Food Science Packaging – Rosemary Tobelmann, Michael T. Goebel, Ann Marie Albertson, General Mills Inc
Abstract for “Systems and methods to determine nutrients in dietary intake”
“The invention uses market research data for accessing nutrient intakes in a population. A 24-hour recall of diet data does not generally reflect a pattern of “usual”. A population’s intake behavior. A unique 14-day food diary method was created to determine the effect of food consumption on nutrient intake. Food industry has used food records for years to track food consumption and monitor growth. However, some of the most valuable databases regarding longer-term (e.g. 14-day) food intake don’t record individual portion sizes. While other databases that are based on short-term (24-hour recall), survey periods can record portion sizes, they do not reflect eating habits. In the preferred embodiment, two of these databases are combined with a third set that provides detailed nutritional information for all foods. This database can be used to create a combined database to process nutrient intake reports by statistical analysis. This flexible system allows users to classify the population based upon?usual?. This flexible system allows the user to categorize the population based on their consumption of specific foods, brands of foods, and/or food categories. Then they can determine dietary differences between them and those who ‘non-use?. counterparts.”
Background for “Systems and methods to determine nutrients in dietary intake”
“Most people shop carefully for the food they feed their families. So that our families have a healthy diet, we look for foods with high nutritional value. We pay particular attention to the food we purchase for our older and younger family members. It is vital to eat the right foods every day for our health and longevity.
“America’s medium-sized supermarket is a marvel of the modern age. Supermarkets have a greater variety of food options than ever before. You can find Texas beef, seafood from Washington State, oranges and New Zealand kiwis, as well as other international foods. There are also a wide variety of packaged and prepared foods. It becomes harder to make sensible and healthy choices for our families and ourselves as the number of products available increases. We often wish we had more direction.
Advertising is a great source of information about food products that we may like to purchase and give to our families. Food product manufacturers want to accurately advertise their products, and they are keen to develop new products that appeal to customers and meet their nutritional requirements. There is strong evidence to support the claim that eating whole-grain foods, including other plant foods, can reduce your risk of developing heart disease. This information is vital for consumers as it allows them to choose healthier foods and make better choices.
Food marketing researchers can use data to determine how to improve a person’s diet by recording detailed information about their food habits. For marketing research purposes, food consumption data is a continuous record of the nation’s food intake. These detailed food records have been used by the food industry to monitor growth and track specific food items.
The NPD survey panel records food eaten at-home and abroad over a period of 14 days. The NET database tracks the consumption patterns and trends for more than 4,000 food and beverage products. It also provides a range of information such as:
The NPD/NET data set can be used to determine what American households eat. Because food is often bought by one household member for the whole household, and because meals are usually eaten together in a given household, it makes sense to place emphasis on household. A practical reason is that only one household member (e.g. the person responsible for food preparation) usually records the survey data required for the entire household. The problem with this emphasis on household data recording, however, is that while the diaries record how many of a particular food item was served to the household, they do not permit any household member to record how many of it he or she eats.
“In detail, each household must fill out a survey/diary asking for information about how much food was served to them, how much was eaten, and who ate that particular food. For example, NPD Group’s Sample Daily Meal Diary is included herein as a reference. This is a common procedure used in panel surveys to reduce the amount of information and increase reliability. It is not necessary to record individual portions. This is because if all participants are required to track how much food they eat over a 14-day period, it can be burdensome and may compromise the accuracy of the recording.
“The NPD/NET dietary intake dataset for certain purposes may have a problem with the quantity of nutritional information it provides. NPD doesn’t attempt to provide nutritional information for every food that was included in its survey. This type of detailed nutritional analysis is usually the work of food researchers and is not provided in the NPD/NET dataset. However, some food research applications may require detailed information about the daily intake of more than 100 nutrients (including individual amino and fatty acid) each day. A company might want to know the daily intake of each nutrient by different demographic groups in order to formulate or reformulate a food product that will ensure Americans get the right essential nutrients. Manufacturers and providers of food products may want to see evidence in order to make claims about their products being part of your daily diet. Health specialists may wish to analyze nutrient consumption or nutrient consumption trends in the population overall, by demographically-specific segments of the overall population, or by household and/or individual, in order to try and discern correlations between nutrient consumption and disease. Other applications require detailed information about the daily intake of nutrients and tracking it over time. The NPD/NET data sets do not address these issues.
There are many good sources of nutrition analysis data for the foods we eat. Many research organizations, including the University of Minnesota have collected the nutritional content of many foods. The University of Minnesota’s Nutrition Database System for Research (NDSR) software provides detailed nutritional information for over 18,000 foods, as well as over 8,000 brand-name items. Although there is a lot of information available about America’s eating habits and nutritional information, it is spread across a variety of data sets that have been created by various entities, including governmental, corporate, and academic. The data sets can be incompatible and each set is designed to accomplish different goals.
“According to an example and an exemplary embodiment, data integration is performed on three separate, special-purpose food research related data sets. One data set includes food consumption data that is based on 14-day diaries. The second data set includes data on portion sizes for large numbers of food types (e.g., more than 8,000). A third set includes nutrient data for many (e.g., more than 18,000) unique food constituents. This integrated data set can be used to analyze the results using standard statistical analysis procedures.
“According to a further aspect of an illustrative or exemplary embodiment, a data set is analysed and processed to determine the mean age and sex specific serving sizes of a number of food items. These portions are then matched to the food in a second set that represents 14-day food intake. A third, nutrient-based data set is used to assign complete nutrient profiles to each food. These data sets are combined into a database. Then, nutrient intake reports can be processed using a standard statistical reporting interface. This flexible system allows users the ability to classify the population based upon their usual consumption of specific food types, brands, and/or brand of food, as well as to determine differences in dietary habits versus non-using. counterparts.”
“In one non-limiting, illustrative embodiment information is combined from three data sources:
This research on dietary intake is of great value. “For example:
“The dietary research results can help to build credibility within scientific and food policy communities.
The illustrative preferred embodiment allows the data sets to explore for new information, trends, and themes that can be transformative and stimulate product development, help to create marketing programs and suggest strategies (e.g. BMI and cereal intake, diet modeling to meet three whole grain per day, etc.). These techniques can also be used in food product marketing and public relation (e.g. sugar defense, whole grains intake, impact of breakfast cereals on diet, calcium intake and breakfast patterns, portion sizes, eating patterns for children, seniors, and other demographic groups etc
“The techniques described herein can also be used for product development and product reformulation (e.g. by identifying the nutrient requirements in a population, such as calcium fortification, folate enrichment and enrichment etc
“The information in the illustrative and exemplary embodiments may be helpful in a regulatory environment to assist with claims documentation, policy development, and data for regulatory comments and fortification review.
