Unit-based Pathways to Reduce InequitieS for familiEs speaking languages other than English (UPRISE)

Abstract

OBJECTIVE

Health care inequities are common among individuals who speak languages other than English (LOE). Within our PICU, LOE families prefer communication via in-person interpreters (IPI). Spanish-speaking patient families are our largest LOE population; therefore, we sought to increase Spanish IPI utilization for medical updates in the PICU.

METHODS

A quality improvement initiative in a 36-bed PICU included: the addition of a dedicated weekday Spanish-speaking IPI, the creation of communication tools, staff education, optimized identification of LOE families, and development of a language dashboard across multiple Plan, Do, Study, Act cycles. The primary outcome was IPI utilization rates for daily medical updates.

RESULTS

Spanish IPI utilization for daily weekday medical updates among 442 Spanish-speaking patient families increased from a median of 39.4% at baseline to a new centerline median of 51.9% during implementation, exhibiting 66.3% (465 of 701) utilization in the final 6 months of implementation. The greatest sustained increases in Spanish IPI utilization occurred after PICU-based IPI implementation, staff education, electronic health record optimization, and a split work week between 2 PICU-based IPIs.

CONCLUSIONS

This quality improvement initiative increased Spanish IPI utilization for daily weekday medical updates in the PICU across multiple Plan, Do, Study, Act cycles. Future work will adapt these interventions to other languages and other hospital-based units.

Communication is central to delivering safe, high-quality care for all families, including patient families who speak languages other than English (LOE).^1–5 In pediatrics, the patient’s family or guardian often holds decision-making authority and thus needs language-access services. The National Standards for Culturally and Linguistically Appropriate Services in Health care calls for equitable care for families with LOE,⁶ emphasizing the need for communicating in preferred languages. Additionally, The Joint Commission published a communication roadmap requiring hospitals to identify patients’ communication needs and provide information in their preferred language.⁷ Despite these national standards, no standardized method for overcoming language discordance in patients and families with LOE exists.⁸ Consequently, only 10% to 30% of patients who speak a LOE receive professional interpretation.⁹ This deficit contributes to worse clinical outcomes in patients with LOE because of delays in care,¹⁰ reduced access to care,^11–13 and increased incidence of adverse events,⁵ length of stay,^14,15 and mortality.¹⁶ These disparities are amplified in the PICU, given the fast-paced environment, high acuity, inability to communicate directly with patients, and substantial cognitive demand on providers.

Interpreter services are crucial to providing culturally-sensitive care.¹⁷ In-person interpreters (IPIs) can equitably satisfy the communication needs of many LOE families compared with English-speaking families and are critical to the efficient and culturally nuanced-transfer of medical information.¹⁸ Providers and families described improved accuracy and efficiency with IPI use.¹⁷ An initiative incorporating IPIs into family-centered rounds in a pediatric ward noted a 63% increase in IPI utilization, associated with improved family satisfaction for those who speak a LOE.¹⁸ This project identified gaps in the admission screening for those speaking LOE and medical record-generated LOE patient lists.¹⁸ Our project was designed to address these knowledge gaps.

Our project was conducted in a 36-bed pediatric PICU at a quaternary academic medical center. Our interpreter services department provides language access for >200 languages. Interpreter modalities include remote audio and video interpreters and IPIs (Spanish, Mandarin, Cantonese, Arabic, Japanese, ASL, and Vietnamese) with only Spanish IPIs available round-the-clock. Traditionally, when an IPI was needed, interpreter services would send the next available IPI to the unit. Despite these valuable resources, patient families reported a lack of daily medical updates and inconsistent use of qualified interpreters for medical updates. Workflow barriers included delayed identification of patient families with LOE and delayed knowledge of preferred interpreting modality or medical update preferences.

