Article Text

Original research
Medical students’ responses to uncertainty: a cross-sectional study using a new self-efficacy questionnaire in Aotearoa New Zealand
  1. Ciara Lee1,
  2. Katherine Hall1,
  3. Megan Anakin2,
  4. Ralph Pinnock2
  1. 1Department of General Practice and Rural Health, Dunedin School of Medicine, University of Otago, Dunedin, New Zealand
  2. 2Education Unit, Dunedin School of Medicine, University of Otago, Dunedin, New Zealand
  1. Correspondence to Dr Ciara Lee; ciara.lee{at}


Objectives Responding well to clinical uncertainty is a crucial skill for any doctor. To better understand how medical students develop this skill, Social Cognitive Theory can be used to explore students’ perceived capability to respond to situations of uncertainty. This study aimed to construct a self-efficacy questionnaire and use it to measure medical students’ responses to clinical uncertainty.

Design A 29-item questionnaire was constructed. For each item, participants rated their confidence in responding to uncertain situations using a scale of 0–100. Data were analysed with descriptive and inferential statistics.

Setting Aotearoa New Zealand.

Participants The questionnaire was distributed to 716 of 852 medical students in second, fourth and sixth year, at the three campuses of the Otago Medical School.

Results The Self-Efficacy to Respond to Clinical Uncertainty (SERCU) questionnaire was completed by 495 participants (69% response rate) and found to be highly reliable (α=0.93). Exploratory factor analysis confirmed a unidimensional scale. A multiple linear regression model predicted self-efficacy scores from year of study, age, mode of entry, gender and ethnicity, F(11,470) = 4.252, p<0.001 adj. R²=0.069. Male students and those admitted to the programme 3 years postdegree or with significant allied health experience were predicted to have significantly higher self-efficacy scores. Year of study was not a significant predictor of average efficacy scores.

Conclusions Our research contributes a novel, highly reliable questionnaire that uses self-efficacy to measure medical student responses to uncertainty. The questionnaire revealed that students’ confidence in responding to uncertainty may be more related to their background and life experience than to progression through the curriculum. Medical educators and researchers can use the SERCU questionnaire to obtain a new perspective on how their students respond to uncertainty, inform future research and tailor teaching about uncertainty.

  • medical education & training
  • education & training (see medical education & training)
  • statistics & research methods

Data availability statement

No data are available. Consent to make data openly available was not sought at the time of data collection.

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See:

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


  • A strength of the Self-Efficacy to Respond to Clinical Uncertainty (SERCU) questionnaire is the use of Social Cognitive Theory to inform its design.

  • The SERCU questionnaire was rigorously developed using principles described by Bandura and found to be highly reliable.

  • Data collection involved travel across three campuses and distribution of the questionnaire in online form, which contributed to a high response rate.

  • Data were limited to a cross-sectional sample of medical students at a single university.


Uncertainty is a part of the clinical world in which we teach, learn and practice. Clinical uncertainty can be defined as the ‘dynamic, subjective perception of not knowing what to think, feel or do’.1 Responses to uncertainty can shape medical education and clinical practice.2–4 For example, maladaptive responses to uncertainty, such as avoidance and deferral of decision-making, can have an impact on career choices and the delivery of patient care.5–7 Poor tolerance of clinical uncertainty by medical students and clinicians has been linked to burnout, stress and anxiety.8 9 An understanding of how medical students respond to uncertainty can help us to envisage and shape the type of doctors that they might become.

Medical students’ responses to uncertainty sit within a broader framework of clinical uncertainty, as illustrated in figure 1.1 This framework was developed through review and synthesis of models of uncertainty in healthcare and medical education. In this framework, uncertainty is represented by three dimensions. The first is the source of uncertainty, which can be knowledge (either uncertainty due to the nature of that knowledge, the application of knowledge, or lack of knowledge), relationships and systems. The second dimension is subjective influencers. These are the unique factors that impact on how students may perceive and respond to uncertainty, including personal characteristics (personality, cognitive capacity and worldview), experience and affect. The third dimension is responses to uncertainty. Responses to uncertainty can be cognitive, emotional or behavioural, and they can be positive or negative. In this framework, the three dimensions of uncertainty are dynamically linked, meaning that sources, influencers and responses to uncertainty are not static and can change depending on the context.