“The information could also be used to prepare abstracts and manuscripts for scientific journals, supplement internal and external laboratory research projects, or for any other scientific purpose.”
“Additionally, information from an illustrative or exemplary embodiment could be useful for data for speeches, presentations, public relations facts and advertising copy as well as consumer information.”
“According to a more specific aspect of an illustrative preferred embodiment, we use a food description reduction algorithm that reduces large amounts of food item data from a 14-day diet database into smaller amounts of useful data for identifying nutrients in foods actually eaten by participants in dietary intake studies. A particular embodiment uses a subset of data fields to identify food items on the order over 5,000 from more than a billion possible. For example, this data field subset could include 8-dimensional coordinates that represent food item identification (e.g. type, form and characteristic, flavor, class, preparation method, packaging type and special label code). In the preferred embodiment, many codes are combined and then grouped according to nutritional factors. These codes, when combined, are mapped to nutrient values that are based on food nutrient profiles and portion sizes.
“In the example and the illustrative embodiment, groupings are performed on the basis of a lookup table that uses four keys:
“In the illustrated and illustrative embodiment, food descriptor mapping proceeds by scanning a table that contains data to determine if the food item within the dietary intake data is defined and has a combination key. Multiple iterative scans may yield additional combination keys, which can be combined to create a combination code for the specific food item in the food intake data. The combination code food descriptor generated is stored in a food-portion data file. This data file contains foods that have been previously described by combining a portion data set with a nutrition information data set. The combination code food descriptor is located within the food portion file. In this example embodiment, it is mapped into a simpler unique food designationator (for storage space considerations). If the code is not located (indicating, for instance, that a new food item has been reported in the dietary intake dataset), an exception is created so that a dietary intake specialist can dynamically update the lookup tables to include this new item. This can be done iteratively to allow for interactively defining new foods items as they are introduced and appear in the dietary intake data.
“In accordance to another aspect of the illustrative embodiment, a household Master Analysis is performed to track individuals consumers?even through multiple intake survey from different times periods. For many other research purposes, it is necessary to have individual data about food intake and nutrition. This invention’s preferred embodiment is exemplary in that it can track individual person’s dietary intake using dietary intake data sets. These data sets are usually designed at the household level, but include enough data to allow individual tracking, if handled properly.
“For example, it is possible to get more precise dietary intake results by using data sets containing dietary intake data from different surveys. These surveys often survey the eating habits of the same households as the same people within the households. But, household compositions can change (e.g. when college students move on to the next stage of their lives), and different households may have different reporters/diarists during different periods. In the exemplary embodiment, it is possible to determine whether a household or individual is included in multiple food intake survey. The preferred embodiment assigns each individual a unique individualID. This individual ID is different from the one used to code them within the food intake data sets. To allow individual tracking of dietary intake, unique individual IDs can be assigned to individuals who have reported on different dietary intake surveys. The exemplary embodiment can analyze the food intake survey results using individual data. This allows for more precise results as the eating habits and dietary intakes of individuals can be weighted to make them one person. The ability to track individual consumption over a longer time span, such as years, can provide significant flexibility and advantages. You can find valuable information by tracking how an individual’s eating habits change over time, such as the age of a person.
“According to a further aspect, an exemplary and ilustrative embodiment, data are combined to create demographic-based (e.g. age and sex), categories for portion sizes determinations.
“In accordance to yet another aspect of an exemplary embodiment, recipe file are used for extracting nutrient information form food descriptors. It is possible to identify a food item in the food intake data and determine the nutrients that are present in that particular food. The example embodiment’s food intake survey data does not include information about an individual’s actual portion size. Instead, the portion size information is obtained using a different set of data based on the age, gender, and other demographics of those who ate the food item. The preferred embodiment uses recipes to estimate the nutrition that the consumer received from the food product once the food item and the portion size have been established.
The nutrient information within the nutrient set may not provide a complete nutrient profile for every one of the thousands of food products. The nutrient set may not include information about the nutrition obtained from eating a combination, but the nutrient set might contain complete information about the ingredients of such foods (e.g. flour, butter, milk, oil, and other components of a pancake batter recipe). This aspect of the preferred and illustrated illustrative embodiment is preserved so that recipe files can be used to separate particular food items into their constituent parts. To dynamically determine the overall nutritional content of the food item, the nutrient data can be used to identify the nutrients in each component.
“FIG. “FIG. System 100 uses a unique methodology that utilizes 14-day food diaries and additional data sets to provide nutrient portion information.
“As shown at FIG. 1 is the 14-day food intake data collection 110. This input is used in the overall dietary intake analysis system 100 in the preferred exemplary embodiment. The 14-day food intake data 110 could be, for instance, the NPD/NET dataset (see FIG. 1A) is a result of extensive data collection from many households, who record exactly what they eat over a 14-day period.
The preferred embodiment system 100, shown in FIG. 1. includes a data integration mechanism 200, which links information from the food consumption data set 110 with information in two other data sets, namely, a food nutrition data set 120 and a portion data set 130 in the exemplary embodiment.
“Food nutrient set 120″ is a collection of detailed nutrition data, which are generated by food researchers. It includes information about thousands (e.g. 120 nutrients) of uniquely identified foods that have been assigned to specific 200 (e.g. 179) food groups. The University of Minnesota’s Nutrition Data System for Research is an example of a food nutrient data collection 120. The preferred embodiment 100 uses the University of Minnesota Nutrition Coordinating Center’s NDS-R software to assign complete nutritional profiles to all foods in the 14-day food intake dataset 110. FIG. 1A.”
“As mentioned above, the food intake data 110 used in the preferred embodiment 110 does not include information about the individual’s portion sizes. Instead, portions are kept on a household basis and not for individual consumption. It is important to know the amount of each nutrient that has been consumed by each person. Not just each household. Many households often consist of several individuals from different demographic groups (e.g., older Americans and younger Americans; children of different genders; etc.). Marketing and health research can benefit greatly from individual-based data.
