| Othera | 2 (0.2) | 2 (0.6) | 0 (0) | 0 (0) | |
| Declined | 45 (3.7) | 6 (1.7) | 20 (4.4) | 19 (4.6) | |
| Race and ethnicity (n = 1187) | .80 | ||||
| Asian, non-Hispanic | 221 (18.6) | 62 (18.2) | 76 (17.2) | 83 (20.4) | |
| Black, non-Hispanic | 72 (6.1) | 21 (6.2) | 23 (5.2) | 28 (6.9) | |
| Hispanic or Latino | 106 (8.9) | 32 (9.4) | 40 (9.1) | 34 (8.4) | |
| White, non-Hispanic | 675 (56.9) | 192 (56.5) | 261 (59.2) | 222 (54.7) | |
| Other, non-Hispanicb | 113 (9.5) | 33 (9.7) | 41 (9.3) | 39 (9.6) | |
| Years of experience (n = 1227) | .16 | ||||
| <1 y | 637 (51.9) | 173 (48.5) | 248 (54.9) | 216 (51.7) | |
| 1–5 y | 586 (47.8) | 184 (51.5) | 201 (44.5) | 201 (48.1) | |
| 6–15 y | 4 (0.3) | 0 (0) | 3 (0.7) | 1 (0.2) | |
| Characteristic | Overall, N (1240) (%) | Pre-implementation, N (363) (%) | Early-implementation, N (457) (%) | Late-implementation, N (420) (%) | P |
|---|---|---|---|---|---|
| Age, y (n = 1221) | .22 | ||||
| 18–34 | 1155 (94.6) | 336 (94.9) | 425 (94.2) | 394 (94.7) | |
| 35–44 | 58 (4.8) | 13 (3.7) | 24 (5.3) | 21 (5.1) | |
| ≥45 | 8 (0.7) | 5 (1.4) | 2 (0.4) | 1 (0.2) | |
| Gender (n = 1225) | .09 | ||||
| Female | 826 (67.4) | 240 (67.2) | 307 (67.6) | 279 (67.4) | |
| Male | 352 (28.7) | 109 (30.5) | 127 (28.0) | 116 (28.0) | |
| Othera | 2 (0.2) | 2 (0.6) | 0 (0) | 0 (0) | |
| Declined | 45 (3.7) | 6 (1.7) | 20 (4.4) | 19 (4.6) | |
| Race and ethnicity (n = 1187) | .80 | ||||
| Asian, non-Hispanic | 221 (18.6) | 62 (18.2) | 76 (17.2) | 83 (20.4) | |
| Black, non-Hispanic | 72 (6.1) | 21 (6.2) | 23 (5.2) | 28 (6.9) | |
| Hispanic or Latino | 106 (8.9) | 32 (9.4) | 40 (9.1) | 34 (8.4) | |
| White, non-Hispanic | 675 (56.9) | 192 (56.5) | 261 (59.2) | 222 (54.7) | |
| Other, non-Hispanicb | 113 (9.5) | 33 (9.7) | 41 (9.3) | 39 (9.6) | |
| Years of experience (n = 1227) | .16 | ||||
| <1 y | 637 (51.9) | 173 (48.5) | 248 (54.9) | 216 (51.7) | |
| 1–5 y | 586 (47.8) | 184 (51.5) | 201 (44.5) | 201 (48.1) | |
| 6–15 y | 4 (0.3) | 0 (0) | 3 (0.7) | 1 (0.2) | |
Provider characteristics derived from self-reported survey data.
The “other” category for gender includes 2 respondents who did not further specify their gender.
The “other, non-Hispanic” category for race and ethnicity includes 48 respondents who selected more than 1 race: white and other (n = 3); Native Hawaiian/Pacific Islander and white (n = 1); Black and white (n = 3); Asian and other (n = 1); Asian and white (n = 20); Asian, white and other (n = 1); Asian and Native Hawaiian/Pacific Islander (n = 2); Asian, Black, and white (n = 2); American Indian/Alaskan Native and other (n = 1); American Indian/Alaskan Native and white (n = 9); American Indian/Alaskan Native and Black (n = 2); American Indian/Alaskan Native and Asian (n = 1); American Indian/Alaskan Native, Asian, Black, Native Hawaiian/Pacific Islander, and white (n = 2). Additionally, 65 respondents provided the following write-in categories for race and ethnicity: Arab (n = 3); Indian (n = 1); ME (n = 1); Middle Eastern (n = 5); mixed (n = 1); Pakistani (n = 1); South Asian (n = 2); N/A (n = 1); unknown (n = 1); if I answer this you will know who I am (n = 1); and I don’t think it’s appropriate to ask about race (n = 1). The remainder of respondents selecting “other” did not further specify their race and ethnicity.
