Development of an emergency medical system model for resource‐constrained settings

Abstract

en

Objectives

An emergency care system is an important aspect for healthcare organisations in low‐ and middle‐income countries (LMICs) with a growing burden from emergency disease conditions. Evaluations of emergency care systems in LMICs in broader contexts are lacking. Thus, this study aimed to develop a comprehensive emergency medical system model appropriate for resource‐constrained settings, based on expert opinions.

Methods

We used the Delphi method, in which questionnaire surveys were administered three times to an expert panel (both emergency medical care providers and healthcare service researchers), from which opinions on the model's components were compiled. The panel members were mostly from Asian countries. In the first round, the questionnaire drew a list of model components developed through a literature review; the panel members then proposed new components to create a more comprehensive list. In the second and third rounds, the panel members rated the listed components to achieve consensus, as well as to remove components with low ratings. Finally, we rearranged the list to improve its usability.

Results

In total, 32 experts from 12 countries participated. The final model totalled 177 components, categorised into 8 domains (leadership, community‐based actions, emergency medical services, upward referral, definitive care, rehabilitation, downward referral, and evaluation and research). No components needed removal.

Conclusions

We developed a comprehensive emergency care system model, which could provide a basis to evaluate emergency care systems in resource‐constrained LMICs; however, field‐testing and validation of this system model remain to be done.

Abstract

fr

Objectifs

Un système de soins d'urgence est un aspect important pour les organisations de soins de santé dans les pays à revenu faible ou intermédiaire (PRFI), qui subissent une charge croissante d'affections de maladie d'urgence. Les évaluations des systèmes de soins d'urgence dans les PRFI dans des contextes plus larges font défaut. Cette étude visait donc à élaborer un modèle de système médical d'urgence complet, adapté aux contextes à ressources limitées, sur la base d'opinions d'experts.

Méthodes

Nous avons utilisé la méthode Delphi, dans laquelle des enquêtes par questionnaire avaient été administrées à trois reprises à un groupe d'experts (prestataires de soins médicaux d'urgence et chercheurs en services de santé), à partir desquelles des avis sur les composants du modèle ont été compilés. Les membres du panel venaient principalement de pays asiatiques. Lors du premier tour, le questionnaire a dressé une liste des composants du modèle développés à travers une revue de la littérature; les membres du panel ont ensuite proposé de nouveaux composants pour créer une liste plus complète. Lors des deuxième et troisième tours, les membres du panel ont évalué les composants énumérés afin de parvenir à un consensus et d’éliminer les composants faiblement cotés. Enfin, nous avons réorganisé la liste pour en améliorer la convivialité.

Résultats

Au total, 32 experts de 12 pays ont participé. Le modèle final comprenait 177 composants, classés en 8 domaines (leadership, actions communautaires, services médicaux d'urgence, référence ascendante, soins définitifs, suivi, référence postérieure, évaluation et recherche). Aucun composant ne nécessitait d’être exclus.

Conclusions

Nous avons élaboré un modèle complet de système de soins d'urgence qui pourrait servir de base à l’évaluation des systèmes de soins d'urgence dans les PRFI aux ressources limitées. Cependant, les essais sur le terrain et la validation de ce modèle de système restent à faire.

Introduction

An emergency medical care system, including emergency medical services (EMS), is an important element of healthcare organisations in low‐ and middle‐income countries (LMICs), where there is a greater burden of emergency disease conditions than in high‐income countries (HICs) 1. Emergency conditions can arise from various disease categories, such as maternity and communicable diseases 2-4. Non‐communicable diseases and injuries are increasing in LMICs 5. Despite this burden, emergency care systems in LMICs remain underdeveloped and underused 1. Thus, the development and strengthening of emergency care systems in LMICs is a pressing issue.

Emergency medical system models devised in HICs may not be applicable to resource‐constrained settings in LMICs, particularly in remote areas. Patients in LMICs frequently have to overcome many barriers to access care: lack of a formal ambulance system, extremely distant locations of healthcare facilities and poor transportation systems 6-9. Overcoming these barriers requires different system models, where formal systems are integrated with informal ones; for example, by task shifting to non‐professional personnel (e.g. lay people training as first responders) and community‐based mutual aids (e.g. cost‐sharing and transport arrangements) 6, 10-12.

