The quality of the interaction between physicians and patients influences patient outcomes in clinical settings (1, 2). A fundamental pillar of the healthcare mission is based upon compassionate care that forecasts greater patient trust and satisfaction, superior patient-physician connections, and better patient outcomes (2, 3). Within healthcare organizations, compassionate care entails diverse responsibilities of healthcare professionals in explaining their roles, establishing rapport, and in spending time on attentively listening to the patients' concerns (4). Compassionate healthcare—a key competency is characterized by reflections of others' concerns, respect for persons, and contextualized understanding of the patient as a key player of healthcare systems (5).

The two leading elements of patient centered care include empathy and compassion. Empathy refers to “a capacity to understand but without joining the feeling of the patient” (6) while compassion refers to “the feeling that arises in witnessing another's suffering and that motivates a subsequent desire to help” (7). In the medical field, empathy pertains to a multifaceted strand with moral, cognitive and behavioral enlightenment (8). When applied with objective reasoning, empathy promotes the impact of medical care and facilitates physician-patient communication (9). Empathy, being a multi-construct concept, includes, but not limited to, physician-patient interactions, interprofessional practice, self-compassion, empathy for others' pain, professional identity formation, stress awareness, and self-reflection and communication. Empathetic communication in patient-physicians interactions fosters information exchange and the impact of understanding and adherence to management plans, which lead to an early return to work, pain relief, mood elevation, and improved functional status of patients (10). Both empathy and compassion are generally considered to be interchangeable terms. However, empathy (understanding of patient feelings) is necessary to trigger compassion (emotional response including actions to alleviate patient sufferings) (11).

There is a compelling evidence in literature that treating patients without compassion can lead to deleterious outcomes (12). An absence of compassionate care results in poor quality of care and higher risk of complications through medical errors (13). Unfortunately, despite an explicit emphasis on the vital role of compassionate patient care, healthcare professionals often miss opportunities to be compassionate, rather they pay attention to biomedical data and management plans. From another perspective, sustaining compassion in medicine is hard as “compassion fatigue” prevails in ~20–70% of healthcare professionals (14). Compassion fatigue refers to loss of compassion in healthcare providers due to work related stress (15). Compassion fatigue certainly impairs the ability of healthcare providers in providing compassionate care which is a fundamental pillar inpatient-centered service (16). Educators have argued that a great majority of medical students enter health care with a strong set of espoused ideals for providing high-quality, patient-centered care (17). Unfortunately, during their training, medical students witness dissonance between the personal and professional conducts of their faculty and clinical teachers, students become more frustrated, less empathetic and more distanced from patients (18).

By and large, the ethos of most health care professional curricula uses typical biomedical models that primarily focus on teaching, training, and practice of clinical medicine with less emphasis on patients' psycho-social well-being (19). In the absence of a rigorous integration of biomedical knowledge with the understanding of human behaviors will potentially further detach medical and health sciences students and healthcare professionals from the patient's emotions and contextualized perspectives (20). Likewise, empathy plays a vital role in interprofessional practice where physicians from various disciplines work together toward the safe and effective patient-related clinical outcomes by shared decision-making and by regular consultations among healthcare teams and patients (21). This interprofessional practice alleviates patients' anxiety, stress, and uncertainty. From another perspective, self-awareness and self-regulation with empathy facilitates patient-doctor therapeutic interactions by evolving physicians' self-discovery (22). Among the practicing physicians, the phenomenon of self-discovery is an ingredient to professional identity formation (23). An empathetic professional identity formation is considered to be an effective tool for therapeutic actions in the healthcare system (24). In summary, empathy embodies a constellation of inter-related and interchangeable attributes that collectively lead to improved positive patient-related outcomes and the quality of care.

In clinical practice, a compassionate care has shown a strong association between improved clinical outcomes, quality of life and well-being (25). Among health professionals, empathetic, and compassionate care helps in mitigating the risk of burnouts (26). From the neuroscience perspective, a study on the functional magnetic resonance imaging has shown that empathy activates distinct pain centers of the brain; whereas focusing on compassion activates the reward pathways (27). These findings propound that a synchronous teaching and practice of empathy and compassion enhance patient and clinician well-being. Such outcomes may be accomplished with great success if educators can inculcate structured training programs for empathy and compassion in undergraduate and residency programs. However, there is a compelling evidence of empathy decline during the course of educational and training programs (28) and there is an urgent need for developing evidence-based curricula that can secure a sustained change in attitudes and behaviors (29).

A wealth of teaching pedagogies has been used to develop empathy, compassion and respect for patients in medical students (30–33) These include, but not limited to, patients' narratives and creative arts, drama workshops, communication skills, reflective writing, video-based learning and experiential learning (34–37). Lastly, the impact of role modeling and hidden curriculum in teaching compassion and empathy is well-established (38–40). There is little evidence about how the learning contexts, including the structure and delivery of medical curricula, influence the understanding of medical students about empathy and compassion (41). However, some patient-reported studies about compassion and empathy have shown the influence of educational interventions on medical students in enhancing their approach toward patients and their families (42).

