In this real world non-randomized intervention to promote healthy eating in military setting, we found significant intervention results mostly in favor of Intervention Group. Measured by food indexes and individual food items, Intervention Group showed positive changes in several cereal fiber, fat and sugar containing foods compared to Control Group. Still, the intervention was not successful in increasing fruit and vegetable intake. These findings were not compromised by some significant differences between study groups at baseline, as these differences were taken into account in statistical analyses .
The positive results supported largely the nutritional goals of the intervention. Increasing fruit and vegetable consumption was not achieved contrary to a Danish military refectory intervention . This result is unfortunate because earlier we have found that before service only 13% of young men ate vegetables daily and as few as 8% fruits and berries . Earlier soldiers’ vegetable consumption has been shown to be low not only in Finland  but also in Norway [39, 40]. One reason to possibly explain this result is the renovation of one of the garrison refectories during the study. This may have affected especially serving fresh salads and fruit because the refectory had to operate under field conditions. Also, fruit and vegetable consumption can be challenging to assess by self-reported measures especially among young men. Nevertheless, several intervention actions aimed to increase the supply and also consumption of fruit and vegetables.
A multitude of actions was developed in the intervention in the two main military eating environments. The actions were creative, diverse and can be divided into two categories: some were apparent and could be measured quantitatively. Others were less-visible to improve the quality of the food supply. The distinction has importance because the first-mentioned were the main focus of this study when using eating frequencies as main outcome measures. The last-mentioned can also improve the diet of conscripts but are not easily assessed by self-report measures. However, all actions aimed to promote healthy food supply and to meet the nutritional goals of the intervention. To examine the complete effect of the intervention, dietary habits are to be explored in more detail and potentially in relation to health risk factors .
Other dietary environmental interventions in the military setting can be considered with regards to these results. In the United States, new recipes for healthier food items were developed in order to decrease soldiers’ fat, cholesterol and sodium intake. This was successful as acceptability was positive for the new foods developed specifically for institutional eating at an army garrison refectory . Sporul et al.  found that it is challenging to influence soldiers’ food choices with nutrition labeling of healthy foods at point-of-purchase. Healthy food labeling was reinforced with promotional posters but still sensory attributes such as taste, quality and appearance influenced meal selection most.
Environmental dietary interventions can be conducted also at the workplace or school setting. Intervening can involve physical and informational environments with the twofold aim of increasing the availability of healthy foods and providing education and support for health choices . These settings lack the controlled environment where physical fitness is of importance. Also, military service is a period of institutional life where actions are externally directed. Individual’s possibilities of making independent decisions are limited. This applies also to eating because completely free dietary choices are unfeasible. The peer group is especially important in adolescence and it has major influence in developing both eating habits and lifestyles . The role of the peer group may be even stronger in the military setting where individuality is diminished and actions directed. As dietary interventions can be more effective for young females , actions need to be targeted directly for males and gender specificity of actions needs to be improved especially in this setting. In all, improving implementation and ensuring maintenance of the positive achievements is important .
Other military interventions have focused on individual dietary education while tackling overweight and obesity. Approaches with dietary education and restriction of energy intake have resulted in reduction of body weight [19, 22]. Also, behavioral internet treatment has been successful in weight loss, increasing fruit and vegetable consumption and decreasing snacking . When dietary education was combined with a fitness program, positive results were achieved. The exercise-plus-diet intervention group lost weight and reduced energy intake, the proportion of fat and saturated fat in the diet. Also they increased fiber intake notably by increasing bread, fruit and vegetable consumption .
Our study confirms earlier findings  suggesting that military service per se is a significant intervention for conscripts also in terms of diet. Both study groups exhibited short and long-term changes including increased consumption of cereal foods and a decrease in fatty foods. However, the effect of service is not only a positive one as consumption of sugar-rich foods increases . This could be difficult to influence because the military affects also food attitudes: craving for sweet foods, using food as a reward and as a source of pleasure increases . However, the intervention was successful in decreasing consumption of several sugar-rich foods.
One issue to be considered is that first measurement time point (T0) was not identical in both groups as in Control Group the study questionnaire was sent home and responding took place before entering service. Thus, the time frame of responding expands from less than one month before military service to entering service. Respectively, Intervention Group responded at the beginning of military service. They were instructed to answer retrospectively regarding civilian life before military service. Advantageously, the retrospective time period is relatively short but this fact may affect and thus limit the interpretation of results. For collecting dietary data, a frequency method is applicable in short-term recall as it produces less excess in overestimation, within-person error and interpersonal variation  and it has also been used [46, 47]. The issue was accounted in analyses when baseline consumption frequency was adjusted for as a covariate in the repeated measures ancova. This is an advisable procedure if there is an association with outcome measures , as was the case here (p < 0.01 for all, data not shown), especially when analyzing change from baseline and with continuous outcome measures .