Characteristics of Sources of Evidence

The flow of selection of publications for this review is outlined in Figure 1. Of the 53 full texts evaluated for eligibility, 31 were excluded as they did not report relevant quantitative data on TAFA or dealt exclusively with telecommunicator-guided assessment of the patient’s condition. The remaining 22 articles were included in the analysis. ^{Reference Lee, Mao and Ng4,Reference Billittier, Lerner, Tucker and Lee14–Reference Idland, Kramer-Johansen and Bakke34}

Figure 1.

Study Selection Diagram.

Key characteristics of particular articles, including authors, years of publication, countries of origin, study aims, addressed health conditions, designs, methods, main findings, conclusions, and recommendations related to TAFA, are available in the online repository. ^{Reference Birkun and Douma13}

The articles were published from 2000 through 2024. Based on the first author’s primary affiliation, the reported studies originated from nine high-income countries, including the United States (n = 10; 45.5%), ^{Reference Billittier, Lerner, Tucker and Lee14,Reference Evenson, Brice, Rosamond, Lellis, Christian and Morris18,Reference Beauchamp, Phrampus and Guyette19,Reference Caceres, Adil and Jadhav22,Reference Goldberg, Dworkis and Liao25–Reference Li, Hu and Gregg29,Reference Davidson, Correll and Gottula33} Norway (n = 3; 13.6%), ^{Reference Melbye, Hotvedt and Bolle23,Reference Bakke, Steinvik, Ruud and Wisborg24,Reference Idland, Kramer-Johansen and Bakke34} France (n = 2; 9.1%), ^{Reference Labarère, Torres and Francois16,Reference Fourny, Lucas and Belle20} the United Kingdom (n = 2; 9.1%), ^{Reference O’Cathain, Turner and Nicholl15,Reference Barron, Clawson and Scott21} Australia (n = 1; 4.5%), ^{Reference Lee, Mao and Ng4} Germany (n = 1; 4.5%), ^{Reference Auerbach, Schreyögg and Busse17} Japan (n = 1; 4.5%), ^{Reference Igarashi, Suzuki and Norii32} Spain (n = 1; 4.5%), ^{Reference Aranda-García, Santos-Folgar and Fernández-Méndez30} and Switzerland (n = 1; 4.5%). ^{Reference Dami, Enggist, Comte and Pasquier31}

Research designs included experimental randomized controlled (n = 4), ^{Reference Scott, Olola and Gardett28,Reference Aranda-García, Santos-Folgar and Fernández-Méndez30,Reference Igarashi, Suzuki and Norii32,Reference Davidson, Correll and Gottula33} non-randomized controlled (n = 1), ^{Reference Melbye, Hotvedt and Bolle23} and non-controlled (n = 1) ^{Reference Ornato, You and McDiarmid27} descriptive cross-sectional (n = 1), ^{Reference Lee, Mao and Ng4} observational cross-sectional (n = 4), ^{Reference Billittier, Lerner, Tucker and Lee14–Reference Labarère, Torres and Francois16,Reference Evenson, Brice, Rosamond, Lellis, Christian and Morris18} prospective observational (n = 5), ^{Reference Beauchamp, Phrampus and Guyette19,Reference Fourny, Lucas and Belle20,Reference Bakke, Steinvik, Ruud and Wisborg24,Reference Goldberg, Dworkis and Liao25,Reference Idland, Kramer-Johansen and Bakke34} retrospective observational (n = 5), ^{Reference Auerbach, Schreyögg and Busse17,Reference Barron, Clawson and Scott21,Reference Caceres, Adil and Jadhav22,Reference Grisanti, Martorano and Redmond26,Reference Li, Hu and Gregg29} and retrospective descriptive (n = 1) studies. ^{Reference Dami, Enggist, Comte and Pasquier31} Five studies were cohort (two prospective ^{Reference Fourny, Lucas and Belle20,Reference Idland, Kramer-Johansen and Bakke34} and three retrospective ^{Reference Barron, Clawson and Scott21,Reference Caceres, Adil and Jadhav22,Reference Dami, Enggist, Comte and Pasquier31} ) studies.