The analysis of existing body of literature emphasizes a need to introduce a standard teaching modalities within medical curricula that can enrich traits of empathy and compassion in undergraduate medical students. Currently, there is scarce evidence of horizontal or vertical integration of teaching programs of empathy and compassion in undergraduate medical curricula (43, 44). Furthermore, there is little evidence about the effectiveness and quality of teaching programs for empathy and compassion. We conducted this systematic review of the literature to summarize and report the published work on educational interventions for empathy and/or compassion curricula in undergraduate medical students. This review also aims to highlight best practices to implement an evidence-based empathy and/or compassion curriculum in undergraduate medical training.

In our study, we used the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (45). The PRISMA tool provides an evidence-based minimum set of data for a standard reporting in systematic reviews and meta-analyses.

Figure 1 outlines the workflow and algorithm of studies selection in this study. Our initial search yielded 2,861 studies. After removing duplicates, we screened 754 titles and abstracts. This led to the identification of 244 articles using PICO for a detailed full-text review. This helped us to exclude another 227 articles whose content did not meet the inclusion criteria. We reviewed reference lists of the included articles to identify additional studies for potential inclusion, used electronic citation tracking, and consulted the librarian. This hand search yielded 15 review articles with 7 additional articles for inclusion. Finally, authors agreed on a list of 24 articles (23, 51–73), that had used 24 unique educational interventions either de novo, validated or adapted from previous publications and explicitly met inclusion criteria of our research. The summarized and comprehensive information about each article is presented in Table 1.

A maximum number of 9/24 (37.5%) articles were published in 2017, while 11/24 (45.8 %) studies originated from the USA. The sample size varied greatly across the selected studies, ranging from 10 to 299 participants. Our research included a total of 2,657 respondents with a mean of 110. Table 2 outlines the range of study designs that were used in the selected 24 studies. Most studies (21/24; 87.5%) were single-center, 11 (50%) were randomized controlled trials (54, 56, 57, 59, 60, 63, 64, 68, 71, 72), 4 (16.6%) controlled trials. (51, 52, 61, 65) Five (20.8%) studies measured outcomes with single group pre- and post-tests. (58, 62, 66, 67, 69) Our research identified 22/24 studies with their reported quantitative data that allowed us to calculate their MERSQI scores as shown in Table 3. All studies used statistical means to evaluate effectiveness. Collectively all studies have a total MERSQI score of 283 with a mean of 12.8. The study by Foster et al. (60) secured the highest MERSQI score of 17 from a maximum of 18. Finally, 10/22 (45.5%) studies scored ≥13 on MERSQI checklist.

Most curricula (20/24, 83.3%) used multiple educational modalities and 90% of studies turned out to be effective in achieving their primary outcomes (Table 4). A variety of teaching modalities were employed; didactics (53, 59, 61, 62, 64, 66, 70, 71, 73), small-group discussions (56–58, 61, 63, 66–68), reflection exercises (23, 56, 58, 66, 70) and simulations. (53, 57, 59, 62, 66, 68, 72) Likewise, virtual hangouts and technology-enhanced interventions (52, 58, 60, 69) were also significantly effective. A great majority of studies were conducted in multiple sessions within one academic year.

Only one study instituted the Balint training groups (57), while four studies (54, 59, 64, 71) used adapted mindfulness based curricular tools (75% effective ratio) developed by Kabat-Zinn (74). We did not observe any clear association between duration, frequency and complexity of an educational intervention and its effectiveness. For example, 75% (3/4) of single modality curricula were effective while multimodality curricula turned out to be effective in 90% (18/20) of the selected studies. Our data showed that single cross-sectional curricula (4/5) were as effective as longitudinal curricula (3/4) as shown in Table 4.

