Introduction

As of 2021, there are over 25 million people with non-English language preference (NELP) in the United States (U.S.) [1]. Of note, NELP is a term that replaces the deficit-centering and ethnocentric label “limited English proficiency” (LEP) to describe an individual who prefers a language other than English in daily life [2]. Throughout this article, we will use “NELP” instead of “LEP” to continue this practice of reorienting language proficiency from a deficit model to a reflection of a person’s social context.

Patients with NELP suffer from an array of disparities, including worse health outcomes, higher rates of hospital readmission, longer hospital length of stay, and higher healthcare resource utilization [3,4,5,6,7]. While patients with NELP are legally entitled access to professional interpretation services [8], there is evidence that U.S. healthcare facilities often lack the resources and personnel to meet the language needs of patients with NELP [9, 10]. There has been a recent surge in literature related to disparities among populations and patients with NELP, including reviews and quantitative studies that examine disparities in numerous hospital settings and medical specialties [4, 5, 11,12,13,14,15,16,17,18].

Spanish-speakers comprise 77% of people with NELP in the U.S. [19]. The Hispanic and Latino community, a predominantly Spanish-speaking group, made up almost one quarter of COVID-19 deaths during the early parts of the pandemic [20], highlighting the disproportionate impact of the pandemic on these communities [21, 22]. Monolingual Spanish-speakers had higher rates of COVID-19 infection, COVID-19 related hospital stays, ICU admission, intubation, and mortality than other Hispanic groups [23, 24]. Those hospitalized with COVID-19, are likely to experience a complex illness, a serious or critical illness that may prompt prolonged hospitalization and presenting challenges in clinical care, medical decision-making, and goals of care conversations [25].

Given that the COVID-19 pandemic resulted in the hospitalization of many exclusively Spanish-speaking patients [24], understanding their experiences during prolonged hospitalization is an important avenue for study. There is considerable evidence that language services were restricted or even curtailed during the COVID-19 pandemic, potentially undermining communication for these individuals, yet few studies have explored this from the patient perspective [26, 27].

Some qualitative studies soliciting the perspectives of patients with NELP have been published but have mainly focused on outpatient settings [28,29,30,31,32]. In addition, there is a lack of qualitative studies that examine the experiences of previously hospitalized Spanish-speaking patients with complex illnesses utilizing community-engaged recruitment methodologies.

Community-engaged research (CER) studies prioritize the community’s needs and interests through robust communication and collaboration between academic institutions and community members [33]. CER can be designed and implemented in numerous ways, including involving the community throughout study design, implementation, dissemination, and solution creation [34]. Community-engagement is vital to health equity advancement because the community holds pertinent insights on the inequities experienced by their own members as well as potential solutions to remediate community mistrust from decades of exploitative research [33].

Our study’s objective was to utilize community-engaged recruitment methods and qualitative research to elicit and elucidate the unique experiences, beliefs, and perceptions of Spanish speakers who had been hospitalized due to complex illnesses, exploring interpreter use during admission, the impact of language barriers on medical decision-making, and the effect of COVID-19 on the hospital experience.

Methods

Study setting and design

The research team conducted a community-engaged, multi-site study between October 2022 and February 2023 in three geographic regions: the Southwest, Southeast, and Midwest of the U.S. We used videoconferencing to conduct virtual focus groups and semi-structured interviews in Spanish. The participants in this study were Spanish-speaking patients, their caregivers (e.g., family, loved ones, friends), and Hispanic leaders within the community. The Mayo Clinic Institutional Review Board approved the protocol (IRB ID: 22-003784). Oral informed consent was obtained from participants at the beginning of the focus groups and interviews.

Participants and recruitment

Inclusion criteria consisted of being 18 years or older and speaking Spanish as a primary language. Patients and caregivers had a recent experience of complex care in any hospital setting with a 10-day or greater stay within the last 3 years. The study also included community leaders, people who held community-based positions (e.g., CEO of safety net clinic, leader of Hispanic community center, religious leader) and had insights and opinions about complex care in hospital settings among Spanish-speaking members of their community.

