Emergency care with lay responders in underserved populations: a systematic review

Abstract Objective To assess the individual and community health effects of task shifting for emergency care in low-resource settings and underserved populations worldwide. Methods We systematically searched 13 databases and additional grey literature for studies published between 1984 and 2019. Eligible studies involved emergency care training for laypeople in underserved or low-resource populations, and any quantitative assessment of effects on the health of individuals or communities. We conducted duplicate assessments of study eligibility, data abstraction and quality. We synthesized findings in narrative and tabular format. Findings Of 19 308 papers retrieved, 34 studies met the inclusion criteria from low- and middle-income countries (21 studies) and underserved populations in high-income countries (13 studies). Targeted emergency conditions included trauma, burns, cardiac arrest, opioid poisoning, malaria, paediatric communicable diseases and malnutrition. Trainees included the general public, non-health-care professionals, volunteers and close contacts of at-risk populations, all trained through in-class, peer and multimodal education and public awareness campaigns. Important clinical and policy outcomes included improvements in community capacity to manage emergencies (14 studies), patient outcomes (13 studies) and community health (seven studies). While substantial effects were observed for programmes to address paediatric malaria, trauma and opioid poisoning, most studies reported modest effect sizes and two reported null results. Most studies were of weak (24 studies) or moderate quality (nine studies). Conclusion First aid education and task shifting to laypeople for emergency care may reduce patient morbidity and mortality and build community capacity to manage health emergencies for a variety of emergency conditions in underserved and low-resource settings.


Introduction
Conditions that could be treated with prehospital and emergency care account for an estimated 24 million lives lost each year in low-and middle-income countries. 1 Training lay providers and volunteer paramedics to respond to health emergencies is among the most cost-effective health interventions globally, at just United States dollars 6-14 per disabilityadjusted life year saved. 2 In May 2019, the World Health Organization (WHO) resolved to improve emergency care in all Member States, including through informal prehospital systems. 3 International guidelines define first aid as "initial care provided for an acute illness or injury" or "help to a suddenly ill or injured person which is initiated as soon as possible and continued until that person has recovered or medical care is available." 4,5 Teaching first aid to laypeople is part of a 150-year medical humanitarian tradition and a vital component of both formal and informal prehospital care systems. 6 Over 15 million people in 52 countries receive education in first aid each year from member organizations of the International Federation of Red Cross and Red Crescent Societies alone. 7 First aid education enhances bystanders' helping behaviours in emergencies, but it remains unclear what interventions can be delivered effectively by laypeople to save lives and reduce morbidity. 8 Laypeople with training in first aid can improve access to care for underserved populations, and often deliver the only emergency health-care services in low-resource settings. 1 Where lay responders are taught first aid to enhance patients' access to essential interventions, first aid education is a part of the broader concept of task shifting. WHO defines task shifting as "the rational redistribution of tasks within health workforce teams," specifically from specialized professionals to providers with less training, lay caregivers and patients. 9 Task shifting improves access to care and outcomes for maternal and child health, chronic and mental health conditions, and communicable diseases. [10][11][12][13] Less is known, however, about task shifting in emergency health care. We adapted our search for grey literature using keywords that targeted the websites of humanitarian and global health agencies, academic grey literature databases, theses and dissertations, clinical trials registries and conference proceedings. 14

Study selection
We trained an international team of 18 reviewers with varied expertise in the subject matter and methods of the review, using a video to familiarize them with the research question and inclusion criteria. All reviewers screened a test set of 70 papers selected from our search that included seven papers that met enough inclusion criteria to proceed to full-text review. We conducted an internal study on this test set to confirm substantial inter-rater agreement (Fleiss' κ > 0.61) between reviewers. 28 More details are in the authors' data respository. 17 The reviewers screened the titles and abstracts of studies retrieved through the electronic and manual searches, independently and in duplicate. We conducted independent and duplicate full-text review of all papers retained through screening. One of the two lead investigators resolved discrepancies. We documented reasons when papers were excluded at this stage. We assessed papers in Dutch, English, French, German and Norwegian languages. We used Google Translate (Google LLC, Mountain View, United States of America, USA) and Cochrane TaskExchange volunteers 29 to review papers in other languages.