The studies investigated hypothetical or actual application of TAFA for the following health conditions: anaphylaxis (n = 3), ^{Reference Grisanti, Martorano and Redmond26,Reference Dami, Enggist, Comte and Pasquier31,Reference Davidson, Correll and Gottula33} chest pain or myocardial infarction (n = 3), ^{Reference Evenson, Brice, Rosamond, Lellis, Christian and Morris18,Reference Fourny, Lucas and Belle20,Reference Barron, Clawson and Scott21} stroke (n = 3), ^{Reference Evenson, Brice, Rosamond, Lellis, Christian and Morris18,Reference Caceres, Adil and Jadhav22,Reference Li, Hu and Gregg29} trauma (n = 3), ^{Reference Auerbach, Schreyögg and Busse17,Reference Bakke, Steinvik, Ruud and Wisborg24,Reference Idland, Kramer-Johansen and Bakke34} bleeding (n = 2), ^{Reference Billittier, Lerner, Tucker and Lee14,Reference Scott, Olola and Gardett28} childbirth (n = 2), ^{Reference Billittier, Lerner, Tucker and Lee14,Reference Aranda-García, Santos-Folgar and Fernández-Méndez30} foreign body airway obstruction (n = 2), ^{Reference Billittier, Lerner, Tucker and Lee14,Reference Igarashi, Suzuki and Norii32} opioid overdose (n = 2), ^{Reference Goldberg, Dworkis and Liao25,Reference Ornato, You and McDiarmid27} and respiratory arrest (n = 2). ^{Reference Billittier, Lerner, Tucker and Lee14,Reference Beauchamp, Phrampus and Guyette19} The type of addressed health emergencies was not specified in four publications. ^{Reference Lee, Mao and Ng4,Reference O’Cathain, Turner and Nicholl15,Reference Labarère, Torres and Francois16,Reference Melbye, Hotvedt and Bolle23}