The study utilized purposive sampling, using numerous recruitment tools. We collaborated with our institutional community engagement staff, attended in-person community events such as Latino heritage celebrations and faith gatherings held in Spanish. The research team also distributed paper and electronic recruitment material such as flyers that intentionally included wording that invited participants with experiences with language barriers and in-patient hospitalization due to complex illness, as well interpreter services, at sites frequented by Hispanic people and our institutions’ patient areas. To identify community leaders, Hispanic study team members leveraged personal and professional social networks, a purposive sampling strategy. We also used other recruitment strategies including electronic record review and portal messaging to identify patients admitted to our institutional hospitals who met inclusion criteria.

Patient and caregiver participants were remunerated with a modest cash incentive. The study did not remunerate community leaders.

Moderator and interview guide

The moderator and interview guide were developed by the multidisciplinary study team, comprised of biomedical ethics and health disparities researchers, an interpreter, physician researchers, and community-based research staff. It included questions based on literature review and the team’s content expertise and experience [27, 35, 36]. The research team explored the impact of language barriers on the hospital experience including communication, health outcomes, and care needs; interpreter access, use and challenges; and decision-making. We also asked about the impact of the COVID-19 pandemic on hospitalization experiences. Community leaders were asked similar question but tailored to evaluate their perspectives about the community. (See Appendix A).

Data collection and analysis

Based on geographical challenges and the ongoing impact of the COVID-19 pandemic on travel and daily life, we used a combination of virtual focus groups and one-on-one semi-structured interviews to offer the greatest accommodation to participants. Participants engaged in only one of the two data collection methods. Three bilingual study team members conducted focus groups and interviews in Spanish that lasted between 15 and 60 min. Interviews were audio-recorded, transcribed, translated and deidentified. When the study reached data saturation, and no new themes were identified, we ceased data collection [37, 38].

Two research team members conducted an initial analysis of transcripts to create a codebook, a framework to highlight common themes amongst participants. Data was analyzed using the codebook informed by grounded theory methodology with open, axial, and selective coding [39]. All data was coded, managed, and analyzed using NVivo Version 13 (Lumivero, Denver, CO) software. All transcripts were coded in duplicate, and coders met weekly to reach thematic consensus on all transcripts, ensuring coder reliability [40, 41]. The codebook was modified and refined during the coding process with the additions of parent and child codes. Coding definitions were progressively clarified during data analysis. Once coding was completed, investigators met to generate consensus on pertinent and representative themes, and to select representative quotes.

Results

We enrolled 27 patients (PT), 10 caregivers (CG), and 3 community leaders (CL) (n = 40). We conducted 1 focus group and 36 one-on-one interviews. The focus group consisted of four participants, 2 patients and 2 caregivers. Additional demographic data for the focus group and semi-structured interview participants are reported in Table 1. Within our cohort, 90% participated in an interview and only 10% participated in a focus group.

Table 1 Semi-structured interview, Focus Group, and aggregate participant demographics and characteristics
Full size table

Theme 1: Disconnected experiences impeding interactions, communication and decision-making (table 2)

Table 2 Disconnected experiences impeding interactions, communication and decision-making
Full size table

Spanish-speaking patients described feeling disconnected from the healthcare team. Patients noted that they could not express themselves to healthcare staff and that the language barrier inhibited their understanding of their medical care. Participants also stated that the language barrier stifled their confidence in medical decision-making and ability to express themselves. Participants reported that professional interpreters helped facilitate confident decision-making among patients and caregivers. In response to a lack of professional interpretation services, participants or their clinicians utilized numerous workarounds, including drawings, electronic translation applications, and body gestures.

Theme 2: Inadequate interpreter services (table 3)

Table 3 Inadequate interpreter services
Full size table

Most participants were offered, had access to, and used diverse interpretation modalities while hospitalized. However, some participants noted they were not offered interpreter services. Participants reported a preference for in-person interpretation over other modalities, such as remote video and phone interpretation. Participants noted that professional in-person interpretation facilitated use of body language and gestures, aiding communication. Furthermore, patients felt more comfortable and trusted in-person interpreters resulting in improved rapport building. When requested, however in-person interpretation services were frequently delayed.