Attending Characteristics (n = 378)
| Characteristic | Overall, n (378) (%) | Pre-implementation, N (194) (%) | Early-implementation, N (74) (%) | Late-implementation, N (110) (%) | P |
|---|---|---|---|---|---|
| Age, y (n = 375) | .01 | ||||
| 18–34 | 89 (23.7) | 50 (26.0) | 13 (17.8) | 26 (23.6) | |
| 35–44 | 184 (49.1) | 100 (52.1) | 28 (38.4) | 56 (50.9) | |
| 45–74 | 102 (27.2) | 42 (21.9) | 32 (43.8) | 28 (25.5) | |
| Gender (n = 374) | .24 | ||||
| Female | 241 (64.4) | 131 (68.6) | 40 (54.8) | 70 (63.6) | |
| Male | 122 (32.6) | 56 (29.3) | 29 (39.7) | 37 (33.6) | |
| Declined | 11 (2.9) | 4 (2.1) | 4 (5.5) | 3 (2.7) | |
| Race and ethnicity (n = 370) | .86 | ||||
| Asian, non-Hispanic | 74 (20.0) | 34 (18.1) | 18 (24.7) | 22 (20.2) | |
| Black, non-Hispanic | 17 (4.6) | 11 (5.9) | 3 (4.1) | 3 (2.8) | |
| Hispanic or Latino | 21 (5.7) | 12 (6.4) | 4 (5.5) | 5 (4.6) | |
| White, non-Hispanic | 228 (61.6) | 117 (62.2) | 41 (56.2) | 70 (64.2) | |
| Other, non-Hispanica | 30 (8.1) | 14 (7.5) | 7 (9.6) | 9 (8.3) | |
| Years of experience (n = 375) | .19 | ||||
| <1 y | 7 (1.9) | 5 (2.6) | 0 (0) | 2 (1.8) | |
| 1–5 y | 86 (22.9) | 48 (25.0) | 10 (13.7) | 28 (25.5) | |
| 6–15 y | 189 (50.4) | 95 (49.5) | 38 (52.1) | 56 (50.9) | |
| ≥16 y | 93 (24.8) | 44 (22.9) | 25 (34.3) | 24 (21.8) | |
| Characteristic | Overall, n (378) (%) | Pre-implementation, N (194) (%) | Early-implementation, N (74) (%) | Late-implementation, N (110) (%) | P |
|---|---|---|---|---|---|
| Age, y (n = 375) | .01 | ||||
| 18–34 | 89 (23.7) | 50 (26.0) | 13 (17.8) | 26 (23.6) | |
| 35–44 | 184 (49.1) | 100 (52.1) | 28 (38.4) | 56 (50.9) | |
| 45–74 | 102 (27.2) | 42 (21.9) | 32 (43.8) | 28 (25.5) | |
| Gender (n = 374) | .24 | ||||
| Female | 241 (64.4) | 131 (68.6) | 40 (54.8) | 70 (63.6) | |
| Male | 122 (32.6) | 56 (29.3) | 29 (39.7) | 37 (33.6) | |
| Declined | 11 (2.9) | 4 (2.1) | 4 (5.5) | 3 (2.7) | |
| Race and ethnicity (n = 370) | .86 | ||||
| Asian, non-Hispanic | 74 (20.0) | 34 (18.1) | 18 (24.7) | 22 (20.2) | |
| Black, non-Hispanic | 17 (4.6) | 11 (5.9) | 3 (4.1) | 3 (2.8) | |
| Hispanic or Latino | 21 (5.7) | 12 (6.4) | 4 (5.5) | 5 (4.6) | |
| White, non-Hispanic | 228 (61.6) | 117 (62.2) | 41 (56.2) | 70 (64.2) | |
| Other, non-Hispanica | 30 (8.1) | 14 (7.5) | 7 (9.6) | 9 (8.3) | |
| Years of experience (n = 375) | .19 | ||||
| <1 y | 7 (1.9) | 5 (2.6) | 0 (0) | 2 (1.8) | |
| 1–5 y | 86 (22.9) | 48 (25.0) | 10 (13.7) | 28 (25.5) | |
| 6–15 y | 189 (50.4) | 95 (49.5) | 38 (52.1) | 56 (50.9) | |
| ≥16 y | 93 (24.8) | 44 (22.9) | 25 (34.3) | 24 (21.8) | |
Provider characteristics derived from self-reported survey data.
The “other, non-Hispanic” category for race and ethnicity includes 14 respondents who selected more than 1 race: white and other (n = 1); Black and white (n = 1); Asian and white (n = 7); and American Indian/Alaskan Native and white (n = 5). Additionally, 4 respondents provided the following write-in categories: Brown (n = 1); Indian (n = 1); and Middle Eastern (n = 2). The remainder of respondents selecting “other” did not further specify their race and ethnicity.