Previous studies that assessed emergency care systems in LMICs focused mostly on the capacities of service delivery by health facilities or EMS. Their evaluations address physical and human resources, such as equipment, knowledge and skills 13-21. Some attempts have been made to assess emergency care systems in wider contexts, which included leadership, financial issues, political commitment and regulatory mechanisms 22-25; however, none of them were comprehensive enough to cover the entire system. Some covered only EMS systems, while others paid little attention to informal systems or aftercare (e.g. rehabilitation and discharge support). Thus a comprehensive system model for LMICs is needed that covers organisational and financial issues, as well as service delivery, and which integrates informal mechanisms. Such a system model would assist the development and evaluation of emergency care in LMICs.

This study aimed to develop a comprehensive emergency medical system model that is appropriate in resource‐constrained settings by compiling the opinions of experts regarding emergency medical care and healthcare services. Specifically, the objectives of the study were as follows: to obtain expert opinions on what components of an emergency care system are essential for an effective model and should be evaluated in LMICs and to create a structured list of components to be used for system evaluation.

Methods

Study design

This study used the Delphi method, whereby opinions are collected and compiled to reach a consensus through repeated questionnaire surveys administered to an expert panel. The panel consisted of emergency medical care providers and healthcare service researchers, from mostly Asian countries 26. The baseline questionnaire was developed from a literature review, which indicated a list of items that were considered important components of an emergency care system. The feedback collected from the panel members was reflected in the next questionnaire sent out. The surveys were repeated three times: first to create a list of items, second to rate the listed items and third to reach a consensus. The final survey results guided the development of an emergency care system model. The questionnaires were in English, Vietnamese, Laotian and Thai. This study was approved by the ethics committees of Teikyo University in Japan, Hanoi Medical University in Vietnam, the University of Health Sciences in Laos, and Khon Kaen Hospital in Thailand; ethical clearance was waived in the other countries.

Participants

The expert panel included physicians and nurses working in emergency departments, paramedic staff from EMS and researchers or administrators of healthcare service systems. The members were purposively selected from Vietnam, Laos, Thailand, Sri Lanka and Japan; but some were also invited from other countries through a snowball sampling procedure. First, the principal investigator invited an expert from each of the countries and asked them to recommend other potential participants regardless of nationality. These potential participants were contacted and asked to participate. In Vietnam, Laos and Thailand, participants did not need to be fluent in English because we prepared the questionnaire in the local languages, but in the other countries, the participants needed to understand English because the snowball sampling did not allow us to determine the countries of experts' origin before starting the surveys. Some of the participants from Malaysia, the Philippines, Taiwan and Sri Lanka were invited from the second‐round survey due to delays in ethical clearance or in the sampling process.

For the first‐round survey, we invited 42 experts from 11 countries and received 27 responses from 8 countries (Figure 1). For the second‐round survey, we invited 27 first‐round respondents, of these 26 responded but one did not; we invited an additional 11 experts from 7 countries, of these five participants from four countries responded, who were new at the second round. In the third‐round survey, we invited the 31 second‐round participants but did not invite new participants; of them, 26 responded. Our analysis included the 31 second‐round participants where we analysed the 26 third‐round responses and the 5 second‐round responses for those who did not participate in the final round.

Figure 1

Participant selection for the Delphi survey. ^aCambodia, Canada, China, Japan, Korea, Laos, Mongolia, Thailand, Turkey, United Kingdom, Vietnam. ^bCambodia, Canada, Japan, Laos, Mongolia, Thailand, Turkey, Vietnam. ^cRespondents in the previous rounds were invited. ^dChina, India, Indonesia, Malaysia, the Philippines, Sri Lanka, Taiwan. ^eMalaysia, the Philippines, Sri Lanka, Taiwan. ^fCambodia, Canada, Japan, Laos, Malaysia, Mongolia, the Philippines, Sri Lanka, Taiwan, Thailand, Turkey, Vietnam.