Remote video interpretation services were perceived to provide some benefits over in-person interpreters, for example convenience and immediacy. One participant noted a benefit of virtual modalities reduced the risk of pathogenic transmission and supported infection mitigation. However, several participants expressed major concerns about using remote interpretation services including frequent technological and connectivity problems, causing frustration. Furthermore, participants conveyed that virtual interpreters were at times distracted and interpreting in non-professional environments. Consequently, patients and caregivers cited inaccurate and lower quality interpretation. Additionally, many participants noted virtual interpretation was impersonal and “flavorless” citing lack of body language and physical presence, leading to a sense of detachment.

Some participants also noted that elderly patients may not feel comfortable with virtual interpreters. Lastly, participants also expressed privacy concerns with tablet remote interpretation due to challenges with appropriately positioning the camera during patient interactions.

More broadly, community leaders commented that interpretation services use language that does not have a direct translation from English to Spanish. This can cause more confusion for patients and their families. An example a community leader cited was “comfort care”: in which they noted that Spanish-speaking patients might have difficulty comprehending the concept.

Theme 3: Benefits and consequences of family at the bedside (table 4)

Table 4 Benefits and consequences of family at the Bedside
Full size table

Family caregivers were an important part of Spanish-speaking patients’ hospital experiences, with participants citing that family has great importance in the Hispanic community. Many participants relied on family members to interpret important medical information to the healthcare team, especially when a professional interpreter was unavailable or not present. Many caregivers expressed a desire to stay with their hospitalized loved ones and patient participants noted that this decreased feelings of loneliness and anxiety. Furthermore, some participants noted that their family members provided additional clinical care, including changing urinary catheter bags. While family functioned as part of a patient’s support system, our study confirmed that family members also frequently served as interpreters when professional interpretation services were unavailable or delayed. The COVID-19 pandemic visitation restrictions however, impacted the role of family in different ways.

Using family as interpreters introduced challenges for patients and their caregivers. Family members expressed that the additional responsibility can be challenging causing discomfort. Some participants also noted that family members lacked training, adequate health literacy, and sufficient language proficiency to accurately interpret for the patient, creating confusion for the patient and healthcare team. Participants also noted that when family members interpreted frequently, they assumed the additional responsibility of being a designated surrogate decision maker.

Theme 4: Community-informed recommendations (table 5)

Table 5 Community informed recommendations
Full size table

When asked to offer recommendations, patient and caregiver participants offered several suggestions, particularly related to professional interpreter accessibility and quality. Overwhelmingly, they wished that hospitals would hire more in-person interpreters to mitigate shortages and delays. They also desired that medical staff become more aware about the need to use and how to access interpreter services. Some participants even recommended that hospitals hire dedicated bilingual health staff to ensure language needs are met.

Patient and caregivers urged interpreters to use accessible language and hospitals to provide additional education and Spanish resources in response to lower health literacy. They advocated for additional training for interpreters regardless of modality, and hospitals to consider accountability measures to bolster the quality of medical interpretation.

In response to conceptually challenging medical terminology, community leaders also recommended that professional interpreters should streamline interpretation and use accessible language. Community leaders encouraged hospitals to provide more education and resources about patients’ rights to free interpreter services, citing patients’ lack of knowledge and awareness.

Regarding decision-making, community leaders encouraged clinicians to employ the teach-back method to ensure patients have a complete understanding of their treatment plans and care more broadly. Community leaders also highlighted the role of professional interpreters in supporting discussions and medical decision-making.

Community leaders recognized that many Spanish-speaking patients worry about the cost of their care, including interpretation services. Therefore, community leader participants recommended that patients connect with community health workers and patient navigators to access resources to ease the burden of care and streamline ongoing care. Finally, community leaders urged patients and health facilities to engage and utilize surrounding community organizations to assist Spanish-speaking patients throughout their hospital experience.

Discussion

This community-engaged qualitative study recruited and highlighted the unique voices of Spanish-speaking community leaders, caregivers, and patients who had been hospitalized for prolonged periods to garner insights about the impact of language barriers on the hospital experience, the use of interpreter services, and the impact of COVID-19. There is limited literature capturing the voices of Spanish-speaking patients and caregivers hospitalized either before or since the start of the COVID-19 pandemic. Most of the existing literature is focused on hospitalized pediatric patients [42,43,44].

