Medical students’ attitudes towards science and involvement in research activities: A comparative study with students from a reformed and a traditional curriculum

Abstract

Background and aims: Little is known about medical students’ attitudes towards science and scientific methodology. We aimed to evaluate these attitudes and students’ involvement in research activities.

Methods: Cross-sectional study comparing fifth-year medical students from the reformed and the traditional curriculum of the Charité University Medical Centre Berlin, Germany. Students filled out a standardised questionnaire containing three domains: research project for a dissertation; self-reported behaviour, knowledge/attitudes towards evidence-based medicine (EBM) and scientific methodology; and attitudes towards science.

Results: Two-thirds of the students had already started research for their dissertation and 70% agreed that reading articles and conducting research was challenging. Reformed curriculum students showed a higher involvement in scientific activities and felt more secure about their own scientific competencies. The odds for involvement in different research activities were significantly higher (odds ratios from 1.8 to 2.4) for students who agreed that ‘science enables medical progress’ or who felt ‘secure in understanding medical articles and statistics’ compared with students without these attitudes.

Conclusion: The general attitude towards science and scientific methodology was positive among students from both the traditional and the reformed medical curriculum. Specific attitudes predicted involvement in research activities; however, they should be examined in other settings and student populations.

Introduction

Healthcare professionals face a considerable challenge in keeping up with knowledge of current clinical practise over the course of their career. The principles of scientific methodology, critical appraisal and evidence-based medicine (EBM) play an important role to handle these challenges for every practising physician (Sackett 2000).

Within the last decade, the teaching of research methodology and EBM has been increasingly introduced as topics in medical education (Norman & Shannon 1998; Taylor et al. 2000; Bradley et al. 2005). It is still unclear, however, which educational intervention prepares medical students best for tracking down the evidence to solve clinical problems or working as a researcher after graduation. Not every medical student will develop skills to the level necessary to pursue a research career. Nevertheless, the universities should assert a claim that those who simply wish to practise their profession have at least an understanding of the research process and know how to interpret research findings.

Table

Practice points

Little is known about students’ attitudes towards EBM, scientific methodology and their involvement in research activities, and about traits or behaviour patterns that predict later professional behaviour. Students from the problem-based medical curriculum showed greater involvement in research activities and felt more secure about their scientific competencies than students from a traditional curriculum. A sound theoretical model to predict physicians’ scientific activities would support this research.

At the moment, there are no traits or behaviour patterns established that would predict later professional behaviour like critical appraisal or research activities. According to the theory of planned behaviour (Ajzen 1998), which provides a framework to study attitudes towards behaviour, we assumed that later physician behaviour is influenced by student behaviour or behaviour intention. Student behaviour is again influenced by personal attitude, perceived social pressure or subjective norms and perceived behavioural control (Figure 1).

Figure 1. Modified model of planned behaviour – physician's behaviour: research/EBM.

The aim of the present study, therefore, was to evaluate the self-reported behaviour, knowledge and attitudes of medical students at the Charité University Medical Centre Berlin, Germany. As there are two different curricula of undergraduate medical education, we compared students’ attitudes towards EBM, scientific methodology and a research project for dissertation. Furthermore, we adapted the above-mentioned theoretical model of how to possibly predict student future behaviour.

Methods

Study sample

In 2004 and 2005, 147 out of 160 students from the traditional medical curriculum (TMC) and 47 out of 56 from the reformed medical curriculum (RMC) agreed to participate in the study and completed the standardised questionnaires.

Out of all applicants interested in joining the RMC at the Charité University Medical Centre Berlin, Germany, 63 each year are randomly selected. In comparative tests, RMC students did not differ from the TMC students regarding cognitive and non-cognitive factors (Kiessling et al. 2002, 2003, 2004; Burger 2006). The teaching modules and research opportunities for medical students in the two curricula are rather diverse; the TMC is lecture and seminar based, whereas the RMC follows an integrated problem-based approach (Burger & Froemmel 2002).

Epidemiology, Biostatistics and EBM courses

Traditional Medical Curriculum

During the second year, TMC students had lectures and seminars on basic research methodology. In the fifth year, students were introduced to major public health issues, epidemiology, and EBM in eight 90 min lectures including a multiple choice test (MCQ) at the end of the last lecture. Subsequently, a 5-day module started with a session on the strengths and weaknesses of the main epidemiological and medical study designs followed by the development of a cohort or case-control study design and the critical appraisal of a randomised control trial (Figure 2).