Studies of Real-World EMS Call Data

Eight studies (36.4%) ^{Reference Fourny, Lucas and Belle20–Reference Caceres, Adil and Jadhav22,Reference Bakke, Steinvik, Ruud and Wisborg24,Reference Grisanti, Martorano and Redmond26,Reference Li, Hu and Gregg29,Reference Dami, Enggist, Comte and Pasquier31,Reference Idland, Kramer-Johansen and Bakke34} involved analyses of real EMS call data with sample sizes ranging from 113 to 52,282 (median [Q1, Q3]: 374.0 [170.8, 37,377.8]). In particular, Fourny, et al (2011) ^{Reference Fourny, Lucas and Belle20} reported the provision of medical advice for patients with myocardial infarction (instead of immediately dispatching a mobile intensive care unit) as a type of inappropriate initial dispatcher’s decision that, combined with other inappropriate decisions, may result in substantial delays in time to reperfusion therapy. Barron and coauthors (2013) ^{Reference Barron, Clawson and Scott21} determined rates of completion of emergency medical dispatch (EMD) protocol for pre-arrival aspirin administration to treat suspected acute coronary syndrome and revealed a substantial number of cases where the protocol was not used by dispatchers, though the patients met the necessary criteria. A study by Caceres, et al (2013) ^{Reference Caceres, Adil and Jadhav22} showed that recognition of stroke by emergency dispatchers significantly increased the likelihood of the provision of telephone guidance to callers. Further, Li and colleagues (2021), ^{Reference Li, Hu and Gregg29} along with reporting rates of the provision of PAI for suspected stroke events, confirmed that the use of instructions was significantly and negatively (inversely) associated with EMS delay. Bakke and coauthors (2017) ^{Reference Bakke, Steinvik, Ruud and Wisborg24} presented rates of emergency dispatch telephone guidance in trauma, revealed low sensitivity and specificity of dispatchers for correctly identifying trauma patients requiring first-aid measures, and found that dispatcher advice did not significantly influence whether first aid was performed correctly or attempted at all, though in regions with a high spontaneous first aid rate. According to Idland, et al (2024), ^{Reference Idland, Kramer-Johansen and Bakke34} video streaming by dispatchers in calls regarding injured patients, although infrequently used, improved dispatcher recognition of the need for first-aid measures but did not influence the quality of first aid compared with audio-only calls. Grisanti, et al (2019) ^{Reference Grisanti, Martorano and Redmond26} demonstrated suboptimal adherence to EMD protocols for allergies by EMS dispatchers, who infrequently asked about epinephrine availability or instructed callers with suspected anaphylaxis and available epinephrine to use it. Subsequent study of suspected anaphylaxis cases by Dami, et al (2022) ^{Reference Dami, Enggist, Comte and Pasquier31} reported rates of accepted and declined dispatcher advice on using epinephrine autoinjector and identified reasons for refusing the use of the drug. Based on the research results, the authors recommended to further investigate a relationship between TAFA and bystander attempts to provide first aid, ^{Reference Bakke, Steinvik, Ruud and Wisborg24} improve accuracy of telephone triage to enhance dispatcher identification of the need for TAFA, ^{Reference Fourny, Lucas and Belle20,Reference Bakke, Steinvik, Ruud and Wisborg24} conduct studies to understand reasons of why dispatchers are not using an EMD protocol, ^{Reference Barron, Clawson and Scott21} make efforts to increase protocol adherence, ^{Reference Grisanti, Martorano and Redmond26} perform protocol improvement, ^{Reference Grisanti, Martorano and Redmond26} enhance public access to aspirin ^{Reference Barron, Clawson and Scott21} and epinephrine, ^{Reference Dami, Enggist, Comte and Pasquier31} implement related public education campaigns, ^{Reference Barron, Clawson and Scott21} provide training for dispatchers, ^{Reference Caceres, Adil and Jadhav22,Reference Grisanti, Martorano and Redmond26,Reference Li, Hu and Gregg29,Reference Dami, Enggist, Comte and Pasquier31} and carry out TAFA performance oversight. ^{Reference Caceres, Adil and Jadhav22,Reference Grisanti, Martorano and Redmond26}

Simulation-Based Experimental Studies

Another eight studies (36.4%) ^{Reference Beauchamp, Phrampus and Guyette19,Reference Melbye, Hotvedt and Bolle23,Reference Goldberg, Dworkis and Liao25,Reference Ornato, You and McDiarmid27,Reference Scott, Olola and Gardett28,Reference Aranda-García, Santos-Folgar and Fernández-Méndez30,Reference Igarashi, Suzuki and Norii32,Reference Davidson, Correll and Gottula33} generated research data in simulated settings. Sample size varied from 12 to 246 (median [Q1, Q3]: 31.5 [21.0, 80.0]). Beauchamp, et al (2009) ^{Reference Beauchamp, Phrampus and Guyette19} demonstrated a high success rate of laypersons in telephone-directed placement of a laryngeal tube during simulated respiratory arrest. Melbye and colleagues (2014) ^{Reference Melbye, Hotvedt and Bolle23} tested mobile phone video conferencing between EMD and bystanders who performed various first-aid tasks according to the provided instructions in different surroundings. The study group concluded that video conferencing can be used during medical emergencies under suboptimal sound and light conditions. According to a study by Goldberg, et al (2018), ^{Reference Goldberg, Dworkis and Liao25} bystanders were able to access pre-stationed intranasal naloxone and administer it in a shorter time than community EMS response times. Later, Ornato and collaborators (2020) ^{Reference Ornato, You and McDiarmid27} showed successful dispatcher-directed bystander administration of naloxone nasal spray delivered by drone to opioid overdose victims. An article by Scott and coauthors (2020) ^{Reference Scott, Olola and Gardett28} reported a randomized controlled study where laypeople provided with telephone instructions from a trained EMD applied a tourniquet and stopped simulated bleeding completely in most cases. A study by Aranda-García, et al (2022) ^{Reference Aranda-García, Santos-Folgar and Fernández-Méndez30} demonstrated the feasibility of remote video assistance to untrained bystanders equipped with smart glasses in a scenario of unplanned out-of-hospital birth. Igarashi, et al (2022) ^{Reference Igarashi, Suzuki and Norii32} confirmed the superior quality of first aid for choking infants with dispatcher advice through video calls compared to voice-only calls. In research by Davidson, et al (2024), ^{Reference Davidson, Correll and Gottula33} bystanders guided by drone-specific or standard EMD instructions successfully interacted with the drone and administered the required interventions for pediatric anaphylaxis, with better adherence to drone safety cues in the group that received the drone-specific guidance. The publications recommended further studies in more realistic scenarios ^{Reference Beauchamp, Phrampus and Guyette19,Reference Davidson, Correll and Gottula33} or clinical settings ^{Reference Goldberg, Dworkis and Liao25,Reference Ornato, You and McDiarmid27} and advised considering relevant ethical aspects of using video communication technologies for TAFA. ^{Reference Aranda-García, Santos-Folgar and Fernández-Méndez30}