“To address the issue with the exemplary data set 110 on food intake, we link an additional data set 130 on food portion sizes obtained from a different source, e.g. the USDA’s Continuing Survey of Individual Food Intakes (CSFII). FIG. 1A. 1A. The 24-hour recall method of collecting data on diet does not always reflect the ‘usual’ pattern. Food intake. The USDA data set is more comprehensive than the 140-day food intake data set 110 in some respects. The USDA survey requires that individuals record how many calories they ate during 24-hour periods. The USDA data set 130 provides detailed demographic information, such as household size, income, age, and sex. The preferred exemplary and illustrative embodiment 100 uses this USDA data set 130 to obtain demographically-based food portion size data. It is possible to calculate statistically from the USDA data set 130 that, for example, a 34-year-old male eating a steak dinner with french fries and steak will likely eat approximately xxxg of steak, and yyyg of french fries. A 12-year-old female eating the same meal would eat zzzg of steak, and aaag grams of French fries. System 100 uses this demographically-based portion size information to estimate individual portion size of food intakes recorded in the dietary intake data set 110. System 100 could also contain the Pyramid Servings data from USDA CSFII using the CSFII Food codes.
“Briefly the food coding structure used in data integration operation 200 is based upon 53 basic food categories as shown in the exemplary embodiment. Each food category has a unique and progressive level of detail that describes the food in that category. For example, food coding could include the following individual codes:
“Example System Implementation.”
“FIG. FIG. 4 illustrates an example block diagram for an illustrative and non-limiting implementation system 100. 1. FIG. Referring to FIG. FIG. FIG. 5 shows how the mass storage devices 152 (which contain the various data sets 110-120, 130) can be input in a variety ways. For example, the portable mass storage device (154), other optical magnetic (155), or another mass storage device (152) may be supplied. The data sets may also be transmitted over a computer network (158) such as the Internet.
“In the example embodiment 100 mass storage device 152 is shown in FIG. 4, which may contain one or more magnetic disk drives, can also store additional data structures such as a recipe file, a food portion link table 150, a household mastertable 160, and other data structures (e.g. 135). These data structures (145, 150 and 160 in the exemplary embodiment) are created and/or updated automatically by system 100. They are used in data integration process 200.
“As also illustrated in FIG. “As also shown in FIG.
“These diverse processes, 400, 500 and 600, are controlled interactively by one or more users via display workstations 162 to generate a statistical data collection called ‘food base? 140 is the combined data from the three input data sets 110 to 120 and 130. This statistical data set 140 can be stored on the same mass storage device 152 or a different one by the computation block 156. This statistical data set 140 can be transferred to another location using optical disk 154 or 155, network 150, or any other means for further manipulation and analysis.
“In the example embodiment, the exact same or different computation block is 156. A personal computer, for example, performs statistical, trend, and other analysis 700 on a data set 140 in response user commands. To generate reports 310, the data set 140 is input via one or more workstations 162. These reports 310 can be printed, displayed on a computer 162, or transmitted over the network 158 (e.g. via email, web pages, etc.).
“As shown in FIG. 4. The exemplary system 100 also includes an additional Health Focus Data set 135 that is directly linked to households in the food intake data sets 110 (and so to household master tables 160). NPD Group may supply the Health Focus data set 135. The Health Focus data set 135 allows system 100 users to examine attitudes towards health and how they relate to dietary intake at a household-level. In the exemplary embodiment, data set 110 on food intake preferably includes results from multiple surveys (e.g., a survey that is about ten years old and a more recent survey), in order to provide long-term trends information.
Referring to FIG. 5. In one specific example, computer 156 executes several processing routines to implement data integration process 200. Computer 200 might execute the following example:
“As also illustrated in FIG. 5 illustrates how exemplary system 100 could include a modeling component in the form a modeler 1060, which can be used to model nutrient data, portion data, and recommended daily allowances. This modeling capability can improve the operation of exemplary 100.
“Example of Software Architecture”
“FIG. “FIG. The FIG. 6 illustrates a sample software architecture diagram that shows how the FIG. The example embodiment’s household master routine 400 processes the dietary information set 110 to determine individual dietary consumption records. (i.e. by resolving households into individuals and matching up individuals in different surveys within the 110 dietary intake dataset 110 from different times periods to unique individual IDs assigned to system 100 so that an individual is not counted twice, but instead has all of his/her dietary intake details from different surveys considered to be associated with the same person?regardless of whether the individual was a reporter on some or all of the surveys).
“In this example, the dietary input data set 110 is processed using the food descriptor reduce algorithm 500. This coded each food in the intake set 100 that a given person eats based on specific fields within the diet data set. The specific exemplary, illustrative, but not-limiting embodiment defines an 8-dimensional coordinate system that specifies the following food item parameters.
“The food descriptor algorithm 500 generates a combination number (?combo_CD?) This is used as input to the food portion link table 150. In the exemplary embodiment, the food-portion link table 150 links specific food items with the recipe information from recipe file 140 and the food portion sizes information from the portion data set 130.
“The exemplary data integration process 200 moves through the dietary input data set 110 dynamically solving individual identifiers using the household master routine 400. Demographic information (primarily gender in the illustrated embodiment) is extracted from the 110 dietary intake data sets 110 and applied to 130 portion size databases. A portion size data set 130 produces a portion code that, statistically, indicates how much of a particular food item the individual is likely to have eaten (based on their demographic information). In the two examples, this portion size information can be used in two different ways. The portion size information is applied in the food nutrients data set 120 to allow for eventual calculation of the amount consumed of nutrients. (Note that the total amount consumed of nutrient depends on the amount of food eaten and the amount consumed). The portion size information from the portion data set 130 is used as an input to food-portion link 150. In the example embodiment, the food portion link table uses the recipe information from recipe file 140 and the combination code to generate the food code to be output to the output data sets 140.
“In this example embodiment, the recipe codes obtained from recipe files145 are used to determine the specific constituent ingredients within each food item. It also indicates the amount of each ingredient within each food item. Recipe files 145 have been created in the preferred illustrative and exemplary embodiments to accurately reflect every nutritionally important ingredient within every one of many thousand foods people eat. The recipe files 145 can be used to very precisely specify the ingredients of food items and the proportions. Other cases, such as homemade food like pancakes or soup, the recipe files might be less precise and may instead use the more detailed ingredient breakdowns from the dietary intake database 110. In any event, the purpose of recipe files 145 is to as accurately as possible resolve individual food items identified in the dietary intake data set 110 into their respective individual nutritionally-significant constituents and corresponding relative amounts of each.”
“The exemplary embodiment uses this recipe information in conjunction with portion size information (e.g., in a straightforward mathematical multiplication or other scaling) to obtain the amount (e.g., in grams or other convenient quantity units) of each nutritionally-significant constituent within the food item that was consumed. This information, along with the constituent identification and quantity information, is used to index nutrition data set 120. This generates a list of nutrients in each constituent component in the same amount. This nutrient list/quantity is output by the nutrition data set 120 to the output data sets 140.