Survey tool development

The principal investigator, with the study group members, conceived a baseline conceptual model and an initial list of system components through a literature review. The review prioritized assessment or evaluation of emergency care systems in order to measure their effectiveness, which facility‐level indicators alone cannot achieve 27. A preliminary framework on access to emergency care, which our previous study proposed 6, led the review to cover several issues. These included formal and informal pre‐hospital care (e.g. lay first responder and informal transport planning) 8, 10-12, 25, 28, 29, community‐based activities improving access to healthcare (e.g. community health workers and micro health insurance) 30, 31, and referral systems (e.g. communication and transport arrangement between facilities) 12, 32-34. We did not disregard the evaluation of care quality, the capacity of healthcare facilities or pre‐hospital care; instead, we integrated facility‐level indicators (mainly structure and process indicators) into our final model 14-18, 21, 35.

A review of key literature added components to the preliminary framework. A component of leadership (i.e. organising, regulating and policy making mechanisms) was introduced from healthcare system models and evaluation tools for trauma care 22, 23, surgical care 24, EMS 15, 25, and primary healthcare 36, 37. We also included components of transport arrangements, rehabilitation, aftercare and downward referral 9, 22, 28, 38.

From the literature review, we developed a baseline conceptual model that guided the creation of a list of essential items that constitute emergency care systems, particularly in LMICs. The model consists of 7 categories of components (referred to as domains hereafter): leadership, community‐based activities, EMS (formal pre‐hospital care), upward referral system (between hospitals), definitive care at tertiary care level, after care and downward referral (back to community) and evaluation and research. We then included them into a framework of 4 levels: policy level, hospital level, inter‐facility referral and community level (formal EMS and community‐based actions).

Data collection

The repeated questionnaire surveys ran from August 2017 to March 2018. The expert panel members were contacted via email and received the questionnaire as an Excel file. In Vietnam, Laos and Thailand, either the English or local language version was used; in other countries, only the English version was used.

The first round aimed to increase the number of model items on the baseline list developed through the literature review to make it more comprehensive. The panel members expressed their opinions on the list and proposed new items, sub‐domains or domains to be added. The first‐round results guided the creation of the second‐round questionnaire. New items were added to the list, and some of the baseline items were divided: for example, ‘lead agency’ was divided into ‘lead agency for emergency medical care’ and ‘lead agency for disaster management’. We also re‐categorised some items and created new sub‐domains as needed.

The second round aimed to obtain ratings on the listed items. The panel members rated the importance of each item on a 5‐point Likert scale from 1 (unimportant) to 5 (very important), disclosed additional opinions, if any, and indicated new elements they thought should be added. When rating the items, the participants skipped the ones with which they were not familiar.

The third round aimed to reach a consensus by reducing variability in the ratings. The third‐round questionnaire included the same list as the second round with no new items added, as well as a summary of panel members' ratings in the second round. The panel members compared their own ratings with the summary results and modified the ratings if necessary. This final procedure in the Delphi survey aimed to narrow the gap between individual ratings and the summary results.

Analysis

Based on the ratings, we created a list of items that are important to the emergency care system model. We calculated the proportions of responses rated 4 (important) or 5 (very important) for each of the items. For the sub‐domains, we indicated the median of the responses of the constituent items. We planned to delete items that had less than 20% of our participants rating them 4 (important) or 5 (very important); such a restrained deletion plan was taken to make the list comprehensive and inclusive. Items and sub‐domains that had less than 60% of those rating them 4 and 5 were made optional, because we considered them irrelevant in some settings. We did not use statistical tests because we did not compare groups or examine associations in this study. Instead, we described distributions of the ratings using the proportions of high ratings and median scores.

The first author, with the advice of the study group members, modified the list in several ways, aiming to improve its usability. First, certain phrasing and terms were simplified, and explanations were added as needed, because we had to clarify descriptions of some items during the surveys in response to questions from panel members. Second, we combined items with the same meanings and divided some items that addressed multiple aspects. Finally, we re‐categorised the items, moved some items to another domain, and combined or divided domains and sub‐domains as needed.