Surveys

Five studies (22.7%) ^{Reference Lee, Mao and Ng4,Reference Billittier, Lerner, Tucker and Lee14–Reference Labarère, Torres and Francois16,Reference Evenson, Brice, Rosamond, Lellis, Christian and Morris18} collected data through surveys, which encompassed from nine to 652 participants (median [Q1, Q3]: 409.0 [47.0, 588.0]). Billittier and colleagues (2000) ^{Reference Billittier, Lerner, Tucker and Lee14} evaluated the lay public’s (n = 524) expectations of PAI when dialing 9-1-1 and found that most respondents expected to receive the instructions for health emergencies regardless of the current availability of such service. O’Cathain, et al (2002) ^{Reference O’Cathain, Turner and Nicholl15} determined the acceptability of TAFA to ambulance callers before (n = 355) and after (n = 297) the introduction of the Advanced Medical Priority Dispatch System. The study showed the effect of an increase in the amount of first aid advice given to callers and caller satisfaction with the service. Positive public rating of telephone guidance provided by EMS call centers and 61% patient compliance with the self-treatment advice given were reported in an article by Labarère and coauthors (2003). ^{Reference Labarère, Torres and Francois16} A survey of 9-1-1 communication centers carried out by Evenson, et al (2007) ^{Reference Evenson, Brice, Rosamond, Lellis, Christian and Morris18} examined personnel training, certification, and PAI provision rates for stroke and myocardial infarction. As part of a survey of EMD services across Asian countries published by Lee, et al (2020), ^{Reference Lee, Mao and Ng4} seven of nine participating countries confirmed the practice of provision of non-resuscitative PAI. The research-synthesized recommendations included future investigations of callers’ compliance, willingness, and ability to follow PAI, ^{Reference Billittier, Lerner, Tucker and Lee14,Reference Labarère, Torres and Francois16} examination of TAFA economic impact and its actual effects on patient outcomes, ^{Reference Billittier, Lerner, Tucker and Lee14} and organization of training and certification for call center personnel. ^{Reference Labarère, Torres and Francois16,Reference Evenson, Brice, Rosamond, Lellis, Christian and Morris18}

Economic Evaluation Research

One study (4.5%) published by Auerbach, et al (2006) ^{Reference Auerbach, Schreyögg and Busse17} represented a cost-effectiveness analysis of two equipment versions of telemedical devices for traffic accident emergency rescue, one of which included a mobile device with a touch-screen display, audio, and video connection for instructing and supporting laypersons in performing first aid. The study found that the application of the telemedical device may considerably reduce the number of traffic-related fatalities, but its application would be associated with a high cost-effectiveness ratio. To facilitate the implementation of the technology, the authors recommended building a coordinated European initiative among European member states.