Figure 2. Synopsis research/EBM education. Traditional vs. reformed medical curriculum at the Charité University Medical Centre Berlin, Germany.

Reformed Medical Curriculum

In the fourth semester, RMC students had to conduct an individual 4-week research project in one of the faculty's departments and write a scientific report. During this special study module (SSM), students could participate in voluntary seminars on research methodology. In the Seventh semester, RMC students took part in a 1 week compulsory seminar on the principles of scientific methods in medicine. Similar to the TMC students, they discussed different study types and evaluated medical publications. At the end of this semester, students had to pass a MCQ test (Figure 2).

Questionnaire on behaviour, knowledge and attitudes regarding science

For the standardised questionnaire, we developed the first two domains and used a validated questionnaire for the third domain. Part one consisted of three questions asking whether the students have already started with a research project for their dissertation, whether they have already cancelled one and what type of study design they had chosen for their research. The second part contained (i) two statements about self-reported behaviour concerning reading articles and conducting literature searches using the database MEDLINE, (ii) two statements about subjective secureness in reading articles and biostatistics, (iii) one item about behaviour intention concerning EBM and (iv) three items covering attitudes towards science and EBM. Students had to indicate their answers on a 7-point scale (1 = strongly agree, 7 = strongly disagree). The two items concerning subjective secureness were aggregated to a score called ‘secureness in reading articles in medical journals and dealing with statistics’ (bivariate correlation coefficient was r = 0.551). The two items covering excitement about reading articles and doing research were also aggregated (bivariate correlation coefficient was r = 0.649).

The third part about attitudes towards science consisted of two subscales (29 items) from a questionnaire with answers on a 5-point scale (1 = strongly disagree, 5 = strongly agree) (Hren et al. 2004). The third subscale, value of science to humanity (20 items), was beyond the objectives of the present study and, therefore, not considered.

In a confirmatory factor analysis, we could not approve the two subscales, value of scientific methodology and value of science for medicine. After a subsequent explanatory factor analysis, four subscales were established and one additional single item was selected for further data analysis: subscale 1: science enables progress in medicine (four items Cronbach's alpha 0.739); subscale 2: scientific methods are important to generate reliable data (six items, Cronbach's alpha 0.747), subscale 3: only scientifically verified methods of treatment are reliable (three items, Cronbach's alpha 0.707), subscale 4: negative effects of scientific methodology for physicians (four Items, Cronbach's alpha 0.680), single item: every physician should have good competencies of scientific methodology.

MCQ on epidemiological and statistical knowledge

For a subgroup analysis, all 147 TMC students filled out a MCQ-test with 30 questions (each with five possible answers) covering the major interventional and observational study designs and basic biostatistics. The high internal consistency (Cronbach's alpha 0.787) allowed us to use an aggregated global score. Each correct answer was rated with one point, an incorrect answer with no point. A maximum of 30 points were possible. Six students from the TMC were excluded from data analysis because it seemed that they purposely ticked the wrong answers, because they had less than 20% correct answers, which is below the likelihood of pure guessing considering five possible answers per question.

Statistical Methods

The statistical analyses were performed using SPSS 12.0 (SPSS, Inc., Chicago, Illinois). The statistical significance of the differences between the two medical curricular was assessed using χ²- and Mann–Whitney U-tests. Level of significance was set using the Bonferroni-method.

To identify aspects that would predict students’ behaviour, logistic regressions analyses were performed according to the aforementioned model of planned behaviour. Odds ratios and 95% confidence intervals (95% CI) were calculated to estimate how much an increase on the Likert-type scale for the independent variable increased the odds of the relevant outcome.

This model led to four logistic regression models with the following factors:

Regression models 1–3 (including 125 students)

Outcome variables (Behaviour)

• ‘Dissertation’ = ‘I perform a dissertation’ (binary)

• ‘Reading articles’ = ‘I regularly read original articles in medical journals’ (binary recoded)

• ‘Searching MEDLINE’ = ‘I have conducted a patient related MEDLINE-search’ (binary recoded)

Independent variables

• Attitude: Zagreb subscale 1–4; excitement about reading articles and doing research

• Subjective norm: role of patient-related literature search on wards

• Perceived behavioural control: secureness in reading articles and statistics

• Knowledge (results in the MCQ-test) is not part of the theory of reasoned action but was also considered to be an interesting independent factor.