Nurse Characteristics (n = 819)
| Characteristic | Overall, N (819) (%) | Pre-implementation, N (410) (%) | Early-implementation, N (174) (%) | Late-implementation, N (235) (%) | P |
|---|---|---|---|---|---|
| Age, y (n = 800) | <.0001 | ||||
| 18–34 | 413 (51.6) | 193 (48.0) | 119 (68.8) | 101 (44.9) | |
| 35–44 | 181 (22.6) | 96 (23.9) | 24 (13.9) | 61 (27.1) | |
| 45–74 | 206 (25.8) | 113 (28.1) | 30 (17.3) | 63 (28.0) | |
| Gender (n = 803) | .26 | ||||
| Female | 746 (92.9) | 375 (93.5) | 158 (91.3) | 213 (93.0) | |
| Male | 38 (4.73) | 16 (4.0) | 13 (7.5) | 9 (3.9) | |
| Declined | 19 (2.4) | 10 (2.5) | 2 (1.2) | 7 (3.1) | |
| Race and ethnicity(n = 794) | <.001 | ||||
| Asian, non-Hispanic | 46 (5.8) | 35 (8.8) | 1 (0.6) | 10 (4.4) | |
| Black, non-Hispanic | 28 (3.5) | 13 (3.3) | 3 (1.8) | 12 (5.3) | |
| Hispanic or Latino | 36 (4.5) | 21 (5.3) | 3 (1.8) | 12 (5.3) | |
| White, non-Hispanic | 623 (78.5) | 296 (74.6) | 151 (88.3) | 176 (77.9) | |
| Other, non-Hispanica | 61 (7.6) | 32 (8.0) | 13 (7.6) | 16 (7.1) | |
| Years of experience (n = 807) | < .0001 | ||||
| <1 y | 70 (8.7) | 28 (6.9) | 32 (18.5) | 10 (4.4) | |
| 1–5 y | 293 (36.3) | 143 (35.3) | 81 (46.8) | 69 (30.1) | |
| 6–15 y | 242 (30.0) | 127 (31.4) | 32 (18.5) | 83 (36.2) | |
| ≥16 y | 202 (25.0) | 107 (26.4) | 28 (16.2) | 67 (29.3) | |
| Characteristic | Overall, N (819) (%) | Pre-implementation, N (410) (%) | Early-implementation, N (174) (%) | Late-implementation, N (235) (%) | P |
|---|---|---|---|---|---|
| Age, y (n = 800) | <.0001 | ||||
| 18–34 | 413 (51.6) | 193 (48.0) | 119 (68.8) | 101 (44.9) | |
| 35–44 | 181 (22.6) | 96 (23.9) | 24 (13.9) | 61 (27.1) | |
| 45–74 | 206 (25.8) | 113 (28.1) | 30 (17.3) | 63 (28.0) | |
| Gender (n = 803) | .26 | ||||
| Female | 746 (92.9) | 375 (93.5) | 158 (91.3) | 213 (93.0) | |
| Male | 38 (4.73) | 16 (4.0) | 13 (7.5) | 9 (3.9) | |
| Declined | 19 (2.4) | 10 (2.5) | 2 (1.2) | 7 (3.1) | |
| Race and ethnicity(n = 794) | <.001 | ||||
| Asian, non-Hispanic | 46 (5.8) | 35 (8.8) | 1 (0.6) | 10 (4.4) | |
| Black, non-Hispanic | 28 (3.5) | 13 (3.3) | 3 (1.8) | 12 (5.3) | |
| Hispanic or Latino | 36 (4.5) | 21 (5.3) | 3 (1.8) | 12 (5.3) | |
| White, non-Hispanic | 623 (78.5) | 296 (74.6) | 151 (88.3) | 176 (77.9) | |
| Other, non-Hispanica | 61 (7.6) | 32 (8.0) | 13 (7.6) | 16 (7.1) | |
| Years of experience (n = 807) | < .0001 | ||||
| <1 y | 70 (8.7) | 28 (6.9) | 32 (18.5) | 10 (4.4) | |
| 1–5 y | 293 (36.3) | 143 (35.3) | 81 (46.8) | 69 (30.1) | |
| 6–15 y | 242 (30.0) | 127 (31.4) | 32 (18.5) | 83 (36.2) | |
| ≥16 y | 202 (25.0) | 107 (26.4) | 28 (16.2) | 67 (29.3) | |
Provider characteristics derived from self-reported survey data.
The “other, non-Hispanic” category for race and ethnicity includes 40 respondents who selected more than 1 race: white and other (n = 1); Native Hawaiian/Pacific Islander and white (n = 2); Asian and white (n = 9); Asian and Native Hawaiian/Pacific Islander (n = 3); American Indian/Alaskan Native and other (n = 1); American Indian/Alaskan Native and white (n = 23); Asian, Native Hawaiian/Pacific Islander, and white (n = 1). Additionally, 4 respondents provided the following write-in categories: East Indian (n = 1); Korean (n = 1); not relevant (n = 1); and unspecified (n = 1). The remainder of respondents selecting “other” did not further specify their race and ethnicity.
Intervention Adherence
At the time of intervention, the proportion of observations demonstrating bedside rounding with families significantly improved (from 76.5% to 89.5%), as did adherence to all 5 I-PASS components on rounds (25.5% to 60.8%). Particularly strong improvements were seen in Illness severity (65.8% to 95.4%) and Synthesis by receiver (53.6% to 84.1%). Immediate improvements in written rounds summary (29.8% to 72.4%), family engagement (70.9% to 87.6%), nurse engagement (33.4% to 63.0%), and plain language (50.2% to 60.6%) also occurred, all P < .05. Except for written summary, improvements sustained during early- and late-implementation (Fig 2). Observed trainee teaching did not change by implementation period.