According to the four-level outcome-based Kirkpatrick model, 14/24 (58.3%) studies had a primary outcome pitching on level one; self-reported changes in attitudes and behavior (23, 53–56, 58, 61, 62, 64, 65, 67, 70, 71, 73). Two (4.8%) studies evaluated level three outcome (behavior), while six (25%) studies evaluated level four (patient outcomes) either by standardized or simulated patients or by a third party observation (51, 57, 60, 63, 66, 72). The leading constructs used in education interventions that showed positive impact on empathy and compassion of medical students in our study include communication skills, mindfulness, early clinical exposure, technology-enhanced learning, and humanities. A maximum of 10 studies used communication skills (53, 54, 56, 57, 60, 62, 63, 66, 71, 72), followed by mindfulness by 5 studies (54, 59, 64, 71), early clinical exposure by four (23, 51, 55, 65), technology-enhanced learning by virtual patient hangouts, computerized tasks, hot spotting by another four (52, 58, 60, 69), and comics (70), and arts and culture (61). Table 5 outlines an inventory of the validated instruments for outcome assessments used in the selected studies. The most commonly used self-assessment outcome tool was the Jefferson's Scale of Empathy by 15/24 (62.5%) studies (52–54, 56–58, 61–63, 65, 67, 69–72) that recruited 1,973 students. Conversely, 6/24 (25%) studies (54, 56, 59, 64, 69, 71) used a combination of tools to probe the impact of their interventions on behaviors and attitudes of medical students toward empathy and/or compassion and to determine the impact of self-compassion on behavior change. (59) Only three studies looked for the long-term effects of educational interventions (53, 62, 71).

Our study elucidates major findings of 24 studies that have collected data from 2,657 medical students using educational interventions on empathy and compassion. Overall, there is a positive impact of teaching empathy and compassion on medical students using a wide range of teaching pedagogies. The body of evidence from our research has deduced that effective communication skills, mindfulness, early clinical experience, comics, arts and culture, and technology-enhanced learning by virtual patients, hangouts and hot spotting had a positive impact on empathy and compassion of medical students.

The overarching findings in our study underpin a need to teach empathy using the concepts of physician-patient interactions, interprofessional practice with professional identity formation, stress awareness, and self-reflection and communication. In our study, a substantial body of evidence has stressed on empathetic communication for better patient outcomes (51, 57, 60, 63, 66, 72). Communication is the foremost determinant of a safe clinical practice and ensures satisfaction of both patients and healthcare providers (75). The outright benefits of timely and professional communication in the medical field are far ranging; greater patient understanding and compliance, superior clinical outcomes, improved patient safety and alleviation of patient stress (76). Kelley and Kelley have argued that verbal communication between a healthcare professional and a patient enacts an empathetic connection to the patient that carries a powerful transformative and healing power (77). Pity, sympathy, fellow-feeling, comfort, commiseration, assuagement, and reflexive encouragement constitute effective strands of practicing empathy and compassion (78). The identification of compassionate opportunities, conformation, and appreciation with pursuit provide a practice of compassionate and empathetic care (79). In the study by Beard et al. the researchers measured patients' satisfaction about clear communication by medical students comparing the Veterans Affairs Longitudinal Undergraduate (VALUE) program with a control group of patients matched with disease severity (51). The results of this longitudinal study on students' involvement using VALUE program (patient education, communication, and collaboration) and a control group without the VALUE program. The VALUE students showed significant improvement in care coordination among their patients and physicians. Educators should provide regular and sustainable opportunities to medical students to develop and validate their interpersonal communication skills that can potentially improve empathetic communication. The study by du Vaure et al. (57) used the Consultation And Relational Empathy Measure (CARE) scale in a two-site randomized controlled trial on medical students in a weekly Balint group forum for 2 months. Results of this group were compared with the group of routine education. There was an insignificant difference in mean CARE score (Intervention vs. control groups) however an increase in Jefferson Scale of Empathy (JSE) score for intervention and decrease in score for control from baseline to follow-up was recorded. LoSasso et al. have studied the impact of SALTED (Set-up, Ask, Listen, Type, Exceptions, Documentation) technique and role-plays in a case controlled trial on medical students using little educational time of 1 h only (63). At the start and end of clerkships, both groups completed the JSE. While faculty and standardized patients examined and graded students' history-taking and communication skills as well as their empathic behaviors using the Jefferson Scale of Patient Perceptions of Physician Empathy (JSPPPE) at the end of clerkships. The mean scores of JSE of both groups increased from pre-test to post-test. Historically, the Balint groups have been used to facilitate physicians and students in promoting and sustaining their empathy skills (12) This analysis reflects how the patient-related outcomes can be improved by using a wide range of educational tools and techniques (VALUE, CARE, and SALTED) for incorporating, measuring, and enhancing empathy and compassion in medical students.

In our research, a great majority of studies have measured self-reported changes in knowledge and skills of participants and only two studies could tap into behavioral changes after an intervention (59, 69). Sustained training sessions are essential for achieving such paradigm change in behaviors. Although literature has shown some controlled trials on empathy, a precursor for compassion, but there is scarce information on compassion training of medical students or healthcare professionals (80). One such compassion enhancing training exercise is provided by mindfulness, “a continuous awareness of present moment experience in a calm and non-judgmental manner” (81). Mindfulness is a multi-factorial construct that contains a host of correlates such as healthy lifestyle, health education, wellness, empathy, compassion, good quality of life, regular exercise, sleep and hygiene, yoga, and guided imagery (82). Training for enhancement of mindfulness can potentially enhance compassionate care in clinical practice (83), close physician-patient relationship and can help alleviate physician burnout (84). Mindful meditation, attention to breathing, didactic learning, reflection, mindful walking, journalism, and sitting meditation are some of the most popular mindfulness training exercises that have shown to enhance empathy and compassion in medical students and healthcare workers (59). In the study by van Dijik et al. (71), the authors incorporated a stress reduction training curriculum based on mindfulness strategies to improve the mental health of medical students during clinical clerkships. This intervention led to an insignificant betterment of mental health of students which was followed up longitudinally over a 20 month period. However, the researchers could not detect a difference in JSE at 12 months in the studied cohort of students. Such findings necessitate the incorporation of a sustained empathy training program, particularly empathetic communication, for undergraduate medical students.