Figure 1

Framework for clinical uncertainty. Image reproduced from Lee et al 2021.1

Our understanding of how medical students respond to uncertainty comes predominantly from research using quantitative instruments that explore students’ perceived tolerance of uncertainty or ambiguity.4 5 10–13 Quantitative instruments used to explore medical students’ perspectives on uncertainty may be inadequate due to lack of conceptual clarity, poor reliability, inconsistent results and abstract items.4 12–14 Another issue is the interchangeable use of the terms uncertainty tolerance and ambiguity tolerance.4 9 14 Concern over the suitability of these instruments have led one group of researchers to rewrite questionnaire items for a student population,15 while others have developed a new instrument, specifically designed for medical students.12

An understanding of how medical students report their tolerance of uncertainty when they respond to general statements or context-free questionnaire items gives us little indication of how they might respond to situations of uncertainty that they might encounter in learning environments which include clinical workplaces. For example, a student may report agreement with the statement that ‘I feel comfortable that in medicine there is often no right or wrong answer’ (TAMSAD item 17).12 The same student may, however, feel less comfortable informing a patient or healthcare worker that they are unsure about a diagnosis. Likewise, a student who answers that they agree with being ‘quite comfortable with uncertainty in patient care’5 may feel less so in the context of a busy ward with a distressed and unwell patient. While providing insights into how medical students tolerate the idea of uncertainty or ambiguity, these instruments leave a gap in our understanding of how capable students might feel when faced with examples of uncertainty in patient care contexts. Capability has been defined as ‘the individual’s psychological and physical capacity to engage in the activity concerned’, (Michie16 p4) which is the definition we use in this research.

Social Cognitive Theory can be used to explore medical students’ responses to situations of uncertainty that they might encounter in learning environments. According to Social Cognitive Theory, the functioning of individuals is determined by an interaction between their environments, their own behaviours and personal factors. These three elements interact in dynamic interacting ways, termed ‘triadic reciprocal causation’.17 The strength of influence that each of these factors may have can change depending on the people involved, the environment or type of activity. The concept of triadic reciprocal causation is compatible with the framework of uncertainty in clinical practice shown in figure 1, where sources of uncertainty including knowledge, relationships and systems can be modulated by subjective influencers (personal factors), resulting in a range of cognitive, emotional and behavioural responses.

Within Social Cognitive Theory, individuals act as agents with a range of capabilities17 and are seen to have a set of ideas about themselves that influence their thoughts, feelings and actions, known as self-efficacy beliefs.18 19 Self-efficacy beliefs are the beliefs that an individual holds about their capabilities to perform an action with the skills that they currently possess.18 Beliefs about self-efficacy are formed from an individual’s evaluation of the task itself, the context and themselves.18 20 Self-efficacy beliefs represent an individual’s confidence to perform specific tasks, not general characteristics such as personality.21–23 Sources of self-efficacy beliefs include enactive mastery (personal experience of a task), observation of others, verbal encouragement and social influence, and affective and physiological states (eg, a stress response or the impact of mood).18 These beliefs influence the effort individuals put into actions and whether individuals persevere with these actions.18 24 In order to attempt a specific task, an individual must believe (at some level) in their capability to achieve the task, making self-efficacy beliefs an important source of agency and driver of action.18 25 Individuals do not need to have experienced a task to possess beliefs about their capabilities to achieve it. For example, a person can possess a self-efficacy belief about their capability to successfully complete a triathlon without having participated in one, or a medical student can hold a belief about their capability to perform cardiopulmonary resuscitation without having previously encountered an emergency situation.

