Research has shown that there is a correlation between medical students’ exposure to the humanities and a higher level of positive personal attributes such as emotional appraisal, empathy and tolerance of ambiguity, and lower levels of negative attributes such as signs of burnout [1]. In particular, higher levels of empathy may lead to better clinical outcomes. Intolerance of ambiguity has been shown to correlate with increased psychosocial distress and poorer clinical performance [2]. These findings underscore the importance of these skills for healthcare professionals and the need to promote them in medical education.

Worldwide, there are several programs involving collaboration between medical faculties and art museums. In the Netherlands, several art-related educational programs have been developed in all medical schools, varying in duration, design, and content [3].

The Amsterdam Universitair Medisch Center (AUMC) has developed an art-based observational training program for medical students and residents, in collaboration with the Rijksmuseum in Amsterdam [4,5]. The program consists of a structured course at the Rijksmuseum by an art educator and a clinical specialist. We investigated whether this form of art-based education improves empathy and judgement skills (tolerance of ambiguity) in participants.

The aim of this study was to evaluate the effects of art-based observational training on medical students, residents, and specialists using quantitative methods to assess the competencies of empathy and tolerance of ambiguity. The hypothesis underlying this study was that art-based observational training would have a positive effect on these competencies.

Informed consent was obtained from the participants. This was not a clinical trial involving a health risk; therefore, institutional review board approval was not required in the Netherlands.

An experimental, single-group, pre- and post-test design was devised. Validated questionnaires were used to examine the outcomes of art-based observational training in medical students, residents, and specialists. The study was described according to the Transparent Reporting of Evaluations with Nonrandomized Designs (TREND) statement available at: https://www.cdc.gov/trendstatement/index.html.

Between December 2023 and June 2024, 15 two-hour courses were held at the Rijksmuseum in Amsterdam, the Netherlands (Supplement 1). Follow-up took place after the sessions were completed. The study’s design did not include a control group and focused only on the participating cohorts.

Participants were recruited voluntarily through the university faculty and the network of the AUMC. The inclusion criteria required that the participants were medical students, doctors, or other health professionals with regular patient contact. All participants had to be proficient in Dutch, and all were exposed to the same, 1-time intervention.

Due to the limited number of participants, it was decided not to exclude the results of the participants with missing responses in either the pre- or post-test. The corresponding question was not included in the analysis.

Groups consisted of 6 to 8 participants. Each group was led by an experienced clinician (T.v.G.) trained in visual thinking strategies, by the Foundation of VTS Netherlands (Stichting VTS Nederland), and a medical doctor who was studying for a master’s degree in History of Art (S.A.B.). VTS (or parts of the VTS program) were used in the structured sessions. The group discussions were adapted to the participants’ medical background [4]. The artworks had different tasks, but all were selected on the basis of an assumed physical abnormality or otherwise medical feature. The 4-stage ABCD method of art-observation was used, as previously reported (Table 1) [4,5].

Pre- and post-intervention tests were administered using validated and standardized scales. The pre-intervention test had to be completed online by the participants before the museum visit and the post-intervention test was administered during the post-intervention evaluation.

Empathy and tolerance of ambiguity were assessed using the Jefferson Scale of Empathy and the Tolerance of Ambiguity in Medical Students and Doctors scale (TAMSAD), respectively [6,7]. The JSE was used in this study with permission from the Thomas Jefferson University. The TAMSAD is in the public domain. The JSE is a 20-item scale designed to measure empathy in health professionals [6] (Supplement 2). Items are scored on a 7-point Likert scale assessing 3 components of empathy—namely, perspective taking, compassionate care, and the ability to put oneself in the patients’ shoes. The JSE is scored between 20 and 140 points, with higher scores indicating higher levels of empathy [6].

The TAMSAD contains 29 items measuring tolerance of ambiguity in medical students and doctors [7]. Most of the items in the questionnaire refer to clinical contexts (Supplement 3). The items are answered on a 5-point Likert scale. The TAMSAD is scored from 29 to 145 points, with higher scores indicating greater tolerance of ambiguity. This score is calculated as a mean score out of 5 across the 29 items and then transformed to a score of 0–100 score using the formula: new score=25(old score–1) [7]. In both questionnaires, half of the items are negatively worded, for which the scores are reversed.

