When to trust our learners? Clinical teachers’ perceptions of decision variables in the entrustment process

Clinical training programs increasingly use entrustable professional activities (EPAs) as focus of assessment, but questions remain about which information should ground decisions to trust learners. Clinical teachers struggle to find ways to assess the preparedness of learners to independently perform patient care tasks. An entrustment decision cannot be credited to a single variable but requires a multitude of variables. We defined decision variables as the characteristics, skills, qualities, previous performance or other information about learners that provide supervisors with input about learners’ performance on an EPA to make valid entrustment decisions. Such decision variables include features that determine the learners’ trustworthiness, the clinicians’ trust propensity and the risks and benefits involved in entrustment decisions. This study provides an overview of decision variables that clinical teachers find relevant in the elaboration of the entrustment decision processes into five categories: ability, humility, integrity, reliability and adequate exposure. The findings can substantiate entrustment decision-making in the clinical workplace. Extensive reflection by different observers on performed tasks is highly relevant: both reviewing of actions after shifts by teachers, reflection by the learner her or himself as well as multi-source feedback were mentioned as relevant decision variables to base entrustment decisions on. Faculty development will be necessary to explain these types of entrustment decisions as being situational versus focused on summative assessment when introducing workplace curricula based on EPAs.

Assessing medical learners in the workplace is notoriously difficult [1‐3]. Clinical teachers struggle to find meaningful ways to assess the preparedness of learners to independently perform patient care tasks [4]. Assessment often relies on subjective impressions of clinical teachers, who are often not specifically trained in assessing learners [5, 6]. The ultimate purpose of assessment is typically not to know how learners have acted in the past, but to predict how they will act in the near future [7]. One way to improve medical learner assessment may be reframing assessment to the question: ‘(When) can we trust this learner to execute this critical task with a certain level of supervision?’. In this approach, the concept of trust is crucial [8‐11].

Using entrustable professional activities (EPAs)—units of professional practice to be entrusted to learners once they have demonstrated sufficient competence—in medical training programs requires elaboration of entrustment decision processes [11‐14]. Entrustment decisions require a grounded entrustment process, based on decision variables collected in the clinical workplace over time. These decision variables should provide evidence to determine the learner’s trustworthiness, the clinicians’ trust propensity and the risks and benefits involved in entrustment decisions [12, 15]. In this study decision variables are defined as any information (e. g. skills, generic qualities, characteristics or previous performance) that provides supervisors with input about learner performance on an EPA in order to make valid entrustment decisions.

Variables that affect the confidence to trust a learner in performing a clinical task at a certain level of supervision have been a topic of research in the past decade [4, 16‐19]. The construct of ‘trusting a learner’ feels better aligned with how clinicians think than regular rating scales in workplace-based assessments [20‐22]. Teachers often attend to or value different aspects of performance, lack a clear standard for judging performance or use variable standards against which to judge learners [23‐25]. Assessors frequently compare learner performance with what they would do (self as standard) or to what they would expect of learners at a similar level of training; some rely on a gut feeling [23]. Clinical teachers regularly make judgements of learners’ competence at the point-of-care, in order to allow them to deliver patient care. They reported incorporating their perceptions of learners’ credibility and willingness to seek help when determining how much they trusted those learners [8]. In making these decisions about the level of supervision required in a given clinical situation, supervisors assess learners’ trustworthiness, a multidimensional construct which includes clinical knowledge and skills, discernment of limitations, truthfulness, and conscientiousness [16] and other features [8, 26]. These variables were recently reframed as: ability (knowledge, skill, experience) to complete the task, but also integrity (truthfulness, benevolence), reliability (conscientiousness, consistency), and humility (observing limits, willingness to ask for help) [27].

Entrustment decisions combine traditional assessment of ability with the right to execute an EPA without (or with indirect) supervision [23]. There are many variables of influence in determining how, when, and whether learners are granted responsibilities under indirect or distant supervision [7, 14]. Clinical teachers should take these variables into account and push learners to stretch their scope of performance with the need for safe, high-quality patient care. A previous study provided an overview of information sources derived from discussions with experts [11]. The aim of the current study is to identify the decision variables clinical teachers use and value in making entrustment decisions. Therefore, we addressed the following research question: Which decision variables are perceived by clinical teachers as being relevant to make entrustment decisions on learners’ progress towards unsupervised practice?

The medical teacher focus group included eight clinical teachers (six males, two females, mean age 45.0 years, mean experience as clinical teacher 13.6 years); the veterinary teacher focus group comprised nine (3 males, 6 females, mean age 44.6 years, mean experience as clinical teacher 14.7 years). Both focus groups consisted of clinical teachers from various disciplines: within the medical focus group representatives from dermatology (n = 2), internal medicine (n = 2), ophthalmology (n = 1), otorhinolaryngology (n = 1), pathology (n = 1) and paediatrics (n = 1) were present. Within the veterinary focus group there were specialists in companion animal health (n = 3), equine health (n = 3) and large animal health (n = 3).

Data collection proceeded in both groups until theoretical sufficiency was reached concerning the current research question. The data were then critically reviewed by the authors for their unique nature. Whenever two decision variables were basically similar, they were combined into one item. This process resulted in a list of 11 and 17 decision variables from the medical and veterinary group, respectively, certified as relevant. The subsequent process of axial coding yielded a total of 21 unique decision variables mentioned as being relevant in one or both groups, as listed in Table 1.