We designed a quality improvement (QI) initiative for improving communication by implementing Unit-based Pathways to Reduce InequitieS for familiEs (UPRISE) with LOE in the PICU between January 2020 and June 2022. We designed this QI initiative using principles from the Institute for Health care Improvement to: (a) make health equity a strategic priority; (b) develop structures and processes supporting health equity; and (c) address the multiple determinants of health.¹⁹

We designed a multilayered communication initiative to increase Spanish IPI utilization for daily weekday medical updates in the PICU from 43% to 80% by July 2022. The initiative implemented dedicated PICU-based Spanish IPI, the creation of communication tools, staff education, optimized identification of patient families, and developed a language access dashboard for unit IPI utilization monitoring.

Methods

Context

Over 30% of our PICU population spoke LOE, 60% of LOE families spoke Spanish as their primary language, and 83% of our LOE families preferred IPI, with 63% wanting medical updates during family-centered rounds. The admitting department assessed demographics and interpreter needs on admission. Language modality and medical update preferences were assessed initially via in-person conversations with Latinx families. Even with the knowledge that these families wanted IPI updates and daily updates, Spanish-speaking families were receiving IPI medical updates only 41.2% of the time.

With the PICU increasing from 24 to 36-beds, interpreter services increased their workforce rather than redistributing existing language services to support Spanish IPI in the PICU and therefore mitigating opportunity costs. Executive sponsors were engaged after educational presentations and in-kind support was garnered from our hospital’s Value Improvement Program. Because of unit expansion, funding for this initiative was readily supported by the hospital, PICU, and Interpreter Services department. Based on the number of PICU families reporting IPI preference during conversations with Latinx families and our fast-paced or critical environment, the hospital leadership recognized gaps in our ability to communicate with Spanish-speaking families and fully supported this initiative.

Spanish-speaking PICU families with interpreter needs were included in this QI initiative. Families with a PICU stay less than 24 hours, not available at the bedside, or admitted and discharged over the same weekend were excluded. We used post-encounter interpreter notes to accurately assess family presence on rounds, type of communication updates provided (medical versus other), communication barriers, and overall IPI utilization. We collected patient demographic and clinical data using the Virtual Pediatric Systems database and electronic health record (EHR) chart reviews. After the baseline for this initiative in January through June 2020, multiple Plan, Do Study, Act (PDSA) cycles were initiated until the final project implementation in June 2022. Figure 1 illustrates key drivers and their corresponding interventions. The local Institutional Review Board deemed this project exempt.

FIGURE 1.

Key drivers toward the incorporation of communication strategies for LOE patients and their families.

Interdisciplinary Team

Leaders from nursing, immigrant and refugee resources, physicians, language access, clinical research, bioinformatics, and graduate medical education collaborated to optimize institutional understanding of workflows, policies, systems, resources, and roles. Our team included several first-generation immigrants with lived experiences of health inequity who met biweekly during the design, preparation, and implementation.

Communication Initiative

Our QI initiative allowed us to identify gaps and solutions including (1) dedicated PICU-based Spanish IPI, (2) communication tools, (3) staff education, (4) optimized screening of LOE patient families, and (5) data visibility using an electronic language dashboard. Multiple PDSA cycles were informed by data from project dashboards and run charts to guide interventions.

(1) Dedicated PICU In-person Interpreter

The dedicated IPI and corresponding interventions are denoted by points C, D, and F on Fig 2. The dedicated IPI was assigned to the PICU in July 2021, staffed weekdays from 9 am to 1 pm. PDSA cycles improved PICU-based IPI workflow by (a) identifying rounding times for specific patients or PICU teams; (b) rounding sequences provided to the PICU-based IPI upon arrival; (c) IPI participation in family-centered rounds or after rounds (Fig 2 and Supplemental Table 3 and Supplemental Fig 4 in Supplement). High-census workflows were created to alert interpreter dispatch to send additional interpreters as needed. Project leaders conducted biweekly check-ins with the PICU-based IPI and medical teams to discuss workflow needs. In February 2022, the PICU-based IPI work week was split between 2 dedicated PICU interpreters to reduce the risk of burnout caused by the cognitive, emotional, and physical workflow demands in the PICU.