Data extraction
For each included paper, two investigators independently extracted information on the study objective, study design, population, details about the intervention and control groups (mode and duration of education; emergency health conditions treated; and role of the lay person), Publication: Bulletin of the World Health Organization; Type: Systematic reviews Article ID: BLT.20.270249 Page 5 of 32 outcomes (type of health outcome; description of health outcome; type of emergency care provided; and effect size and confidence interval) and key conclusions. Where multiple publications reported on the same underlying study, we extracted data from all related papers and reported results from the most definitive paper.
We performed independent and duplicate assessment of study quality, including internal and external validity, selection and measurement biases, and confounding factors, using the Effective Public Health Practice Project quality assessment tool. 30 This tool permits the appraisal of multiple types of studies and is designed and validated for the assessment of studies concerning health systems and population health interventions. We resolved discrepancies through consensus among the lead investigators.

Synthesis
We prepared a narrative and tabular synthesis of our findings. We grouped studies qualitatively according to the illnesses or conditions addressed, the role of lay providers, the type of educational intervention provided and the type of outcomes reported. We distinguished individual health outcomes such as survival to hospital discharge; community health outcomes such as all-cause mortality; and measures of community capacity to manage emergencies such as cardiac arrest response times. Our rationale for these groupings was first to underscore the emergency health conditions for which studies had been identified, and then to provide information to guide future task shifting and first aid training interventions. We drew on Cochrane Collaboration guidance on syntheses without meta-analysis to assess the risk of bias across studies, and considered the number of studies, consistency of effects and directness of findings to develop plain-language summary statements of the effects of interventions. 31

Results
Our database searches yielded 19 308 unique papers. We retained 415 papers for full-text review, resulting in 43 eligible papers from 34 unique studies (Fig. 1). Grey literature and manual searches did not yield additional publications. Interrater agreement between the screening authors was good for study inclusion (Fleiss' κ = 0.75 17 ). Studies excluded at fulltext review are described in the authors' data respository 17 . reported on community health outcomes (Fig. 2).

Outcomes
Most studies reported small effect sizes (Table 1). Some studies reported statistically significant and clinically important effects on measures of individual and community health.
For example, one study reported an absolute reduction of 20.4 per 1000 in all-cause under-5 mortality in a randomized trial of malaria peer education for mothers in Ethiopia. 38 Another study comprising 2788 patients treated for trauma reported a reduction in mortality from 17% to 4% in a before-and-after study of a community first-responder programme in Iraq. 64 Through a cohort and modelling study, researchers estimated that opioid overdose education and naloxone distribution in British Columbia, Canada, averted 1650 deaths in 20 months. 51 Two included papers reported null or equivocal results. A cluster randomized controlled trial of a malaria education and management programme for women's groups observed no effect on the prevalence of severe malaria-associated anaemia in children. 40

