Our study presents several important findings as the first paper to examine the impact of medical interpreters on eyecare visits. Patients needing an interpreter spent slightly longer with both their ophthalmic technician and eye care provider compared to patients without an interpreter. The longer time could be accounted for by time waiting for an interpreter to be available, time for interpretation itself, and/or adjustments in communication strategies and behavior. The need for an interpreter had greater impact on time with technician than time with provider; this is intuitive since our technicians are required to collect more history (such as reviewing medication lists), and in our study LEP patients were more likely to have their allergies, medication list and problem list reviewed (although the data does not distinguish between review by provider or review by technician, this task is usually performed by the technician). The technicians are also often responsible for the initial refraction, which can be difficult even for English speakers. In this clinic, which sees many vulnerable populations, poor literacy may further contribute to this language barrier, making refraction as well as medical interpretation even more challenging. Unfortunately, literacy status is not routinely recorded in the EMR so this study could not adjust for this, which may have influenced the results. Interestingly, many of the differences in encounter times were not significant when speakers of the same language who did not require an interpreter were used as the comparison. This may be accounted for in part by the smaller sample size when dividing each language into two groups for comparison. However, it could also be that there are cultural differences, aside from language, which impact the encounter length.
We also showed that patients who need an interpreter spent more time waiting for their provider after the technician had completed their initial work-up, however the only language it reached statistical significance for was Spanish. Some possible explanations for this include providers prioritizing patients who do not need an interpreter, and/or dialing the number for an interpreter then moving on to do something else while waiting for the interpreter to come on the line. Fortunately, our data suggest that needing an interpreter or speaking a language other than English did not change the time waiting to be taken into an exam room. Long wait times are often cited as barriers for LEP patients seeking care,16 one hypothesis for this could be that they are not prioritized in the waiting room since staff believe their visit will be more difficult or take longer. However, our data contradicts this, and suggests that if waiting room times are long, they are experienced equally by all patients in the clinic.
This study found that LEP patients were more likely to keep an appointment once it had been made. It is possible this reflects lack of interpreter use by front desk staff. For example, patients may not want to call again once an appointment has been made. Another possible explanation is that LEP patients were more likely to receive their After Visit Summary (AVS) than English speakers, and this AVS includes their next appointment time. Prior studies have suggested even when interpreters are used in a clinic by nursing staff and providers, patients will often try and “get by” at the front desk without an interpreter.18 Like our technicians, almost all of our clinic schedulers are bilingual in Spanish and English. However, this does not help patients who speak a language other than Spanish or English, and the effect was seen for both Spanish and other languages. This suggests that there may be other factors contributing, such as value placed on the appointment or cultural differences.
Limitations
Our study has several limitations. First, patients self-identify their primary language and whether they need an interpreter or not. In our study, almost ¼ of patients whose primary language was not English did not require an interpreter. This could result an underestimation of effect if patients falsely identify their primary language as English (for example to avoid perceived provider bias). Additionally, if a patient denies need for an interpreter it raises the question of whether some patients with limited fluency in English may not understand details of all discussions or risks of surgery, but are embarrassed to admit this lack of understanding which could affect quality of care. Alternatively, patients sometimes come to their appointments with English-speaking relative family or friends, and prefer to have their companion translate for them, which is not recorded typically. Prior studies have reported that interpreter utilization changes with how they are offered: "In what language do you prefer to receive your medical care?" appears to be mostly likely to result in appropriate interpreter utilization.19 Second, our EMR only records whether an interpreter was needed, we cannot to be sure that a qualified medical interpreter was used for the entire visit in every case where it was needed. Also, these results highlight the impact of primarily remote interpreters as are found in our clinic, and results may be different than in person interpreters. Although one study during the COVID-19 pandemic reported no difference between remote and in person interpreters,20 it has not been widely examined. Finally, most of our technicians are bilingual in English and Spanish which could have impacted our results for time with technician for Spanish speakers. The data collection method we used does not specify which technician took part in the encounter and so we are unable to account for bilingual staff. This is an important area for future studies to examine as prior reports suggest significant cost savings with bilingual staff,9 and it is likely to impact time as well.
In addition to understanding the impact of bilingual staff, our study raises the question of whether the same care is being delivered if the visit is only 60 seconds longer and it takes an average of 16 seconds21 to get an interpreter on the line. This average phone waiting time for an interpreter is reported by the interpreter service contracted by our hospital and is likely a generous estimate. Nonetheless, it seems unlikely that the provider is communicating the same amount of information to the patient through an interpreter using only 44 extra seconds of time as compared to a visit without an interpreter. This an important point that future studies must evaluate. Further, it is still to be determined whether this difference is associated with patient outcomes or satisfaction, and these are significant questions that should be also targeted by future research.