Data set 140 also receives in the exemplary embodiment the person identifier of the dietary intake data sets 110. This allows for individual tracking of consumption trends. To facilitate demographic-based analysis, the output data set 140 also contains demographic information, such as age, gender and other factors. The output data set 140 may store the food code, which can be used to identify the specific food. It also has the ability to store information about serving sizes, obtained from the portion data set 130.
“FIG. “FIG.7” is an example diagram that illustrates a specific, non-limiting, but exemplary data set linkage. This FIG. This FIG. 7 diagram shows exemplary links between an example household Master Table 160, the dietary intake information set 110, portion size data set 130, and the food portion table 150. The illustrative and exemplary embodiment shows a human dietary researcher adding the portion code to food portion link 150 after a specific recipe has been finalized. When the food descriptor algorithm 500 routines are completed, the portion code (portion_cd), is added to the diary 110. After this, the serving sizes of the person are added to the diary 110.
“As illustrated in FIG. “As shown in FIG. Once assigned, the person ID from the household master table 160 is included in the individual dietary intake records of dietary database 110. This allows system 100 to associate data set 110’s dietary intake information with the person identification assigned by exemplary systems 100. System 100 can personalize the dietary intake data 110 to reflect individual information. It is also able to assign the exact same person ID to multiple participants in multiple dietary intake survey.
The exemplary system 100 focuses primarily on the eight dimensions of food coordinate data as illustrated in this exemplary portion of a dietary intake dataset 110. The exemplary system 100 converts the information in the eight fields of dietary intake data sets 110 into a 32-digit combination codes that are then used to index into the food-portion-link table. This unique code uniquely defines thousands upon thousands of foods and associates them with the recipe file 145 as well as the portion information in the portion data set 130. The food-portion table 150 links the dietary intake data sets 110 and 130. It is possible to calculate the statistically correct portion size using the portion data set 130 once the individual’s demographics and the specific food item have been determined. This information is combined with the recipe information used by System 100 to calculate the nutrient levels from the nutrient set 120.
“Exemplary & Illustrative Process Flow”
“FIGS. 8A?8E shows an example process flowchart showing the steps taken by the preferred embodiment. The process flow shown in the example begins at a “start?” FIG. 8A (block 1100). In the exemplary embodiment, the first step is to read in a record of dietary intake from data set 110. Then add a unique person identifier to an incoming dietary intake household filing 1102 for a specific year (block 1104). The dietary intake data set 110, as mentioned above, is organized by household but also contains individual demographic information for each household member.
“FIG. “FIG. 8A Block 1104. 8A block 1104. This exemplary embodiment creates a unique identifier for each person in order to better track them. The 110 exemplary dietary intake data sets are more concerned at the household level. Different household members could act as reporters at different times within the exemplary diet intake data sets. This can be problematic if one wants to determine individual food consumption. FIG. 100 is an example of the system 100. 8B routine 1104 is used to identify individuals who participated in the dietary intake study in previous years and to assign the same ID (e.g. a sequential or another value) to them. The example embodiment will match individual consumption data from data set 110 that correspond to different survey times periods and record them in output data set 140 with the same unique individual ID to enable individual tracking of intake and nutrition. Routine 1104 adds household master files 160 to accommodate new people, and generates new identifiers for them on an as-needed base.
Referring to FIG. “Referring to FIG. If there is a match (?”yes? Exit to decision block 1108), then it is not necessary to update the master family file 160 with a new ID. If there is not a preexisting identifier in the master household 160 that corresponds to this individual (?no?) Exit to decision block 1108, the preferred embodiment adds a record to the master file 160 and increments the?last name identifier? Counter to generate a unique person identifier. The example embodiment shows routine 1104 updating the dietary intake file household file 1102 for the year with the new/existing individual ID (block 1112) and returns (block 11114).
Referring again to FIG. 8A. Once routine 1100 has determined that the incoming diet intake household file 1102 contains a unique ID for each person associated to a dietary intake record (FIG. ?away from your home? diaries 1116 and 1118 to create an individual-specific common diary (block 1120). This is the exemplary and illustrated embodiment 100. It uses these diaries 1116 for identifying and filtering out those who have not reported correctly eating food items. This information is not included in the dietary intake data 110. The example embodiment 100 describes the meals that have been reported as eaten, but contains some fields that are null (i.e. it fails to indicate what was eaten at each meal). The example embodiment codes non-reported home meals as one value (e.g.?10?). Away from home, non-reported meals will be coded with a different value (e.g.?10?). These records will be filtered later. Non-reported meals can be distinguished from valid skipped meals. The preferred embodiment 100 tracks for certain analysis.
“FIG. 8C is a flowchart showing illustrative and exemplary steps taken by the diary creation program 1120 in FIG. 8A. Referring to FIG. Referring to FIG. 8C, the example embodiment first verifies, for each in-home record with a particular value indicating away home eating that there are at least one or more food records in the associated away home file 1118 (block 1102). If this is not true, the example embodiment 100 codes each in-home record with a specific coding that indicates a non-reported dinner (block 1122). Block 1124 merges the away from home and in-home files 1116, 1118 to create a diary for each time period. This example uses the diary file 1126 as a tool to identify each person (here identified with household ID, member number) and determine the day and meal. If there is no valid helping code for a food record, each food item in the record that corresponds to that person, day, and meal will be coded to indicate an unreported in-home meal (block 1128). This data processing eliminates non-reported meal records to improve the accuracy of the resulting data set 140.
“Referring again to FIG. 8A. Once the food intake diary 1126 is created, the exemplary embodiment matches household records with diary records to obtain the personal identifier and gender from the updated household tableau 1102 (block 1102). Next, the process 1100 exemplar performs the food descriptor removal algorithm 500. This is the process of generating a code value for each record in the diary file 1126 that either identifies an item from the food portion table 150 or a new item (i.e. an item not yet identified within the food-portion table 150).
“In the example embodiment, reduction routine 500 allows preferred embodiment to tie together data sets 110 and 130 regarding dietary intake. The reduction algorithm 500, in the exemplary embodiment, is not able to generate all possible combinations (eight factororial), but it can find an existing food code and create a place for a food code using only few thousand records (e.g. on the order 2000). The reduction algorithm 500 significantly reduces the number of food items that the dietary intake information set 110 can theoretically identify into a smaller number of food items most people eat.
