Key Question 1. Summary

We assessed the impact of remote triage on utilization of health care, case resolution, patient safety, patient satisfaction, and costs. Studies were too heterogeneous to conduct meta-analysis. We conducted narrative synthesis, focusing on larger and higher quality designs. Overall, these studies tested 3 major comparisons: (1) mode of interaction between patient and practitioner (ie, telephone vs in-person consultation); (2) triage conducted by staff of different professional types (eg, nonclinical call handler, nurse, general practitioner [GP]); and (3) level of triage organization (eg, national triage systems, local in-practice triage systems). None of the studies that met KQ 1 eligibility criteria addressed modalities of triage delivery other than in-person and telephone. As a result, we were unable to address KQ 1B on the impact of remote triage by different modalities such as video, web, and SMS.

Overall, the majority of included studies did not demonstrate a decrease in primary care or ED use; however, the current evidence is limited and of marginal quality. Only 1 high ROB study found a significant decrease in primary care utilization when comparing a national telephone triage system to a more local telephone triage system,⁴⁵ and no study found a decrease in ED utilization. Instead, 4 studies reported a significant increase in utilization.^38,39,41,42

Evidence from 2 studies suggested that local, practice-based telephone triage services have higher case resolution outcomes and refer fewer patients to emergency or primary care services compared with regional/national telephone-based remote triage.^40,43,44 Yet only 1 of these studies assessed the rate of referral to emergency services; both local and regional/national telephone-based remote triage services referred very low numbers of patients to emergency services.⁴⁴

We also explored safety outcomes including ED visits, emergent hospitalization, and death; patient satisfaction with the provide remote triage service; and cost of providing remote triage. Only 2 studies^38,43 addressed safety outcomes, and neither identified statistically significant differences in safety outcomes among study arms. No clear pattern emerged about the effects of remote triage on patient satisfaction. Some evidence supports that patient satisfaction is affected to the degree that patients perceive the service they receive to differ from the service they expected (eg, seeking same-day appointment vs after-hours advice). Last, we addressed the comparative costs of a telephone triage system. Two studies evaluated the costs of in-person primary care to either GP-led or nurse-led telephone triage and found no difference in overall cost of care.^38,41 A third study compared a national telephone triage system to a local triage system, finding that overall cost was not different when controlling for the final point of contact.⁴⁰

Our stakeholders identified health care utilization, patient safety, and patient satisfaction as the outcomes critical to decision-making. Thus, these are the outcomes for which we conducted certainty of evidence (COE) ratings. These assessments reflect the degree of confidence we have in our summary findings. We focused on rating the COE for the randomized study designs, since the nonrandomized studies had consistently discordant confidence ratings from the randomized designs.⁸⁷ For each outcome of interest, we present the COE by the 3 comparisons of remote triage services: mode of triage delivery (ie, telephone, in-person), triage professional type (eg, nonclinical call handler, nurse, GP) and organizational level of triage system (eg, national triage systems, local in-practice triage systems). These ratings are summarized below, with supporting information provided in Table 13.

• We found moderate COE to support that remote triage has no effect on ED utilization among the studies comparing in-person and phone modalities and call professional type.
• We found moderate COE for no effect on ED utilization among the studies comparing between local in-practice triage and regional/national triage call centers.
• We found moderate COE for an increase primary care visits among the studies comparing between in-person and phone modalities.
• There is low COE that remote triage operated by different call professionals increases primary care utilization.
• We found low COE to support that remote triage has no effect on primary care visits among the studies comparing between local in-practice triage and regional/national triage call centers.
• There is low or very low COE that remote triage has no effect on reducing patient deaths or improving patient satisfaction.
• Of the included studies, only 3 were high-quality, randomized studies that were rated with an overall low ROB that reported each of these outcomes.

Table 13

Certainty of Evidence for Primary Outcomes of Effect of Remote Triage.

Key Question 2. Summary

No studies identified best practices but brought forth multiple considerations for promising practices when implementing a remote triage system in a large-scale health system such as the VA. Thematic synthesis of the abstracted data identified 11 themes across all KQ 2 studies that conceptualized considerations for best practices that impact the planning, execution, and evaluation of remote triage services for adults seeking clinical care advice. Overall, the planning phase contained the greatest number of studies (n=19), followed by evaluation (n=14), with execution phase having the fewest number of included studies with relevant findings (n=11). Across aspects of remote triage implementation, the process domain contained the largest (78%) and technology domain had the smallest (25%) volume of included studies.

