Through a bundle of multifaceted interventions, we were able to increase our NICU’s utilization of in-person Spanish interpreters five-fold above the baseline while simultaneously increasing overall usage of all modes of medical interpretation. Importantly, several of our interventions took little effort & cost to implement and had immediate effects.

Significant improvements in in-person interpreter utilization were seen with each intervention we implemented. Staff educational sessions & health equity grand rounds focused on language access were associated with a three-fold increase in in-person interpreter utilization. These interventions had no associated cost beyond time and effort. These educational interventions primed our unit to maximize the impact of a dedicated in-person interpreter once their services were secured for daily patient rounds.

Removing barriers to use of high-quality medical interpretation is a powerful tool to ensure that families receive communication from the medical team in their preferred language [13, 16]. Installation of a dedicated unit interpreter led to a significant increase in interpreter utilization. Additionally, having a dedicated in-person interpreter had a profound impact on unit culture regarding language access. The interpreter was perceived as part of the team, rather than a cumbersome service. This intervention created a feedback loop, where, as the team realized the ease of use, built trust with, and experienced superior communication with families through the in-person interpreter, there was a natural desire to continue working with the interpreter. This feedback loop steadily increased interpreter utilization. We believe this slow but steady culture change regarding language access is the most impactful aspect of our work.

Our work aligns with prior limited literature investigating methods to increase interpreter utilization. A similar multifaceted quality improvement project at a large children’s hospital in the Pacific Northwest increased telephonic interpreter utilization rates from 0.38 uses per person day to 0.58 [15]. A health system-wide initiative focused on improving language access in Michigan for adult patients with diabetes resulted in an increase from 19% to 83% in the proportion of outpatient visits for patients with PLOE with a qualified language services provider present [22]. Our work demonstrates a similarly low initial level of interpreter usage with a marked increase following focused interventions. Concordant with our work, both cited studies relied heavily on staff education regarding the importance of medical interpretation and removal of barriers to existing language access resources.

Our efforts have a few distinguishing features from prior work. First, we concentrated on increasing in-person interpretation (rather than telephonic interpretation), as this modality is widely considered the gold standard of medical interpretation and our hospital had existing resources in place to allow for frequent use of in-person interpreters. Second, our work was in an ICU setting, where complex medical discussions with families routinely occur and optimal communication is of utmost importance. Last, we made a conscious effort to educate families on their right to medical interpretation and empower them to request interpreters.

It is critical to note that our interventions led to an overall increase in use of any interpreter modality (i.e., both in-person and video). Thus, we did not find evidence of staff limiting interactions with families when it was not feasible to use an in-person interpreter. Rather, the in-person interpreter was preferentially used, but video interpretation remained widely employed. We hypothesize that the initial surge in video interpretation in the beginning of our intervention period was related to increased awareness of need for medical interpretation but limited ability to access in-person interpretation. Our video interpretation utilization returned to baseline levels as we removed barriers to in-person interpretation (i.e., implemented a dedicated NICU medical interpreter). This pattern of interpreter modality use was consistent with our suggested language access guidelines presented during staff educational sessions.

We encountered several challenges during our work that may be informative for other hospitals focusing on improving language access. First, we had initial difficulty understanding our pattern of interpreter utilization at baseline, as there was no dedicated process in place to track interpreter use. We were able to better study our interpreter use after creation of monthly reports of total in-person and tablet requests by using available data from the electronic medical record and video tablet interpretation company. Second, cost and staffing were two major constraints that initially prevented us from obtaining a dedicated in-person interpreter for the NICU. A dedicated NICU interpreter was approved on a pilot basis; given her impact and overwhelmingly positive feedback from NICU staff and families, members of our task force were successful in obtaining grant funding for a full-time NICU interpreter.

There are several limitations to our work. First, we did not track interpreter requests for individual patients. For example, our current data shows only the total number of interpreter requests for any given day – not if each individual family received an update via a medical interpreter. Future work may focus on understanding patterns of interpreter utilization for individual patients (rather than aggregate interpreter utilization in the NICU). Second, our work did not delineate what type of communication occurred when the in-person interpreter was requested. For example, we could not ascertain from our data whether the interpreter request was for a direct medical update from a provider or for another reason (i.e., orientation to the unit, equipment education, etc.). Lastly, many of our neonatal health outcomes are not systematically tracked by language, limiting our ability to describe associations between our work and other health outcomes such as receipt of human milk, length of stay, & readmission after discharge, among others.

In conclusion, we substantially increased our unit in-person interpreter utilization through a series of multifaceted interventions, many of which were low-cost. The ability to communicate with the medical team is a patient right and an important driver of health equity. We encourage NICUs to systematically study patterns of interpreter utilization to identify areas of improvement. Finally, we advocate for NICUs to view an in-person interpreter as a dedicated and integral part of the NICU team.