FIGURE 2.

Monthly in-person interpreter utilization for medical updates with hospital-wide and PICU-based IPIs. centerlines (median) for baseline 6-month period, shifted with 8 consecutive months on 1 side of the median. Hollow circles indicate preintervention phase. Squares indicate rollout of communication tools. (A) Staff education. (B) LOE patient lists added to electronic health record. (C) Implementation of PICU-based in-person interpreter for family-centered rounds. (D) Biweekly PICU-based interpreter check-ins by the project team. (E) Electronic health record modifications to the admission database. (F) Splitting the work week between 2 PICU-based IPIs. (G) Interpreter notes to track families not at the bedside.

(2) Communication Tools

Various communication tools were implemented at the square points on Fig 2. Spanish-speaking families received a welcome packet on admission in Spanish with information on visitation, daily activities, and overall care coordination. Interpreter requests and “basic needs” (blanket, water, food) cards were created in the 4 most common languages (Supplemental Figs 5 and 6), allowing families to request interpreter services and basic items easily, and unit signage was improved to include Spanish.

(3) Staff Education

Staff education was completed at various intervals, with initial education at point A on Fig 2. The project began with team members providing daily reminders to staff on the importance of updating families in their preferred language. The project leader assessed daily PICU IPI utilization, circulated performance metrics weekly, and provided IPI workflow orientations for new providers. In October 2020, the PICU medical faculty were educated on the importance of integrating non-English speaking families into rounds and updating in their preferred language and modality. Nurses and physicians were educated via tutorials for contacting and requesting IPIs; didactic health equity rounds were added for PICU physician trainees; PICU fellowship orientation included the PICU-based IPI workflow and communication initiative.

(4) Optimized Screening or Identification

Various screening or identification interventions were implemented at points B, E, and G on Fig 2. Nurses completed the Communication Preference Plan flowsheet during admission intake to identify bedside family members, rounding preferences, and preferred interpretation modality. This generated our automated LOE patient list and scheduled the Spanish PICU-based IPI for medical updates. The post-encounter IPI note assessed family presence for rounds, medical versus other, communication barriers, and the PICU-based versus hospital-wide IPI utilization. This note was further enhanced to identify reasons a session did not occur. These EHR enhancements (Supplemental Fig 7) provided critical data for assessing IPI needs and utilization.

(5) Language Dashboard for Data Visibility

A language dashboard displayed data extracted from these EHR enhancements to improve data visibility (Supplemental Fig 8). Monolingual Spanish was the default language used to monitor this initiative; however, the dashboard also identified bilingual and non-Spanish LOE families. Interpretation data were collected based on IPI notes to determine the type of update and interpretation barriers. Dashboards were refreshed daily at midnight to follow Spanish IPI utilization trends. A project member reviewed the data daily, sent weekly updates to the QI team, and monthly updates to all PICU providers.

Outcome Measures

The primary outcome measure was the number of Spanish-speaking families in the PICU who received daily in-person medical updates given by a medical provider either during or after rounds during weekdays. Our process outcome measure was the post-IPI documentation in the EHR. We collected data from PICU staff both quantitatively and qualitatively throughout this project. We used the data to develop our QI interventions and to detect any unintended consequences as a balancing measure.

Analysis

Run charts were used to track quality improvement data, allowing the team to monitor progress and plan PDSA interventions, using the median IPI utilization for the 6-month pre-implementation period as a centerline. We calculated a new centerline median IPI utilization for any 8 month period of consecutive values on 1 side of the centerline. We used Fisher’s exact test to compare IPI usage between baseline and full initiative implementation. To quantify the overall change in IPI utilization during the intervention, we used a multivariable logistic regression model that adjusted for unit- and patient-specific characteristics hypothesized to affect IPI utilization.

Results

Patient Characteristics

Demographic characteristics are summarized (Table 1) for 442 patients, with a median age 9.4 years (interquartile 1.6–14.5 years), 251 male (57%), and 434 (98%) of Hispanic ethnicity.

