Medical school fails to improve Mediterranean diet adherence among medical students

Abstract

Background: The Mediterranean Diet (MeDi) is associated with a reduced risk of several non-communicable chronic diseases. High levels of nutrition knowledge are associated with greater MeDi adherence irrespective of education and other socio-economic factors. The present study investigates whether attending medical school changes students’ adherence to the MeDi. Methods: A cross-sectional study was conducted in November–December 2014 using the 16-item KIDMED questionnaire, which inquiries into eating habits. Each respondent’s KIDMED index (range 0–12), whose scores reflect ‘poor’, ‘average’ or ‘good’ MeDi adherence, was calculated. Scores were analyzed for gender, age and early/late medical school year. The Mann–Whitney U -test and the chi-square test were used for group comparisons. Ordinal logistic regression was applied to estimate the odds ratio (OR) and 95% confidence interval (CI) for changes in MeDi adherence from early to late medical school after adjustment for potential confounders. Results: Participants were 1038 medical students (573 females) aged 18–34 years (mean 21.9 ± 2.4). There were 20.8% ‘poor’, 56.5% ‘average’ and 22.7% ‘good’ scores. Gender significantly affected scores ( P < 0.01). In the ordinal logistic regression model the odds for greater MeDi adherence increased for the female gender ( P < 0.05); females were 48% more likely to have a ‘good’ KIDMED score than their male peers, whereas early/late medical school year did not affect results [adjusted odds ratio (OR) = 0.95; P = 0.15]. Conclusions: More than 70% of our medical students know very little about healthy eating and the MeDi. However, this knowledge is essential in their future profession both to support therapeutic decision-making and effective preventive actions.

Introduction

Preventing or reducing the risk factors for pathological conditions associated with unhealthy eating is among the World Health Organization’s strategic objectives. ¹ Physicians are increasingly called upon to provide dietary advice to prevent or counter a wide range of non-infectious chronic conditions. ^2,3 Most recently, the new EU health policy framework, ‘Health 2020’, has set its Member States the goal to achieve a situation in which the negative impacts of preventable diet-related non-communicable diseases and malnutrition in all its forms—including overweight and obesity—have been dramatically reduced, and all citizens adopt healthier eating habits throughout their lives.

The Action Plan has adopted Objective 3 of the WHO European Food and Nutrition Action Plan 2015–20, ‘Reinforce health systems to promote healthy diets’, which undertakes to improve the food system governance and the overall quality of the population’s diet and nutritional status to achieve the ultimate goal of greater health and wellbeing. ⁴

Teaching medical students to help future patients adopt a healthy eating style and follow dietary advice ^5,6 is therefore an important goal whose achievement can also be promoted by changing the medical school curriculum. In addition, high levels of nutrition knowledge have been demonstrated to be significantly associated with greater adherence to the Mediterranean Diet (MeDi) and lower prevalence of obesity irrespective of level of education and other social and economic factors. ⁷

At present, very little information is available regarding MeDi compliance by health science students. ⁸ The fact that these students are seldom asked to take part in such studies is not wholly surprising, because they are generally held to possess a thorough knowledge of the diet and its implications, and their responses are felt to be biased by their training. ⁹ Yet this information can help develop or improve nutrition-related education programmes and promote healthier food choices and eating styles.

The present study investigates whether attending medical school changes students’ adherence to the MeDi.

Methods

Study design and participants

A cross-sectional study was conducted at the School of Medicine of the University of Catania (Sicily, Italy) in November and December 2014. The study was monocentric, because all Italian medical schools have similar curricula. ¹⁰

All undergraduate students enrolled in the 6 years of medical school were offered to participate in the study, which involved self-administration of a questionnaire based on the KIDMED tool (MeDi Quality Index). Briefly, all students going to classes on the days the study was scheduled to take place were invited to participate. The aims of the study and the questionnaire, which inquired about eating habits and also collected data on age, gender and medical school year, were explained to participants; the KIDMED items were clarified as necessary. The questionnaire was distributed by the Hygiene professor and students were given 20 min to complete it. The professor remained at a distance, to guarantee anonymity.