Regression model 4 (including 128 students)

Outcome variable (Behaviour intention)

• ‘EBM-course’ = I would participate in an EBM course (binary recoded) Independent variables (all behaviour)

• ‘Dissertation’

• ‘Reading articles’

• ‘Searching MEDLINE’

Results

Dissertation

More TMC than RMC students (73% vs. 66%) had already started with research for their dissertation; however, this difference was not statistically significant (p = 0.368). Almost a fifth of all students had already cancelled at least one of these research projects (19% TMC vs. 18% RMC students, p = 0.899). Most commonly, the projects had experimental (66%) or empirical designs (22%) with no significant differences between TMC and RMC students.

Self-reported behaviour, knowledge and attitudes towards EBM and scientific methodology

RMC students had significantly more often read original medical articles, performed MEDLINE searches, and felt more secure with statistical terminology. TMC students expressed a higher interest in voluntary EBM courses for their last year of medical school compared to RMC students. There were no significant differences between the two groups regarding the excitement of scientific work and literature, and the impression that patient-related literature searches seem to be less important in everyday clinical routine (Table 1).

Table 1.  Self-reported behaviour, knowledge and attitudes towards EBM and scientific methodology among students from the TMC and RMC (statistically significant results are printed in bold)

1 = strongly agree 7 = strongly disagree	TMC (N = 146) Median; Mean (SD*)	RMC (N = 47) Median; Mean (SD*)	U-test p-value**
I regularly read original articles in medical journals.	3	5	0.006
	3.5 (2.0)	4.4 (1.8)
During my clinical clerkships, I have prepared MEDLINE searches for patient-related problems.	1	3	0.001
	2.2 (1.6)	3.4 (2.1)
Excitement about doing research and reading scientific literature (aggregated score)	5	5.5	0.165
	4.9 (1.4)	5.1 (1.4)
Secureness in reading scientific articles and in dealing with statistical terminology (aggregated score)	4	4.5	0.001
	3.7 (1.3)	4.4 (1.2)
During my last year of clinical clerkships, I would voluntarily participate in an EBM course.	5	4	0.001
	5.1 (1.4)	4.2 (1.4)
I have the impression, that patient-related searches play a minor role in everyday clinical routines on the ward.	5	5	0.867
	4.8 (1.5)	4.9 (1.5)

*SD = standard deviation; **Alpha was set at p < 0.008 according to Bonferroni.

Attitudes towards science and scientific methodology

Of all students, 89% agreed that science enables progress in medicine, 79% that scientific methods were important to generate reliable data, and 8% that scientific methodology can be negative for physicians by deceiving or distracting them from natural treatment ways. There were no significant differences between TMC and RMC students’ attitudes towards science/scientific methodology (Table 2). The unconditional scientific foundation of treatment methods was seen critical: 72% did not agree that “only scientifically proven therapies are reliable” or that “treatments without scientific verification are dangerous and unreliable”.

Table 2.  Attitudes towards science and scientific methodology among students from the TMC and RMC

1 = strongly disagree 5 = strongly agree	TMC median; Mean (SD*)	RMC median; Mean (SD*)	U-test p-value**
Subscale 1: Science enables progress in medicine	4	4	0.502
	4.0 (0.7)	3.9 (0.7)
Subscale 2: Scientific methods are important to generate reliable data	3.7	3.7	0.894
	3.6 (0.6)	3.6 (0.6)
Subscale 3: Methods of treatments needs to be scientifically proved	2.7	2.8	0.144
	2.7 (0.9)	2.9 (0.9)
Subscale 4: Negative effects of scientific methodology for physicians (recoded)	3.8	3.9	0.992
	3.9 (0.7)	3.8 (0.7)
Single item: Every physician should have good competencies of scientific methodology	4	4	0.501
	3.7 (1.0)	3.9 (0.7)

*SD = standard deviation; **Alpha was set at p < 0.01 according to Bonferroni.

Prediction of students’ behaviour

Research for dissertation

Students who agreed that science enables progress in medicine or that treatment methods need to be scientifically proven were more likely to conduct research for dissertation compared to students without these attitudes. Other factors were not associated with this outcome. The variables in this model accounted for 16% of the variance: χ² = 14.2, p = 0.076 (Table 3).