Harms
Resident-reported overall, minor, and major harms were unchanged by implementation-period (Fig 3, Supplemental Table 5). However, there were differential decreases in overall harms in larger-sized hospitals and hospitals with greater nurse engagement and I-PASS adherence, compared with their counterparts. In larger hospitals, overall harms per 1000-resident days fell from 116.9 to 86.3 to 72.3 (P = .006), a 38.2% reduction. In smaller hospitals, harms were unchanged: 74.9 to 88.8 to 86.4 (P = .85). There was a significant differential change in harms by hospital size (P = .03). In hospitals with greater nurse engagement, overall harms fell from 110.6 to 73.3 to 65.3 (P < .001), a 41.0% reduction. In hospitals with lesser nurse engagement, harms were unchanged: 80.0 to 102.8 to 93.9 (P = .39). There was a significant differential change in harms by nurse engagement (P < .001). In hospitals with higher adherence to I-PASS structure, overall harms fell from 95.3 to 73.6 to 72.3 (P < .05), a 24.1% reduction. In hospitals with lower adherence to I-PASS structure, harms were unchanged: 91.8 to 100.6 to 85.9 (P = .62). There was no significant differential change in harms by I-PASS adherence (P = .26).
Resident-reported harms by implementation time period, overall and by hospital characteristics.
Resident-reported harms by implementation time period, overall and by hospital characteristics.
Safety Climate
Adjusted top-2-box staff safety climate scores significantly improved pre- vs early- vs late-implementation for all 12 items (P < .05), including working together as a team (90.4% to 93.1% to 94.4%), staff freely speaking up (83.1% to 84.5% to 89.0%), and overall safety grade (80.5% to 86.3% to 88.0% excellent or very good) (Fig 4). Staff safety grade improved in smaller hospitals (74.6% to 86.2% to 87.8%, P < .001), but not in larger hospitals (86.2% to 86.0% to 88.1%, P = .44) (P = .02 for difference by size) (Supplemental Fig 7).
Top-2-box patient and family safety climate scores for overall safety grade remained unchanged, though ratings for staff are afraid to ask questions (never or rarely [reverse-coded]) went from 68.0% to 72.0% to 73.2% (P = .06) (overall ratings in Supplemental Fig 8; ratings by provider type in Supplemental Fig 9).
Experience
Top-2-box patient and family experience did not significantly change by implementation-period (Supplemental Table 6). However, top-2-box patient and family experience for nurses making you feel part of the team improved for larger hospitals (90.4% to 93.2% to 95.1%, P = .02) versus smaller hospitals (96.8% to 95.7% to 94.8%, P = .19) (P = .01 for difference by size).
Discussion
In this 21-center hybrid-effectiveness-implementation study, a novel parent-nurse-physician Mentor-Trio implementation strategy was associated with sustained high adherence to core PFC I-PASS components. Tests of “real-world” effectiveness of PFC I-PASS showed that resident-reported patient harms improved for larger-sized hospitals and those with greater PFC I-PASS adherence and nurse engagement. Staff safety climate improved, without negatively affecting patient and family experience or trainee teaching. Our findings indicate that the coproduced Mentor-Trio strategy was a promising strategy to implement PFC I-PASS on a larger scale and in more diverse contexts. Our study also provides further evidence about feasibility and effectiveness of structured PFCR in diverse settings and the value of authentic partnership between patients and families, nurses, and physicians in successfully implementing complex hospital-based QI interventions.
Our novel implementation approach for PFCR, in which parents, nurses, and physicians led together, was feasible and well-received, even during a global pandemic, when many hospitals limited PFCR and bedside caregivers.32 –34 Our study demonstrates how family partnership can be accomplished virtually (not just in-person), allowing family coinvestigators to more easily participate in hospital-based projects. Coproduction between patients and families, clinicians, and hospitals is important24 but underutilized, particularly in QI efforts. Our Mentor-Trio approach, which embraces coproduction principles called for by the Institute for Patient and Family-Centered Care and American Academy of Pediatrics,3,7,8 can be applied to other projects. By authentically involving physicians, nurses, and patients and families together rather than utilizing siloed QI efforts, hospitals may more successfully implement interventions. Nursing and parent participation in Mentor-Trios also helped address nursing-specific and family-specific implementation challenges (eg, nursing bedside presence, family perceptions of trainee redirection on rounds).
Robust studies examining PFCR implementation are lacking: most are single-site QI projects.11 Our findings and prior research show that adherence to complex I-PASS-structured communication interventions during resident handoffs2,35 and PFCR14 is possible. Compared with the prior 7-center PFC I-PASS Study, adherence to PFC I-PASS components, like synthesis by receiver (teach-back), a health literacy best-practice that improves outcomes, was higher in the current study.36 This may be because of several modifications to the intervention, including use of the Mentor-Trio model, and better training about teach-back. The growing evidence-base around PFC I-PASS may also have contributed, as providers may be more willing to implement interventions with a stronger evidence-base. Harms improved more in hospitals with better nurse engagement and patient and family experience with nursing improved in larger hospitals; this result is consistent with prior research showing structured communication interventions with modest nursing investments significantly improve nursing-related measures.14,37,38 Indeed, nurse engagement in QI efforts is crucial in promoting hospital safety and quality.39 –41
We again found no decrease in resident teaching, a balancing measure and common criticism of PFCR.6,13 Although this study did not measure rounds duration and results of other PFCR studies vary,11 our 7-center study found no increase in rounding duration with PFC I-PASS.14 Ensuring all team members, including nurses and families, are present and engaged in rounds can promote shared understanding and save time12 by obviating downstream clarifications and streamlining communication throughout the day.