Consequently, examining medical students’ self-efficacy beliefs may be a useful means of understanding their capability to respond to clinical uncertainty. To date, there are no known instruments which have been designed to explore medical students’ self-efficacy beliefs about responses to uncertainty. Therefore, we aimed to use Social Cognitive Theory to develop a questionnaire to investigate how capable medical students feel to respond to situations of uncertainty in their learning environments. We also aimed to use this new questionnaire to measure the self-efficacy beliefs of medical students regarding their responses to uncertainty.


This cross-sectional study took place at the University of Otago, New Zealand. The MB ChB programme is structured so that all medical students study at the Dunedin campus until the end of their third year. From fourth to sixth year, students are based in either Dunedin, Wellington or Christchurch. Students enter the MB ChB programme through one of three pathways. The first pathway is through grades obtained in a 1-year health sciences course, termed Health Sciences First Year (HSFY). Second, students may enter via the graduate pathway, if they have graduated within 3 years of their application. Third, students may enter the programme through an alternative pathway, where applicants no longer qualify for the graduate category and/or may possess health-related professional experience.

To address the first aim of our study, questionnaire items were developed using principles suggested by Bandura.22 According to these principles, an effective self-efficacy questionnaire should have content validity, be specific to a domain of functioning, have gradations of challenge, have an adequate response scale, minimise responder bias through confidentiality and items should be analysed for reliability and correlation. The development of each item was also guided by the theoretical framework outlined in the introduction to describe a range of possible responses to situations of clinical uncertainty.

The framework of clinical uncertainty illustrated in figure 1 was used to inform the development of each questionnaire item. From this framework, we understood the sources, influencers of and responses to uncertainty as being dynamically linked. Responses to uncertainty can be cognitive, emotional, or behavioural, or a combination of these three. These responses can stem from sources including knowledge, relationships and systems and be modulated by subjective influencers. Questionnaire items were designed to include a combination of cognitive, emotional or behavioural responses evoked by the sources of uncertainty above. For example, in questionnaire item 3 (see online supplemental appendix A) the response to uncertainty described is ‘tell a patient when you don’t have an answer to their question’. This details a cognitive (acknowledging) and behavioural (communicating) response to uncertainty. The sources of uncertainty in this item include knowledge (not knowing answer to question) and relationships (the relationship held with the patient). Details of the sources of and responses to uncertainty for each of the final questionnaire items can be viewed in online supplemental appendix A.

For each item, participants were asked to use a scale of 0–100 to rate how confident they felt about enacting the response described. Self-efficacy questionnaires with a 0–100 response format have been shown to be more effective than those with smaller interval scales.22 26 Initial items were constructed by the first author and then refined following discussion with the other three authors. The first author brought perspectives from clinical experience in the NHS and as an undergraduate and postgraduate health professional educator in the UK and New Zealand. The second author brought experience from clinical practice, insights from teaching undergraduate and postgraduate healthcare students, and formal training and research interests in medical humanities, ethics and clinical reasoning. The third author brought perspectives from primary, secondary and tertiary educational practice and research, and personal experience working with clinical educators as a faculty developer. The fourth author brought experience from their roles as clinician, clinical educator and medical programme leader, coupled with a research interest in clinical reasoning.

A 32-item pilot questionnaire was completed by 22 medical students from the second, third, fourth and sixth year of the MB ChB programme. Data were analysed using descriptive statistics to ensure that participants were responding differently to questionnaire items. This allowed the authors to ensure gradations of challenge within the items where some items we viewed as more challenging, thus preventing a collection of items which were perceived as ‘easy’, producing an artificially high efficacy score from participants.22 Seventeen students attended focus group interviews where they were asked to give feedback on the questionnaire design and content. The questionnaire was also reviewed by an education researcher with expertise in self-efficacy and an experienced medical educator, both external to the research team. The focus group participants and expert reviewers provided feedback on their understanding of the pilot questionnaire items and their relevance, item wording, and questionnaire design. Feedback from these sources was also interpreted to support the construct validity of the questionnaire. This pilot phase led to the removal of four items, rewording of 12 items and addition of 1 item.