Participants noted that sometimes clinicians did not utilize or try to engage interpreters. Participants perceived that interpreter accessibility challenges stemmed from clinicians’ lack of knowledge and training or initiative to request these services for patients and caregivers, as well as a general lack of interpreter availability [10, 45, 46]. Similar findings were noted by Schenker et al. in a cross-sectional study that documented clinician’s frequent underutilization of interpretation services. Other studies have also identified this trend, and some have shown worse patient outcomes related to lack of access [47,48,49,50].

In our study, we found that participants favored in-person interpreters over remote modalities due to perceptions that remote interpretation was impersonal and more prone to inaccuracy. However, existing literature on modality preference demonstrates a variety of findings. A study using survey methodology among Spanish-speaking pediatric patients and their parents noted that parents were more satisfied with in-person than telephone interpreters [51], and a prospective cohort study by Stevens also found that patients exhibited more comfort with an in-person interpreter [52]. A study involving hospital personnel and Spanish-speaking patients documented participants’ perceptions that remote interpretation was detached but warranted in some situations such as emergencies, coinciding with our participants’ perspective that remote interpretation offers immediacy and convenience [53].

Another study evaluating the implementation of video remote interpretation in a metropolitan children’s hospital found that benefits included decreased wait times and that it supplemented in-person interpretation, affirming our study’s participants perception of remote interpretation’s convenience [54]. Although the challenges of remote interpretation, such as connectivity issues, have been previously documented, these can potentially be remedied through education and implementation rounds, suggesting that intentional implementation and troubleshooting may address some challenges faced by patients [54].

However, other studies report no difference in patient satisfaction between in-person and video interpreter modalities [55, 56]. Notably, those studies were conducted before the COVID-19 pandemic and did not focus on the in-patient experience. Given the reduced family visitation and potential for isolation, the desire for in-person connection may have caused a preference for in-person interpreters in our study [57, 58]. Furthermore, our study focuses on patients with complex illness that require challenging conversations that particularly benefit from in-person interpretation [35, 59, 60]. Studies suggest that ICU clinicians support in-person interpreter engagement for complex patients not only for verbatim language interpretation but also cultural brokering and to act as “health literacy guardians”, countering the so-called “Triple threat” [35, 36, 61].

Our study highlights caregiver and family members’ sense of duty to interpret and care for their hospitalized loved ones, citing that a family’s role in one’s care is an integral part of Latino culture. A qualitative study by Pines et al. reported that some patients with NELP prefer family interpretation over professional interpreters in certain situations, such as providing comfort to the patient [62]. Rosenberg et al. also documented that family members felt an obligation to support and interpret for hospitalized loved ones [63]. However clinicians may over-rely on family interpreters [64]. Literature suggests family interpreters act as a supplement to professional interpreters in addition to serving as an advocate and emotional support [63, 65]. In fact, participants in our study articulated concerns about relying on family members whose poor interpretation might lead to miscommunication or even potential harm, a finding not previously documented about patient perspectives. Despite these worries, in response to a lack of available interpreters, patients often relied on family members to interpret when presented with no alternatives.

Other scholars have raised ethical concerns about family interpretation; these include concerns about bias, accuracy, incomplete interpretation of all medical information, and intentional withholding of information that might be perceived to upset their loved one (for example non-disclosure of a poor prognosis or diagnosis) [66,67,68]. These concerns were also identified in a qualitative study assessing physician perception of family interpreters’ skill-level and accuracy [69]. No participants in our study articulated concerns about non-disclosure specifically but clearly some participants voiced concerns about accuracy impeding their care while hospitalized. Notably, although some studies have found Spanish-speaking patients feel like a burden due to their language needs, none of our patient participants expressed this concern [70].

