| In recent years,the rapid development of the economy has had a huge impact on the lifestyle and dietary structure of residents,resulting in corresponding changes.Through monitoring,it has been found that there have been significant changes in the chronic disease spectrum of the population,which has also made people increasingly concerned about diet and health.Total Diet Study(TDS),as a representative method for studying and estimating the various dietary chemical components(nutrients,pollutants)ingested by a certain population through cooking and processing,including drinking water,can not only effectively address the targeting of the population,but also measure food samples after cooking and processing,and cover a large number of survey subjects.It is the most economical and effective dietary exposure assessment method.It is also an internationally recognized,most cost-effective and reliable method recommended by the World Health Organization to assess the intake of nutrients and exposure to chemical hazards in a country or region across different population groups,as well as the potential health risks associated with the intake of these substances.China started relatively late in TDS,and this study was conducted for the first time in Shandong Province using the method of TDS.Objective:1.This study aims to carry out TDS to know and catch the dietary status of residents in Shandong Province,and to evaluate dietary exposure to common food pollutants among residents.2.To analyze the dietary patterns of residents in Shandong Province,studying the relationship between dietary patterns and common chronic diseases,and establishing a BP neural network model for common chronic diseases through R software,a new and more effective approach is provided for predicting and warning chronic diseases.Methods:1.Dietary status of residents in Shandong province and assessment of dietary exposure to common food pollutants:⑴According to the method of TDS,survey groups were designed and six city-level survey points representing the average dietary composition and level of residents in Shandong Province were selected,including Yantai City,Weifang City,Jinan City,Liaocheng City,Tai’an City,and Linyi City.Two counties and one district with moderate economic level were selected as county(city,district)level survey points in each city,with a total of 18 county-level survey points for research.Using multi-stage stratified cluster random sampling method,a total of 3667 households and 11821 Shandong residents were selected to participate in this research survey.⑵The point assessment method was used to assess dietary exposure of lead,cadmium,mercury,arsenic,chromium,aluminum,fluoride,nitrite,in 12 major food categories including grains,meat,beans,etc.To use the median,rate or composition ratio,average consumption((?)±s)are used for statistical description,and the comparison between counting data is usedχ2 test,P<0.05 was statistically significant.Data collection and analysis were conducted using Epiinfo7.2software and Excel 2007,using descriptive epidemiological analysis for relevant indicators.2.The relationship between dietary pattern and common chronic diseases:⑴Establishing a database of dietary patterns using data obtained from the food frequency survey questionnaire.The 106 types of food in the food frequency survey questionnaire were divided into 12 categories:grains and Potatoes,beans,fungi and algae,fruits,vegetables,nuts,meat,eggs,aquatic products,dairy,snacks,and beverages.The sum of the average daily intake(g/d)of various foods in each type of food is the intake of that type of food.The division of dietary patterns is performed using principal component analysis,with Kaiser Meyer Olkin(KMO)values greater than 0.6 and Bartlett’s sphericity P-value less than 0.05.To analyze a the association between different dietary patterns and common chronic diseases,and analyze the dietary patterns among different regions and urban/rural populations participating in the survey,comparing the intake of heavy metals in different dietary patterns.The quantitative data comparison between different dietary patterns was conducted using the Kruskal Wallis test,while the count data comparison was conducted usingx~2test.The statistical tests are all bilateral tests,with a significance level of P<0.05.