The final version of the questionnaire consisted of 29 self-efficacy items and included demographic questions about year of study, gender, age, mode of entry and ethnicity. Qualtrics software (version February 2019, Provo, UT, USA) was used to administer an online version of the questionnaire. This online questionnaire could be completed on student owned devices and responses were anonymised at the time of data entry. Anonymised data were stored on password-protected computers and servers. The Qualtrics software was configurated so that only one attempt to complete the survey could be made per device, and each unique attempt was assigned a code. Participants rated their confidence to respond to each item using a slider which illustrated the score out of 100. To provide participants with an understanding of how to rate the self-efficacy items, the questionnaire was prefaced with the following text: What follows are a number of responses to challenging situations that medical students face. Please rate how confident you are that you can do each of the things listed by selecting an appropriate number on the slider. When selecting a number, please rate your confidence to act with the skills that you have now. If you have not encountered one of the scenarios before, imagine how you would act if you encountered the scenario today.

The questionnaire was distributed to 716 of 852 medical students in second, fourth and sixth year, at the three campuses of the Otago Medical School between February and May 2019. Second-year, fourth-year and sixth-year students were chosen to participant as at the time of data collection they possessed a range of preclinical and clinical experiences and represented educational transition points. Second-year students were beginning their first year of the medical programme, fourth-year students were starting their first majorly clinical year and sixth-year students were in their final year of the programme and preparing for practice as junior doctors. Not all medical students could be contacted due to absence during the sessions when the questionnaire was distributed, including one-quarter of sixth-year students (approximately 67 students) who were overseas on their electives. The first author travelled to each campus to inform students about the study in person at lectures and tutorials. Students were provided with a link to access the online questionnaire. Prior to commencing the questionnaire, students were required to read an information sheet and provide consent via an online form. Where travel was not possible, the first author followed the same process using video conferencing software.

Questionnaire data were anonymised and analysed with descriptive and inferential statistics using IBM SPSS (V.25). Only participants who completed all self-efficacy items were included in the final analysis. An alpha level of 0.05 was used for all statistical tests. The Kaiser-Meyer-Olkin measure was performed to assess data suitability for factor analysis.27 An exploratory factor analysis was performed to explore the self-efficacy items using principal axis factoring and an oblique rotation. The decision on which factors to retain was made by examining the scree plot28 and consideration of conceptual interpretability.29 An item was considered relevant to a factor if the loading value was 0.32 or higher.30

To address the second aim of our study, a multiple linear regression was conducted to investigate if year level, age, gender, ethnicity and student mode of entry predicted the average self-efficacy scores of students participating in the study. To conduct this investigation, the categorical variables of year, gender, age groups, ethnicity and mode of entry were recoded as dummy variables. These variables were selected for exploratory purposes. As this was the first time that the questionnaire was used with medical students at the University of Otago, we did not want to hypothesise about the influence of variables. Instead, we wished to obtain as broad a picture as possible of the relationships between these variables and student responses to uncertainty.

Patient and public involvement

No patient or public involved.


Of the 716 students to which the questionnaire was distributed, 529 attempted the questionnaire and 495 completed it. The response rate of 69% was calculated using only those participants who completed all self-efficacy items. Participant demographics are presented in table 1. Nearly two-thirds of participants were female. Over two-thirds of participants entered the medical degree through the HSFY pathway. The median age of participants in year 2 was 19 (range 18–51), year 4 was 21 (range 19–40) and year 6 was 23 (range 22–40). Demographics were compared with data provided by the Otago Medical School describing the student population in 2019 (communication from Otago Medical School, November 2019, see online supplemental appendix B).