TABLE 1.

Patient-level Characteristics

Patient-level Characteristics	Patients (n = 442)
Age at admission (years), median [Q1, Q3]	9.4 [1.6, 14.5]
Ethnic group, N (%)
Declines to state	5 (1.1)
Hispanic	434 (98.2)
Not Hispanic	3 (0.7)
Sex, N (%)
Male	251 (56.8)
Female	191 (43.2)
Diagnosis category, N (%)
Cardiovascular	20 (4.5)
Endocrine or metabolic	17 (3.9)
Gastrointestinal	25 (5.7)
Genetic	17 (3.9)
Heme, oncology, or immunology	38 (8.6)
Infectious	39 (8.8)
Injury or poisoning	30 (6.8)
Neurologic	77 (17.5)
Orthopedic	19 (4.3)
Other medical	18 (4.1)
Renal or genitourinary	15 (3.4)
Respiratory or ENT	126 (28.6)
Outcome, N (%)
Died	7 (1.6)
Survived	435 (98.4)
Physical length of stay (days), median [Q1, Q3]	2.4 [1.6, 5.5]
Language, N (%)
English or Spanish	79 (17.9)
Spanish	356 (80.5)
Spanish or Non-Spanish	7 (1.6)
PIM 3 ROM, median [Q1, Q3]	0.7 [0.2, 1.7]
Patient-day-level characteristics	Rows (n = 3391)
Unit census families with LOE, median [Q1, Q3]	6.0 [5.0, 8.0]

ENT, ear, nose, and throat

Staff Education

Increases in IPI utilization were noted immediately after providing staff education (Fig 2; Point A); however, effectiveness of education alone appeared insufficient, with a subsequent gradual drop in IPI utilization. Only when structures and processes were in place (Fig 2; Points B, C, E, and G), supported by education, did we begin to see longer lasting improvements in IPI utilization.

In-person Interpreter (IPI) Utilization

The project was associated with higher IPI utilization for medical updates, with a large gain after splitting the PICU-based IPI work week (Fig 2; Point F). Weekday medical updates with Spanish-speaking IPI increased from a centerline median of 39.4% at baseline to a new centerline median of 51.9% during implementation. Overall, IPI utilization was 66.3% (465 of 701) in the final 6 months of full initiative implementation. The majority of IPI medical updates were completed by the PICU-based IPI after their introduction, increasing from a baseline centerline median of 0.0% to a new centerline median of 46.1% during implementation. Despite education, IPI usage on weekends did not increase during the study period (+0.4% relative to baseline) (Fig 3).

FIGURE 3.

Comparison of IPI medical updates for weekdays and weekends throughout the study period.

In October, December 2022, and January 2023, vacation time was taken by the PICU-based IPI, leading to intermittent dips in IPI utilization. Unit-specific training was offered to hospital-wide IPIs to help alleviate this gap. Shortly after deployment of the project, the sole PICU-based IPI identified burnout risks associated with the cognitive, emotional, and physical demands of the PICU. Interventions, including debriefing opportunities and biweekly check-ins occurred. Splitting the work week between 2 PICU-based IPIs (Fig 2; Point F) contributed most to mitigating these risks. In April 2022, a decrease in IPI was observed after a nursing work stoppage with temporary staff unaware of unit processes. The program recovered when regular staffing resumed.

In logistic regression models, IPI use was higher in the post-intervention group when compared with baseline (odds ratio 1.64, 95% confidence interval 1.36–1.98, P < .001), after adjustment for unit- and patient-specific characteristics. Lower odds of IPI utilization was associated with LOE families on the days of admission or discharge, female patients, and patients admitted for genetic diagnoses (Table 2). No differences in IPI utilization were observed by weekday, daily census of LOE families, risk of mortality (PIM3), or mortality (versus survival) of PICU patients.

TABLE 2.