Our findings differ somewhat from our 7-center study, in which harmful errors decreased overall post-implementation.14 The resident-reported, communication-related harms effectiveness measure in this study is less sensitive than the intensive systematic safety surveillance measure in the prior study, but was chosen as a more practical, real-world alternative. The lack of overall harm improvement in this study may be because of decreased measure sensitivity, or differential effectiveness of the intervention across different centers or situations where nurse engagement, intervention adherence, or other unmeasured factors vary. We also may have been underpowered to detect more subtle improvements in harms. Additionally, COVID-19 may have adversely affected safety,42 –45 further confounding findings. That staff safety climate improved post-implementation across centers is promising, as safety climate is associated with directly measured safety outcomes.46 –48 Given provider burnout49 –51 after COVID-19, improving safety climate is important in its own right.
We observed no improvements in patient and family experience post-implementation. However, hospital experience was high at baseline. Additionally, our intervention focused on communication structure, not other aspects of interpersonal communication like trust-building, which may influence experience. Although patient and family experience did not improve, it did not worsen post-implementation, despite the pandemic; it is possible (though unknown) that the intervention might have mitigated a pandemic-driven deterioration in patient and family experience.32,52,53
Operationalizing effective PFCR interventions requires support from front-line staff and leaders, key elements of our implementation process. Barriers to nursing PFCR involvement exist, like competing time-demands (eg, 8 am medications).11,12 Frontline nursing and leadership support and nurse-centric PFCR workflow changes may be needed (eg, rounding by nursing assignment, notifying nurses of rounds 1 to 2 patients in advance).
Our study had limitations. Our design—appropriate for a hybrid effectiveness-implementation study—limits conclusions about causality. However, our interrupted-time-series design strengthens rigor of analyses compared with simple pre and post designs and our study remains one of the most robust PFCR studies. Although this study provides evidence supporting broad PFC I-PASS implementation, we focused on teaching hospitals’ pediatric services. Additional research is required to determine whether PFC I-PASS generalizes to other specialties or hospital types. Although we surveyed Arabic, Chinese, and Spanish-speakers, patients speaking less common languages may have different experiences. Similarly, in our study (and, sadly, much of medicine), physicians’ and nurses’ racial and ethnic composition was different than patients’. Hospital experience and other outcomes might be improved with a more diverse physician and nursing workforce with greater patient-provider racial and ethnic concordance.54,55 This is an important area for future study by our study group and the field. Rounds observations were primarily conducted by physician observers, so may have been subject to bias. Measurement of resident-reported harms was incomplete and potentially subject to reporting and social desirability bias but pragmatic based on study resources and the desire to more closely mimic usual operational conditions in data collection; similar resident-reported harm measures have been used successfully previously.21 COVID-19 may have confounded findings because of provider burnout, patient experience, and staffing; if so, our intervention might have been more beneficial than the current study suggests. Although we made strides in family and nurse engagement, further improvement in nursing engagement on rounds is required, perhaps by understanding attitudes, beliefs, and competing demands.12 Finally, we may have been underpowered to detect differences in harms, patient and family safety climate, and experience, and a ceiling effect may have existed with the latter 2. Future work should study nurse engagement as well as PFCR disparities and best practices for patients speaking other languages (an ongoing focus of our study group), adults, and subspecialties.
Conclusions
A novel Mentor-Trio modification of mentored implementation was an effective strategy to implement PFC I-PASS in pediatric inpatient units. This larger-scale implementation was associated with significant improvements in provider and patient and family communication processes and safety climate. Additionally, in larger hospitals and hospitals with better nursing engagement and adherence to PFC I-PASS structure, PFC I-PASS was associated with reductions in resident-reported harms. The current intervention was successful in diverse teaching hospital settings despite the challenges of a global pandemic. Our findings suggest that hospitals should better engage patients and families and nurses at the bedside during rounds. Doing so can improve patient safety, safety climate, and patient and family and staff engagement, increasing self-efficacy, coproduction,24 and family-centeredness. Lastly, our findings provide evidence that coproduction of QI efforts by patients and families, nurses, and physicians working as a team can drive sustained transformational changes in hospitals.