A novel finding of our study is that the role of surrogate decision-maker fell upon the family member who had the most advanced English skills. While no studies, to our knowledge, have documented this family dynamic in medical care, higher quality clinician-family communication results in more confident decision-making [71]. In our study, the primary family interpreter served as a language broker between the healthcare team and family; therefore, this person had the most information to make an informed decision on behalf of the incapacitated patient. This finding raises questions about how clinicians may best initiate conversations about surrogate decision-making among family members and caregivers. In cases where a patient lacks a legally designated durable power of attorney, decision-making authority may be determined by state laws [72]. Clinicians may need to consider how language proficiency may influence who is chosen to have decision-making authority on behalf of the incapacitated patient.

Our study confirmed that language barriers make decision-making more difficult for Spanish-speaking patients. There is limited literature that seeks the perspective of Spanish-speaking patients about language barriers and medical decision-making. A recent multisite cohort study by Piscitello et al. noted an increase in unilateral do not resuscitate orders among Spanish speaking patients during the pandemic which may reflect communication difficulties that impacted decision-making and subsequent care [73]. An additional qualitative study also found that immigrant patients and physicians perceive that the language barrier poses difficulty during shared decision-making [74]. Other studies have also demonstrated the association between language and differences in decision-making [11, 27]. Given that our study focuses on complex patients with prolonged stays and challenging decision making as well as limited interpretation services, these findings are concerning.

Many participants recommended that hospitals should hire more in-person interpreters, echoing previous calls for hospitals to expand interpreter accessibility to address disparities in communication and care [75]. To address shortages and delays within professional interpretation services more broadly, patients and caregivers suggested that hospitals hire English-Spanish bilingual healthcare staff (e.g., physicians, nurses). This finding resonates within the literature. Work by Diamond et al., and a perspective by Knuesel et al., supports language concordant care provided by bilingual clinicians for increased patient satisfaction and improved health outcomes [13, 75,76,77]. Unfortunately, increasing language diversity in medical schools and the medical workforce and hiring sufficient bilingual clinicians may continue to be a challenge [78]. Additionally, there are other concerns in providing Spanish-language concordant care; for example, there are no standardized processes to assess Spanish proficiency or good training for healthcare professionals to provide Spanish-language concordant care [79,80,81,82]. However, potentially in the future health systems will be able to leverage these resources when available to provide optimized care to this vulnerable population. Otherwise, providing consistent professional interpretation services should be the norm in practice, meeting ethical and legal standards [83].

Other approaches to meeting the language needs of patients is to focus on optimizing interpreter services effectiveness and workflows to address barriers in the provision of interpreters [84]. A strategy the research team has developed and are currently testing with a pragmatic clinical trial is using artificial intelligence and informatics to identify patients with NELP and complex care needs to prioritize in-person interpreter use among patients experiencing complicated interventions or having challenging preference-sensitive decision-making. The team has integrated the algorithm into the clinical and interpreter services workflow and provide active outreach to clinicians with a secure chat through the electronic medical record nudging them to use an interpreter. Moreover, this strategy was considered acceptable and helpful to clinicians and other stakeholders [85]. Interim results from this work have shown a 37% increase in interpreter use in our intervention group [86]. Other use of artificial intelligence and machine learning to provide interpretation or translation (written word) require further study and cannot yet be recommended as acceptable modalities [87]. To address shortages of interpreters, career investment programs can train already existing healthcare staff (such as desk or housekeeping staff) to become interpreters in languages the institution may need that are spoken by the local population [88].

Clinicians also need better training about how to effectively work with interpreters as robust and effective training programs are lacking for both trainees and clinicians [89]. When managing patients with NELP and complex care needs we also recommend pre-encounter huddles between clinicians and interpreters prior to the discussion with patients and family. This strategy can increase bidirectional information sharing between interpreters and clinicians and improve the communication and the quality of the patient/family interpreted discussion [90,91,92].

Community leaders provided important insights towards mitigating disparities experienced by Spanish-speaking patients, such as engaging patient navigators and community health workers to educate patients and optimize their care transitions. They also supported hospital collaboration with local Spanish-speaking organizations. Several studies note that patient navigator and community health worker programs contribute to improvement in patient health outcomes [93, 94]. Hospital collaboration with community organizations and faith-based groups have also been shown to improve community health [95,96,97]. Community leader insights as key informants are fundamental for understanding the needs of Spanish-speaking-patients [98, 99]. Patient education and community engagement efforts can reduce readmission rates and hospital care costs, thus beneficial for patients, healthcare staff, and hospital administrators [100]. Additionally, in our study, community leaders noted a need for patients to advocate for themselves by requesting professional interpreter services.