“FIG. 8D illustrates steps that the food descriptor algorithm 500 of the illustrated embodiment performs. This particular example uses eight fields from the dietary intake database 110 (i.e. Type, Form and Characteristic, Flavors, Classification, Preparation Type, Packaging Type, Special label Code). The exemplary embodiment uniquely identifies over 5000 food items (block 112). Each of the fields in the example embodiment contains a numeric number between 1 and 3 characters. In the example embodiment, many of these codes are combined and grouped according to dietary intake factors that relate to the nutritional makeup of the food. These groups are stored in the food-portion link lookup table 150.
“A COMBO KEY file 1134, which correlates combination key information and column name/column value information, is kept. The COMBO KEY in the example embodiment may be a four-digit sequential number that identifies one eighth of a 32-byte character cod. The column name field can be used to identify the column in the dietary intake data 110. One or more values may be provided in the column value field that can apply to this specific column and its associated food category. You may also want to keep a category code that specifies a certain number of unique codes (e.g. 53) that identify a specific type of food group (e.g. 10=cereals and 3=milk, 53=baby foods, etc ).
“As an illustration, consider the following values:
“Type: 0\nForm: 12\nCharacteristic: 0\nFlavor: 249\nClassification: 2\nPrep Method: 0\nPackage Type: 3\nSpecial Label: 34”
“Assume that the following rows are found in the COMBO KEY 1134 file:”
“COMBO? “COMBO?” 34nLABEL CODE CD
“In the above example, the reduction algorithm 500 would take every dimension and scale cmbntn_val columns for the values given for each column name and category code. The COMBO KEY will be converted into a four-character string, and then inserted in the appropriate section of a 32 character code.
“000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000” pkg specialnmthd. type label”
“Not all dimensions are necessary to identify a food. The algorithm can be flexible enough to assign column names 0 (null), which are not found in the scan of the preferred exemplary embodiment.
“Thus, in the above example combo key file, 134 fragment, one scans for cmbntn_val=249 and col_nm=flvr_Cd. The algorithm will then find the corresponding unique combo_key equal or greater than 299. In the exemplary embodiment, that value is converted to?0299 by the algorithm. The algorithm in the exemplary embodiment converts that value to?0299?. Next, SPC_LBL_CD=34 for category=10 is found. The unique combination key for that record’s record is then converted to string?3243? and then inserted into the eighth of the octet. These are the only dimensions required for this particular food item. There are no other column names (not illustrated) for this category code. The 32-character code corresponds to this:
FIGS. FIGS. 9A?9L illustrate an example of the progression of data structures as described above. FIG. FIG. 9A illustrates an example excerpt from a food diary 1126. Not all fields are shown due to space limitations. The exemplary system 100 will begin to process the diary 1126 information. A first record will be pointed to by the FIG. 9B). The first record is read in, and then the steps 1132 and 1136 of FIG. 8D is used to create a food code that can be rewritten to the journal 1126 (see FIG. 9C). FIG. FIG. 9D shows an excerpt of the exemplary diary 1126, taken after the FIG. 8D routine 500 is executed for all records in the extract.
“FIG. FIG. 9E illustrates an example combo key file 1134, which is used by FIG. 8D block 1132 is used to convert the eight dimensions of a food item in the dietary intake data 110 into a 32-character unique number that maps to the food portion table 150. FIG. 9E is an example of the records returned by the combo key file 1132. 9E shows an excerpt from the combo key file 1134 records for an exemplary diary entry. 9F containing a category code for a particular value (in our case,?27?). FIG. FIG. 9F highlights the relevant data fields. The key columns in this example are: TY_CD, FRM_CD(Form), FLVR_CD(Flavor), FLVR_CD [Flavor], PKG_TY_CD (?Packaging Type?), SPC_LBL_CD (?Special Label?). Other highlighted columns are not relevant for this particular category code.
“FIG. 9G illustrates an example in which the Type (TY_CD=1) was included in a CMBNTN_VAL within a specific record number 1017. This is an example where the type column name can be parsed to create a combination key that is based on the?Type? information. FIG. FIG. 9H illustrates a similar example, where another combination key (??1020?) is used. 9H shows a similar example where another combination key (?1020? in this instance) is created using the?Form. information in the diary 1126. FIGS. FIGS. 9I, 9J, and 9K provide additional examples of combination codes that can be used for flavor, package type, and special label coding. These are just examples. A different diary entry can generate different combination codes depending on the specific values of the eight food descriptor dimensions.
“In this case, the 32-digit food code obtained by adding the different combination codes in a specific order is defined as:
FIG. 9L highlights the matching records that make the combination key. 9L. FIG. FIG. 9M shows the corresponding diary record 1126, which is returned with the 32-digit combination codes obtained from the 500 food descriptor reduction process.
“Once the data-reduction algorithm has determined the 32-character key using the record number of found items in the combo key (block 1136), then the reduction algorithm searches for the 32 character code in the food portion line file 150 (block 1138). The data reduction algorithm 500 scans the file 150 containing food-portion links. This file contains information about foods that have been previously defined (i.e. by previous operations of System 100). It combines the portion size data set 120 and the nutrition data sets 130 to generate this 32-character code. If found (?yes? If found (?yes?), exit to decision block 1140. The food diary record 1126 will be updated with the current food key, portion code, and recipe code (block 1142), which were obtained from the food-portion table 150 (see FIG. 7). 7). Storage space consideration. Routine 1100 increments a counter that tracks the number of times this food item has been used in the diary 1126. The diary 1126 record also includes the food_cd value as well as the portion code from the portion data set 120. This data can be later merged into the final output data sets 140.
“If a 32-character code for food is not found in food-portion link file 150 (???), then the new code is added along with all eight dimensions (i.e., type, form and characteristics of the food item) to identify it. Exit to decision block 1140. The new 32-character food code is then added together with all eight dimensions (i.e. type, form and characteristics, flavor classification, packaging method, special label) to identify the food item. This creates a?new foods’. Table 1144 (block 1146). FIG. 8D continues until all entries in diary 1176 are processed (block 1148).
