Development of a Machine Learning Model for Classifying Cooking Recipes According to Dietary Styles.

To complement classical methods for identifying Japanese, Chinese, and Western dietary styles, this study aimed to develop a machine learning model. This study utilized 604 features from 8183 cooking recipes based on a Japanese recipe site. The data were randomly divided into training, validation, and test sets for each dietary style at a 60:20:20 ratio. Six machine learning models were developed in this study to effectively classify cooking recipes according to dietary styles. The evaluation indicators were above 0.8 for all models in each dietary style. The top ten features were extracted from each model, and the features common to three or more models were employed as the best predictive features. Five well-predicted features were indicated for the following seasonings: soy sauce, miso (fermented soy beans), and mirin (sweet cooking rice wine) in the Japanese diet; oyster sauce and doubanjiang (chili bean sauce) in the Chinese diet; and olive oil in the Western diet. Predictions by broth were indicated in each diet, such as dashi in the Japanese diet, chicken soup in the Chinese diet, and consommé in the Western diet. The prediction model suggested that seasonings and broths could be used to predict dietary styles.

Impact of an online nutrition management application service on glycaemic management in individuals with diabetes: A propensity-score-matched retrospective cohort study.

AIMS: To assess the impact of 'Oishi Kenko', a nutrition management application (app), on glycaemic control in patients with diabetes. MATERIALS AND METHODS: A propensity-score-matched retrospective cohort study was performed using data from the KAMOGAWA-DM cohort study conducted between January and June 2022 in Japan. We analysed data from patients with type 1 and type 2 diabetes, comparing users who used the Oishi Kenko app (app group) with non-users (control group) over 3 months. RESULTS: Among the 50 participants who actively used it, 47 participants in both the app and control cohorts were selected from the KAMOGAWA-DM cohort according to propensity-score matching. Within the app group, the median glycated haemoglobin (HbA1c) level was 51 mmol/mol (6.9%) at baseline, which slightly decreased to 50 mmol/mol (6.8%) at the 3-month mark (median change 0.0%). Conversely, in the control group, the baseline HbA1c level of 51 mmol/mol (6.9%) exhibited a marginal increase of 52 mmol/mol (7.0%) after 3 months (median change 0.20%). The median HbA1c level change between the groups was statistically significant, with the app group showing a significant positive change compared with the control group (p = 0.012). CONCLUSION: The Oishi Kenko app effectively improved glycaemic control in patients with diabetes; hence, it may be a promising tool for patient-driven dietary management.

Breakfast Skipping is associated with More Deleterious Lifestyle Behaviors among Japanese Men: The TRF-Japan Study Using the Original "Taberhythm" Smartphone Application.

Background: Time-restricted eating has been increasingly recognized as a promising option to reduce food intake and combat obesity. Especially in Asian countries such as Japan, because of the wide variety of food choices available, a dietary approach that emphasizes meal timing can be more practical and easier to implement and adhere to, compared with approaches that focus on specific dietary content, such as low-fat or low-carbohydrate diets. Objectives: We aimed to identify eating patterns among Japanese men and women using a smartphone application (app) called "Taberhythm." In addition, we sought to evaluate the relationship of breakfast eating habits with lifestyle behaviors and body mass index, and determine whether sex differences were present. Methods: A total of 3369 smartphone users were eligible to participate in this observational study. Users recorded 1 mo of lifestyle logs using the app; 254 participants (178 women, 38 ± 12 y old, body mass index 23.3 ± 4.9 kg/m2) had sufficient records to calculate daily fasting duration and sleep duration, and were eligible for the analyses. Results: Fasting duration was ∼12.6 h and was longer in women than men, among participants who never skipped breakfast. Breakfast skipping was associated with longer screen time, and more frequent snacking, only in men. Men with irregular breakfast eating patterns had a longer duration of fasting after awakening that was associated with obesity. Conclusions: We investigated eating patterns among Japanese people using a smartphone app and revealed that skipping breakfast was more deleterious in men than in women.

Individuals with Type 2 Diabetes Mellitus Tend to Select Low-Carbohydrate, Low-Calorie Food Menus at Home on Diet Application.

