Over the past few years, a great deal of research has been done on the relationship between doctors' and medical students' own health and how this affects the health care of their patients. It has become clear that healthy physicians set a strong example for their patients and have an improved ability to motivate their patients to change unhealthy behaviors ^[1]–^[4]. The same has been proven for medical students: their personal healthy habits influence their counseling practices in a positive way, for physical activity ^[5], as well as for smoking and drinking ^[6], ^[7]. These findings suggest that in promoting health in our medical students, we have found a novel approach to promote health in the general population.

Yet, from experience we know that students' health is hardly addressed in medical school and that the medical training is demanding and could even cause the student to neglect healthy habits. Aspects of medical students' health have been monitored in several countries, with results showing that although the medical students are generally healthier than their non-medical equivalents, unhealthy behaviors are prevalent and medical students are at a higher risk for specific psychological ailments. Studies show that medical students are in suboptimal physical shape ^[8]–^[12], and that they smoke ^[6], ^[13]–^[15], drink ^[7], ^[16], ^[17], and take drugs ^[18], ^[19] like their peers in the general population. Additionally, the high risks of burnout and depression that medical students face were revealed by multiple studies ^[20]–^[23], as well as the risk of sleep deprivation, which is not only unhealthy but also potentially hazardous ^[24]. Furthermore, medical students appear to eat irregularly, which might be another potential cause of fatigue, but also vice versa ^[25].

Additionally, if the medical students face health issues, they appear to have difficulty finding the appropriate help ^[26], ^[27]. This is partly due to a performance-driven medical culture in which work attendance as well as portraying competence and good health are demanded ^[28], ^[29], and in which the student may have trouble adapting to the patient role ^[30]. Embedded in medical culture is an unprofessional, inappropriate and potentially hazardous ^[31] practice of self-care and self-prescription. Numerous studies have shown a high prevalence of self-care, e.g., 54% of Norwegian doctors had self-prescribed over the course of a year ^[32], and up to 90% of Australian physicians reported that they can treat their own conditions ^[33]. Importantly, it seems that these practices are already acquired in medical school and increase with availability, increased clinical access and closer contact with seniors, from whom students receive ‘curb-side’ consults and drug prescriptions ^[27], ^[34].

There are no data available on these topics for Dutch medical students. We might expect them to have similar behaviors and issues as their international colleagues. Information on their health behaviors can be valuable for medical schools (and eventually, the patients) to target and prevent unhealthy habits in more effective ways, as well as to offer their students more appropriate care.

We were also interested in possible differences between senior clinically active students and junior preclinical students. The last two years of the Dutch curriculum is the clinical phase, which is commonly regarded as demanding and stressful, both physically and emotionally. Students often work over 50 hours per week, not including home study, without financial compensation. It could be that clinically active students have less healthy behaviors, as the intensive course could make students omit healthy habits or induce unhealthy coping styles. We might also expect senior students to be more tolerant towards self-prescription, as their attitudes are influenced by the clinical environment. Such effects would be valuable information for any medical curriculum comprised of a preclinical and a clinical phase. Thus, we designed this study to investigate these parameters in medical students from a major Dutch medical school. We assessed three main subjects: (1) health behaviours, (2) care needs and (3) acceptance of self-prescription, and compared answers from the preclinical and clinically active students.

A total of 954 students filled out the questionnaire (relative response rate = 35%). Data of 902 students could be used for this study as some incidental questions were left blank and only complete individual sets were used. Student demographics (Table 1) did not differ significantly between preclinical and clinical students, with male sex of 33% vs. 30% (p = 0.37) and Dutch natives 76% vs. 79% (p = 0.21). Percentages of 30–35% male and ca. 33% clinically active participants correspond with the actual ratio of sexes and build of the curriculum of the course.

In total, 9% of students did not meet the minimum amount of MET-minutes per week, and 6% of students had a current smoking habit. Unhealthy eating habits were prevalent with 51% of students not meeting the criteria set. Additionally, unhealthy alcohol consumption was prevalent with 46% of students meeting the AUDIT-C criteria for risk of alcohol-related health problems. Underweight or overweight was present in 20% of students. Health behaviors are shown in Table 2.

The classification in healthy, moderately healthy and unhealthy categories is shown in Table 3. There were no significant differences in these categories between the preclinical and clinical group (p = 0.54), with 17% of preclinical and 20% of clinical students reported not to have any unhealthy behaviors.

The forms of care most needed were help with career choices, physical health care, help with study problems and mental health care. In preclinical students, the need for physical health care, help with study problems, help with career choices and help with financial problems was significantly higher (p<0.05), as can be seen in Table 4.

Finally, the summed score of acceptance of self-prescription was significantly higher in the clinical subgroup over the preclinical subgroup (181±4 vs. 142±5 p<0.001). The scales for acceptability of self-prescription were tested for internal reliability using Cronbach's alpha and were found reliable at 0.83.