Development of an evaluation tool

We made a short list with the aim of using it for evaluation activities because the final model list, which is comprehensive and inclusive, is not easy to use for actual evaluation. We selected highly rated items that had 90% or more of our participants rating them 4 (important) or 5 (very important).

Results

Participants

Thirty‐two participants responded at least once in the three‐round surveys, 22 of whom responded to all three rounds (Table 1). Most of the participants were male medical doctors aged 35–54 years working in public tertiary or secondary hospitals.

Table 1. Participant characteristics (n = 32)

	n
Age, years
25–34	4
35–44	13
45–54	12
55–64	3
Sex
Female	7
Male	25
Specialty
Medicine	25
Nursing/midwifery	2
Pre‐hospital care	4
Other	1
Facility
National hospital	9
Provincial/district hospital	9
Private hospital/clinic	5
Pre‐hospital care	4
University/research institute	2
Other	3
Participation
First round** Of the 27 participants, one did not respond to the subsequent rounds.	27
Second round†† Of the 31 participants, five participated from the second round; two responses were not included in the second‐round analysis because of a delayed response.	31
Third round‡‡ Five of the 31 second‐round participants did not respond to the third round. Thus, the analysis included 26 third‐round responses and 5 second‐round responses.	26
Country§§ The countries listed are where the participants were based at the time of the survey rather than their nationalities.
Vietnam	6
Laos	5
Japan	4
Mongolia	4
Thailand	3
Cambodia	2
Canada	2
Sri Lanka	2
Malaysia	1
Philippines	1
Taiwan	1
Turkey	1

• * Of the 27 participants, one did not respond to the subsequent rounds.
• † Of the 31 participants, five participated from the second round; two responses were not included in the second‐round analysis because of a delayed response.
• ‡ Five of the 31 second‐round participants did not respond to the third round. Thus, the analysis included 26 third‐round responses and 5 second‐round responses.
• § The countries listed are where the participants were based at the time of the survey rather than their nationalities.

Survey results

In the first round, the questionnaire included a list of 118 items for evaluation (baseline model). The items were categorised into 34 sub‐domains and seven domains. Based on the participants' comments on the list, we added 58 items and two sub‐domains (‘policy/planning’ under the domain of leadership and ‘lead agency for EMS’ under the domain of EMS) to the list. The second‐ and third‐round lists (revised model) included seven domains and 36 sub‐domains with 176 items (Table 2).

Table 2. Structures of baseline (first round), revised (second and third rounds) and final model lists