Table 3.  Medical students’ attitudes towards science/scientific methodology and involvement in research activities (‘already started with research for dissertation’, ‘regularly reading original articles in medical journals’, ‘performing MEDLINE searches’) (statistically significant results are printed in bold)

	Dissertation		Reading articles		Searching MEDLINE
Independent variables	Odds ratio (95% CI*)	p-Value**	Odds ratio (95% CI*)	p-Value**	Odds ratio (95% CI*);	p-value**
Attitude: Zagreb 1: Science enables progress in medicine	1.77 (1.07–2.94)	0.027	1.06 (0.62–1.83)	0.832	2.32 (1.11–4.83)	0.025
Attitude: Zagreb 2: Scientific methods are important to generate reliable data	1.19 (0.73–1.94)	0.492	0.80 (0.48–1.33)	0.381	0.99 (0.54–1.81)	0.963
Attitude: Zagreb 3: Methods of treatments need to be scientifically proven	0.48 (0.28–0.81)	0.006	0.82 (0.51–1.32)	0.419	0.90 (0.52–1.58)	0.719
Attitude: Zagreb 4: Negative effects of scientific methodology for physicians	0.98 (0.60–1.61)	0.937	0.63 (0.38–1.03)	0.068	0.64 (0.36–1.15)	0.133
Attitude: Excitement about reading articles and doing research	1.41 (0.85–2.37)	0.145	3.06 (1.67–5.65)	0.001	1.12 (0.60–2.10)	0.716
Perceived control of behaviour: Secureness in reading articles and statistics	0.98 (0.61–1.57)	0.936	1.95 (1.17–3.25)	0.010	2.38 (1.22–4.65)	0.011
Subjective norm: Role of patient-related literature search on wards	1.15 (0.71–1.84)	0.560	1.09 (0.69–1.72)	0.712	1.53 (0.89–2.64)	0.126

*CI = confidence interval; **Alpha was set at p < 0.01 according to Bonferroni.

Reading original medical articles regularly

Students who found medical publications and research exciting and those who felt secure in understanding statistical methods were significantly more likely to regularly read original articles in medical journals. The variables in this model accounted for 34% of the variance: χ² = 36.3, p = 0.001 (Table 3).

Performing patient-related MEDLINE searches

Students who agreed that science enables progress in medicine and those who felt secure in reading articles and understanding statistics were more likely to perform MEDLINE searches. The variables in this model accounted for 21% of the variance: χ² = 15.8, p = 0.045 (table 3).

Interest in EBM course

Students who already started research for their dissertation (OR = 3.36, 95% CI 1.38–8.18, p = 0.008), and those who had independently performed MEDLINE searches (OR = 1.81, 95% CI 1.04–3.17) showed significantly more interest in an EBM course. Reading medical articles regularly was not associated with interest in such a course (OR = 1.12, 95% CI 0.71–1.75). The variables in this model accounted for 17% of the variance (χ² = 16.7, p = 0.001).

Epidemiological and statistical knowledge (subgroup analysis with TMC students)

The mean score ± standard deviation in the MCQ test was 17.5 ± 4.2 (range 8–28). There was no statistically significant difference between students’ who started research for dissertation and those who did not (17.6 ± 4.2 vs. 16.7 ± 4.5; p = 0.242). Students conducting experimental projects scored significantly better than other students (18.6 ± 4.1 vs. 16.5 ± 4.0; p = 0.005). There was no correlation (r = 0.09) between the MCQ-test results and the aggregated score of students’ subjective assessment of secureness in understanding medical articles and statistics.

Discussion

The findings of this study showed that fifth-year German medical students (RMC and TMC) had overall positive attitudes towards science and scientific methodology. About two-thirds had already started research for their dissertation. Students were critical concerning their own abilities of methodological and statistical understanding. Furthermore, the study revealed that patient-related MEDLINE searches played a minor role during clinical clerkships. TMC students were more likely to participate in an EBM course during their last year of medical school compared with RMC students. We assume that the RMC provides different opportunities for students to develop an investigative approach to medical problems not only in the research course; therefore, these students may not feel the need to take another course.

The majority of medical students in an American study on attitudes towards EBM and research methodology appreciated the importance of biostatistics and research methodology within the context of critical appraisal of the medical literature (Caspi et al. 2006). Although German and American curricular and healthcare systems differ, perceived competence and performance of EBM seem to be similar among medical students. Critical appraisal skills are valued by medical students but their attitude towards statistics and methodology is often less enthusiastic which corresponds with the application of EBM in the clinical training (Caspi et al. 2006).