Multivariable Analysis of In-person Interpreter Usage

	Oddsa	P
Intercept	0.66 (0.48–0.92)	.01
	aOR (95% CI)	P
Implementation phase (versus Jan–Jun 2020)	1.64 (1.36–1.98)	<.001
Day of week (versus Monday)
Tuesday	1.24 (0.99–1.56)	.06
Wednesday	1.10 (0.88–1.38)	.41
Thursday	1.09 (0.87–1.37)	.43
Friday	1.10 (0.88–1.37)	.42
Daily census of families with LOE (per 1 family increase)	1.01 (0.97–1.04)	.97
Language preference (versus Spanish)
English or Spanish	0.45 (0.36–0.55)	<.001
Spanish or Non-Spanish	1.02 (0.56–1.88)	.94
Female sex (versus male)	0.79 (0.68–0.92)	.002
Admission diagnosis group (versus respiratory or ENT)
Cardiovascular	2.06 (1.37–3.12)	.001
Endocrine or metabolic	0.73 (0.35–1.53)	.40
Gastrointestinal	1.11 (0.85–1.45)	.44
Genetic	0.28 (0.14–0.54)	<.001
Heme, oncology, immunology	3.00 (2.27–3.96)	<.001
Infectious	1.21 (0.95–1.55)	.12
Injury or poisoning	1.72 (1.21–2.45)	.002
Neurologic	1.75 (1.37–2.23)	<.001
Orthopedic	1.85 (1.12–3.05)	.02
Other medical	2.32 (1.58–3.41)	<.001
Renal or genitourinary	0.73 (0.5–1.07)	.11
Transplant	1.99 (0.99–3.99)	.05
Day of ICU stay (versus routine daysb)
Admission day	0.67 (0.54–0.83)	<.001
Discharge day	0.70 (0.56–0.87)	.001
PIM 3 risk of mortality (per 1% increase)	1.01 (1.00–1.02)	.17
Died before discharge (versus survived)	1.17 (0.77–1.78)	.47

Model via logistic regression with in- person interpreter utilization as the dependent variable. N = 3391 patient-days. aOR, adjusted odds ratio; ENT, ear, nose, and throat; LOE, languages other than English.

^a Calculated as exp(logit)

^b Routine days are defined as all ICU days between the day of ICU admission and the day of ICU discharge

Discussion

This multilayered communication initiative was designed to increase Spanish IPI utilization for weekday medical updates in the PICU. Our primary outcome of Spanish IPI utilization rates for weekday medical updates increased by 25 percentage points. The biggest impact was found with the availability of PICU-based IPIs, specifically when the work week was divided between 2 PICU-based IPIs, with the majority of IPI updates being completed by a PICU-based IPI.

Although we did not achieve 80% IPI use for medical updates as originally intended, we found significant improvements in IPI utilization during the project period. These improvements have been sustained after the project ended in July 2022, surpassing the original 80% target by achieving 84% IPI utilization in February 2024. We will continue to monitor if longer periods of the complete intervention will assist us in achieving the ideal of 100% IPI updates.

IPI provides improved communication and higher satisfaction among LOE families over telephone interpreters²⁰ by offering timely and effective communication, nuanced by the body language, facial expression, and other nonverbal cues of LOE family members.^18,21,22 Although consistent IPI use may be central to achieving language access equity,^2,3,8,18,22 barriers to interpreter involvement remain pervasive. Prior studies reported that time-intensive rounds, unreliable interpreter integration or availability, and lack of recognition of LOE families were barriers to utilizing interpreter services.^5,8,23

We operationalized the recommendations of National Standards for Culturally and Linguistically Appropriate Services and The Joint Commission by targeting our interventions to overcome these communication barriers, specifically maximizing interpreter availability through PICU-based IPIs, identifying families with LOE, documenting their preferences, improving the EHR documentation, educating staff, adjusting PICU-based IPI workflows, and creating a language access dashboard.^6,23 This QI initiative formalized the PICU evaluation of daily interpreter needs and created a minimum standard for Spanish-speaking families in the PICU. This minimum standard provides a practical starting point toward meaningful interactions and language-appropriate health care.