Table 1

Frequencies and percentages of participant demographics by year level

The results of the Kaiser-Meyer-Olkin and Bartletts’ tests indicated that the data were factorable. The overall Kaiser-Meyer-Olkin measure was 0.93 which is a classification of ‘marvellous’ according to Kaiser.27 Bartlett’s test of sphericity was statistically significant (p<0.0005). The correlation matrix was inspected, and all variables had at least one correlation coefficient greater than 0.3. Principal axis factoring was chosen as the factor extraction method due to the inter-correlation of the items and theoretical underpinnings of the item design.29 As factors were known to be correlated, an oblique rotation was applied. Principal components analysis revealed six components that had eigenvalues greater than 1 and which explained 36%, 7%, 5%, 4%, 4% and 4% of the total variance, respectively. Eigenvalues of less than 1 may reflect unstable factors.29 Inspection of the relative values of these eigenvalues using a scree plot indicated that one factor should be retained and the analysis was repeated to extract this one factor. Factor loadings for each item are illustrated in table 2.

Table 2

Self-efficacy to respond to clinical uncertainty questionnaire items and factor loadings

All items were retained and thus formed a unidimensional scale. This scale reflects and upholds the theoretical framework presented in the introduction where cognitive, emotional and behavioural responses to uncertainty are interlinked. This factor accounted for 33% of the total variance. The principal axis factoring was repeated with no rotation and varimax rotation to identify if a different factor structure was present, however, the structure remained very similar. The resultant unidimensional scale had high reliability (α=0.93). As the questionnaire items formed a unidimensional scale, it was possible to obtain an average self-efficacy score for each participant. Average scores based on the included demographic categories are described in table 3.

Table 3

Means and SD of average self-efficacy for medical students (N=495) by gender, ethnicity, mode of entry and year level

A multiple regression was performed to predict self-efficacy scores from year of study, gender, ethnicity, age and mode of entry to the medical degree course. Dummy variables were used in the analysis for the categorical independent variables. The baseline values for analysis were male gender, participants aged 18–23, entry to the programme through HSFY and New Zealand European ethnicity. There was linearity as assessed by a plot of studentised residuals against the predicted values. A Durbin-Watson statistic of 1.936 indicated independence of residuals. Visual inspection of a plot of studentised residuals versus unstandardised predicted values suggested homoscedasticity, which indicated that the variance of the predicted dependent variable was the same for all predicted scores.30 There was no evidence of multicollinearity, as all tolerance values were greater than 0.1. There were two cases where the standardised residual was greater than ±3 SD. The original data for each case was inspected and it was felt that the values were due to individual variation. The assumption of normality was met as assessed by a Q-Q plot. The multiple regression model statistically significantly predicted self-efficacy, F(11,470)=4.252, p<0.0005, adj. R2=0.069. Regression coefficients and SEs can be found in table 4.

Table 4

Multiple regression results for average self-efficacy scores

Year of study was not a significant predictor of average efficacy scores when fourth-year students (β=−0.095, t(469)=−1.931, p>0.05 and sixth-year students (β=0.076, t(469)=1.423, p>0.05) were compared with second-year students. Two variables added statistically significantly to the prediction, gender and mode of entry. Female students were predicted to have lower average self-efficacy scores than male students (β=−0.13, t(469)=−2.82). Students who entered medicine through the alternative pathway were predicted to have higher self-efficacy scores (β=0.19, t (469)=2.54).


