Introduction

Sepsis is characterized by life-threatening organ dysfunction resulting from the body’s response to infection, estimated to affect 48.9 million patients and account for 11 million deaths each year [1]. The emergence of the coronavirus disease 2019 (COVID-19) pandemic also accentuates the disease burden of sepsis, as sepsis is a frequent complication resulting in high morbidity and mortality [2]. Early recognition and treatment is vital to increase survival and reduce morbidity. As sepsis may result from infection in any organ system, healthcare professionals working in any specialty may need to manage septic patients [3]. Thus, improving survival from sepsis requires all healthcare workers to be educated in its recognition and management.

In 2017 the World Health Assembly adopted a resolution to improve the prevention, diagnosis and management of sepsis and suggested educational development for health care workers [3]. More recently, the development of a core competency framework on sepsis for health care workers was identified as a priority at a WHO Sepsis Technical Expert Meeting [4]. Teaching medical students about sepsis may be an effective way of ensuring that all future doctors have sufficient ability to diagnose and treat septic patients. Medical school curricula are increasingly designed using the competency-based medical education framework [5]. However, there is currently no consensus on what competencies medical students should achieve regarding sepsis recognition and treatment. Determining appropriate competencies is a crucial first step for development of appropriate curricula and for evaluating adequacy of training. Although previous studies have investigated the knowledge of medical students and junior doctors regarding sepsis, the usefulness of these data are limited by the absence of external criteria to judge adequacy [6,7,8,9]. Furthermore, a MedLine search using the medical subject headings “sepsis” and “students, medical” and “curriculum” revealed no publications that addressed the issue of what should be included in medical student sepsis curricula.

Most studies of sepsis and recent management guidelines have focused on high income countries/regions but most of the worlds’ population lives in low and middle income countries/regions [10]. In 2017 85% of sepsis cases occurred in areas with low, low-middle or middle standard development index (SDI). Furthermore, the highest age-standardised sepsis-related mortality occurred in areas with the lowest SDI, highlighting the high burden of sepsis in low and middle income countries/regions [1]. The challenges to providing high quality sepsis care likely differ substantially between high income and low/ middle income healthcare systems. For example, 15% of Latin American intensivists reported inadequate conditions to manage patients with septic shock due to insufficient technology, laboratory support, imaging and drug resources [11]. In Africa, only 1.5% of respondents to a survey in 2011, felt they had resources to implement the entire Surviving Sepsis Campaign guidelines [12]. Consequently, it is likely that training requirements for medical students should also differ [9].

We therefore carried out a study, using a Delphi technique, to determine what sepsis-related competencies medical students should achieve by the end of their medical student training in both high or upper-middle incomes countries/regions (HUMIC) and in low or lower-middle income countries/regions (LLMIC) [12]. This study is intended to be the first step in a multi-stage process. Obtaining consensus on what should be included in a sepsis medical school curriculum is the first stage. Further work and studies are required to design and implement a curriculum, develop assessment tools and validate the competencies/ curriculum suggested by this study.

Method

Approval to carry out the study was obtained from the Survey and Behavioral Research Ethics Committee of The Chinese University of Hong Kong (SBRE 221-17).

An initial search of EMBASE and PubMed using the medical subject headings “sepsis” and “students, medical” and “curriculum” was conducted to identify studies for determining medical student sepsis curricula. No studies were found.

A Delphi method was used to determine sepsis-related competencies which medical students should have achieved by the end of their undergraduate training. The Delphi process involved several stages [13]. Firstly, panels of experts were assembled: one consisting of panelists from HUMIC (as defined by the World Bank for the 2018 fiscal year) and one consisting of panelists from LLMIC [14]. When forming the panel the following factors were considered. First, there should be a balance between panelists with and without a specific interest in sepsis. Second, panelists should come from a range of geographical areas and specialties that commonly manage patients with sepsis. Third, the panel should include doctors who had recently qualified and not yet completed specialist training. Finally, the panel should include members with responsibility for the overall medical school curriculum at their institution. Medical students were not included as panelists as they were deemed to lack sufficient experience and insight regarding sepsis education. However, we believe that recent graduates retain an appreciation of the scope and volume of the whole medical curriculum while understanding the aspects of medical education necessary for patient care. Digital written informed consent to participate was given by all panelists, who volunteered their time without funding.

During the second stage, panelists were asked to individually suggest a broad list of up to 20 competencies that they thought final year medical students should have regarding sepsis. From these suggestions a list of competencies from each of the two panels was compiled. Two researchers (GELM and LL) then individually eliminated duplicates from these lists, with disagreement between the two researchers being resolved by discussion.

During the third stage, the lists of competencies were circulated to the respective panelists, who independently rated the importance of each competency, without awareness of other participant’s ratings, on a numeric scale of 1 to 5 as essential (1), very important (2), moderately important (3), slightly important (4) or unimportant (5) [15].