Implementing a remote triage service is a multifactorial process. There are several key findings for emerging practices for the implementation of a remote triage system that may improve efficiency and outcomes. First, the decision to implement a remote triage system influences the entire health care system. At the individual level, considerations must be made for individuals serving as remote triage staff, including (1) adequate initial and ongoing training and (2) a work environment that supports physical and emotional well-being, patients who use triage, and ancillary staff who assist in the daily functioning of triage. At the clinic level, considerations must be made for how remote triage influences the clinic workflow, scheduling and availability of appointments, and workload among clinical and nonclinical call handlers. At the system level, considerations must include how remote triage is influenced by, and also influences, the availability and accessibility of health care services (ie, clinic appointments, ambulance services). Attention must be paid to the health care resources in the external environment that impact both remote triage decisions and patients’ ability to adhere to advice. Purposeful planning prior to, and throughout the implementation of, remote triage is important in ensuring the success of remote triage. Next, educating patients and their family members on the purpose of remote triage may promote appropriate use of remote triage services. Third, involving call handlers with clinical experience in the planning, execution, and evaluation of remote triage services may facilitate future implementation and use by ensuring that remote triage programs enhance the patient-provider relationship. Last, implementing a remote triage system is perceived as safe, has the potential to reduce medical workload, and can produce a high rate of call resolution. Yet, it remains unclear whether a reduction in workload is actual or only a delay in the provision of health care services.

Key Question 3. Summary

There are multiple considerations for how to evaluate a remote triage system. We sought to curate a list of possible metrics for operations leaders to consider when assessing the implementation of such a system in the VHA. Thus, we curated the outcomes from the comparative studies in KQ 1. We grouped these metrics based on the 6 categories adapted from Carrasqueiro et al.³⁴ These categories include (1) enhanced access to care; (2) change in rates or trends of services use or change in professionals’ workload; (3) adverse events (deaths, emergency department attendance, hospital admissions) and delayed care; (4) clinical outcomes after triage; (5) patient satisfaction measured via Likert scales; and (6) savings from avoided services use and triage costs.

Prior Systematic Reviews

Most prior literature reviews of telephone triage have included primarily observational studies,^23,66,88 with only 1 (completed for the Cochrane Database) limiting the studies reviewed to those meeting EPOC criteria. Like our review, none found sufficient homogeneity to allow for meta-analysis.⁸⁹ Overall, few conclusions can be drawn from the prior systematic reviews. The most consistent findings appear to be that telephone triage did not decrease emergency room visits,⁸⁹ a finding that is consistent with what we report. There was evidence that telephone triage may reduce GP workload in the near term, although some studies suggested these visits may just be delayed.^88,89 In contrast, we report no positive impact on decreasing primary care utilization and found 4 studies that reported increase rates of primary care use in patients experiencing remote triage.

Across previous reviews, low power to detect effects and heterogeneity in study designs, interventions, and outcomes limited the conclusions that could be drawn about key constructs including cost, safety, access to care, patient satisfaction, or differences in effectiveness of triage by different health care professionals.^23,66,88,89 Several reviews concluded that the approach to outcome evaluation critically impacted the results and the conclusions that were drawn.^23,66,89 For example, observational studies, studies that surveyed patients, or studies that reviewed medical records were likely to conclude that remote triage was safe with no significant increase in adverse events (AEs). Studies using high-risk simulated patients, however, suggested that about 50% of patients were likely to receive unsafe advice, significantly increasing the risk of AEs.^23,66 In our review, we only included EPOC designs and objective measures of safety (ie, accident and emergency visits, emergent hospitalization, death). Only 2 studies addressed these outcomes and neither identified statistically significant differences in safety outcomes among study arms.^38,43 Finally, 1 review concluded that the safety and quality of remote triage appeared to be linked to the properties of the broader system in which the triage was rooted, including policy priorities, health care costs, demographic and cultural factors, and technical infrastructure.²³ We were unable to address this contextual finding in our review.

Clinical and Policy Implications

Remote clinical triage centers are an increasingly prevalent feature of health care delivery, particularly among large health care organizations. Their growing popularity is matched by a myriad of applications that can vary by contact modalities, staffing models, technologies, expected users, and outcomes, among other features. In the VHA, remote clinical triage centers are similarly fragmented in that there is no standard model for such a center, and any particular process is often based on local decisions, needs, and resources. A stated goal of the VHA is to develop a model for a 24/7 clinical contact center that allows for a variety of contact options. The multi-modal contact feature is important and is meant to include telephone, chat, text, and video. Existing programs in the VHA provide patient-to-clinician communication. For example, My HealtheVet (https://www.myhealth.va.gov/mhv-portal-web/home) is a web-based portal that, in addition to other characteristics, allows patients to send secure messages to their clinicians or clinical teams (eg, patient aligned care team or PACT). While My HealtheVet is available and functions VHA-wide, telephone contact with those same clinicians and clinical teams often follows local mechanisms as prescribed by the clinic or local telephone triage agreements. One goal of our review is to synthesize the evidence and best practices that VHA can use in developing and optimizing a multi-modal clinical contact center model with the potential for national rollout. Our review, however, mostly only focuses on telephone triage, at that is the predominate modality in the literature.