In addition to the standardized questionnaires, participants were asked to complete a qualitative evaluation form with 9 questions that could be answered on a 5-point Likert scale, with the possibility to elaborate on the answer (Supplement 4).

Participants were recruited voluntarily, which may introduce selection bias and confirmation bias. This is considered in our discussion.

Post-hoc power analysis was conducted to evaluate the difference between 2 dependent means (matched pairs) for the total scores of the JSE and the TAMSAD scale using G*Power (Heinrich-Heine-Universität Düsseldorf). The calculated power (1-β error probability) was 0.851 for the JSE and 0.893 for the TAMSAD. The analysis used the following parameters: a one-tailed test, effect size of 0.550 for the JSE and 0.342 for the TAMSAD, and total sample sizes of 29 (JSE) and 62 (TAMSAD). The effect size for the JSE was derived from the mean and standard deviation of the pre-test group (110.83±12.56) and the post-test group (114.53±12.04), with a correlation coefficient of 0.851 between the paired groups. Similarly, the effect size for the TAMSAD was calculated based on the pre-test mean and standard deviation (100.62±8.99) and post-test values (102.71±9.99), with a correlation coefficient of 0.798 between the groups. These parameters demonstrated robust statistical power for detecting differences in both scales, ensuring the reliability of the findings.

Participants were recruited among students of the medical faculty of the University of Amsterdam and the network of the AUMC and assigned to the museum for a planned course. The first group received the JSE, and the second group completed the TAMSAD scale.

A blinded group was not possible in this study, nor was a control group included.

The unit of analysis was each group of participants who took 1 of 2 surveys.

Statistical analyses of the collected data were performed using IBM SPSS Statistics ver. 29.0 for Windows (IBM Corp.). SPSS was used to analyze the pre- and post-tests for both the JSE and TAMSAD using the paired t-test. Significance was defined as P<0.05, and results were rounded to 2 decimal places.

In total, 92 medical students, residents and specialists participated in the art-based program at the Rijksmuseum, with 1 participant being excluded. This resulted in 91 participants being included in the analysis: 63 women and 28 men. The participants represented various levels of medical training and career stages.

Thirty participants were included in the JSE group; 1 participant did not complete the post-test and was excluded from this study. Twenty-nine participants completed the JSE, and 1 participant had a missing response on the pre-test. Sixty-two participants completed the TAMSAD. Six participants had a missing response on either the pre- or post-test. The characteristics of the participants are presented in Table 2.

Overall, the post-tests showed significant increases in mean and total scores for both the JSE and TAMSAD (Figs. 1–3). There were no statistically significant differences in total scores by gender, medical speciality, or function.

The response rate to the online evaluation form was 100%, and all 91 participants completed the questionnaire. Overall, the participants reported positive reactions towards the art-based program (Fig. 4). Some of the participants’ responses are highlighted below:

Our study, which was conducted among medical students, residents, and specialists, showed that the art-based observational training in the Rijksmuseum in Amsterdam significantly increased empathy and tolerance of ambiguity. The mean difference between the pre- and post-intervention tests was 3.71 points for empathy, ranging from 20 to 140, and 1.86 points for tolerance of ambiguity, ranging from 0 to 100. We did not observe any significant differences in the results between participants in terms of gender, medical speciality, or function.

Our qualitative analysis showed positive responses; most participants reported that the art-based observational training program was useful in terms of medical education and professional development. The majority of the participants felt that the course helped them to become more aware of their own judgements and biases, and to improve their ability to delay judgement in healthcare situations. In addition, many participants found that the course improved their ability to empathize with patients, while at the same time promoting communication and teambuilding.

This experimental study lends support to the view that the inclusion of art-based observational training in medical education may be valuable. Although the results showed slight increases in both empathy and tolerance of ambiguity, the results were statistically significant. This small difference may be explained by the fact that there was a single intervention that lasted for only 2 hours, providing only brief exposure to the art-based observational training program.