“In general, if the 32 character code is not found, an exception record will be created and one of two things will occur. If the 32-character code cannot be found because of a new or previously undefined value in the dietary intake database 110 (and assuming that the new value has validity), then the new value will go into the appropriate group according to a (human) dietetic intake scientist. If the new item is not for a food previously tracked, then a human dietary intake scientist will create a new recipe in file 140 based on nutrition data set 120, the values of the eight dimensions fields within new foods data file 144, and insert the new item into food-portion link 150. This allows later encounters with the same food item to get the appropriate nutrition information (FIG. 8A, blocks 1154 and 1152. FIG. 8A shows the exception handling that is possible after the data reduction algorithm 500 has been completed. It is based on the contents of the new foods files 1144. If there are any new foods in the new foods files 1144 after the data reduction algorithm 500 (?yes?) Exit to decision block 1150. Next, for each new food, a dietician creates or updates a recipe from the dietary intake database 110. Assigns a portion code and then adds the new (or updated record) to the food-portion link table 150 (block 1552). The data reduction algorithm 500 should be run again until there are no more exceptions. This is the end (?no). At this point (?no?), exit to decision block 1100. The preferred exemplary routine 1100 exports 2 deliverables (the diary files 1126 and 160, respectively) in the form an output data set 140. This is for final processing (block 1156).
“FIG. 8E is an illustration of a flowchart that illustrates the final merge procedure 1156 as shown in FIG. 8A. 8A. FIG. 8E: To create a temporary diary file 1126, we remove all records that indicate non-reported meals (block 1160). Is this temporary diary file 1126? This temporary diary file 1126 is combined with the nutrient information set 120 (or at most portions thereof) to create a food base data set 140. The specific embodiment is described for illustration purposes only. It initializes cereal flags, lunch, dinner, and breakfast flags. The routine 1156 calculates the correct amount for each nutrient based upon serving size (block 1164).
“GMI FOOD FOODnMEAL DAY HLP GRP MJRnperson ID NBR G FOOD CD YEAR FOOD NM CD FOOD NM, Plain Oats, NSF7800n108538 2 119 111106 Regular, Caffeinated NSF118? 1100\n108538 1 49805 2 1 19 31867 Skim milk ? ?24C 2200\n108538 1 49805 2 1 19 41375 Tomato Juice 62 1800\n108538 1 49805 2 1 19 101136 Raisin Bran ? ?77C 1900n108538? 49805? 2 19 101136 Raisin Bran 62 1800n108538? 49805? 2 19 51375 Tomato Juice 62 1800n108538? 49805? 2 19 101136 Raisin Bran 74 1900nNR Untoastedn108538? 49805? 2 19 5012361 Bagels Regular White or 74 1900nNR Regular Sugar Hyd Banana? ? ?63A 1800nor.NR, Fresh, Reg Sugarn108538 3 49805 2 19 441882 Peanut butter, Reduced Fat ?60B 1700\n108538 2 49805 2 1 19 5213731 Cracker, Regular Butter, 90 2300\nRegular\n108538 3 49805 2 1 19 253275 Vegetable, Plain, Black 70 2600\nBeans, Cooked\n108538 3 49805 2 1 19 253931 Combination Vegetable 69 2600\nDish, Plain Mixed Veg,\nCkd\n108538 3 49805 2 1 19 253310 Vegetable. Plain, Beets, 68 2600nRawn108538 2 49805 2 19 412062 Ketchup Regular 125? 2300\n108538 3 49805 2 1 19 331864 Apple Pie 102? 1500n108538 2 49805 3 1 19 331864 Apple Pie 102? ?75A 1900\nToasted\n108538 3 49805 2 1 19 284168 Ice Cream, 27 2200\nDiet/Lowcal/Sorbet, Van or\nNR, Reg Fa\n108538 4 49805 2 1 19 387384 popcorn light 84 2300”
“GminMeal Portion Cereal Sz Ind Grn Ind Nutrient Cd Srvng Sz Ind Grn Ind Grn Nutrient Cd Consumed Breakfast Lunch Dinner Snack”
“GminMeal TotnPerson Smry Cereal Grm Yogurt Rte Coffee Tea Presweet Energie Tot Carb Protein 137.43 2.30 00 5.82 n108538 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 138.86 9.35 10.13 7.09 8.87 1.59 n108538 0 0 0 0 0 0 0 0 0 0 0 0, 0 0 0 0 0 0 0 0 0 0 0 0 0 0 05 0 0 0 0 0 0 0 0 0 0 0 0 0 0 e 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1088538 0 0 0 0 0 0 0 0 0 0 0 0 0 001 0 0 0 0 0 0 0 0 0 0 0 0 0 0 sn108538 0 0 0 0 0 0 0 0 0 0 0 0? 0 0 0 0 0 0 0 0 0 0 0 0 0 0???
“Gmi Tot Tot\nMeal Anml Veg Tot Sat Musat Pusat\nPerson Smry Protein Protein Alcohol Cholest Fat Acid Fat Acid Fat Acid Fructose Galactose Glucose Lactose\n108538 1 0.00 5.82 0.00 0.00 0.40 0.73 0.84 0.05 0.00 0.00 0.00\n108538 1 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00\n108538 1 3.27 0.00 0.00 1.73 0.12 0.05 0.01 0.00 0.00 0.00 4.24\n108538 1 0.00 0.00 0.00 0.00 0.02 0.02 0.04 3.99 0.00 2.94 0.00\n108538 1 0.00 5.05 0.00 0.00 0.13 0.22 0.71 2.96 0.00 3.45 0.00\n108538 1 0.00 5.89 0.00 0.00 0.12 0.07 0.39 0.11 0.00 0.22 0.00\n108538 2 0.00 0.88 0.00 0.00 0.16 0.04 0.08 2.38 0.00 3.70 0.00\n108538 2 0.00 7.01 0.00 0.00 1.47 4.28 3.03 0.00 0.00 0.06 0.00\n108538 2 0.00 1.00 0.00 0.00 0.99 2.30 0.53 0.03 0.00 0.06 0.00\n108538 3 0.00 9.09 0.00 0.00 0.16 0.05 0.26 0.84 0.00 0.63 0.00\n108538 3 0.00 2.79 0.00 0.00 0.02 0.01 0.11 0.56 0.00 0.62 0.00\n108538 3 0.00 0.82 0.00 0.00 0.02 0.02 0.03 0.05 0.00 0.10 0.00\n108538 3 0.00 0.19 0.00 0.00 0.01 0.01 0.02 0.46 0.00 0.96 0.00\n108538 3 0.00 3.78 0.00 0.00 5.03 8.79 5.29 6.84 0.00 2.41 0.00\n108538 3 1.52 2.67 0.00 31.25 3.36 2.88 0.81 0.16 0.00 0.37 1.08\n108538 3 0.00 0.40 0.00 0.00 0.00 0.01 0.05 0.56 0.00 0.62 0.00\n108538 4 0.00 8.69 0.00 0.00 3.17 8.14 2.76 0.07 0.00 0.07 0.00”