The first major result of this study is that it was possible to use Social Cognitive Theory to design a questionnaire to measure self-efficacy beliefs about responses to clinical uncertainty. The Self-Efficacy to Respond to Clinical Uncertainty (SERCU) questionnaire is the first questionnaire, as far as we are aware, to use self-efficacy as a means of exploring medical students’ responses to uncertainty. Medical students are assessed on knowledge and competencies throughout their education. What we do not know, however, is how efficacious students feel to put their knowledge and skills into practice when faced with uncertainty. By exploring self-efficacy beliefs about responses to uncertainty, this questionnaire provides educators with insights that are not achievable with other quantitative instruments currently available. For example, knowledge of medical students’ tolerance of uncertainty or ambiguity5 10 does not tell us how capable these students feel to enact that tolerance. Using items which describe uncertainty-provoking scenarios that students are likely to face in clinical learning environments, our questionnaire helps to build a picture of how students perceive their capability to respond to uncertainty, rather than tolerance alone. Our questionnaire can complement instruments which have been designed to explore tolerance of ambiguity or uncertainty,12 as an understanding of self-efficacy beliefs may help to explain why students who are less tolerant of uncertainty might behave in maladaptive ways. Individuals with low self-efficacy scores may be more likely to avoid tasks.18 31 In the case of uncertainty, we know that avoidance is a common maladaptive response4 and may perpetuate a mindset where students ‘…act on their established self-beliefs without further reappraising their capabilities’.18 p19 This feedback loop may explain why students with low efficacy in responding to clinical uncertainty may avoid it, respond negatively or express a low tolerance; further decreasing their efficacy. The strategy of combining uncertainty tolerance and self-efficacy measures has recently been employed in a study which looked at how uncertainty impacts doctors’ perceptions of genomic testing.32 By combining elements of uncertainty tolerance measures and a specifically designed self-efficacy scale for genomic tumour testing, the authors were able to gain a broader perspective of doctors’ responses to uncertainty. Our questionnaire adds to the tools available to medical educators to help understand how their students respond to uncertainty.

Previously, self-efficacy scales used in medical education have been criticised for a lack of conceptual clarity and issues with measurement fidelity.20 Our questionnaire is strengthened through a design which follows the principles recommended by Bandura,22 as outlined in the methods. One criticism of the use of self-efficacy scales in education is the over-generalisation of such scales.23 Though a generalised scale has been recently used in the context of medical students and uncertainty,33 the appropriateness of these scales has been debated,22 31 as the items have not been designed around specific tasks within a domain of functioning, which is not consistent with Social Cognitive Theory. In our research, a clear conceptualisation of what uncertainty in clinical practice means (figure 1) was used to ensure that the developed questionnaire was domain-specific, rather than a generalised assessment of self-efficacy.1 Within this conceptualisation, we understood that students can face uncertain situations with a dynamic combination of cognitive, emotional or behavioural responses. The interlinked nature of these responses to uncertainty is reflected in the factor analysis results which indicate a unidimensional scale, despite questionnaire items encompassing a wide range of responses to uncertainty (table 2). The high response rate indicates that participants engaged with the questionnaire content. The strong theoretical framework, rigorous content development and refinement, response rate, and evaluation results including high internal consistency, provide support that this questionnaire is feasible to administer and for medical students to interpret to meaningfully provide a useful measure of clinical uncertainty to researchers.

The second major result of our study was that our use of this new questionnaire produced three noteworthy findings about students’ efficacy beliefs in relation to uncertainty in their learning environments. First, there was no statistically significant difference in self-efficacy scores in response to uncertainty when second-year students were compared with fourth-year and sixth-year students. The lack of difference in efficacy scores between year groups suggests that medical students’ confidence to respond to uncertainty may not change as they progress through their medical education. Our results can be interpreted to suggest that there could be missed opportunities for teaching and learning about responses to clinical uncertainty within the complex environments that students are learning to navigate. An alternative explanation could be that the impact of subjective influencers on responses to uncertainty (personal factors, experience, affect) is not accounted for when medical students are taught about uncertainty. These findings show how our questionnaire can be used to identify gaps in curricula, in this case, a potential missed opportunity for devising strategies to enhance students’ self-efficacy with respect to uncertainty. Second, medical students who entered our MB ChB programme through the alternative pathway had significantly higher average self-efficacy scores than students who entered via other pathways. Alternative pathway entrants at our university are generally older and bring with them a range of life experiences, such as previous qualifications and careers in other health-related professions. These students may, therefore, have already developed skills and techniques to manage uncertainty and perceive themselves as more capable to respond to uncertainty. It has previously been shown that mature medical students are more comfortable transitioning to clinical learning environments,34 which can be interpreted as a time of great uncertainty, which reflects the higher self-efficacy scores. Third, male students had higher average self-efficacy scores than female students. It is unclear why this difference in efficacy beliefs exists. This finding merits further exploration because studies which have explored gender differences in students’ responses to uncertainty and ambiguity have shown inconsistent results.15 35–39 Overall, our findings show how this questionnaire can be used as a first step in exploring student perspectives on uncertainty. These findings can then be used as a springboard for further research, or as an indication of areas to target future teaching.