Baseline model			Revised model (Second and third rounds)			Final model
Code** An alphabet alone indicates a domain and an alphabet with a number indicates a sub‐domain.	Domain/sub‐domain	N †† Number of items in each sub‐domain.	Code** An alphabet alone indicates a domain and an alphabet with a number indicates a sub‐domain.	Domain/sub‐domain	N †† Number of items in each sub‐domain.	Code** An alphabet alone indicates a domain and an alphabet with a number indicates a sub‐domain.	Domain/sub‐domain	N †† Number of items in each sub‐domain.
A	Leadership		A	Leadership		A	Leadership (mainly government functions)
A1	Lead agency assigned	2	A1	Lead agency assigned for emergency medical care system	5	A1	Lead agency assigned for emergency medical care system, including EMS	10
A2	Legislation/regulatory mechanisms	6	A2	Legislation/regulatory mechanisms	9	A2	Legislation/regulatory mechanisms	8
A3	Financing to emergency care system	5	A3	Financing to emergency care system	9	A3	Financing for the emergency care system	8
			A4	Policy/planning	5	A4	Policy/planning/guidelines (described based on existing documents)	11
B	Community		B	Community		B	Community‐based (mostly informal) actions
B1	Community‐level cost‐sharing	4	B1	Community‐level leadership/organisation (in rural or resource‐constrained areas)	9	B1	Community‐level leadership/organisation (in rural or resource‐constrained areas) (OPTIONAL)	9
B2	First responder training	3	B2	First responder training	11	B2	First responder training/first‐aid training	11
B3	Health centre roles	5	B3	Health centre roles in rural areas	6	B3	Roles and abilities of health centres or primary care clinics	6
B4	Health centre care capacities	1	B4	Health centre care capacities	1	B4	Health centre care capacities	1
B5	Physical access to medical care	5	B5	Physical access to medical care	5	B5	Physical access to medical care (geographical and transport environments)	5
B6	Emergency transport plan (community level)	4	B6	Emergency transport plan in remote areas (community level; if there is no formal EMS)	4	B6	Community‐level emergency transport plan in remote areas (if there is no formal EMS)	4
C	Emergency Medical Services (EMS)		C	Emergency Medical Services (EMS)		C	Emergency Medical Services (EMS)
			C1‡‡ This sub‐domain was combined with A1 sub‐domain when creating the final model.	Lead agency assigned for EMS	5
C1	EMS organisation	5	C2	EMS organisation	6	C1	EMS organisation	6
C2	Coordination of the EMS system	4	C3	Coordination of the EMS system	5	C2	Coordination of the EMS system	5
C3	Contacting with the EMS system	4	C4	Contacting with the EMS system	6	C3	Contacting the EMS system	6
C4	Human resources	3	C5	Human resources	7	C4	Human resources	7
C5	Access to EMS	6	C6	Access to EMS (at national/regional/community level)	11	C5	Access to EMS (at the national/regional/community level)	11
C6	Vehicle arrangement/safety	5	C7	Vehicle arrangement/safety	8	C6	Vehicle arrangement/safety	8
C7	Communication	4	C8	Communication	5	C7	Communication	5
C8	Transfer arrangement	2	C9	Transfer arrangement	4	C8	Transfer arrangement	3
C9	Prehospital care skills	1	C10	Prehospital care skills	1	C9	Pre‐hospital care skills	1
D	Upward referral within medical care system (between hospitals)		D	Upward referral within medical care system (between hospitals)		D	Upward referral within the medical care system (between different levels of care)
D1	Referral arrangement	5	D1	Referral arrangement	6	D1	Referral arrangement	6
D2	Human resources	3	D2	Human resources	3	D2	Human resources	3
D3	Communications	5	D3	Communications	5	D3	Communications	5
D4	Transportation arrangement	3	D4	Transportation arrangement	4	D4	Transportation arrangement	4
E	Definitive care		E	Definitive care		E	Definitive care (tertiary care)
E1	Regionalization	2	E1	Regionalization	2	E1	Regionalization	2
E2	Supply demand balance	2	E2	Supply demand balance	3	E2	Supply demand balance	3
E3	Patient flow management	3	E3	Patient flow management	3	E3	Patient flow management	3
E4	Discharge support	4	E4	Discharge support	4	E4	Discharge support	3
E5	Hospital capacities (resources/skills)	1	E5	Hospital capacities (resources/skills)	1	E5	Hospital capacities (resources/skills)	1
F§§ This domain was split into F and G in the final model, and the order of sub‐domains was changed when creating the final model. The original codes are used in the baseline and revised models.	Aftercare and downward referral (back to community)		F§§ This domain was split into F and G in the final model, and the order of sub‐domains was changed when creating the final model. The original codes are used in the baseline and revised models.	Aftercare and downward referral (back to community)		F	Rehabilitation (returning to society)
F1	Rehabilitation services	3	F1	Rehabilitation services	3	F1	Rehabilitation services	3
F5	Support for returning to society/job at community	3	F5	Support for returning to society/job at community	3	F2	Support for returning to society/job in the community	3
						G	Downward referral within the medical care system (between different levels of care)
F2	Downward referral arrangements	3	F2	Downward referral arrangements	5	G1	Downward referral arrangements at tertiary or secondary care	5
F3	Transportation	3	F3	Transportation	3	G2	Transportation arrangement	2
F4	Communication	2	F4	Communication	2	G3	Downward communication (from tertiary to secondary or primary care)	2
G	Evaluation/research		G	Evaluation/research		H	Evaluation/research
G1	Record keeping	4	G1	Record keeping	4	H1	Record keeping	4
G2	Analyses/evaluation/research	3	G2	Analyses/evaluation/research	3	H2	Analyses/evaluation/research	3
Total		118	Total		176	Total		177

• * An alphabet alone indicates a domain and an alphabet with a number indicates a sub‐domain.
• † Number of items in each sub‐domain.
• ‡ This sub‐domain was combined with A1 sub‐domain when creating the final model.
• § This domain was split into F and G in the final model, and the order of sub‐domains was changed when creating the final model. The original codes are used in the baseline and revised models.