Discussion
We found that first aid education and task shifting to laypeople may reduce morbidity, mortality and community capacity to manage health emergencies for a variety of emergency conditions. The studies include cardiac arrest, burns, malaria, malnutrition, opioid poisoning, paediatric communicable diseases, snakebites and trauma. All of the included studies evaluated targeted training for priority local emergency conditions; there were no eligible studies concerning courses with general, untargeted first aid curricula. The overall weak quality of studies in our review underscores the limitations in the available science, the need for rigorous studies in this field, and the challenges inherent in evaluating complex population health interventions such as task shifting. 67 The widespread practice of training laypeople to deliver lifesaving interventions for acute health emergencies in underserved settings arises from sound logic and humanitarian principles. 2,6,7 Our review shows that there is limited empirical evidence to demonstrate an individual or community health benefit arising from this practice.
Previous reviews demonstrate the effectiveness of first aid education by reporting on knowledge, skills, helping behaviours or confidence among trainees. 8 Guidelines and curricula for first aid generally derive interventions for the lay public by adapting practices from professional prehospital practice and health-care research. 21,26 Database searches that rely on the keywords "first aid" or "layperson" to retrieve studies concerning first aid may overlook papers concerning interventions that do not use these terms but are thematically aligned with first aid. In comparison with other reviews on first aid, ours covers a greater breadth of research concerning interventions provided by laypeople to address emergency medical problems in underserved populations and low-resource settings. 68,69 Like other task-shifting strategies, first aid education is a complex, system-level intervention that requires its own foundation of evidence. 70 Clinical interventions that may be effective when implemented by professionals may not produce the same results when implemented by other providers. The adaptation of professional practices and the assessment of educational outcomes among people trained in first aid is insufficient to establish the effectiveness and safety of first aid interventions for target patients, programmes or communities.
Our review advances novel conceptual ties between first aid and task shifting. Lay emergency care and volunteer paramedic interventions are among the most cost-effective ways to reduce avoidable mortality worldwide, but unlike other task-shifting interventions, first aid has not been widely characterized or evaluated based on broad public health impacts. 2 The connection between first aid education programmes and task shifting underscores how first aid interventions and lay emergency care might contribute to addressing priority global health challenges such as opioid overdose, trauma or malaria.
We have summarized the breadth of contexts and conditions where lay responders, bystanders or friends and family can provide first aid. Leading international guidelines define first aid as "the initial care provided for an acute illness or injury" that "can be initiated by anyone in any situation." 5 The interventions and acute conditions included in this review conform with this definition, but many of the included studies concerned conditions and interventions that are mostly absent from conventional first aid training, such as lay assistance for acute malnutrition, opioid poisoning or paediatric communicable diseases. The appropriate scope of first aid and the set of interventions captured in those studies may be determined based on the care that can be initiated by anyone to provide initial care for an acute illness or injury in a safe and clinically effective manner. First aid need not be defined based on the set of interventions included in standard courses or curricula.
The strength of this review is its breadth, including a search of multiple databases and inclusive search terminology to synthesize the wide range of experimental and observational research concerning task shifting for emergency care in low-resource and underserved settings worldwide. By training and assessing interrater reliability of an international team of reviewers we were able to achieve a manual review of over 19 000 studies. Our approach to populations including both low-and middle-income countries and underserved subpopulations in high-income countries is a conceptual strength aligned with global approaches to health equity. 71 Our review also has limitations. We excluded studies conducted in well-resourced populations because interventions that are effective in wellresourced settings cannot be presumed to work in contexts with fewer resources. For example, systematic reviews have demonstrated the efficacy of mental health first aid in high-income countries. 72,73 Although mental health first aid has been studied in lower In conclusion, first aid for laypeople may have its greatest impact when approached as a series of targeted interventions that equip the public to respond to the health emergencies that they are likely to encounter in their everyday lives and communities. More work is needed to orient first aid education to deliver the greatest effects on patient and community health, and to identify the modalities that are best suited to specific contexts, populations, clinical conditions and public health priorities. Task shifting to laypeople for emergency care may save lives, reduce morbidity and enhance community capacity to address acute health problems in low-resource settings.

Funding:
This project received financial support from the Northern Ontario Academic Medicine Association Innovation Fund (Project #A-15-07 Underserved or low-resource population: A group that faces any barrier to accessing organized prehospital emergency medical services, including geographical, financial, occupational, sociopolitical, ethnocultural, infrastructural or informational barriers. 14 We excluded people serving in the military or populations living in war zones from this definition. Emergency health condition: Health problem(s) where treatment should occur within minutes or hours to reduce suffering, morbidity or mortality. Task shifting for routine intrapartum and perinatal care has been reviewed systematically elsewhere and we therefore excluded it from our definition of emergency health conditions. 10,19,20 Intervention criteria First aid or prehospital emergency care: Any effort to identify, care for or treat an emergency health condition in a prehospital or out-of-hospital setting. First aid may be definitive care or may involve transition to more advanced care. 4,5,21 Training or education: Any effort intended to confer knowledge or skills to a person, or change their attitudes and behaviours.
Laypeople trainees: Any community member who has no health professional designation or certification and who is not primarily employed in health-care delivery. This definition of laypeople excludes paraprofessional cadres such as community health workers, where emergency care formed part of the workers' practice.

Outcome criterion
Individual or community health effects: Any quantified effect on morbidity, mortality or community capacity to manage a health problem. We considered willingness to provide emergency care as a health outcome when measured at the community or population level and not when measured only among trainees.