APPENDIX 1 Patient and Family Centered I-PASS SCORE Study Group Author Affiliations
In addition to all co-authors named in the byline (Khan, Patel, Anderson, Baird, Johnson, Liss, Graham, Calaman, Fegley, Goldstein, O’Toole, Rosenbluth, Alminde, Bass, Bismilla, Caruth, Coghlan-McDonald, Cray, Destino, Dreyer, Everhart, Good, Guiot, Haskell, Hepps, Knighton, Kocolas, Kuzma, Lewis, Litterer, Kruvand, Markle, Micalizzi, Patel, Rogers, Subramony, Vara, Yin, Sectish, Srivastava, Starmer, West, Spector, Landrigan), the Patient and Family Centered I-PASS SCORE Study Group includes the following group authors: Erin Abu-Rish Blakeney, PhD, RNa; Bridget Allard, DOb,c; Karla J. AuYeung, MDd; Kathleen Bartlett, MDe; Angela Byrd, MDf; Tim Capecchi, MDg; Cynthia Castiglioni, MD, MSh,i; Weijen Chang, MDj; Marco Costilla, MSN, RN, CPNk; Jeff Cousins, BAl; Ebony Cunningham, MPH, RN, CPNm; Mike Daly, BAn; Sarah F. Denniston, MDo,p; Carla Dyer, MDq; Jonathan S. Farkas, MDr,s; Shaun Fitzgerald, MD, MPHt; Morgan Foster, MSN, RNu; Anna Gary, MDv; Marquita Genies, MD, MPHw; Alexander F. Glick, MD, MSs,x; Laura Hagemeyer, MDg; David J. Hall, MDy; Jennifer Hall, MDv; Nathalee Harris, RN, BSNz; Rohini Harvey, MDj; Erin Hauck, MDf; Marie Haugen, RNaa; Vicky Holman, MBA, RNCbb; Helen Hughes, MD, MPHw; W. Charles Huskins, MD, MScy; Satoko Igarashi, MDj; Ashley Joshi-Patel, DOcc; Evdoxia Gloria Karagounis, MSN, RN, CPNm,n,dd; Mikelle Key-Solle, MDe; Jennifer Kleven, MD, MPH, FAAPaa,ee; Erin Knoebel, MD, MPHy; Jessica Kosut, MD, FAAPb,c; Shannon Koty, MSN, CPN, CPHONff; Shawnice Kraeber, MDt; Jeffrey Lancaster, MDgg; Lauren Lay, MSN, RN, CPNhh,ii; Tracie Major, DNP, APRN-CNS, CPNjj; Janet Manegoldkk; Angie Marin, MSN, RN-Cll; Jordan Marmet, MDg; Christine Marrese, MDj; David Marseille, MDdd; Meghan E. McClure, MDcc; Stacy McConkey, MDmm; Sue Mead, MSnn; Heidi Nolen, BSN, RNoo; Lauran Odom, RN, BSN,CPNm; Jenna Pellowski, MSN, RN, CNLaa,pp; Timothy Porter, MDo,p,qq; Heather Rayburn, MHA, BSN, RNrr; Stefanie Reed, DO, FAAPdd,ss; Melissa Reichenbach, MSN, RN, C-PNP, CRRNkk; Gwendolyn Reyes, MDtt; Manda Robinson, BSN, RN, CPNuu; Eileen Romano, DNP, MBA, RNvv; Melissa Roy, MDf; Jennifer P. Rubin, MDh,i; Andrew Saxon, MDt,ww; Gretchen Schwindt, RN, MSNxx; Rina Shah, MDt; Mahesh Sharman, MDtt; Erin E. Shaughnessy, MD, MSHCMyy; Carmela E. Sosa, MD, FAAPzz,aaa; Hannah B. Stevenson, BSN, RNbbb; Chickajajur S.Vijay, MDff,gg; Rick Visconte, MSN, RNo; Eric Volz-Benoit, RNccc; Jillian Wall, MD, MPHv; Nicole Webb, MDddd; Sharon Welsch, BSN, RN, LCNeee,fff; Mary Ann Westcott, BSN, RN, CNMLggg; Jessica Witkowski, MDhhh; Natalie Woodward, MSf; and Clifton Yu, MD, FAAPiii,jjj
Patient and Family Centered I-PASS SCORE Study Group Author Affiliations:
a Departments of Biobehavioral Nursing and Health Informatics, University of Washington, Seattle, WA
b Department of Pediatrics, University of Hawaii John A. Burns School of Medicine, Honolulu, HI
c Kapi'olani Medical Center for Women and Children, Hawaii Pacific Health, Honolulu, HI
d Division of Pediatric Gastroenterology, Hepatology, and Nutrition, Valley Children’s Healthcare, Madera, CA
e Department of Pediatrics, Duke University School of Medicine, Durham, NC
f Department of Pediatrics, Our Lady of the Lake Children’s Hospital, Baton Rouge, LA
g Departments of Pediatrics, University of Minnesota Masonic Children’s Hospital, Minneapolis, MN
h Department of Pediatrics, Ann and Robert H. Lurie Children’s Hospital of Chicago, Chicago, IL
i Department of Pediatrics, Northwestern University Feinberg School of Medicine, Chicago, IL
j Department of Pediatrics, Baystate Children’s Hospital, Springfield, MA
k Department of Nursing, Texas Children’s Hospital, Houston, TX
lPatient Experience, AdventHealth for Children, Orlando, FL
m Department of Nursing, Atrium Health Levine Children’s Hospital, Charlotte, NC
n Department of Patient Experience, Atrium Health Levine Children’s Hospital, Charlotte, NC
o Department of Pediatrics, Tufts Children’s Hospital, Boston, MA
p Department of Pediatrics, Tufts University School of Medicine, Boston, MA
q Departments of Child Health and Medicine, University of Missouri School of Medicine, Columbia, MO
r Department of Pediatrics, NYC Health + Hospitals/Bellevue, New York, NY
s Department of Pediatrics, NYU Langone Health, New York, NY
t Department of Pediatrics, Kaiser Permanente Oakland Medical Center, Oakland, CA
u Duke University Hospital, Durham, NC
v Division of Pediatric Hospital Medicine, Phoenix Children’s, Phoenix, AZ
w Department of Pediatrics, Johns Hopkins School of Medicine, Baltimore, MD
x Department of Pediatrics, Bellevue Hospital Center, New York, NY
y Department of Pediatrics and Adolescent Medicine, Mayo Clinic, Rochester, MN