In the second round, most of the items were highly rated, and no new items were added. Though the third‐round list was the same as that in the second round, the ratings of the third round were slightly changed because a few participants changed their ratings after seeing the summary results of the second round (Table S1). The items with relatively low ratings (proportion <0.7) were mostly related to informal community activities (e.g. funding, cost‐sharing, and transport plan), private sector roles (e.g. private EMS organisations and funding activities), and social support (e.g. social work to support discharge and return to social activities). Rehabilitation during the acute phase of disease, as well as incentives in the referral system, also had relatively low ratings.

Development of the system model

We reviewed the third‐round list (revised model) to develop the final system model. No item was excluded from the list because none met the deletion criteria (the proportion ≤0.2). We made some lower‐rated items and sub‐domains (proportion <0.6) optional; for example, some community‐based informal activities may not be necessary in urban areas with a functioning and formal EMS system.

Then we modified the listed items as needed. To specify the meaning, the expressions were modified and explanations were added. Table S2 shows the details of how we divided and combined the items (to code the items, the revised model used serial numbers and the final model used alphabets and numbers). We divided items addressing multiple aspects: legislation of EMS was divided into national level and local level legislation (revised model code 16 was divided into final model codes A2.3 and A2.4); guidelines/policy/regulations were divided into legislation/regulations and guidelines/policy (code 13 into A2.5 and A4.10; code 14 into A2.6 and A4.11); national or regional plan was divided into that of emergency medical care and that of EMS (code 29 into A4.1 and A4.2; code 30 into A4.3 and A4.4). We integrate half of some divided items with other items: policy/plan/statute regarding emergency medical care (codes 9 and 10) were divided into legislation and policy/plan, and the latter half was integrated into items addressing the same issues (A4.1 and A4.3, respectively) and vehicle arrangement/maintenance (code 166) was divided and the maintenance part was integrated into item D4.3. We made combinations of redundant items regarding informal cost‐sharing schemes (codes 25 and 37 made B1.2), physicians' instruction (codes 139 and 143 made C7.5), prior agreement on referral (codes 185 and 197 made G1.3), and transportation cost (codes 202 and 203 made G2.2).

We divided the domain of ‘aftercare and downward referral’ into 2 because ‘aftercare’ covers hospital level and community‐based care, whereas ‘downward referral’ covers inter‐facility interface; aftercare was renamed as ‘rehabilitation’ because its elements are rehabilitation services and support to return to society (Table 2; Table S3). Sub‐domains assigned for ‘lead agency for emergency medical care system (A1)’ and ‘lead agency assigned for EMS (C1)’ were combined. Consequently, the final system model included 8 domains, 35 sub‐domains and 177 items (Table 2; Table S3). The main frame of the model with the four levels (i.e. policy level, hospital level, inter‐facility interface and community level) was retained (Figure 2).

Figure 2

Emergency care system conceptual model.

Development of the evaluation tool

We developed a short list, which includes 77 items, as the evaluation tool from the final model list (Table S4).

Discussion

In this study, expert opinions from various countries, including LMICs and HICs, yielded a comprehensive list of components constituting an emergency care system model. The list covered all the aspects of emergency care systems; for example, policies, medical care delivery at hospital, referral, formal EMS and informal community‐based actions. The list was expanded, and some re‐categorisation was required after three rounds of surveys. However, the structure of the baseline system model was maintained, suggesting its relevance.