Questionnaire

The KIDMED is a 16-item questionnaire that inquires about eating habits and is a widely used tool in the literature. ^11,12 Replies allow calculating a MeDi Quality Index, a practical and sound instrument that is particularly suited to explore our study population. In fact, despite the wide range of dietary patterns, the Mediterranean regions substantially share a diet based on olives, olive oil, fruit, vegetables, fish, seafood, pulses and cereals. ^13–16

The questionnaires were collected and divided into two sets: those compiled by first- to third-year students, who study basic science subjects (early medical school group), and those compiled by fourth- to sixth-year students, who study clinical subjects (late medical school group). The questionnaires were then used to calculate each participant’s KIDMED index. Each reply consistent with MeDi adherence is scored as + 1 and each reply reflecting behaviours inconsistent with MeDi patterns is assigned a score of −1 ( table 2 ). The final score therefore ranges from 0 to 12 (Serra-Majem et al. ¹¹ ). Final scores reflect different degrees of MeDi adherence: scores ≤ 3 indicate very low diet quality (poor adherence); scores 4–7 indicate that improvement is needed to adjust the diet to MeDi eating patterns (average adherence) and scores ≥8 reflect an excellent diet (good adherence).

Ethical considerations

This type of study, an anonymous survey of volunteers, merely required informal approval by the local ethics committee; this was obtained in a meeting chaired by the Dean. There was no need for written consent from individual students. The study complies with the principles of the Helsinki Declaration.

Statistical analysis

The descriptive statistics used were absolute frequencies and percentages for categorical variables and mean and standard deviation (SD) for quantitative variables. Descriptive statistics of the three KIDMED scores based on gender was applied to the two groups of participants. The Mann–Whitney U -test and Pearson’s chi-square test were used for bivariate analysis, to assess differences in quantitative and categorical variables, respectively. Ordinal logistic regression was used to estimate OR and 95% confidence interval (CI) for changes in MeDi adherence from early to late medical school years after adjustment for gender and medical school years.

A significance level of 0.05 was set. All data analyzes were conducted with SPSS 20.0 (SPSS, Chicago, IL).

Results

Of the 2000 students enrolled in Medical School, 1065 (53.3%) agreed to participate and filled the questionnaire. Their age range was 18–34 years (mean 21.9 ± 2.4 years), 55.2% were females. Their main characteristics are listed in table 1 . Incomplete questionnaires (2.5%) were discarded, leaving 1038 questionnaires for evaluation.

Table 1

Characteristics of the medical students

Academic year	Age (mean ± SD)	Females % ( n )	Males % ( n )	Number
First year	19.3 ± 1.0	51.6% (99)	48.4% (93)	192
Second year	20.4 ± 1.5	58.6% (116)	41.4% (82)	198
Third year	21.6 ± 1.4	56.2% (95)	43.8% (74)	169
Fourth year	22.2 ± 1.3	59.5% (69)	40.5% (47)	116
Fifth year	23.7 ± 1.6	53.8% (98)	46.2% (84)	182
Sixth year	24.8 ± 1.5	53.0% (96)	47.0% (85)	181
Total	21.9 ± 2.4	55.2% (573)	44.8% (465)	1 038 *

*Only complete questionnaires were assessed; incomplete questionnaires were 2.5% and were discarded.

There were 20.8% ‘poor’, 56.5% ‘average’ and 22.7% ‘good’ scores ( table 2 ). Male students showed a greater proportion of ‘poor’ scores (25.2% vs. 17.3%), whereas female students had a greater percentage of ‘good’ scores (25.2% vs. 19.6%). The mean KIDMED scores of male and female students were significantly different. Their evaluation showed that male students eat less vegetables than their female peers (48.4% vs. 71.6%) and eat fast-food meals more frequently than female students (13.1% vs. 6.6%). A smaller proportion of female than male students eat pasta or rice (65.1% vs. 78.4%). Male students skip breakfast more frequently than females (30.5% vs. 19.7%) ( table 2 ).