⑵Establish logistic regression models and BP neural network models for common chronic diseases using data from medical examination forms and personal health status questionnaires,and compare them.Logistic regression model was established by using R software version 3.3.1,with statistically significant indicators selected from single factors as independent variables,and the disease attribute of the study object as dependent variables.The maximum likelihood ratio forward method was used for stepwise regression analysis.The inclusion criteria of the selected variables are P<0.05,the standard of excluding variables is P>0.10.Then the BP artificial neural network prediction model is established by R software version 3.3.1.Modify the connection weight and activation threshold of each layer,and finally achieve the stable convergence of the network.Use the tanh function as the activation function.The criterion for selecting the prediction probability is 0.6.The value≥0.6,the prediction result is Positive,otherwise it is negative.Results:1.Dietary status of residents in Shandong Province and assessment of dietary exposure to common food pollutantsThis survey was conducted using the method of TDS.It began on September 1,2015 and was gradually launched at various survey points.By the end of March 2016,on-site investigations were completed,with a total of 3667 households and 11821people surveyed.Among them,the 3-day 24-hour dietary review survey was conducted on 1206 households and 3963 people,the food frequency method survey was conducted on 1225 households and 3929 people,and the ready-to-eat food survey was conducted on 1236 households and 3929 people.⑴The results of the 3-day 24-hour dietary survey revealed that:(1)The survey was conducted on the consumption of 12 types of food among residents,including grains,beans,nuts,potatoes,meat,eggs,aquatic products,dairy,vegetables,fruits,sugars,beverages,and alcohol.From the perspective of the change in the standard daily consumption of all kinds of food,the average consumption of food in cities is higher than that in rural areas,with the exception of cereals and their products,vegetables and their products,and alcohol.The average consumption of grains,potatoes,dairy,vegetables,fruits,sugars,and other foods in urban areas has increased,while the average consumption of sugars in rural areas has increased.,the other food categories showed a decline or basically no change.(2)Compared with 2012,the daily intake of dietary energy and nutrients per standard person increased by 28.3 kcal,urban residents increased by 248.9 kcal,and rural residents decreased by 98.6 kcal.(3)Cereals are still the main food source of energy,accounting for 47.2%of the total energy,including 44.1%in urban areas and 49.3%in rural areas.(4)The total energy supply ratio of protein is 13.3%,including 13.7%in urban areas and 13.0%in rural areas.(5)The total energy supply ratio of fat in food is 35.0%,including 36.5%in urban areas and 34.0%in rural areas.(6)Cereals provide 41.2%of dietary protein,36.4%in urban areas and 44.5%in rural areas.(7)The sources of dietary protein and beans accounted for 6.6%,including 6.5%in urban areas and 6.7%in rural areas.(8)Animal food sources of dietary protein accounted for 22.7%,including 26.6%in urban areas and 19.9%in rural areas.(9)Animal foods,beans and other foods account for 29.3%of the protein source.⑵The results of the food frequency questionnaire survey showed that:(1)The largest daily average consumption of staple foods is wheat flour and products(raw weight)(202.5 g).(2)The largest daily average consumption of beans is soybean milk(51.2g).(3)The largest daily average consumption of vegetables is leafy vegetables(76.9 g).(4)The largest daily average consumption of bacteria and algae is kelp(fresh weight)(16.1g).(5)The largest daily average consumption of fruit is the fruit of kernel(59.8g).(6)The largest daily average consumption of milk is full-fat liquid milk(100.8g).(7)The largest daily average consumption of meat is fresh(frozen)pork(36.2g).(8)The largest daily average consumption of aquatic products is molluscs(31.1g).(9)The largest daily average consumption of eggs is fresh eggs(46.5g).(10)The largest daily average consumption of snacks and snacks is bread(16.3g),and the largest daily average consumption of beverages and condiments is tea drinks(36.0g).⑶Investigation of ready-to-eat food::(1)The largest daily average consumption of cooked meat products is cooked mutton(prepackaged)(10.6g).(2)The largest daily average consumption of foods such as rice noodles and dairy products is bulk dairy products(95.6g).(3)The largest daily average consumption of aquatic products is prepackaged crab(16.1g).(4)The largest daily average consumption of raw aquatic products is other aquatic products(9.0g).(5)The largest daily average consumption of raw meat is pork(23.5g).(6)The largest daily average consum Ption of raw vegetables is tomato(24.3 g).(7)The largest daily average consumption of fruits is apple(57.4g).(8)The highest heating rate before eating is cooked meat products(91.8%),and the median heating time is 10 minutes,the heating rate of aquatic products is the lowest,and the median heating time is 20 minutes.