Telephone triage utilization among patients with limited English proficiency

Systems for telephonic triage of health concerns have become increasingly common as a cost-effective means of improving access to medical advice and care [1]. In the primary care setting, these triage call centers are typically staffed by nurses who use computer- assisted decision-making tools to provide care recommendations to callers and manage patient consultation requests [2]. This consultation management has been shown to effectively manage patient access to providers while reducing workload and maintaining costs [3‐6]. In the broader healthcare context telephonic triage systems may reduce costs by directing patients to less resource intensive encounters (e.g., primary care appointment instead of an emergency department (ED) visit) [7, 8]. Computer-assisted triage call centers have also been shown to maintain safety standards in providing medical recommendations over the telephone and are effective in assessing the appropriateness of care decisions [9, 10]. Questions still remain, however, about patient adherence rates with nurse recommendations [11], which may be affected by the quality of communication between the patient and the triage nurse [12]. For instance, some callers may misinterpret nurse recommendations while others choose not to adhere to recommendations due to their own intentions, health beliefs, or social circumstances [12, 13].

Quality of healthcare communication is a particularly salient issue among patients with limited English proficiency (LEP). LEP is defined as “speaking English less than very well” in any person 5 years of age and older, and it is associated with disparities in healthcare utilization and health outcomes [14, 15]. Telecommunication studies examining interactions between healthcare systems and patients with LEP have demonstrated barriers and negative outcomes related to emergency medical service calls [16, 17]. Disparities in utilization of telephone triage have also been demonstrated among non-native language speakers in countries outside the U.S. Non-native Swedish speakers utilized a national phone triage service less frequently compared to native speakers, while non-native Norwegian speakers reported less trust in recommendation/advise given to them by nurses through telephone triage [18, 19]. To help mitigate these barriers, healthcare systems typically contract with professional medical interpreters and telephonic or video interpretation services. Interpreter use reduces, but does not eliminate health disparities among patients with LEP, partly due to the type of interpreter used and the quality of interpretation services [20, 21].

Communication between patients with LEP and primary care telephone triage services which use computer-assisted decision making algorithms has not been previously examined. As the use of triage call centers becomes more prevalent and the proportion of patients with LEP continues to grow, it is important to understand utilization patterns within these populations. Moreover, given that significant disparities in care exist within face-to-face healthcare encounters/systems, the potential exists for exacerbating disparities when communication is by phone. Therefore, the purpose of this study was to determine the utilization of a primary care triage call center by patients who require interpreter services.

This was a retrospective cohort study of the utilization characteristics and outcomes of computer-aided nurse telephone triage calls by English proficiency status of patients empaneled to a large primary care practice network in the Midwest United States between 1/1/2012 and 6/30/2013. Patient data were accessed only if they had an active Minnesota research authorization on file, allowing use of their medical records for research, [22] and all study procedures were approved by the Mayo Clinic Institutional Review Board.

In this study of a computer-aided nurse-led telephone triage system in a primary care network, we found similar proportions of callers making one or repeat calls between patients who required IS and those who did not. However IS patients were more likely to receive recommendations for higher acuity care and urgent visits compared to non-IS patients. Furthermore, IS patients were much less likely to follow the recommendations received. To our knowledge, this study is the first to describe call characteristics and adherence to triage recommendations among patients with LEP. These results highlight the need for health systems to examine why LEP calls result in significantly lower adherence to triage recommendations and whether this lower adherence results in poor health outcomes or, perhaps, reflects a need to refine the triage process for patients with LEP.

Previous studies have shown patterns of inefficient healthcare utilization among patients with LEP compared with patients who are English-proficient. These studies reveal higher rates of use for diagnostic studies and increased ED visit times for patients with LEP, higher rates of outpatient and inpatient health care utilization [25‐27], and longer hospital stays [28]. The results of this study support these previous findings, and suggest that these inefficiencies may be mediated, in part, by the tendency of primary care systems to disproportionately recommend higher acuity services to patients with LEP.