According to Social Cognitive Theory, self-efficacy beliefs are not fixed,18 22 25 and this perspective can provide educators with opportunities to facilitate students to feel more capable to respond to uncertainty. A practical use of our questionnaire is to direct educators to areas where more teaching about uncertainty is required. Educators could then incorporate knowledge about the sources of self-efficacy beliefs (mastery, modelling, verbal persuasion, affective states) into their teaching about clinical uncertainty. Mastery experiences are believed to be the most powerful source of self-efficacy beliefs.18 Educators could support medical students to enact responses to uncertainty by inviting them to voice uncertainty on a ward round (questionnaire item 16) or encouraging students to discuss uncertainties with other health professional staff (items 7, 11, 18 and 22). Or educators could model how they respond constructively to uncertainty in clinical learning environments by admitting uncertainty during patient consultations (questionnaire items 3, 21 and 25) or to other members of the clinical team (item 13). Supporting the development of self-efficacy beliefs through verbal persuasion could take the form of constructive feedback on the above behaviours. Being mindful of the impact of affective state on self-efficacy, educators could provide students with opportunities to practice their responses to clinical uncertainty in a safe and supportive environment. This environment could be created in clinical settings, for example, by using the modelling behaviour described above and addressing questionnaire items 4, 14, 19, 23 and 27, or during simulation exercises.40

By incorporating knowledge about self-efficacy beliefs into teaching strategies and learning opportunities for students, educators can help students become aware of their own beliefs about their own capabilities and help them improve their responses to uncertainty. By informing the development of the questionnaire with Social Cognitive Theory, the results can be used to inform theory-based curriculum change. A recent review highlights the universality of uncertainty in health professional education and the limited teaching tools available to medical educators.41 We believe that a further practical use of our questionnaire could be as a useful measurement tool to assess the impact of teaching interventions or curriculum changes to the self-efficacy beliefs of medical students about their responses to uncertainty, and we will investigate this in the future. The insights gained from our research about how Social Cognitive Theory and self-efficacy can be used to understand responses to uncertainty can be used by undergraduate and postgraduate educators in other health professions. The items in our questionnaire could be adapted to suit different specialties and clinical settings. To do this, it is important to consider and adhere to Bandura’s principles of self-efficacy questionnaire design. If questionnaire items require adaption, the theoretical underpinnings of each item should remain. The dimensions of uncertainty (sources of uncertainty and type of response) described for each item in online supplemental appendix A should remain unchanged. The context, however, can be adjusted to make the item more relevant to the intended population. For example, item 24 could be rewritten for nursing students to read: ‘Tell a senior nurse when you think that the wrong medication has been prescribed for a patient.’ Item 17 could be rewritten for physiotherapy students to: ‘Manage feelings of anxiety when practising a clinical skill (eg, complete a subjective or objective assessment) in front of other students.’ Revised versions of the questionnaire could be strengthened by piloting to avoid assumptions about specialties, alongside testing of internal consistency. A further consideration is how medical students from different cultural backgrounds may respond to uncertainty differently. For example, medical students in Aotearoa New Zealand and Australia may be less avoidant of uncertainty than students from other parts of the world, such as China or Indonesia.42 These differences could impact the results of the SERCU questionnaire or influence how the questionnaire may need to be modified for specific populations.