Table 2

KIDMED scores of medical students by gender and medical school year

					Medical school years
					Early (year 1 through 3)			Late (year 4 through 6)
		Females % ( n )	Males % ( n )	Total % ( n )	Females % ( n )	Males % ( n )	Total (%)	Females % ( n )	Males % ( n )	Total (%)
	KIDMED scores
	Poor (≤3)	17.3 (99)	25.2 (117)a	20.8	15.2 (47)	25.3 (63)	19.7	19.8 (52)	25.1 (54)	22.2
	Average (4–7)	57.5 (329)	55.2 (256)	56.5	60.5 (187)	54.2 (135)	57.7	54.0 (142)	56.3 (121)	55.0
	Good (≥8)	25.2 (144)	19.6 (91)a	22.7	24.3 (75)	20.5 (51)	22.6	26.2 (69)	18.6 (40)	22.8
	Total	100.0 (572)	100.0 (464)	100.0	100.0 (309)	100.0 (249)	100.0	100.0 (263)	100.0 (215)	100.0
	Mean KIDMED score (SD)	5.7 (2.4)	5.2 (2.6) b	5.4 (2.2)	5.8 (2.3)	5.2 (2.6) b	5.6 (2.5)	5.7 (2.5)	5.1 (2.6) b	5.5 (2.6)
KIDMED evaluation	KIDMED test items
+1	Takes a fruit or fruit juice every day	78.0 (447)	76.1 (354)	77.2	78.7 (244)	75.5 (188)	77.3	77.2 (203)	76.9 (166)	77.0
+1	Has a second fruit every day	47.1 (270)	46.2 (215)	46.7	47.7 (148)	47.8 (119)	47.8	46.4 (122)	44.4 (96)	45.5
+1	Has fresh or cooked vegetables regularly once a day	71.6 (419)	48.4 (225)a	61.2	68.7 (213)	45.8 (114)a	58.5	74.9 (197)	51.4 (111)a	64.3
+1	Has fresh or cooked vegetables more than once a day	45.2 (259)	25.4 (118)a	36.2	43.2 (134)	22.1 (55)a	33,8	47.5 (125)	29.2 (63)a	39.2
+1	Consumes fish regularly (at least 2–3 times per week)	43.8 (251)	45.4 (211)	44.5	45.5 (141)	46.2 (115)	45,2	41.8 (110)	44.4 (96)	43.6
−1	Goes more than once a week to a fast-food (hamburger) restaurant	6.6 (38)	13.1 (61)a	9.39	7.1 (22)	13.7 (34)a	9.9	6.1 (16)	12.5 (27)a	8.7
+1	Likes pulses and eats them more than once a week	33.2 (190)	33.5 (156)	33.4	34.3 (106)	28.9 (72)	31.9	31.9 (84)	39.1 (84)	35.1
+1	Consumes pasta or rice almost every day (five or more times per week)	65.1 (373)	78.4 (364)a	70.9	66.1 (205)	80.3 (200)a	72.5	63.9 (168)	75.9 (164)a	69.3
+1	Has cereals or grains (bread, etc.) for breakfast	69.3 (397)	63.4 (295)	67.0	73.2 (227)	63.1 (157)a	68.7	64.6 (170)	64.2 (138)	64.4
+1	Consumes nuts regularly (at least 2–3 times per week)	8.9 (51)	12.3 (57)	10.7	8.4 (26)	12.9 (32)	10.4	9.5 (25)	11.6 (25)	10.4
+1	Uses olive oil at home	97.9 (561)	97.4 (453)	97.7	97.1 (301)	96.8 (241)	97.0	98.9 (206)	98.1 (212)	98.5
−1	Skips breakfast	19.7 (113)	30.5 (142)a	24.7	17.7 (55)	29.3 (73)a	22.9	22.1 (58)	31.9 (69)a	26.5
+1	Has a dairy product for breakfast (yoghurt, milk, etc.)	76.1 (436)	62.8 (292)a	70.6	78.4 (243)	65.5 (163)a	72.6	73.4 (193)	59.7 (129)a	67.2
−1	Has commercially baked goods or pastries for breakfast	47.5 (272)	46.5 (216)	47.2	47.1 (146)	41.8 (104)	44.7	47.9 (126)	51.9 (112)	49.7
+1	Takes two yoghurts and/or some cheese (40 g) daily	46.2 (265)	47.5 (221)	46.8	50.3 (156)	48.6 (121)	49.6	41.4 (109)	46.3 (100)	43.6
−1	Takes sweets and candy several times every day	33.0 (189)	27.8 (129)	30.4	34.5 (107)	25.7 (64)	30.6	31.2 (82)	30.1 (65)	30.7

a: Pearson’s Chi-square test, P < 0.01.

b: Mann–Whitney U -test, P < 0.05.