The reasons for the difference in recommendations given between the IS and non-IS groups that were demonstrated in this study are likely multifactorial. Although there were no significant differences in the chief symptom for calls between the two groups, we were unable to determine the severity of symptom complexes for each call; since IS patients were slightly more medically complex, it is possible that severity contributed to recommendations for more acute and aggressive care among the IS cohort. However, our findings did not change after adjusting for potential confounders, including CCI. Furthermore, previous studies have reported significant differences in care recommendations and patterns for patients with LEP despite similarities in presentation with English-proficient patients. For example, patients with LEP who presented to the ED with abdominal pain received three times as many abdominal computed tomographic scans as patients with English proficiency, while infants of parents with LEP were more likely to be prescribed antibiotics compared to those with parents proficient in English [29, 30]. One study of pediatric patients showed higher rates of hospital admission among patients with LEP compared to patients with English proficiency, although the acuity level at presentation was similar [31]. Communication may play the primary role as the mediator of the differences observed in the care recommendations between the two groups. The decision to recommend higher acuity of care may be influenced by a need on the part of the nurse to compensate for communication barriers by recommending more aggressive therapy options. This may be even more compelling in the case of telephone communication, where the healthcare provider does not have the benefit of examining the patient and where the patient may be communicating through a surrogate.

Our finding that patients with LEP were less likely to follow recommendations may reflect a disproportionate lack of understanding of those recommendations and their rationale among patients with LEP. The communication challenges of linguistic and cultural non-congruence between patients and nurses may be compounded by a relatively low health literacy, which is often interrelated with LEP, and has been independently associated with healthcare utilization and higher medical costs. [32‐35] In this particular group of patients with LEP, the influence of health-seeking norms specific to their countries of origin may also play a significant role in their reduced acceptance of phone triage recommendations. For instance, in some recent immigrant groups, patients are only familiar with acute care models, which may prompt delays in seeking care until the perceived acuity escalates [36].

In this study, IS patients showed a higher rate of using a surrogate for triage calls compared to non-IS patients. Most surrogates are family members who serve as ad-hoc interpreters with no professional training. Several studies have examined the use of ad-hoc interpreters in healthcare settings and found an increased risk for poorer health outcomes due to interpretation errors and to family members imposing their own agendas on the care plan. [37, 38] The use of surrogates to make triage calls in the general population has also been shown to increase utilization rates for higher acuity services possibly due to surrogates anticipating these endpoints or exhibiting a higher level of concern for the patient’s symptoms. [39, 40] These findings in the general population may be potentiated among patients with LEP due to the added language barrier. While well-trained and qualified interpreters are the recommended standard of care among patients with LEP, further research is needed to explore the use of surrogates for triage calls in this population in order to better understand how their use affects outcomes and healthcare utilization.

Interestingly, our findings that IS patients were less likely to go to the ED when instructed but more likely to choose a routine appointment within 1 week suggests that IS patients may be choosing continuity with their primary care provider (or care team) over emergency or same day services with less familiar teams. This finding is encouraging given that previous research has found higher ED utilization among patients with LEP for dental, eye, skin, and ear, nose and throat concerns; most of which can be typically addressed in the outpatient setting [26]. Furthermore, encouraging patients with LEP to utilize primary care services for acute care needs may help address the existing inequities documented for management of chronic diseases among patients with LEP such as mental health, asthma, and diabetes mellitus [41‐43]. However, it is also possible that this finding of less likelihood of visiting the ED when advised to may be because of fear of the financial implications of the ED visit and/or the ambulance transportation cost on the part of the caller.

Our study has several limitations. First, the data were collected from a single primary care triage call center, which limits the generalizability of our results. Furthermore, the demographics of patients with LEP in this sample may be different from other primary care centers serving patients with LEP. Second, the data collected was based on the assumption that those callers, who identified as needing IS in the electronic health record, were patients with LEP. Using IS as a proxy recognizes only a subset of patients with LEP [23]. Third, the study incorporated retrospective data that may be influenced by unidentified confounding factors or may be skewed due to missing data, such as patients seeking care in clinical sites other than the primary care practice where this study was done. However, measurable confounding factors were controlled for and there were minimal missing data. Finally, although institutional policy requires that interpreters are used for all communication with patients with LEP, we were unable to ascertain that this policy was adhered to in all of the calls in this study, and we could not ascertain the quality of interpretation in each encounter.

In summary, among patients empaneled in a large primary care setting, IS callers who utilized a computer-aided, nurse-led telephone triage system were more likely to receive recommendations for higher acuity care visits compared to non-IS callers. IS callers were also less likely to follow the recommendation and had a higher rate of surrogate caller use. Additional research is needed to further understand why telephone recommendations to IS patients have more urgent endpoints and why the IS patients are less likely than non-IS patients to follow the triage recommendations. Finally, efforts to refine or tailor telephone triage to better serve patients with LEP seem warranted through quality improvement and research initiatives, such as use of video technology to augment objectivity in system assessment and communication.