(1) Background: From the perspective of patient-centered care, it is important for medical professionals involved in diabetes care to know the role of choice behavior when individuals with type 2 diabetes mellitus select their meals at home. In Japan, online meal management applications are widely used to help individuals to prepare healthy, colorful, and tasty meals. (2) Objective: To assess menu selection from an online diet management application in individuals with type 2 diabetes mellitus over a period of 24 months. (3) Method: The saved data of the selected food menus on the online diet management application were analyzed. We identified specific nutritional groups of the food menus, called nutritional clusters, by clustering the multidimensional data of the nutrients after de-dimensioning them. Then, we analyzed the constitutional nutrients of each nutritional cluster with the highest and lowest frequencies of selection by the users of the application. (4) Results: In all, 9674 food menus made by 3164 people were included in the analysis, and 12 nutritional clusters were identified. Low-carbohydrate and low-calorie food clusters showed the highest selection frequency. The average caloric value of 149.7 kcal and average carbohydrate ratio of 47% in the cluster with the highest selection frequency were significantly lower than the average caloric value of 435.2 kcal and carbohydrate ratio of 63% in the cluster with the lowest selection frequency (p < 0.001, respectively). (5) Conclusion: Individuals with type 2 diabetes in this population preferred to select lower-carbohydrate and lower-calorie food menus at home using online diet management applications. To improve sustained self-management and quality of life, medical professionals may consider incorporating preferred dietary behaviors into medical management of type 2 diabetes mellitus.

Status of online diet management program users in Japan during the 2020 Coronavirus disease 2019 pandemic.

The spread of coronavirus disease 2019 (COVID-19) has led to drastic changes in people's lifestyles, including teleworking and restrictions on socializing. In the context of observing social distancing for preventing infection, the need to maintain fitness and health has attracted particular attention. We aimed to determine the relationship between the increase in the number of active users of online diet management applications and COVID-19 infection rates in Japan. A total of 1.5 million rows of log data was analyzed. The active number of users of online diet management applications increased with increase in the number of COVID-19 infections. The active user number in Kanagawa Prefecture, where the first cases of COVID-19 were reported in Japan, was particularly high. Moreover, there was a significant increase in the user number in prefectures under a state of emergency when compared to that in prefectures not under a state of emergency. The social anxiety caused by COVID-19 is expected to increase the demand for online health management applications further. The use of such programs can aid in achieving social distancing while enabling users to maintain healthy lifestyles.

Possible favorable lifestyle changes owing to the coronavirus disease 2019 (COVID-19) pandemic among middle-aged Japanese women: An ancillary survey of the TRF-Japan study using the original "Taberhythm" smartphone app.

Coronavirus disease 2019 (COVID-19) has had a global effect on people's lifestyles. Many people have become physically inactive and developed irregular eating patterns, which leads to unhealthier lifestyles and aggravation of lifestyle-related diseases; these in turn increase the severity of COVID-19. Prior to the COVID-19 pandemic, we developed a smartphone application called "Taberhythm" to investigate eating patterns, physical activity, and subjective feelings of happiness. We aimed to compare lifestyle data before and during the first phase of the COVID-19 pandemic to objectively assess lifestyle changes during quarantine. A total of 464 smartphone users (346 women, 35 ± 12 years old, body mass index [BMI] 23.4 ± 4.5) participated in Period A (January 7 to April 28, 2019) and 622 smartphone users (533 women, 32 ± 11 years old, BMI 23.3 ± 4.0) participated in Period B (January 6 to April 26, 2020). Compared with Period A, there was a sharp decline in physical activity during Period B (4642 ± 3513 vs. 3814 ± 3529 steps/day, p<0.001), especially during the final 9 weeks in both periods (4907 ± 3908 vs. 3528 ± 3397 steps/day, p<0.001); however, there were large variations in physical activity among participants. We found a surprising trend during Period B toward increased happiness among women aged 30-50 years, the group most affected by stay-at-home policies that led to working from home and school closure. Moreover, daily eating duration declined in this population. Additionally, there was a positive association of happiness with steps per day in Period B (ρ = 0.38, p = 0.02). Despite the many negative effects of the COVID-19 pandemic, subjective feelings of happiness among middle-aged Japanese women tended to increase, which indicates that some favorable lifestyle changes that could be adopted during quarantine in the ongoing COVID-19 pandemic.