In contrast, many of the previously developed tools for evaluating emergency care in LMICs focused on hospital or pre‐hospital care capacities. For example, the WHO Tool 35 for Situational Analysis to Assess Emergency and Essential Surgical Care, and Surgeons OverSeas'Personnel, Infrastructure, Procedures, Equipment, and Supplies (PIPES) Surgical Capacity Assessment Tool 39 assess surgical care capacities of medical care facilities. WHO's 14, 15 Essential Trauma Care Guidelines and Pre‐hospital Trauma Care Guidelines, Wong et al.'s 16 International Assessment of Capacity of Trauma (INTACT) index and Baker et al.'s 21 Emergency and Critical Care (EaCC) indicators provided frameworks to assess trauma care or emergency and critical care in medical facilities or EMS settings.

Although Sassers et al. 15 and Mould‐Millman et al. 25 point out the importance of considering wider contexts, such as leadership, financing, politics and regulations, they focus on EMS systems and do not cover medical care at hospitals. Despite their emphasis on wider contexts, Sassers et al.'s evaluation list covers only physical and human resources.

Meanwhile, few attempts have been made to evaluate entire emergency care systems, rather than focusing on abilities of each health facility or EMS alone. Latifi et al. 23 used the basic trauma care system criteria of American College of Surgeons, Spiegel et al. 24 developed a conceptual model in which surgical care is integrated into health systems and Remick et al.22 developed the Global Trauma System Evaluation Tool (G‐TEST), which indicates a comprehensive model to evaluate the trauma care system. Although these models (particularly G‐TEST) have many aspects in common with the model we developed as some of their components were borrowed, important differences exist. We evaluated care for all types of acute diseases, not only trauma. While our model emphasised informal mechanisms within the communities, which are important in resource‐constrained areas without well‐functioning formal systems, Latifi et al., Spiegel et al. and Remick et al. paid insufficient attention to these aspects.

Informal mechanisms, including task shifting to non‐professional personnel and community‐based actions, can complement underdeveloped formal systems and help improve access to healthcare services 37. This is a core concept of primary healthcare that is also applicable to improving access to emergency care. Lay first responder training improves access to first‐aid 10 and informal trauma care systems based on lay first responder, and non‐professional health workers trained for damage control procedures can improve survival rates among severe trauma patients in more remote areas with no formal EMS systems 11. Community‐based mutual aid mechanisms, such as cost‐sharing schemes and emergency transport arrangements, can tackle barriers to health care 6.

We believe that system models and evaluation tools that articulate the importance of informal mechanisms would help policy‐making in developing emergency care systems that appropriately cover entire countries. Without guides from such models, development of emergency care systems tends to focus on densely populated areas. Western models might work efficiently in these areas; however, places with more dispersed populations are usually left behind. In remote areas with scarce resources, where Western models do not work, different models that include informal mechanisms have potential and should be integrated into formal healthcare systems 40.

In addition to its comprehensiveness, the evaluation tool is flexible; the list is divided into small categories (domains and subdomains), some of which are optional. Therefore, the tool is applicable globally, either by using it in its entirety or just parts of the list. This means that its validation, which is to be done before its actual application, can be assessed in any setting, either totally or partially. Testing in various settings would serve to sophisticate and strengthen this model.

Limitations

There were some limitations in this study. First, even though we tried to include participants from various backgrounds, it is unlikely that our participants represent all types of stakeholders, from all regions. For example, primary care practitioners and nurses were underrepresented in our surveys, and most of the participants were physicians in tertiary hospitals in urban areas. This may explain why the informal components were rated relatively low. Even though different stakeholders may have had different opinions, this would not distort the results of our survey, given that no items were excluded owing to extremely low ratings.

Second, the model and tool have not yet been field‐tested. Despite our efforts to create a comprehensive list, there is a possibility that some important and relevant items may be missing. Therefore, conducting validations in various settings would help fill out any potential gaps, and the list can be updated as needed.

Conclusion

We created a conceptual model for emergency care systems and developed a list of system components through utilising a consensus‐building survey method. The model and list have potential to provide a basis with which to evaluate emergency care systems in resource‐constrained settings of LMICs; although its field‐testing and validation in various settings remain to be done.

Funding

This work was supported by JSPS KAKENHI (Grant Number JP16K11422).

Conflicts of interest

The authors report no conflicts of interest, and they alone are responsible for the content and writing of the paper.