Our study suggested that students of the RMC felt more comfortable with their methodological and statistical understanding and read significantly more often scientific articles in medical journals on a regular basis than TMC students. This corresponds with the findings by Gurpinar et al (2005). who also compared students from a problem-based curriculum with students from a traditional curriculum.

For the three regression models predicting student behaviour, some specific positive attitudes towards science and critical appraisal and perceived control of behaviour (feeling secure about critical appraisal competencies) were able to account for a significant percentage of the variance. The model of planned behaviour was mainly developed to predict and understand intentions, behaviours and outcomes of health-related behaviours. The theory is based on the assumption that human beings act rational and make systematic decisions based on available information. This might be overoptimistic. In our study, the subjective norm as well as knowledge about epidemiology and statistics was not relevant for predicting students’ behaviour like conduction research for a dissertation, MEDLINE searches or reading medical articles. Furthermore, factors such as personality and demographic variables were not taken into consideration. Nevertheless, using a theoretical model with different distinct dimensions to predict student's behaviour is crucial to future research in this field. It may be worthwhile to revise and improve our model and lay emphasis on subjective statements as well as objective performance.

There are some possible limitations in our study. The structure and teaching methods of the two curricula, TMC and RMC, differed; causal relations, therefore, between a single course and a single outcome should be drawn with caution. Nevertheless, for the evaluation of curricular changes comparison with the best possible control group is necessary.

In the subgroup analysis, we tested knowledge of medical and epidemiological study designs and biostatistics in a MCQ test, which might not have been a valid tool to assess critical appraisal skills.

Conclusions

Students from the RMC showed a higher involvement in scientific activities and felt more secure about their own scientific competencies than TMC students. Attitudes towards science and scientific methodology were overall positive and did not differ between TMC and RMC students. In the present study, we identified several statistically significant predictors for involvement in research activities during medical school. Our findings, however, need to be confirmed in other student populations and settings. Triangulation of evaluation methods including a sound theoretical model of predicting later physician behaviour will help to meet this challenge.

Acknowledgements

We thank Prof. M. Marusic and Prof. D. Hren from the Zagreb University School of Medicine who generously provided us the use of their questionnaire on attitudes towards science. We also thank the students of the Charité, Universitaetsmedizin Berlin who took part in the study. We would also like to thank Jacqueline Mueller-Nordhorn, Benno Brinkhaus Michel Knigge and Heike Kanter for their support. The reformed curriculum has been sponsored by Bundesministerium f. Bildung u. Forschung, Robert-Bosch-Stiftung, Carl-Gustav-Carus-Stifting, VolkswagenStiftung and Deutsche Aerzteversicherung.

Declaration of interest: The authors report no conflicts of interest. The author alone is responsible for the content and writing of this article.

Additional information

Notes on contributors

Susanne Pruskil

SUSANNE PRUSKIL, MD, is a resident in the Department of Primary Care at the Charité Universitaetsmedizin Berlin. In addition to supporting medical student learning in the community, her research interests include communications skills training, peer group teaching and learning and curriculum evaluation.

Philip Burgwinkel

PHILIP BURGWINKEL, is a final year medical student at the Charité, Berlin. He studies in the reformed medical curriculum and is employed as student assistent at the faculty. Main focus of his work was planing and organisation of special study modules concerning scientific methods and public health related topics

Waltraud Georg

WALTRAUD GEORg, MD and adult educator, works as an educationalist at the Charité. Her main focus of educational work is the development and implementation of assessment structures for undergraduate medical education. She is also involved in the curriculum development of the Reformed Medical Curriculum at the Charité.

Thomas Keil

THOMAS KEIL, MD, MSc, is an epidemiologist at the Institute for Social Medicine, Epidemiology and Health Economics, Charité Universitaetsmedizin, Berlin, Germany. His research includes pediatric epidemiology and allergology. He teaches basic epidemiology and evidence-based medicine, and is currently evaluating the obligatory research electives as part of the medical curriculum.

Claudia Kiessling

CLAUDIA KIESSLING, MD, MPH, is member of a group running the reformed medical curriculum at the University of Basel, Switzerland. Her main interests in medical education are medical humanities, psychosocial aspects, evaluation and medical education research.