During the fourth stage, the collated results for each panel were sent to all members of that panel. Panelists were shown their own rating and the distribution of ratings by all other panelists for each competency. Participants were able to change their rating with the understanding that the purpose of the round was to achieve consensus. Consensus was defined as ≥ 75% of participants rating the importance within two adjacent categories; e.g. 75% of participants rating a competency as either essential or very important [16]. After this fourth stage the results were examined by three investigators and an independent adjudicator to determine whether discussion of the precise wording of the competency followed by a further round of voting was likely to result in consensus. This process was completed by 9th May 2020. To facilitate translation and adoption into medical school curricula, competencies were classified according to the 8 commonly accepted competency domains in competency-based medical education: patient care, knowledge for practice, practice-based learning and improvement, interpersonal and communication skills, professionalism, systems-based practice, interprofessional collaboration, and personal and professional development [17]. Median was used to summarize importance ratings.

Submission of suggestions, rating of competencies and revision of ratings were all carried out electronically using REDCap electronic data capture tools hosted at the Department of Anaesthesia & Intensive Care, The Chinese University of Hong Kong [18, 19]. There was no direct face to face discussion between panelists, who all worked independently.

Results

We invited 19 participants from HUMIC and 15 participants from LLMIC. All HUMIC invitees accepted but one did not respond to any further correspondence leaving 18 panelists. One of the LLMIC invitees did not respond to our invitation and one accepted the invitation but did not respond to any further correspondence, leaving 13 panelists. In the HUMIC group, 14 panelists were from high income countries/regions and four from upper-middle income countries/regions: six from North America, one from South America, four from Asia, four from Europe, two from Oceania and one from Africa. In the LLMIC group, five were from low-income countries/regions and eight from lower-middle income countries/regions: one from North America, three from Asia, one from Oceania and eight from Africa. The specialty areas for which the panelists were invited are given in Table 1. Members of the panel included representatives from World Health Organization (WHO), Global Sepsis Alliance, Latin American Institute of Sepsis, Chinese Society of Critical Care Medicine, African Sepsis Alliance, Asia Pacific Sepsis Alliance, and panelists with overall responsibility for medical school curricula.

Table 1 Number of participants representing various specialties in high income and low/ middle income countries/regions
Full size table

Panelists in the HUMIC group made 239 suggestions, which were reduced to 109 after removing duplicates. Panelists in the LLMIC group suggested 195, which were subsequently reduced to 88.

After two rounds of rating, consensus was reached for all but one competency in the LLMIC group and all competencies with a median rating greater than very important in the HUMIC group (Table 2). Examination of the distribution of the ratings (by JGM, GCD, LL and GELM) for those competencies for which consensus was not achieved (Supplementary Table 1) suggested that re-wording of the competency was unlikely to achieve consensus if subjected to another round of rating. Competencies rated below moderately important are given in Supplementary Table 2. There was no missing data.

Table 2 Competencies rated moderately important or above with consensus
Full size table

Discussion

Our study is the first study to identify sepsis-related competencies that medical students should achieve by graduation. This is in line with the priority identified at the WHO Sepsis Expert Technical meeting in 2018 to develop core competency frameworks on sepsis for health care workers. We identified 38 and 33 essential competencies which medical students should achieve in LLMIC and HUMIC, respectively. Consensus was reached for all essential competencies. In the LLMIC group, complete consensus was also achieved for competencies ranked as very important and was achieved in 4/5 competencies rated as moderately important. In the HUMIC group, consensus was achieved in 41/57 competencies rated as very important but only 6/11 competencies rated as moderately important. These competencies have been endorsed by the Global Sepsis Alliance to guide development of sepsis curricula for medical students.

Given the burden of sepsis and the need for early intervention, doctors working in all specialties should have the basic competencies required for diagnosing and managing septic patients. It therefore behoves medical schools to provide training in sepsis recognition and management. Indeed, it has been suggested that member states of the WHO should mandate this training for all healthcare workers [4].

We believe the competencies identified in this study provides a useful framework on which to develop sepsis training for medical students. The aim of modern competency-based medical education is to train “health-professionals that can practice medicine at a defined level of proficiency” [20]. Most of the identified competencies were within the domains of patient care, knowledge for practice and interpersonal and communication skills [17]. This reflects the objective of this study which was to identify core competencies that medical students should achieve by graduation to care for patients with sepsis.