z Department of Quality Improvement, Hurley Children’s Hospital, Flint, MI
aa Department of Pediatrics, Gundersen Health System, La Crosse, WI
bb Department of Maternal Child Health Nursing, NYC Health + Hospitals/Bellevue, New York, NY
cc Department of Pediatrics, Baylor College of Medicine, Houston, TX
dd Department of Pediatrics, Atrium Health Levine Children’s Hospital, Charlotte, NC
ee Department of Research and Grants, Gundersen Medical Foundation, La Crosse, WI
ff Department of Pediatrics, West Virginia University Medicine Children’s Hospital, Morgantown, WV
gg Division of Hospital Medicine, West Virginia University Medicine Children’s Hospital, Morgantown, WV
hh Department of Pediatrics, University of Missouri Health Care System, Columbia, MO
ii Department of Adolescent and General Pediatrics, Children’s Hospital of Philadelphia, Philadelphia, PA
jj Department of Nursing, Our Lady of the Lake Children’s Hospital, Baton Rouge, LA
kk Phoenix Children’s, Phoenix, AZ
ll Patient Care Services-Inpatient Pediatrics, UC Davis Health and Children’s Hospital, Sacramento, CA
mm Department of Pediatrics, AdventHealth for Children, Orlando, FL
nn Department of Nursing, Johns Hopkins Children’s Center, Baltimore, MD
oo Clinical Resource Nursing Pediatrics, UC Davis Health and Children’s Hospital, Sacramento, CA
pp Department of Nursing, Gundersen Health System, La Crosse, WI
qq Department of Pediatric Newborn Medicine, Brigham and Women’s Hospital, Boston, MA
rr Department of Nursing Administration, Hurley Children’s Hospital, Flint, MI
ss Department of Pediatric Hospital Medicine, Wake Forest School of Medicine, Winston-Salem, NC
tt Department of Pediatrics, Hurley Children’s Hospital, Flint, MI
uu Department of Pediatrics, University of Missouri Children’s Hospital, Columbia, MO
vv Department of Nursing, Ann and Robert H. Lurie Children’s Hospital of Chicago, Chicago, IL
ww Division of Hospital Medicine, Children’s Hospital Los Angeles, Los Angeles, CA
xx Department of Nursing, Phoenix Children’s, Phoenix, AZ
yy Department of Pediatrics, University of Alabama at Birmingham
zz Department of Complex Primary Care Pediatrics, Valley Children’s Healthcare, Madera, CA
aaa Department of Pediatrics, Stanford University School of Medicine, Palo Alto, CA
bbb Nursing Quality & Safety, UC Davis Health and Children’s Hospital, Sacramento, CA
ccc Department of Patient Experience, Baystate Children’s Hospital, Springfield, MA
ddd Division of Hospital Medicine, Valley Children’s Healthcare, Madera, CA
eee Department of Pediatrics, Johns Hopkins Hospital, Baltimore, MD
fff Johns Hopkins Children’s Center, Baltimore, MD
ggg Women and Children’s Division, Baystate Children’s Hospital, Springfield, MA
hhh Department of Pediatrics, UC Davis Health and Children’s Hospital, Sacramento, CA
iii Department of Education, Training, and Research, Walter Reed National Military Medical Center, Bethesda, MD
jjj Department of Pediatrics, Uniformed Services University, Bethesda, MD
Acknowledgments
We wish to thank the Patient and Family Centered I-PASS SCORE Study Group and participating investigators at the Patient and Family Centered I-PASS SCORE Study sites. We also give special thanks to the patients, families, and interpreters who participated in this study. We also wish to thank the following individuals: Ancil Abney, Touleen Aly, Lezlie Andersen, Rathi Asaithambi, Tracy Barry, Barbara Bayldon, Annette Beland, Jordan Beland, Courtney Brantley, Megan Collins, Mellissa Coomb, Margarita Corredor, Lauren Crenshaw, Andrea Dean, Loretta Dickes, Eleni Duni, Anahita Emamian, Erin Fairchild, Akram Farran, Nakia Fazenbaker, Marleen Fernandez, Phillip Frayre, Sandra Gage, Stephanie Goodall, Stephanie Hagerman, Michelle Hamline, Rashed Hasan, Cara Herbener, Lauren Hess, Peter Holmberg, John Hong, Kayla Horne, Linda Horsfall, Sara Horstmann, Jason Hostetler, Heather Howell, Kum Su Hwang, Tyler Jansen, Emily Jeome, Kristin Jones, Julie Campe, Reva Kleppel, Spencer Lambdin, Jana Leary, Jonathan Lebowitz, Levine Children’s Hospital 9 and 10 Admission Discharge Nurses, Levine Children’s Hospital Administration, Levine Children’s Hospital Family Advisory Council, Matthew Ludwig, Jean Luong, Isaura Macias, Cheri Malinski, Alison McWilliams, Audrey Merrill, Charles Mullett, Kelsey Nariani, Roger Nicome, Michael Olsen, Nikechi Onwuzurike, Victoria Parente, Ishani Patek, Jessica Peuterbaugh, Guillermo Pi, Katerine Piwnica-Worms, Hope Pogemiller, Yaseen Rafee, Deann Rafee, Daniel Rauch, Amanda Rawn, Hannah Riggins, Anisha Rimal, Gretchen Roberts, Jane Rosenman, Lisa Sandoval, Paula Sims, Belinda Smith-Ramsey, Jennifer Snider, Molly Taylor, Megan Thorvilson, Toni Tildon, MaryAnn Tsourounakis, Tippy Ventress, Gurpreet Vidwan, Jill Walters, and Clarissa Workman.