An important goal for remote clinical triage centers is the ability to provide case resolution in the first contact. Such resolution means that a telephone call is managed without triage for other services, or a caller is connected with the appropriate individual with only one call transfer. On call case resolution reflects how well the initial contact serves the needs of the caller by matching care with medical need. In the 2 studies comparing regional or national triage systems to local in-practice triage systems, the practice-based triage system resulted in a greater percentage of case resolution with initial contact.^40,44 A third study found the opposite result but was comparing remote triage with a same-day in-person appointment mechanism.⁴¹ This situation illustrates an important policy consideration: should the remote triage system goal be to resolve cases without triage to in-person services as opposed to determining the appropriate triage destination for that patient?

Designing a remote triage system has implications related to staffing (clinical vs nonclinical staff), setting caller expectations, and other considerations. For example, having a clinician (eg, RN) as the first point of contact could allow for dispensing of medical advice, reducing the need for further triage. An alternative staffing model might involve training non-clinicians to make triage decisions (at sites of care that do engage clinical staff). These different design options come with implications for first-contact outcomes and staffing costs. Findings from our exploration of the best practices provides insights into considerations for implementing staffing structures to optimize outcomes.

Lastly, multi-modal contact is another important consideration for the VHA when designing a remote triage system. The ability to enter the triage system by a means other than a telephone call will be important to study and understand. In particular, smartphone-based mechanisms such as texting, messaging, and chat as alternatives to telephone calls may be a preferred means of contact for many Veterans. While the mode that Veterans use to contact the system may be transparent to the recipient, it is nonetheless important to also consider how the recipient receives these requests and what the expectations are for processing requests. Unfortunately, our review identified only 1 non-telephone-delivered study, and so we were not able to provide evidence to support the development of multi-modal contact centers.

Limitations

Our review has a number of strengths, including a protocol-driven design, a comprehensive search, inclusion of EPOC designs best suited to assess organizational-level interventions, and careful quality assessment. For KQ 2, in addition, we conducted rigorous qualitative synthesis that combined thematic analysis and matrix analysis. Both our review and the literature, however, have limitations. Our review was limited to English-language publications, but the likelihood of identifying relevant data unavailable from English-language sources is low. The number of identified studies for many outcomes was small, and most of the comparative literature for KQ 1 and KQ 3 had design limitations that affected study quality. Other limitations are detailed below.

Research Gaps/Future Research

This comprehensive review of the literature identified several gaps in the current evidence that warrant future investigation. We used the framework recommended by Robinson et al⁹¹ to identify gaps in evidence and classify why these gaps exist (Table 14). This approach considers the population, intervention, comparator, outcome, timing, and setting (PICOTS) to identify gaps and classifies them as due to (1) low strength of evidence or imprecise information, (2) biased information, (3) inconsistency or unknown consistency, and (4) not the right information. VA and other health care systems should consider their clinical and policy needs when deciding whether to invest in research to address gaps in evidence.

Table 14

Evidence Gaps and Future Research.

Conclusions

The US health care system faces several challenges, including an aging population, patients with multiple chronic conditions, and an uneven distribution of primary care providers across the country. These conditions create a shortfall in access to primary care, pushing some patients to seek care in urgent or ED settings. Remote clinical triage systems have the potential to reduce medical workload, improve access to primary care advice, and reduce inappropriate use of urgent care and ED services. Our review provides evidence that the remote triage systems we studied may be falling short of these goals, although the identified literature was sparse and of variable quality. We found limited evidence to support that remote triage reduces the burden on primary care utilization or subsequent use of ED services. However, remote triage by telephone can produce a higher rate of call resolution than regional or national systems, and appears to be safe in the 2 studies that assessed these outcomes. Yet it remains unclear whether this rate of case resolution results in an actual reduction in use of primary and ED services or only a delay in the provision of services. Last, our study underscores several key considerations for implementing a remote clinical triage system, including the careful consideration of organizational and stakeholder buy-in prior to remote triage launch, physical and psychological workplace environment, staff training and ongoing support, and careful consideration of what metrics best assess the effectiveness and efficiency of remote triage implementation. Further study is needed to assess the promise of remote triage in optimizing health care outcomes while maintaining patient-reported satisfaction.