Several studies investigating empathy with the use of the JSE in art-based education have found no significant changes, while some studies found differences in subcomponents of empathy [8]. We found a significant increase in the total score, as well as increases in the perspective taking and compassionate care subscales. A possible explanation for these differences is that the sample size for testing empathy was relatively small. In addition, our course was designed differently from previous research, which may have led to different results.

In terms of tolerance of ambiguity, most studies have used Geller’s Modified Tolerance for Ambiguity Scale [9-11]. Some of these studies found significance, while others did not. More recently, the TAMSAD has been developed. Although evidence of validity has been reported in 3 studies [12], Stephens et al. [13] found the evidence for validity to be weak but growing. As our study is the first to use the TAMSAD in this application, we cannot compare our results with previous research. Nonetheless, our study’s results are similar to those obtained using Geller’s scale [9].

The findings from our qualitative analysis align with previous research [9,14], in which the participants reflected positively on the art-based program.

First, we did not use a specific scale to quantify whether art-based education would have a positive effect on various situations in healthcare-related skills of medical students and doctors. We decided to focus on 2 parameters that have been used in literature on art-based observational training: empathy and tolerance of ambiguity. We chose these 2 competencies because they have been shown to improve after art-based education using standardized and validated questionnaires. Another skill that has been often described in the literature is observational skills. Our course aimed to improve observational skills, but as there are no validated scales or other methods to quantify this outcome, we chose to include it in our qualitative questionnaire, although the latter is subjective to several factors. The responses may have been influenced by participants remembering previous answers.

In particular, most of the participants were female (63 women and 28 men), which is in line with the female predominance in medical training and specialization in the Netherlands. In addition, only 29 participants were included in the group in which empathy was tested. One participant was excluded, meaning the predetermined target of 30 participants could not be reached.

As the participants were recruited voluntarily, there was selection bias. Most of the participants (75.8%), had an affinity with art before joining the program. People who regularly visited museums were more likely to be enthusiastic and to enroll in this program. Another possible limitation of this study is that the participants’ functions and levels of training were different; some were bachelor students, and others were already professionals. However, no statistically significant differences were found between functions.

Because we did not include a control group, we could not test our results against a group that did not receive art-based education.

Finally, the participants were only exposed to a single 2-hour intervention. The effects of the intervention were assessed immediately after the session. The results therefore show the short-term effects of a short intervention.

The 2 measurement tools used in this study allowed us to draw conclusions about both empathy and tolerance of ambiguity in both medical students and doctors. The quantitative results confirmed the qualitative responses, adding to the strength of this study. Our art-based program could be used in other medical universities and local museums, anticipating similar outcomes.

Future research should increase the number of interventions and follow-up participants over a longer period of time, to assess long-term effects. The inclusion of a control group not exposed to the program would increase the strength of the study.

Random allocation of participants would result in an unbiased group population. In order to increase the number of participants in order to make a stronger claim about the effects of art-based observational training in the Netherlands, it is recommended that the programs of the various faculties and museum collaborations be expanded and coordinated, for which the Rijksmuseum program can be used as a model.

The ultimate aim, albeit remote at present, is to include art-based observational training in all the curricula of all medical faculties in the Netherlands, ideally, at the beginning of the master’s program, before the start of clinical internships. In this way, students will be trained at an early stage to approach patients in an open-minded and empathic way.

We evaluated art-based observational training among Dutch medical students, residents, and specialists in the Rijksmuseum in Amsterdam. Two measurement tools, the JSE and the TAMSAD, were used in pre- and post-intervention tests. Our results showed small but statistically significant increases in both empathy and tolerance of ambiguity after the intervention. In addition to the quantitative analyses, the qualitative evaluations also showed positive results as well. Participants were enthusiastic and found the course meaningful and a welcome addition to their regular medical training, where they are constantly challenged to observe, think, and communicate unambiguously. These findings suggest that incorporating art-based observational training into medical education has a beneficial effect on empathy and tolerance of ambiguity.