Telephone Triage & Telehealth

A Brief History

An internet search of the term “telehealth” illustrates how a new industry commingles and conflates terminology, be it clinical, technologica or commercial. Search results include definitions of telemedicine, telephone triage, telehealth, nurse triage, phone triage, virtual visits, televisits and other forms of high-tech remote clinical encounters between patient and clinician.

The term Telehealth has come to stand for both the industry-at-large, as well as pre-hospital or telephone triage — an earlier form of remote, virtual care — in place of a a face to face visit or ED visit. Even before the pandemic, telehealth already had many facets and subsets – including home telemonitoring and management of chronically ill patient groups.

When the COVID epidemic hit, contagion was a risk for face to face visits for patient and clinician alike. Literally overnight, remote encounters became a new form of healthcare.

However, Telehealth now requires research on safe outcomes, and validating safe methods, training, processes and tools. Some experts believe that remote and even on-site triage may be at increased risk for system error (Wachter, 2017). Research points to pitfalls including remote nature of the work, subspecialty underdevelopment, incomplete systems and inadequate, poorly designed computerized decision support software (CDSS) (Wheeler, 2015).

Telephone triage is described as a form of symptom risk assessment and triage by phone by clinicians. Essentially, it involves making decisions under conditions or uncertainty and urgency (Patel, 1995). Face-to-face triage can be challenging, and remote clinical decision-making (when one cannot see the patient or even by televisit) is more difficult. Clinicians must gather adequate, relevant information in order to estimate symptom urgency.

Telephone triage predates telehealth by 50+ years. Still an informal subspecialty, clinicians perform this task Informally and formally, in ambulatory care settings ranging from physician offices, clinics, student health centers, disease management and ambulatory surgery, to Urgent Care, emergency department (ED) and Labor and Delivery settings, Remote triage is ubiquitous.

Soon, the broad, high-tech field and industry of telehealth (video visits, biotelemetry, patient wearables) will subsume telephone triage – a technologically limited field based on outdated technology.

Both telehealth and telephone triage are remote encounters, but differ in several ways. Although evolving into possibly emergent encounters, telehealth is typically a pre-scheduled encounter about a non-urgent matter. Whereas, telephone triage calls are typically about acute, time-sensitive symptoms. Calls are brief (2-15 minutes), unscheduled, initiated by patients seeking clinicians’ help to decide symptom acuity. Patient want clinicians to evaluate and instruct them about when,where and why they need to be seen - in the ED, Urgent care, Office or simply Home Treatment.

The next generation of telephone triage will require a wide range of high-tech features (video, biotelemetry, Predictive analytics, AI and patient wearables). Advanced technology will transform remote encounters into data driven virtual visits.

What Lies Ahead?

Currently, telephone triage rightfully qualifies as a form of “pre-hospital care“ — and requires serious attention. As an emerging nursing subspecialty and initial entry point into the continuum of care, telephone triage requires “guard rails” — a system. Traditionally, a system is composed of structure and process. Typically, systems have standard components - clinical training, standards, qualified, experienced staff, documentation and Guidelines — EMR and CDSS.

Software developers can take steps to reduce "patterns of recurrent mistakes" (Gawande) by developing systems to reduce recurrent error. The foundation — the Duty of Due Care — requires both clinicians and employers to “do what a reasonable, prudent clinician or employer would do under the same or similar conditions”.

If remote triage is ever to be regarded and embraced as an established professional clinical subspecialty, it must meaningfully demonstrate its allegiance to healthcare traditions and values. Those values include accessibility to care, serving broad patient populations and patient safety.

If any CDSS is to be accepted, trusted and quickly adapted --by nurses, physicians, and healthcare administrators, then developers must meaningfully demonstrate two things: That both the system and decision support tool are valid and reliable. Safety is at the heart of remote triage – uncertain and urgent.

TIME-SENSITIVE CLINICAL DECISION-MAKING

Triage has many hallmarks of an emerging nursing subspecialty. It requires repeated time-driven clinical decision making under conditions of urgency and uncertainty. National Institute of Health research indicates that triage error is best mitigated by  Planned Error Recovery. Clinicians often work in a vacuum, but this critical work requires learning from mistakes and working to correct them and that requires feedback about clinical outcomes.

Ms. Wheeler’s system is informed by research, and serving as an expert witness on 35 triage malpractice cases. Ms Wheeler incorporates shared experience, rules of thumb and wisdom of nurse-participants from her clinical training program “The Fine Art of Telephone Triage”. 

Standard triage system components are complete and integrated, as well as designed to reduce recurrent triage error.