“Gmi Tot Sol\nMeal Diet Diet Insol Bc\nVperson Smry Maltose Sucrose Starch Fib Fib Diet Fib Pectins Tot Vitm A Equiv Retinol Vitm D\n108538 1 0.00 0.52 21.02 4.08 1.90 2.21 0.00 853.87 0.00 601.61 0.00\n108538 1 0.00 0.00 0.10 0.38 0.21 0.21 0.00 0.00 0.00 0.00 0.00\n108538 1 0.00 0.00 0.00 0.00 0.00 0.00 0.00 204.00 1.93 57.75 0.98\n108538 1 0.00 0.00 0.00 0.84 0.46 0.38 0.46 556.00 698.52 0.00 0.00\n108538 1 0.90 8.54 18.52 7.35 0.92 6.43 0.11 1364.00 0.40 224.43 1.24\n108538 1 0.00 0.15 25.73 1.29 0.50 0.79 0.00 0.00 0.00 0.00 0.00\n108538 2 0.00 5.73 3.52 2.11 0.53 1.59 0.53 81.00 43.18 0.00 0.00\n108538 2 0.19 2.34 6.40 0.68 0.20 0.48 0.22 0.00 0.00 0.00 0.00\n108538 2 0.00 0.75 6.75 0.26 0.15 0.11 0.00 0.00 0.00 0.00 0.00\n108538 3 0.00 1.57 9.72 6.69 2.51 4.19 0.00 2.00 1.26 0.00 0.00\n108538 3 0.00 0.65 0.82 3.61 1.41 2.20 0.91 3209.73 2389.84 0.00 0.00\n108538 3 0.00 4.18 0.34 0.98 0.59 0.39 0.22 33.60 10.30 0.00 0.00\n108538 3 0.35 1.42 0.00 0.17 0.04 0.13 0.00 1016.00 77.87 0.00 0.00\n108538 3 0.00 19.29 23.91 2.50 1.00 1.37 0.37 20.47 21.11 0.00 0.00\n108538 3 0.08 8.78 19.90 1.36 0.25 1.11 0.00 211.03 42.02 18.97 0.31\n108538 3 0.00 21.18 0.00 2.51 0.28 2.23 0.28 34.20 20.52 0.00 0.00\n108538 4 0.00 0.29 38.82 10.94 0.37 10.57 0.00 163.17 85.01 0.00 0.00”
“GminMeal Tot Alpha Beta Gamma Delta 1 0.00 0.00 0.04 0.13 0.09 1.41n108538 0.00 0.02 0.01 0.18n108538 0.00 0.09 1.69 0.00 0.04 0.08 0.37 0.41 1.47n108538 0.00 0.02 0.02 1.16n108538 0.00 0.59 1.62 1.47 0.00 2.64 0.00 0.00 1.07 0.00 9.67 19.92 1.51 0.28 2.25n108538 1.35 1.39 0.23 1.57n108538 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 5.00 0.35 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 n108538 n108538 1.25 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 1.25 66 0.00 0.00 0.00 0.00 0.00 0.00 5.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00???
“GminMeal Vitm Vitm & Person Smry Acid B6 Folate, B12 Calcium Phosphorus Magnesium Selenium and Iron”
“GminMeal Caproic Capric Lauric Palmitic Stearic Sodium Potassium Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid 1 380.08 159.39 0.02 0.00 0.00 0.01 0.00 0.02 0.00 0.00 n108538 131.65 0.00 0.00 0.00 0.00 0.00 0.00 0.11 0.00 0.53 n108538 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 5.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00??? 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 9.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00?? 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00?n108538???
“Gmi Meal Arachidic Behenic Myristoleic Palmitoleic Oleic Gadoleic Erucic Linoleic Linolenic Parinaric Arachidonic\nPerson Smry Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid\n108538 1 0.00 0.00 0.00 0.00 0.73 0.00 0.00 0.80 0.05 0.00 0.00\n108538 1 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00\n108538 1 0.00 0.00 0.00 0.01 0.04 0.00 0.00 0.00 0.00 0.00 0.00\n108538 1 0.00 0.00 0.00 0.00 0.02 0.00 0.00 0.04 0.00 0.00 0.00\n108538 1 0.00 0.00 0.00 0.00 0.09 0.00 0.00 0.32 0.03 0.00 0.00\n108538 1 0.00 0.00 0.00 0.01 0.07 0.00 0.00 0.37 0.02 0.00 0.00\n108538 2 0.00 0.00 0.00 0.01 0.03 0.00 0.00 0.05 0.03 0.00 0.00\n108538 2 0.05 0.10 0.00 0.00 4.16 0.11 0.00 2.97 0.06 0.00 0.00\n108538 2 0.00 0.00 0.00 0.00 2.30 0.00 0.00 0.53 0.00 0.00 0.00\n108538 3 0.00 0.00 0.00 0.00 0.05 0.00 0.00 0.14 0.12 0.00 0.00\n108538 3 0.00 0.00 0.00 0.00 0.01 0.00 0.00 0.02 0.07 0.00 0.00\n108538 3 0.00 0.00 0.00 0.00 0.02 0.00 0.00 0.03 0.00 0.00 0.00\n108538 3 0.00 0.00 0.00 0.00 0.01 0.00 0.00 0.02 0.00 0.00 0.00\n108538 3 0.00 0.00 0.00 0.00 8.79 0.00 0.00 4.96 0.32 0.00 0.00\n108538 3 0.00 0.00 0.00 0.30 2.57 0.00 0.00 0.75 0.04 0.00 0.01\n108538 3 0.00 0.00 0.00 0.00 0.01 0.00 0.00 0.04 0.02 0.00 0.00\n108538 4 0.04 0.00 0.00 0.06 8.06 0.02 0.00 2.69 0.08 0.00 0.00”
“Docosp Docosh\nGmi Meal Eicos Enoic Enoic\nPerson Smry Acid Acid Acid Tryptophan Threonine Isoleucine Leucine Lysine Methionine Cystine Phenylalanine\n108538 1 0.00 0.00 0.00 0.07 0.19 0.23 0.45 0.23 0.12 0.14 0.31\n108538 1 0.00 0.00 0.00 0.00 0.00 0.00 0.03 0.00 0.00 0.00 0.00\n108538 1 0.00 0.00 0.00 0.05 0.14 0.20 0.32 0.26 0.09 0.03 0.15\n108538 1 0.00 0.00 0.00 0.00 0.04 0.04 0.04 0.04 0.00 0.00 0.04\n108538 1 0.00 0.00 0.00 0.08 0.15 0.18 0.32 0.15 0.07 0.11 0.22\n108538 1 0.00 0.00 0.00 0.07 0.17 0.22 0.42 0.14 0.11 0.12 0.29\n108538 2 0.00 0.00 0.00 0.01 0.03 0.03 0.06 0.04 0.01 0.02 0.04\n108538 2 0.00 0.00 0.00 0.09 0.25 0.31 0.51 0.37 0.09 0.09 0.37\n108538 2 0.00 0.00 0.00 0.01 0.03 0.03 0.07 0.02 0.02 0.02 0.05\n108538 3 0.00 0.00 0.00 0.10 0.39 0.40 0.73 0.63 0.14 0.10 0.49\n108538 3 0.00 0.00 0.00 0.02 0.10 0.10 0.12 0.14 0.04 0.02 0.09\n108538 3 0.00 0.00 0.00 0.01 0.03 0.03 0.04 0.03 0.01 0.01 0.03\n108538 3 0.00 0.00 0.00 0.00 0.01 0.00 0.01 0.01 0.00 0.00 0.00\n108538 3 0.00 0.00 0.00 0.05 0.11 0.14 0.25 0.09 0.05 0.07 0.18\n108538 3 0.00 0.00 0.00 0.05 0.15 0.18 0.39 0.18 0.09 0.08 0.21\n108538 3 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00\n108538 4 0.00 0.00 0.00 0.06 0.32 0.31 1.06 0.24 0.18 0.16 0.43”
Summary for “Systems and methods to determine nutrients in dietary intake”
“Most people shop carefully for the food they feed their families. So that our families have a healthy diet, we look for foods with high nutritional value. We pay particular attention to the food we purchase for our older and younger family members. It is vital to eat the right foods every day for our health and longevity.
