The dietary assessment exercise was performed during 2011–2013 at the Sahlgrenska Academy by 204 medical students (109 women and 95 men) and 75 nursing students (61 women and 14 men) with an average age of 23±4 years. Mean body weight in male students was 75.5±8.8 kg, corresponding to a BMI of 22.9±2.4 (range 18.7–29.1). Mean body weight in female students was 60.4±8.4 kg, corresponding to a BMI of 21.8±2.6 (range 16.9–28.1). In addition, 65 medical students were given the four questions on two different occasions to check the reproducibility.

Four questions regarding consumption of 1) vegetables and root vegetables, 2) fruits and berries, 3) fish and shellfish, 4) rolls, sweets and chocolates, and sugar-sweetened drinks (1) were answered during an introductory lecture before completing the food diary. Each question could yield 0–3 points depending on the frequency of consumption, with higher scores for more frequent consumption on the first three questions, and the other way round on the fourth question. Depending on scoring from 0 to 12 points, subjects were classified as proposed by SoS to have low (0–4 points), average (5–8 points), or high (9–12 points) adherence to dietary recommendations.

During 3 consecutive days (2 weekdays and 1 day in the weekend), the students registered their food intake in a food diary, where they reported type and amount of food eaten. They measured (or when not possible; estimated) the amount as volume or weight of each foodstuff which were entered into the computer programme DIETIST XP version 3.2 (Kost och Näringsdata, Bromma, Sweden).

Basal metabolic rate (BMR) and predicted total energy expenditure (TEE) were calculated by the students during the lab exercise, using equations from Table 9.1 in NNR 2004 (4). TEE was calculated as BMR×physical activity level (PAL), which was selected by each student from Table 9.9 in NNR 2004 (4). Energy intake from 3-day food records was validated by comparison to the calculated BMR. We used a lower cut-off limit of BMR×1.04 as suggested by Goldberg et al. (5).

Mean values of dietary intake were calculated in Microsoft® Office Excel 2010 Pro for all the selected nutrients for both genders and presented as mean±standard deviation. Statistical analyses were made using SPSS® software version 20.0 (SPSS Inc., Chicago, IL, USA). A p-value of <0.05 was considered statistically significant. Medians and 25–75 percentiles were used for non-normally distributed variables. In comparisons, nutrient density per 10 MJ was used, and differences assessed by ANOVA with Bonferroni post hoc test (Table 1). Non-normally distributed variables were log-transformed before statistical testing. Correlations between adherence groups and dietary intake of nutrients and food items were assessed by Spearman's correlation coefficient.

Calculated BMR and calculated TEE were on average 7.6±0.5 MJ and 13.8±1.9 MJ in male and 5.9±0.4 MJ and 10.3±1.8 MJ in female students, respectively. Estimated PAL was thus 1.80 in male and 1.77 in female students. Energy intake in male and female students was 11.8±3.0 MJ and 8.5±2.2 MJ, corresponding to PAL values of 1.55±0.39 and 1.46±0.34.

Reproducibility between assessments with the four questions tested twice in 65 medical students showed no significant difference in mean score, 6.92 versus 7.05 (p=0.63). Fifty-five per cent had the same number of points on both occasions, whereas 20% increased their scores and 25% decreased their scores. Four individuals changed from lowest to middle adherence group, and three from middle to highest group, whereas two changed in the opposite direction for each cut-off point.

No significant differences in scores were found between men and women or between medical and nursing students for each gender. Most students, 64%, were classified as having average adherence to dietary recommendations, whereas only 6% were classified into low adherence, and 30% into high adherence. Dietary intake of fibre, ascorbic acid, and folate was significantly higher in the high adherence group compared to both the other groups (p<0.01). In the low adherence group, median intake of ascorbic acid was 63 mg (44–78), compared to 100 mg (63–147) and 149 mg (112–202) in the average and high adherence groups, respectively. One student in the average adherence group reported an intake of only 8 mg ascorbic acid, which is below lower intake level (LI) of 10 mg/day according to NNR 2004 (4). In the low, average, and high adherence groups, median intake of folate was 211±64 mg, 312±131 mg, and 413±166 mg, and the proportion of students reporting folate intake below LI (<200 µg/day) was 31%, 9%, and 0%, respectively.

Intake of fibre, ascorbic acid, and folate were all fairly well correlated to the adherence groups, with increasing intake from low to high adherence, r_Spearman=0.37, 0.35, 0.39, respectively (p=0.01 for all). Intake of vitamin D was less well correlated, with r_Spearman=0.15 (p<0.05). The proportions of students reporting vitamin D intake below LI (<2.5 µg/day) were 6%, 12%, and 7% in the low, average, and high adherence groups, respectively.

Intake of sucrose was negatively correlated to adherence group, r_Spearman= − 0.22 (p<0.05), whereas intake of fish and fruits and vegetables was positively correlated to adherence groups, r_Spearman=0.32 and 0.52, respectively (both p=0.01). The original score (0–12 points) was slightly higher correlated with the intake of fish and fruits and vegetables, r_Spearman=0.34 and 0.58, respectively (p<0.0001).

Intake of fruits and vegetables was significantly different between groups, with increasing amounts with increasing adherence. The group with high adherence had higher intake of fish and vitamin D than the average group (p<0.018), whereas the low adherence group had higher intake of sucrose than the other groups (p<0.005).

This is the very first attempt to validate the four questions from SoS (1), suggested to be able to discriminate individuals with low compliance to dietary guidelines. Medical and nursing students are supposed to be more health conscious than the general population, which is mirrored in the high scores attained on the four questions. In these students, only 6% were classified as low adherers (which could reduce the statistical power to detect any differences) and 31% as high adherers, whereas outcome in the general population is reported to be approximately 20% and 10%, respectively (6).

We found the questions to be reproducible and to correspond well to the dietary intake of fibre, ascorbic acid, folate, and the intake of sucrose, fish, as well as fruit and vegetables reported in the food diaries. This in spite of the very simple method based on four questions with four pre-defined alternatives for consumption frequency.

Students in the low adherence group did not report different energy intake compared to the average or high adherence groups, nor did their actual mean intake of vitamin C, vitamin D, and folate fall below LI according to NNR, although it was lower than in the higher adherence groups.

The proportion of students reporting vitamin D intake below 2.5 µg/day was actually higher in the average adherence group. This could possibly be due to a high proportion of vegetarians and vegans among these students who, by not eating fish, would score 3 points less on the ratings, and thus be more likely to end up in the average rather than high adherence group. This could be taken as a weakness in the method, but will also be a consequence of the very limited amount of vitamin D in food items other than fatty fish.

Thus, based on these findings, these four questions could possibly be used to quickly and easily rank individuals according to their general adherence to dietary guidelines. The questions on fish and sweets regard weekly consumption, which a 3-day food record cannot be expected to cover fully. Still, we found significant differences in sucrose intake between the low and average adherence groups, and also significant differences in fish intake between the low adherence group and the other groups.

Four questions on the consumption frequency of vegetables, fruits, fish, and sweets are reproducible and correlate well with the in dietary fibre intake of fibre, ascorbic acid, folate, vitamin D, fish, sucrose, and fruits and vegetables as assessed by 3-day food records in health-conscious medical and nursing students.