Strengths and limitations

This study has several strengths. We leveraged community-engaged recruitment methodology as an important tool to include the most relevant perspectives in understanding equity issues associated with language barriers, specifically among Spanish speaking patients with complex medical needs. Our qualitative methods using a focus group and semi-structured interviews were conducted in Spanish by bilingual personnel, bridging the language gap between participants and English-centric research. Qualitative studies can deeply characterize the experiences of participants by utilizing a holistic perspective to explore phenomena [101, 102]. Additionally, the purpose of qualitative methods is to understand the meaning of participants’ perspectives and the context surrounding them [102]. We were able to conduct the study across multiple sites in three regions. We used robust data analysis approaches and triangulated our data by interviewing diverse groups (patients, caregivers, community leaders). These measures bolstered the study’s scientific rigor and validity and ensured the trustworthiness of our findings and strengthened the relevance of the discoveries [103].

This study also has some limitations. Firstly, we had initially planned to conduct focus groups for data collection, but scheduling challenges and a substantial proportion of no-shows compelled us to switch to one-on-one semi structured interviews to ensure consistent data collection [104]. The advantage of one-on-one semi-structured interviews is that they may foster a more comfortable environment for the participant, facilitating more dialogue between the interviewer and the interviewee however focus groups can foster interactive dialogue between participants [104]. Many participants were patients within our health system, an institution that tends to have more robust patient resources, although some were hospitalized in other institutions. Therefore, some findings may not be generalizable to other institutions and settings [105]. As with similar studies, this study may be prone to selection bias with those having strong opinions or an affinity with this topic agreeing to be involved. Although, the interview guide included questions about COVD-19 experiences, participants did not provide many perspectives on this, despite prompting. It is important to note that disentangling the contribution of poor language proficiency and low health literacy is sometimes challenging particularly with complex illnesses that may require complicated interventions however the focus of this study was intentionally language barriers [106, 107].

Further directions

Spanish is the most common language spoken in the United States after English [108] and much of the literature surrounding language barriers describes languages frequently spoken throughout the U.S., such as Spanish, Mandarin, Vietnamese and others [109]. However, future work is needed to understand the experiences of patients who speak languages of lesser diffusion such as indigenous Central American languages and Karen, a group of languages spoken in Myanmar [110, 111]. The development of guidelines for bilingual clinicians when interacting with patients is needed as well as metrics to evaluate clinician proficiency. Additionally, while inclusion of non-English speaking participants is especially important in research at the intersection of health and language, generally, research in all disciplines can benefit from inclusion of non-English participants. Increased protections for vulnerable populations (e.g., incarcerated people, pregnant people) is necessary to prevent exploitative research; however, these protections can also be a deterrent to conduct research with these populations. Further work is needed to provide guidance on how to include participants from disenfranchised backgrounds that prevents exploitation but facilitates more inclusive research opportunities and can make research more generalizable [112]. If legal and ethical imperatives are not sufficient for institutions to provide access to language services, the economic benefits may be. Work examining the costs of providing interpreter services in diverse settings is needed to demonstrate how optimizing communication can positively impact length of stay, healthcare utilization and other measures as well as patient satisfaction [113,114,115,116,117].

Conclusion

This qualitative study that utilized community-engaged recruitment strategies sought to deeply characterize the hospital experience of Spanish-speaking patients with NELP and complex health needs. This vulnerable patient group were more prone to poor communication with their healthcare team, resulting in some distress and challenges during medical decision-making. If access to professional interpretation services occurred, it was usually remote video interpretation and participants often found the interpretation to be inaccurate and low-quality. Many participants relied on family members for interpretation, care, and support. Recommendations for hospitals from participants included expanding in-person interpreter accessibility, training bilingual clinicians, enhancing remote interpretation quality, and developing resources towards patient education.