Score distribution by medical school year ( table 2 ) showed a significant increase in the percentage of students (both genders) who skip breakfast (22.9% vs. 26.5%) and eat vegetables (58.5% vs. 64.3%) and a reduction in those who consume pasta and rice (72.5% vs. 69.3%) as they progress through medical school. Further score stratification by gender disclosed an increase in the percentage of females with poor scores and a small increase in the proportion of those with good scores ( table 2 ).

Ordinal logistic regression for the outcome ‘good MeDi adherence vs. poor/average adherence during the degree course’ showed that female medical students are significantly more likely to show good MeDi adherence than males (adjusted OR = 1.38, P < 0.05), whereas comparison of early vs. late medical school groups did not have a significant effect on MeDi adherence (adjusted OR = 0.98, P = 0.66).

Overall, data analysis disclosed a poor adherence to the MeDi by medical students, particularly males. The questionnaire findings also indicated that attending medical school does not significantly affect MeDi adherence by students.

Discussion

The aim of this cross-sectional study was to investigate whether attending medical school changes students’ adherence to the MeDi.

Although medical students are expected to possess and apply advanced notions on healthy eating, our data demonstrate that only 22.6% showed good MeDi adherence. Because all Italian medical schools share a similar curriculum, it is conceivable that the data we found apply to medical students all over Italy. ¹⁰ Interestingly, the proportion of our students reporting MeDi adherence is quite similar to that found among Italy’s young people ¹⁷ and is in line with data reported for students attending other university courses in various Mediterranean countries, in particular with regard to consumption of fruit and vegetables. ^18,19 Our findings also agree with those of a Greek study, the sole previous investigation exploring eating habits of medical students. ²⁰ Similar to our study, a paper by Baldini et al. ⁹ found that Italian and Spanish university students eat large amounts of fatty food and too few vegetables, and noted that overweight seemed to be related not only to low levels of physical activity, but also to poor MeDi adherence.

The greater MeDi adherence found among female students in our and other studies is a reasonable finding, since women’s greater preoccupation with body weight and fatty food is highly likely to affect their nutritional habits. ^21–23 Wong et al. ²⁴ found higher nutrition knowledge scores and more correct nutrition behaviours among female than male students, and concluded that gender affects eating habits. They also noted that the students who had higher nutrition behaviour scores also tended to eat more green vegetables and skipped breakfast less often. In our survey, males skip breakfast significantly more frequently than female students (30.5% vs. 19.7%).

Several standalone studies have highlighted a lack of confidence and inadequate nutritional knowledge among European physicians. Several survey have found that physicians agree on the importance of nutrition in their medical practice, but do not feel comfortable and adequately trained to provide nutrition counselling to their patients. ^25–28 Indeed, low self-reported competence in this regard has also been described; ^28–30 this involves that the patients of these physicians are unlikely to receive a thorough nutrition assessment and interventions. Notably, Chung et al. ³¹ calculated that European medical curricula provide on average 23.7 h of required nutrition education. A stronger nutrition curriculum is clearly warranted, to provide a firm and broad knowledge basis for dietary advice and nutrition-related health promotion. In addition, one of the tenets of social cognitive theory is that following a given behaviour requires a person to know what the behaviour is and how to follow it. ³² In our case, medical students would therefore need to be aware of what a healthy or a Mediterranean-style diet involves, and be taught its principles and how to follow it. However, social cognitive theories are often insufficient to explain or address why knowledge is not turned into action. A possible relationship has also been hypothesized between diet quality and other social and cultural factors, such as nutrition knowledge and beliefs, which are considered as important factors affecting food choices. ^33,34 Paradoxically, the insufficient attention paid to nutrition and the relationship between diet and its long-term health effects in the medical curriculum may result in a situation where the dietary advice given by medical professionals is influenced to some degree by cultural or local beliefs. Indeed, culturally based beliefs may be so deeply rooted that the education received by healthcare professionals may filtered through them. Leeman et al. ³⁵ investigated beliefs and attitudes about the relationship between diet, health and other health-related issues in physicians from different countries and in physicians and lay individuals from the same country (France, Germany, Italy, UK and USA). Interestingly, they found a stronger resemblance of beliefs and attitudes among individuals from the same culture or country, whereas the resemblance between lay individuals from different countries and between doctors of different cultures or nations was much weaker. These interesting findings reflect the importance of culture over and above formal training.