Overall, the data gives an equivocal message on the healthiness of the students. It is satisfying that over 90% of students appear to get frequent and sufficient physical exercise, and over 80% of students have a BMI within the normal range. However, their eating habits are predominantly unhealthy and/or irregular. Smoking prevalence is relatively low, as are abnormal amounts of sleep per night. It is certainly worrying that almost half of the students meet the AUDIT-C criteria for risk of alcohol-related health problems, and over 6% used stimulants, tranquilizers or sleep medications in the past 30 days.

Contrary to expectations, the demands of the clinical phase of the curriculum appear to have no significant impact on the health behaviors of the students.

The form of care most needed was help with career choices. Students also expressed high needs for easily accessible mental health care, physical health care and help with study problems. The need for physical health care is significantly lower in the clinical phase, whereas the need for mental health care was equally high. The need for help with study problems, career choices and financial problems was also significantly lower. Among students from the clinically active subgroup, self-prescription was considered more acceptable compared to their preclinical colleagues.

The strength of this study is the large absolute number of participants. However, the relative response rate is somewhat low. This may be explained by the fact that, unlike other studies, participation in this study was completely voluntary. Although we are aware of the potential source of bias that the non-response may initiate, we have no reason to suspect the participants not to be representative of the total population because of the comparable numbers of participants from every study year and the variation in the observed responses. Still, it could be that unhealthy behavior is underreported, as is often the case with this kind of research. These findings then reflect too favorable results. Another strength is that, to our knowledge, the curriculum of the investigated medical school hardly changed over the past years, so that found differences are unlikely caused by changes in education or services offered.

Eating habits of the students were mostly irregular and/or unhealthy, as was alcohol consumption. Perhaps this is to be expected from young college students moving out from their parents. Both behaviors were less prevalent among senior clinically active students.

Need for help with study problems, financial problems and career choices was significantly lower among senior clinically active students. Perhaps this can also be ascribed to student maturation and growing competence. Interestingly, a significantly lower need for physical health care was found, whereas the need for mental health care was equally high. Roberts et al. and Hooper et al. already reported senior students to have more ‘curb-side’ consults from colleagues, seniors or themselves, which could explain the lower care needs ^[27], ^[34]. As these informal consults are more adequate for physical than mental health problems, and given the greater stigma surrounding mental health issues, mental health curb-side consults are probably obtained less. Indeed, Roberts et al. also reported student fear of academic jeopardy to be much higher with mental health problems ^[30].

The higher acceptance of self-prescription we found in clinically active students in our study could reflect students' growing confidence in their own medical skills, yet also suggests that exposure to the clinical medical culture has a role herein. The higher acceptance worries us as this could mean that senior students will seek less formal consultation and have a growing tendency to manage themselves. Finally, not shown above but worth mentioning is that the use of sleep medications was three times as high among clinically active students (1.8% vs. 5.6% P = 0.004). The higher use of sleep medications could be indicative of sleeping problems due to nights shifts, but also higher amounts of stress in this stage of the training.

By comparing this study's results with other figures, the prevalence of over or underweight (20%) among our medical students was lower than in the general population ^[40], Greece and Malaysia ^[11], ^[12], but higher than Colombia ^[10]. Smoking prevalence (6%) was also lower than in the general population ^[40], comparable to medical students from the US ^[6], ^[9] and lower than medical students from Colombia, Vietnam and Poland ^[10], ^[13], ^[14].

The results indicate that over eighty percent of medical students, future promoters of public health, act against their own physical wellbeing in at least one of the ways investigated. In particular, the high rates of alcohol abuse are troubling, which makes it incomprehensible that this is not at all addressed in medical school. When alcohol abuse is learned young, students could well develop a disposition for dependence; rates of alcohol and/or drug dependence among doctors are reported to be as high as 9–12% ^[18], ^[41].

The need for specific forms of care is high, which is interesting for faculties to meet their students' needs. If these forms of care are offered and easily accessible, this could contribute to an environment in which help-seeking behavior is normal and encouraged, and could serve as an alternative to curb-side consultations, self-care and self-prescription. Senior students growing towards higher acceptance of self-prescription can be seen as a growing mandate to offer them an appropriate and comfortably accessible alternative. For physical health care, a school physician could provide in this need, whereas help with career choices, mental health care and study problems could be met by a school psychologist.

Importantly, greater school support for health promotion was already found positively associated with a higher counseling frequency ^[42]. Potential supportive measures would be more education on smoking, alcohol and drugs ^[43] physical activity programmes ^[8] and a programme teaching students the basic of preparing healthy food ^[44]. An integrated lifestyle curriculum could deliver all these components at once, and is highly recommended. An valuable example of such a course was recently designed in the United Arab Emirates ^[45]. As our findings suggest that there is no significant deterioration of health behaviors over the years of medical training, life-style changes achieved in the first year of medical school might last throughout the entire curriculum. It can be concluded that there is ample room for improvement of health behaviors of medical students. In the light of evidence, medical faculties have a responsibility in pro-actively promoting health of their students and eliminating the barriers that stand in the students' way to appropriate care. If done so, it is likely that the benefit of healthier and happier medical students will be measurable in a higher wellbeing of the physicians and better preventive care for their patients, and therefore better health for all.

Competing Interests: The authors have declared that no competing interests exist.

Funding: The authors have no support or funding to report.

We acknowledge the respondents in this study.