The decision variables were organized around five categories, four of which were recently proposed as a framework [27]: ability, humility, integrity, reliability, supplemented with a new category ‘adequate exposure’. In both groups 14 similar decision variables were mentioned, 2 decision variables were only mentioned by the medical group and 6 only in the veterinary group.

Most of decision variables were categorized as ‘ability’. The two highest ranked decision variables belonged to this category. Clinical teachers in the FVMU focus group mentioned ‘insight in the learner’s level of task-related knowledge’ as the most relevant variable. The most relevant decision variable for the medical group (UMCU) was ‘the learner’s ability to use the right indication to execute [a task]’. Furthermore, decision variables belonging to reliability and improvability were highly ranked by the participants. The decision variable ‘have knowledge about the judgement of other supervisors’ was the only variable mentioned but not ranked.

The categories ability, integrity, reliability, and humility are the conditions that must be met before someone trusts another person [7, 16, 19, 27]. Despite of this knowledge, our participants mentioned five variables that do not fit within these categories e. g.: personally observing the task execution as often as needed and knowing how often a task has been performed under which level of supervision and its results (see Table 1). These decision variables contain information about the learner based on similar or prior encounters. To categorize these variables we added a new category named ‘adequate exposure’.

This study aimed to increase insight into decision variables that clinical teachers apply when making entrustment decisions in the clinical workplace. In total 21 decision variables—divided over the categories ability, humility, integrity, reliability [27] and adequate exposure—were perceived as relevant for entrustment decision-making. These decision variables could help to build arguments for the multifactorial process of making an entrustment decision and deciding whether a learner is able to execute an EPA at a certain entrustment level.

Results from our study elaborate and expand on findings about entrustment decision-making in the literature [11, 16, 27], and add new insights and not yet listed decision variables. From previous research we know that decisions about supervision level are usually based on more than assessments of clinical skills [11, 16, 23, 27]. Clinical teachers assess the ‘trustworthiness’ of learners to act independently based on four categories: knowledge and skills, discernment, conscientiousness, and truthfulness [16]. We had to add a new category adequate exposure; these decision variables contain information about the learner based on similar or prior encounters, not specific about the learner or characteristics of the learner.

Previous studies have investigated variables needed to trust a learner to perform a critical task at a specific level of supervision [4, 16‐19, 33]. To gain sufficient trust to perform a task at a specific level of supervision (direct supervision, indirect supervision, distant supervision or no supervision), learners need to demonstrate these qualities and skills. This trust cannot be credited to a single quality or skill, but requires a multitude of variables [4, 8]. Extensive assessment and feedback by different observers on performed tasks is highly relevant: trust is multifactorial and highly contextual, it occurs under the broad constructs of task, supervisor, learner, and environmental factors, and is well described in prior work [4]. In line with previously published evidence on workplace-based assessment, direct observation of a task by a teacher was perceived as relevant [23, 30]. On the other hand, this study indicates that extensive reflection by the learner and feedback from different observers on the performed tasks are highly relevant, also with respect to the formative function of EPAs.

The need for robust information for entrustment decision-making aligns well with current thinking about programmatic assessment [34‐36]. In line with previously published evidence on workplace-based assessment, direct observation of a task by a teacher was perceived as relevant [30, 37]. On the other hand, this study indicates that extensive reflective behaviour by the learner and feedback from different observers on the performed tasks are highly relevant with respect to the formative function of EPAs.

A possible limitation of our study is that the focus groups were held in two schools of one university. On the other hand, the fact that it was held at both the veterinary and medical school broadens the generalizability of the results. Both groups of clinical teachers are from a similar (Dutch) educational culture, which is not necessarily transferable to other countries. The number of participants in each focus group was appropriate to allow for sufficient and balanced expression of views [28, 29]. Another limitation of this study could have been the trigger cases presented in the focus groups as they could have influenced the clinical teachers during the sessions. However, during the introduction of the trigger case, the moderator emphasized that this was just one example of an authentic clinical task and participants were explicitly encouraged to think beyond the trigger case.

The feasibility of the listed decision variables needs to be tested by their actual use in educational practice, i. e. whether teachers take these variables into account in entrustment decision-making. The variables could be used in ad-hoc entrustment decisions and as part of summative entrustment decisions. While we did not explicitly present this distinction during the focus groups, we found hesitation among participants to make critical decisions to grant responsibility when insufficient information was available [13]. Faculty development will be necessary to explain these types of entrustment decisions as being situational versus focused on summative assessment, when introducing workplace curricula based on EPAs [12, 14]; clearly in the latter case, our participants felt a need to have sufficient information to ground entrustment decisions. Further research could optimize the decision variables in the clinical workplace and could outline decision variables specific in ad hoc situations and variables that ground summative entrustment.

This study aimed to increase insights about relevant decision variables which are or can be used by clinical teachers to make an entrustment decision about learners in the workplace. These decision variables can assist clinical teachers in the entrustment process by supporting the multifactorial process of making an entrustment decision. Our participants felt a need to have sufficient information to ground entrustment decision, thus faculty development will be necessary to explain entrustment decisions as being situational versus focused on summative assessment, when introducing workplace curricula based on EPAs. Faculty development will be necessary to explain these types of entrustment decisions as being situational versus focused on summative assessment when introducing workplace curricula based on EPAs.