While we understand that different medical schools will have different priorities for teaching, we strongly suggest that, as a minimum, training to achieve the competencies rated essential be incorporated into the curriculum of all medical schools. We also suggest that medical students should achieve competencies rated very important or moderately important for which consensus was obtained (Table 2). For example, both LLMIC and HUMIC panels ranked “know the definition of sepsis” and “know that early recognition of sepsis is important” as essential competencies. Meanwhile, “know the SIRS criteria and recognize them when present in a patient”, “know how to calculate and utilize qSOFA score” and “know rationale to justify revising definitions of sepsis from previous SIRS based definition” were only ranked as very important rather than essential. Whilst these seemingly similar competencies resulted in different importance rankings, the need to know the definition of sepsis and importance of early sepsis recognition are key concepts of sepsis management. Instead, SIRS or qSOFA criteria are imperfect tools for sepsis recognition which may be updated or replaced over time. Indeed, differences in use of SIRS and qSOFA to diagnose and prognosticate sepsis when compared to Sepsis-3 criteria based on SOFA has been well documented [21]. This distinction in ranking may help medical schools prioritize essential learning objectives in curriculum development.

A guiding principle of competency based-medical education is to achieve competencies that are “in accord with local conditions to meet local needs [20]”. We anticipated that LLMIC would have different priorities and requirements for medical students compared with HUMIC. Therefore, LLMIC and HUMIC had separate panels and thus separate results. While many of the suggested competencies overlapped, there were some suggestions that differed between the groups (Table 2 and Supplementary Table 1). For example, an essential competency for LLMIC is to know how to diagnose malaria, while this was not even suggested HUMIC group. In addition, some basic skills such as being able to measure blood pressure and blood glucose were included by the LLMIC panel but not the HUMIC panel. Furthermore, an essential competency included by the LLMIC but not HUMIC panel was to be able to prioritize patients needing critical care when resources are scarce. While we attempted to address heterogeneity by having separate LLMIC and HUMIC panels, this does not address the heterogeneity in diseases and resources within LLMIC or HUMIC.

Studies based on Delphi methodology are entirely dependent on the opinion of their expert panels [22]. By their very nature these panels are not randomly selected and the selection process may introduce bias. We used specific criteria to determine the composition of the panels to minimize bias. In particular, we deliberately included panelists who did not have a particular interest in sepsis and panelists whose professional position allowed them an overview of their entire medical school curriculum. There was, perhaps, an over-representation of Intensive Care specialists in the HUMIC group, but the inclusion of life-threatening organ dysfunction in sepsis criteria means that Intensive Care specialists are inevitably major stakeholders. Furthermore, the relatively high proportion of Intensive Care specialists on the panel does not appear to have been reflected in those competencies considered essential, none of which are specific to Intensive Care practice. In retrospect, the failure to include public health specialists was a weakness, which may have resulted in an absence of suggestions for public health interventions such as vaccination.

We defined consensus as 75% of panelists rating the importance of a competency within two adjacent categories. Percentage agreement is the most common method of defining consensus and the median threshold for agreement in 25 Delphi studies that were recently reviewed was 75%, with a reported threshold agreement range of 50–97% [16]. There is no consensus on the number of panelists required, but our number of panelists falls within the commonly used range. In a review of 76 healthcare related Delphi studies the median number of individuals invited to participate was 17 (interquartile range of 11–31), and in a review of 100 studies the most common number of panelists in the final round was between 11 and 25 [23, 24].

While the Delphi method is a widely used method for developing guidelines and curricula, there are no published data to demonstrate that Delphi-based curricula improve students’ learning outcomes. This is partly due to difficulties in separating the effects of curriculum and curriculum content.

Another potential limitation is that panelists didn’t meet to discuss results. It is possible that consensus wasn’t achieved for some competencies because panelists weren’t able to discuss the reasoning behind their suggestions. In contrast, an advantage of the panelists not meeting was the absence of social pressure to agree with a suggestion that they fundamentally did not agree with, thus avoiding false consensus.

It is possible that greater consensus would have been achieved if a further round of rating had been undertaken after discussion on re-wording competency statements. However, after careful assessment, the adjudicating group believed that this was unlikely, and that considering the level and nature of competencies for which consensus was not reached, greater consensus would not substantially increase the usefulness of this document.

The results of this study were collected prior to the 2021 update to the Surviving Sepsis Campaign International Guidelines for the management of sepsis. While it is possible that these newer guidelines might have altered the panelists’ recommendations, the relatively non-specific nature of the recommendations mean it is unlikely that they would be significantly impacted by changes in guidelines regarding the specific management of sepsis [25].

As mentioned above, the results of the study are based on expert opinion not high level evidence. However, we are not aware of any high level evidence on which to base a medical student sepsis curriculum. Once a curriculum and assessment based on these competencies has been developed, further research to validate our data as a tool to strengthen medical students’ experience and clinical performance may be possible.

Conclusion

We have identified essential sepsis-related competencies for medical students in both LLMIC and HUMIC. Consensus on their importance was achieved for all these competencies. We suggest that medical schools develop curricula to address these competencies, as a minimum, but also consider addressing competencies rated as very or moderately important.