Our commitment to language access equity magnified the need for unit-based language services, and the interpreter’s role as health care communication professionals who are critical to the interdisciplinary care team. Familiarity with the dedicated PICU-based IPI allows for a layer of intentional advocacy for LOE families whose hospitalizations are at risk for poor communication.²⁴ Specifically, the PICU-based IPI addressed multiple gaps, including unit-focused timely language access, continuity of care during medical rounds, and greater IPI use during weekdays versus weekends. Consequently, weekday interpreter utilization rates improved despite higher unit census and more critically ill children because of the PICU-based IPIs activating high-census workflows and coordinating with ICU providers. The use of hospital-wide IPIs in the PICU diminished because PICU-based IPIs were readily available.

Weekend IPI services required a provider-initiated request for a hospital-wide IPI, and utilization rates did not improve despite our educational efforts highlighting hospital-wide IPI availability. Given the contrast between weekdays and weekends (Fig 3), the direct impact of having PICU-based IPI was notable and ultimately strengthened language access.

Our communication initiative provides a pathway for implementing language access equity at other institutions. This initiative’s implementation is feasible and sustainable with hospital support and resource allocation. Developing a team of individuals from different cultural backgrounds, professional experiences, age generations, research knowledge, and skills created a rich environment for innovation and implementation. Having an on-site project member to assess education needs, adjust workflows, provide check-ins, and monitor the project was essential to this project’s success.

This project was conducted in a single center, specialized quaternary hospital, limiting the generalizability of these findings. Our data analysis relied on EHR accuracy, requiring various modifications to improve data capture from the nursing and interpreter flowsheets because initial flowsheets did not explain why IPI updates did not occur. Variability in the length of rounds depended on the nature of the discussion and the number of consulting services. PICU-based IPI workflow limitations occurred when the PICU-based IPI was on vacation and during work stoppage. Education and training of hospital-wide IPIs were needed to learn the PICU-based workflows when the PICU-based IPI was unavailable. Hospitals with limited interpreters or resources may need to elicit other modes of interpreter utilization to provide language access. Although this communication initiative improved daily medical updates, not all families received IPI updates. Our future efforts must focus on IPI utilization gaps during admissions and discharges when PICU-based IPIs are not available. Additional resources are needed to extend the Spanish PICU-based IPI model to weekends, night shifts, and non-Spanish languages. Supplemental communication tools, such as language cards and basic needs cards, may complement in-person or remote professional interpretation as viable resources for institutions, including hospitals with limited resources. Although PICU-based IPI had the greatest impact, other interventions can still play a key role for LOE families.

Improving language equity requires resources, consistent, sustained efforts, and partnerships between language access professionals, clinical teams, ancillary providers, and executives. This work focused on the communication initiative effectiveness, whereas future efforts to assess health care outcomes are also important. Future efforts will evaluate and support the high demands placed on medical interpreters in the PICU, given the fast-paced environment, high acuity, cognitive demands, unpredictable workflow, and the communication needs of multiple ICU teams. Our intervention of dividing the week between 2 PICU-based IPIs helped to alleviate the risk of burnout and moral distress associated with repeated high-intensity workflows. However, this area would benefit from further research.

Conclusions

This communication initiative increased IPI utilization for Spanish-speaking families. Language-appropriate health care access for LOE families has the potential to promote positive therapeutic alliances between families and staff, thus optimizing clinical care. The burden of communicating should not fall on families whose primary focus is their critically ill child’s recovery. Partnering with language access professionals to care for and engage with LOE families builds a foundation for equitable communication, thus making strides toward health care equity. Our multidisciplinary communication initiative provides a template to improve language-sensitive care in other hospitals and health care settings and shows the power of quality improvement for reducing health care inequities. This analysis is one step in our Unit-based Pathways to Reduce InequitieS for familiEs (UPRISE) journey toward equitable care for all.