Dietary behaviours including food choices are determined by a myriad environmental and individual factors. In an interesting study, Beydoun et al. ³⁶ tested the possibility that nutrition knowledge and belief are an effect modifier in the relationship between socio-economic status factors, fruit and vegetable consumption, and diet quality among US adults. They concluded that an overall better dietary quality is achieved with increased socio-economic status among subjects with healthier nutrition knowledge beliefs. The above considerations suggest that knowledge is often not the main factor affecting changes in behaviour.

The limitations of our study include the facts that findings are based on self-reported data; that only the qualitative data related to respondents’ dietary habits were examined; and that participants were not asked to report whether their food choices were related to pathological conditions. Finally, this is a cross-sectional study; as a consequence it has all the limitations of this type of investigation, especially with regard to causality. However several studies have shown that self-reported eating habits correlate with objective eating habit measurements. ¹³ We therefore believe that the present data provide valuable indications for future investigations.

A strength of our study is sample size. Moreover, it is the sole investigation targeting Italian medical students, since existing studies of the nutritional habits of university students describe the main characteristics of the diet in populations that do not include medical students. ^18,37 Another strength is that it evaluates overall eating style; it thus enables overall diet evaluation compared with studies focusing on a single food item or individual nutrients. ³⁸ Furthermore, the anonymous, self-administered questionnaire favoured participation and truthfulness, and required limited resources. The KIDMED index ¹² is an interesting instrument that has been used throughout the Mediterranean, enabling comparison of the eating habits of young people from Spain, Greece, Turkey, Portugal and Cyprus.

Notwithstanding its limitations, our study raises a number of issues that should be tested in the teaching of medical students.

Universities are ideal places to reach large numbers of young adults. Nutrition education is particularly important for medical students, since physicians have a crucial role and unique opportunities to advise patients on diet (it has been estimated that they have annual contact with 80% of their patients) ³⁹ and because of the severity of the diseases associated with poor MeDi adherence, such as obesity, cardiovascular disorders, cancer, neurodegenerative diseases. Nutrition education programmes directed at university, especially medical students, should thus be encouraged, because a greater dietary knowledge appears to be related to healthier eating habits. ^40–45

In conclusion, our findings suggest that medical students know little about healthy eating and the MeDi. This situation clearly needs to be reversed, because a healthy diet is a mainstay of any medical treatment and a key factor in preventing disease. Improving the nutrition curriculum will help future physicians offer effective treatment and prevention and contribute to achieving Objective 3 of the WHO European Food and Nutrition Action Plan 2015–20.

Funding

This study was supported by grants provided by the University of Catania, Catania, Italy.

Conflicts of interest : None declared.

Key points

• A similar proportion of our medical students and lay peers have non-healthy eating habits.

• Stronger nutrition education should be provided in medical curricula.

• Improving physician nutrition education can have important implications for long-term population health.

• Healthier eating can prevent or reduce the risk factors for a range of debilitating conditions.

Acknowledgements

The authors are grateful to Prof. Gaetano Catania, President of the Medical Degree Course of Catania University for his help, and to Marine Castaing for her assistance with the multivariate analysis. The authors also acknowledge Word Designs ( www.silviamodena.com ) for the language revision.