  • SYSTEM COMPONENTS: CLINICAL TRAINING - GUIDELINES - EMR ENHANCEMENTS - STANDARDS  - QA AUDIT

  • TRIAGE PROCESS - DECISION SUPPORT TOOLS -  ACRONYMS - RULES OF THUMB - RED FLAGS - RED HERRINGS

  • HEURISTICS, PATTERN RECOGNITION - ASKING THE RIGHT QUESTIONS - HOW URGENT ARE THESE SYMPTOMS? & HOW HIGH RISK IS THE PATIENT?

AI and Remote Triage: Assumptions & Questions

Results of recent Question to Google AI:

Which task is more difficult for AI— remote symptom diagnosis or remote estimation of symptom urgency (Triage)?

____________________________________

AI Overview

“Based on current research, remote estimation of symptom urgency (triage) is generally considered a more difficult task for AI than remote symptom diagnosis because it requires not only identifying potential medical conditions from symptoms but also accurately assessing the severity and time-sensitive nature of those conditions to determine the appropriate level of care needed. 

Why triage is harder:

·       Complexity of urgency assessment: Determining whether a symptom requires immediate medical attention, urgent care, or can be managed at home is a nuanced decision that often involves considering multiple factors, like patient history, context, and potential complications, which can be challenging for AI to fully capture. 

·       High stakes decision making: A wrong triage decision can lead to delayed or unnecessary medical intervention, potentially causing harm to the patient. 

·       Variability in symptom presentation: Similar symptoms can manifest differently across individuals, making it difficult for AI to consistently assess the urgency of a situation. 

Why diagnosis might be easier:

·       Pattern recognitionL: AI can excel at identifying patterns in symptom data to suggest potential diagnoses based on established medical knowledge. 

·       Large datasets: Access to extensive medical databases allows AI to learn common symptom-disease associations. 

However, it's important to note that:

·       Advancements in AI: Research is actively developing AI models specifically designed for triage, demonstrating potential for improved accuracy in assessing symptom urgency. 

·       Data quality matters: Both diagnosis and triage accuracy heavily rely on the quality and completeness of the data used to train the AI model.” 

Generative AI is experimental.

__________________________________________

Thoughts and Observations

The above AI responses indicate that clinicians will always be essential, and that the best outcomes might result from a Human-Machine Hybrid in remote triage.

·      Quality in AI-Supported CDSS for triage is dependent upon adequate amounts and quality of data (input)

·      In the AI era, what is the Duty of Due Care requirement for remote triage? Do software developers have a similar duty of Due Care for AI-CDSS?

Is technology outpacing professional standards in remote triage?

In remote triage, are emergent symptoms the low-hanging fruit? Is this true: “Triage, defined as making decisions as to whether to send (EMD) resources or not, requires a professional judgment, since ruling out urgency is more difficult than identifying it” Champagne F,(1991).?

·       If “First do no harm” is a medical imperative, then given the time-sensitive nature of remote triage, would “When in doubt, err on the side of caution” be analogous? Then, is it reasonable and prudent for institutions to provide an after hours “venue of caution” (instead of the ED - the overcrowded “venue of last resort”) to which to send patients with borderline symptoms?

Access to Care, Delay in Care, Timely Care in Remote Triage

“Standards for access to care are hotly debated for scheduled surgical procedures using delays of days, weeks, and months to determine what is acceptable or reasonable. No such time objectives exist in Emergency Departments (or for telephone triage)  where delays of minutes or hours for unrecognized problems can be the difference between life and death. Without using a standard measure such as a national triage scale it will be difficult to measure acuity, perform case mix comparisons, or develop ED operational standards.” Adapted from American College of Emergency Physicians (1999)

https://www.acep.org/siteassets/uploads/uploaded-files/acep/clinical-and-practice-management/resources/administration/triagescaleip.pdf

Some research exploring the above issues would be useful.

Rules of Thumb for Triage Clinical Decisionmaking

Research in clinical decision-making has identified heuristics (rules of thumb) as shortcuts. Rules of Thumb used for clinical decisionmaking for triage may apply to remote triage as well as face-to-face triage.

·       When in doubt, always err on the side of caution, and bring the patient in sooner rather than later.

·       “When in doubt, send ‘em out”.  (Clawson, 1998).

·       Always speak directly with patients; A patient “too sick to talk” is a “Red Flag” and should be evaluated on site

·       Beware of the middle-of-the-night call

·       Beware the “Recent or Previous Diagnosis”

·       Remain alert for atypical, silent or novel presentations.

·       Beware of “failure to improve” on current Rx or Home Treatment.

·       “Never abandon the caller in crisis” (Clawson, 1998).                                          

·       The vaguer the symptoms, the greater need for good data collection.

·       Speed does not equal competence; avoid premature closure or jumping to conclusions.

·       Beware the “Non-Diagnostic Diagnosis” (patient interpretation of symptoms)

·       Beware the “Out of Protocol” experience.