⑷The dietary exposure assessment found that:(1)The average MOE value of dietary lead exposure for residents of all genders and age groups in Shandong Province is greater than 100,indicating that the im Pact of dietary lead on human health can be almost negligible.(2)The dietary cadmium exposure of each sex and age group did not exceed the provisional monthly tolerable intake(PTMI)established by the Joint Expert Committee on Food Additives of the Food and Agriculture Organization of the United Nations and the World Health Organization(JECFA).(3)The exposure of dietary mercury and aluminum is far lower than the provisional tolerable weekly intake(PTWI)established by FAO/WHO.(4)Except for females aged2～17,the MOE values of dietary arsenic ex Posure in different populations are greater than 100.(5)The risk index of dietary chromium and fluorine exposure is0.04～0.25.(6)The exposure of dietary nitrite in each sex and age group is about 10%of the daily allowable intake(ADI).2.The relationship between dietary pattern and common chronic diseases⑴Establishment of dietary patterns and related analysis:(1)Establish a dietary database using the food frequency survey questionnaire obtained from 1225households and 3929 people.Through the principal component analysis method,three dietary patterns are constructed.Dietary pattern 1 is mainly composed of grains,potatoes,fruits,vegetables,meat and eggs in dietary intake.Dietary pattern 2 is a pattern with a higher proportion of intake of aquatic products,bacteria,beans,and other foods,while dietary pattern 3 is a pattern with a higher proportion of intake of dairy,snacks,beverages,and other foods.(2)The overweight and waist to hip ratio exceeding standard population are mainly based on dietary pattern 1.There is a statistical correlation between different dietary patterns and the prevalence of common chronic diseases(P<0.05).(3)The intake of lead,mercury and chromium in the dietary pattern 1 was significantly higher than that in dietary the dietary pattern 2and the dietary pattern 3,while the intake of cadmium in the dietary pattern 1 was significantly lower than that in the dietary pattern 2 and the dietary pattern 3.The intake of cadmium and arsenic in the dietary pattern 2 was significantly higher than that in the dietary pattern 1 and the dietary pattern 3,and the intake of lead and mercury in the dietary pattern 2 was significantly lower than that in dietary the dietary pattern 1 and the dietary pattern 3.The intake of arsenic and chromium in the dietary pattern 3 was significantly lower than that in the dietary pattern 1 and the dietary pattern 2.⑵BP neural network models for hypertension and obesity:(1)The BP neural network model of hypertension was established using the data of the medical examination form.The accuracy of the BP neural network model of hypertension was71.1%,and its confidence interval was(0.68,0.74).After adding the indicators of personal health survey,the accuracy of BP neural network model for hypertension increased to 72.5%,and its confidence interval was(0.67,0.75),higher than the accuracy of the logistic regression model for hypertension was 70.1%.(2)The accuracy of BP neural network model for obesity was 88.7%,and its confidence interval was(0.86,0.92)when using the variable index in the medical examination table.After adding the indicators of personal health survey,the accuracy of BP neural network model for obesity is 75.3%,and its confidence interval is(0.73,0.78),higher than the accuracy of traditional regression model for obesity was 71.5%.Conclusions:1.There are differences in the types and quantities of food consumption between urban and rural residents in Shandong Province,factors such as age and region affect the food choices of residents,the dietary structure is unreasonable,the proportion of fat energy supply is high,and the intake of the three major nutrients and food sources of residents are different between regions and populations.The intake and frequency of multiple types of foods such as beans,aquatic products,and dairy products did not meet the corresponding recommended standards.2.The intake of common food pollutants in the diet of residents in Shandong province is generally at a safe level.Attention should be paid to the intake of heavy metals in the age group of 2～17 years.The main dietary sources of heavy metal and harmful elements are cereals,vegetables and aquatic products.3.Three dietary patterns were constructed using principal component analysis,and it was found that there were regional differences in dietary patterns,which were associated with the occurrence of common chronic diseases and showed significant differences.4.The BP neural network model is more suitable for predicting common chronic diseases than the logistic regression model. |
PDF Full Text Request