A complete list of study group members appears in Appendix 1.
Dr Khan conceptualized and designed the study, coordinated and supervised data collection, designed study instruments, drafted the initial manuscript, obtained funding, and participated in acquisition, analysis, or interpretation of data, and critical revision of the manuscript for important intellectual content; Dr Patel conceptualized and designed the study, supported implementation of the intervention, and participated in acquisition, analysis, or interpretation of data, and critical revision of the manuscript for important intellectual content; Ms Anderson, Ms Fegley, Ms Goldstein, Ms Alminde, Ms Caruth, Ms Coghlan-McDonald, Ms Cray, Ms Litterer, Ms Kruvand, Ms Markle, Ms Micalizzi, Ms Rogers, Ms Helen, Ms Vara, Drs Baird, Calaman, O’Toole, Rosenbluth, Bass, Bismilla, Destino, Dreyer, Everhart, Good, Guiot, Haskell, Hepps, Knighton, Kocolas, Kuzma, Lewis, Patel, Subramony, Yin, Sectish, Srivastava, Starmer, and West helped support implementation and data collection efforts, and participated in acquisition, analysis; Ms Johnson and Ms Liss provided administrative, technical, or material support, and participated in acquisition, analysis, or interpretation of data; Dr Graham conceptualized and designed the study, performed statistical analysis, and participated in acquisition, analysis, or interpretation of data and critical revision of the manuscript for important intellectual content, and had full access to data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis; Dr Spector conceptualized and designed the study, supervised the study, and participated in acquisition, analysis, or interpretation of data, and critical revision of the manuscript for important intellectual content; Dr Landrigan conceptualized and designed the study, supervised the study, obtained funding, and participated in acquisition, analysis, or interpretation of data, and critical revision of the manuscript for important intellectual content; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.
COMPANION PAPER: A companion to this article can be found online at www.pediatrics.org/cgi/doi/10.1542/peds.2023-063619.
FUNDING: This project was supported by grant PCORI DI-2017C3-9232 from the Patient-Centered Outcomes Research Institute (PI Landrigan). Dr Khan’s time was supported by grant K08HS025781 from the Agency for Healthcare Research and Quality (PI Khan).
CONFLICT OF INTEREST DISCLOSURES: Drs Landrigan, Patel, Sectish, Spector, Starmer, and West hold equity in and have consulted with the I-PASS Patient Safety Institute. Drs Baird, Kuzma, and Rosenbluth have consulted with the I-PASS Patient Safety Institute. Drs Calaman and O’Toole have consulted with the I-PASS Patient Safety Institute and hold stock options in it. Dr Srivastava is a physician founder of the I-PASS Patient Safety Institute. His employer, Intermountain Health owns his equity in the I-PASS Patient Safety Institute. The I-PASS Patient Safety Institute is a company that seeks to train institutions in best handoff practices and aid in their implementation. The I-PASS Patient Safety Institute was in no way involved in this study. Analyses were conducted by a statistical team who do not have any involvement with the I-PASS Patient Safety Institute. Drs Sectish, Spector, and Starmer have received monetary awards, honoraria, and travel reimbursement from multiple academic and professional organizations for teaching and consulting on physician performance and handoffs. In addition, Dr Landrigan has received monetary awards, honoraria, and travel reimbursement from multiple academic and professional organizations for teaching and consulting on sleep deprivation, physician performance, handoffs, and safety, and has served as an expert witness in cases regarding patient safety and sleep deprivation; Dr Srivastava is supported in part by grants from Agency for Healthcare Research and Quality, Centers for Disease Control and Prevention, National Institute for Health and PCORI for federal grants to his institution. Dr Srivastava has also received monetary awards, honoraria, and travel reimbursement from multiple academic and professional organizations for teaching about pediatric hospitalist research networks and quality of care. Dr West has received monetary awards, honoraria, and travel reimbursement from multiple academic, regulatory and professional organizations for teaching and consulting on handoff safety and competency-based medical education. Dr Knighton reports grant support unrelated to this research from the Moore Foundation and the National Institutes of Health. Dr Knighton owns shares in UnitedHealth Group, a publicly traded health and wellness company not related to this research. Dr Spector has received grant funding from the US Department of Health and Human Services and Agency for Healthcare Research and Quality during the conduct of the study. Dr Baird has received grant funding from the Lucile Packard Foundation for Children’s Health and Arthur Vining Davis foundation during the conduct of the study. All other authors have no conflicts of interest to disclose at this time.
Comments