This study is limited by being a cross-sectional design at a single university. The division of students across three different campuses, however, meant that a broader set of experiences related to encountering uncertainty was obtained despite this limitation. A longitudinal follow-up of participants could be conducted to perform a confirmatory factor analysis to verify the factor structure. We debated the exclusion of item 1, which had a factor loading of 0.32. Though a loading of 0.32 is acceptable,30 other studies use a rule of thumb where factors with a loading of 0.40 are excluded.43 Theoretically, item 1 explores a similar response to uncertainty as items 7, 13, 21 to satisfy Bandura’s development criteria of including different levels of challenge within a questionnaire. As items 7, 13, and 21 had acceptable factor loadings (0.58, 0.65 and 0.56, respectively) we chose to include item 1. The high Cronbach’s alpha score, although an indicator of excellent reliability, could mean that some items were redundant.44 As the response rate was high,45 and this was the first time the questionnaire was used, we decided not to reduce the number of items. We did not follow up with participants who did not complete the questionnaire to understand their reasons for non-completion. This may have been a missed opportunity to further improve the design of the questionnaire. This study took place prior to the emergence of COVID-19, which has been a source of uncertainty for students and educators around the world. A future study could explore how medical students’ self-efficacy to respond to uncertainty has been impacted by the pandemic. There have been recent calls for more qualitative research on uncertainty in medical education.13 Our questionnaire could be used to identify where qualitative enquiry might yield further detail on student perspectives, for example, to further investigate subsets of students who show significant differences in self-efficacy scores. The findings from our questionnaire provide a starting point towards understanding medical students’ beliefs about how they might act when faced with situations of uncertainty in their learning environments.


We used Social Cognitive Theory to inform the development of a novel and highly reliable self-efficacy questionnaire to explore medical student responses to uncertainty, and initial evidence to support its content and construct validity. The results from the first investigation using this questionnaire can help educators to understand how capable medical students feel to respond to situations of uncertainty with the knowledge and skills that they currently possess. Alongside providing a new perspective on clinical uncertainty, this questionnaire can be used to complement existing tools or guide qualitative enquiry, helping to build a clearer picture of the impact of uncertainty on medical students. When used with medical students at the University of Otago, this questionnaire identified a lack of difference in self-efficacy scores of medical students when responding to situations of uncertainty at three points in their medical education, with two notable exceptions. Students who enter the MB ChB programme via an alternate pathway and male students had higher mean self-efficacy scores for clinical uncertainty. These findings may be used to guide further inquiry about the impact of prior experiences and gender differences in medical students’ responses to uncertainty. Our questionnaire provides us with a starting point to consider how we might improve and tailor teaching about uncertainty, which students are certain to encounter in their education and future careers.

Data availability statement

No data are available. Consent to make data openly available was not sought at the time of data collection.

Ethics statements

Patient consent for publication

Ethics approval

This study involves human participants and ethical approval for the study was sought and granted from the University of Otago Human Ethics Committee (D18/365). Participants gave informed consent to participate in the study before taking part.


The authors wish to thank Dr David Berg and Professor Tim Wilkinson for their insights during development of the questionnaire items. The authors also wish to thank Michel De Lange for his advice on statistical analysis.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.


  • Twitter @ciaralee

  • Contributors All authors contributed to the conception and design of the work. CL distributed the questionnaire. MA advised on analysis, which was performed by CL and all authors contributed to interpretation of results. CL drafted the manuscript and KH, MA and RP were involved in critical review and the revision process. CL is the guarantor of the article. All authors gave final approval to the submitted version of the manuscript.

  • Funding CL was supported by a University of Otago doctoral scholarship. Award/Grant number: N/A.

  • Competing interests None declared.

  • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.