Technology-enhanced learning has gained popularity in several branches of medical education. Likewise, virtual patients (VPs) and standardized patients (StPs) are used as attractive alternatives to real patients with promising results (58). Unfortunately, research has shown an empathy decline in medical students', particularly in the third year of medical school (85). At this stage, medical students enter clinical clerkships that witness their encounters with patients and their relatives. A suitable ramification to sustain empathy and compassion is the use of VPs and StPs that can provide a constant source for medical students' experiential learning and behavioral development without endangering patient safety. Experiential learning such as patient shadowing or wellness programs can potentially enhance cognitive and behavioral dimensions of empathy. The long-lasting impact of experiential learning is further endorsed by a study conducted by Modi et al. which showed that volunteering students had better empathy than the non-volunteered medial students (65), who showed a decline in empathy over time.

The intelligent use of arts, culture, and humanities in medical education is an emerging and promising approach that can revitalize the declining empathy in medical students (70). Comics “the combined use of images and text, sequentially, to tell a story, where the images complement and/or enhance the text” (86) is a powerful means of portraying complexities of medical knowledge. Graphical illustration with juxtaposed texts, depicts stories in a tangible manner which help to amalgamate the pictorial and textual cues highlighting an innovative use of technology enhanced learning and comics. Our research has shown some evidence that comics may serve as a distinct tool to promote empathy in medical education (87). Finally, interprofessional education and practice carries great potential to promote empathy and compassion by alleviating anxiety, stress and burnout associated with medical education (21, 88).

Our results highlight the fact that the educational interventions pitched at improving the knowledge outcomes can be significantly effective, followed by behaviors and finally, patient outcomes. This is understandable for some reasons; medical students are attuned to work hard to acquire knowledge which can be easily measured as an educational outcome. Behavioral change in students and improvement in patient outcomes demand a sustained interplay of generic and acquired traits and, therefore, it's hard to achieve. Another interesting finding of our research springs from the fact that frequency, duration and teaching modalities didn't have any bearing on the effectiveness of the educational intervention. Even single session interventions were as effective as longitudinal curricula and such curricula did not have a sustained and long-lasting impact. However, experiential learning drills lead to emotional and behavioral remodeling that can result in durable personality developments.

There are number of limitations of this review. The first relates to the comprehensiveness of the search and included articles. We searched four databases quite rigorously, but there remains a chance that certain pertinent studies are not captured by our search of databases, time and language restrictions. Second, though we aimed at highlighting the best practices in teaching empathy and/or compassion, this turned out to be difficult due to profound heterogeneity in the educational interventions and measurement tools used and the types of accomplished outcomes. Third, we planned to find a common curriculum for teaching and assessing empathy and compassion in medical education, its diverse and heterogeneous nature did not allow us to achieve our goal.

In our systematic review, the identified 24 studies evaluated the empathy and/or compassion curricula for undergraduate medical students. There was a great diversity of teaching pedagogies, curriculum design, and duration of teaching that did not let us secure a single best-evidence teaching modality for empathy and/or compassion. Keeping the multidimensional construct nature in mind, a blend of teaching pedagogies is needed. However, major educational constructs of communication, mindfulness, self-care, self-regulation, reflective practice, early clinical exposure, technology-enhanced learning, comics and arts and culture should be targeted for teaching empathy and compassion. We found that even short standalone curriculum was as effective as longitudinal curriculum. In order to mitigate the risk of decline of empathy and compassion, a sustainable program rather than a single training activity is essential.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.

PM and SSG contributed substantially to conceiving the idea, created, reviewed, and validated the search strategy, hand searched and screened the titles and abstract, extracted, analyzed and interpreted the data and came up with the initial draft of manuscript. Later SYG individually evaluated the search strategy, tweaked the data mining process, revised and improved the intellectual content of the initial draft. PM, SYG, and SSG agreed to take responsibility for the final draft. All authors contributed to the article and approved the submitted version.

This research was funded by the School of Postgraduate Studies and Research of the Royal College of Surgeons Ireland—Medical University Bahrain.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.