·       Two or more calls within hours or days should be seen urgently (1-4 hours)

·       Patients described as “too sick to talk on the phone” may require an urgent appointment

·       All severe symptoms must be seen in 1-8 hours or less.

Age-Based Rules of Thumb

·       Assess all sick children and  elderly for possible dehydration and sepsis

·       All new or worsening confusion in children or elderly is considered emergent.

·       Elderly and pediatric patients may present in silent, atypical or novel ways

·       Elderly and pediatric populations are at greater the risk of hypo- or hyperthermia during weather extremes.  

·       All Teens (including College Student/Young Adult)  are at increased risk for suicide/violence when stressed or depressed due to impulsivity

·       Elderly are at higher risk for completed suicide, especially males over 65 years. who are white, widowed, retired and unemployed.

Trauma-Based Rules of Thumb

·       Trauma + Suspicious History: Consider Possible Domestic Abuse

·       Always consider head and neck injury when there is face or jaw trauma

·       All snakes are considered poisonous until proven otherwise.

·       All chemical and electrical burns have the potential for progressive injury

·       Never remove impaled objects.

Symptom-Based Rules of Thumb

·       All severe pain must be seen within 8 hours or less.

·       Beware of any pain that awakens patient or prevents sleep at night

·       Headaches described as: “First, Worst, Burst, Cursed (other accompanying symptoms), or 51st (age over 50 years)” = Emergent

·       “Temperature extremes often trigger medical problems”. (Clawson, 1998)

·        “Epigastric pain in males > 35 and females > 45, is considered an MI until proven otherwise”.  (Clawson, 1998).

·       Any pain between the navel and nose is chest pain until proven otherwise

·       The first symptom of an MI may be denial.

·       Once an ectopic, always an ectopic

·       Any bleeding in pregnancy is an ectopic until proven otherwise

·       All first-time seizures must be evaluated .

·       All rashes are contagious until proven otherwise

·       Beware of any atypical, silent or novel presentations.

·       Beware all  “flu” symptoms that can “mask” MI, Infection  or sepsis.

·       Pregnancy and breast feeding may be risk factors for domestic violence

  • All High-Risk Patients with moderate symptoms of any kind may require higher triage acuity.

Rule of Thumb for Myocardial Infarct Risks: “The Eight E’s”

Extremes of:

  • Emotion

  • Extremes of Weather/Temperature

  • Exertion

  • Extreme Age:  >75

  • Eating (“Holiday Heart” ?)

  • Epigastric Distress

  • Essential Hypertension

  • Early AM

© Sheila Quilter Wheeler, 2025

References for Rules of Thumb  

1.     John Saunders, The practice of clinical medicine as an art and as a science West J Med. 2001 Feb;174(2):137–141 1

2.     Malin André 1, L Borgquist, M Foldevi, S Mölstad  Asking for 'rules of thumb': a way to discover tacit knowledge in general practice. Fam Pract. 2002 Dec;19(6):617-22. doi: 10.1093/fampra/19.6.617.

 3.     Schmidt HG, Norman GR, Boshuizen HP. A cognitive perspective on medical expertise: Theory and implications. Acad Med 1990; 65: 611–621

4.     Gigerenzer G, Todd P and the ABC group. Simple heuristics that make us smart. Oxford: Oxford University Press, 1999: 3–34.

5.     O’Neill E. Heuristic reasoning in diagnostic judgement. J Prof Nurs 1995; 11: 239–245.

 6.     Essex Ben. Doctors, dilemmas decisions. London: BMJ Publishing Group, 1994.

7.     Morgan D, Kreuger R. The Focus Group Kit, Vol 1. London: Sage Publications, 1998: 9–15.

8.     Morgan D, Kreuger R. The Focus Group Kit, Vol 2. London: Sage Publications, 1998: 55–71.

9.     Crabtree B, Miller W. Doing qualitative research. Newbury Park: Sage Publications, 1992: 13–21.

10.  McWhinney I. A textbook of family medicine. Oxford: Oxford University Press, 1997: 123–155.

12.  Wright GH. Norm and action. A logical enquiry. London: Routledge and K Paul, 1963: 1–16.

13.  Malterud K. The issue of validity regarding the contents of consultations. Scand J Prim Health Care 1993; 11 Suppl. 2: 64–67.

 14.  van Leeuwen YD, Mol SSL, Pollemans MC, Drop MJ, Grol R, van der Vleuten CP. Change in knowledge of general practitioners during their professional careers. Fam Pract 1995; 12: 313–317.

15.  Westin S. Problem solving styles of general practitioners in simulated clinical situations. Fam Pract 1984; 1: 92–99.

16.  Greer AL. The state of the art versus the state of the science. I J Tech Ass Health Care 1988; 4: 5–26.

17.  Soumerai S, Avorn J. Principles of educational outreach (‘Academic detailing’) to improve clinical decision making. JAMA 1990; 263: 549–556