“America’s medium-sized supermarket is a marvel of the modern age. Supermarkets have a greater variety of food options than ever before. You can find Texas beef, seafood from Washington State, oranges and New Zealand kiwis, as well as other international foods. There are also a wide variety of packaged and prepared foods. It becomes harder to make sensible and healthy choices for our families and ourselves as the number of products available increases. We often wish we had more direction.
Advertising is a great source of information about food products that we may like to purchase and give to our families. Food product manufacturers want to accurately advertise their products, and they are keen to develop new products that appeal to customers and meet their nutritional requirements. There is strong evidence to support the claim that eating whole-grain foods, including other plant foods, can reduce your risk of developing heart disease. This information is vital for consumers as it allows them to choose healthier foods and make better choices.
Food marketing researchers can use data to determine how to improve a person’s diet by recording detailed information about their food habits. For marketing research purposes, food consumption data is a continuous record of the nation’s food intake. These detailed food records have been used by the food industry to monitor growth and track specific food items.
The NPD survey panel records food eaten at-home and abroad over a period of 14 days. The NET database tracks the consumption patterns and trends for more than 4,000 food and beverage products. It also provides a range of information such as:
The NPD/NET data set can be used to determine what American households eat. Because food is often bought by one household member for the whole household, and because meals are usually eaten together in a given household, it makes sense to place emphasis on household. A practical reason is that only one household member (e.g. the person responsible for food preparation) usually records the survey data required for the entire household. The problem with this emphasis on household data recording, however, is that while the diaries record how many of a particular food item was served to the household, they do not permit any household member to record how many of it he or she eats.
“In detail, each household must fill out a survey/diary asking for information about how much food was served to them, how much was eaten, and who ate that particular food. For example, NPD Group’s Sample Daily Meal Diary is included herein as a reference. This is a common procedure used in panel surveys to reduce the amount of information and increase reliability. It is not necessary to record individual portions. This is because if all participants are required to track how much food they eat over a 14-day period, it can be burdensome and may compromise the accuracy of the recording.
“The NPD/NET dietary intake dataset for certain purposes may have a problem with the quantity of nutritional information it provides. NPD doesn’t attempt to provide nutritional information for every food that was included in its survey. This type of detailed nutritional analysis is usually the work of food researchers and is not provided in the NPD/NET dataset. However, some food research applications may require detailed information about the daily intake of more than 100 nutrients (including individual amino and fatty acid) each day. A company might want to know the daily intake of each nutrient by different demographic groups in order to formulate or reformulate a food product that will ensure Americans get the right essential nutrients. Manufacturers and providers of food products may want to see evidence in order to make claims about their products being part of your daily diet. Health specialists may wish to analyze nutrient consumption or nutrient consumption trends in the population overall, by demographically-specific segments of the overall population, or by household and/or individual, in order to try and discern correlations between nutrient consumption and disease. Other applications require detailed information about the daily intake of nutrients and tracking it over time. The NPD/NET data sets do not address these issues.
There are many good sources of nutrition analysis data for the foods we eat. Many research organizations, including the University of Minnesota have collected the nutritional content of many foods. The University of Minnesota’s Nutrition Database System for Research (NDSR) software provides detailed nutritional information for over 18,000 foods, as well as over 8,000 brand-name items. Although there is a lot of information available about America’s eating habits and nutritional information, it is spread across a variety of data sets that have been created by various entities, including governmental, corporate, and academic. The data sets can be incompatible and each set is designed to accomplish different goals.
“According to an example and an exemplary embodiment, data integration is performed on three separate, special-purpose food research related data sets. One data set includes food consumption data that is based on 14-day diaries. The second data set includes data on portion sizes for large numbers of food types (e.g., more than 8,000). A third set includes nutrient data for many (e.g., more than 18,000) unique food constituents. This integrated data set can be used to analyze the results using standard statistical analysis procedures.
“According to a further aspect of an illustrative or exemplary embodiment, a data set is analysed and processed to determine the mean age and sex specific serving sizes of a number of food items. These portions are then matched to the food in a second set that represents 14-day food intake. A third, nutrient-based data set is used to assign complete nutrient profiles to each food. These data sets are combined into a database. Then, nutrient intake reports can be processed using a standard statistical reporting interface. This flexible system allows users the ability to classify the population based upon their usual consumption of specific food types, brands, and/or brand of food, as well as to determine differences in dietary habits versus non-using. counterparts.”
“In one non-limiting, illustrative embodiment information is combined from three data sources:
This research on dietary intake is of great value. “For example:
“The dietary research results can help to build credibility within scientific and food policy communities.
The illustrative preferred embodiment allows the data sets to explore for new information, trends, and themes that can be transformative and stimulate product development, help to create marketing programs and suggest strategies (e.g. BMI and cereal intake, diet modeling to meet three whole grain per day, etc.). These techniques can also be used in food product marketing and public relation (e.g. sugar defense, whole grains intake, impact of breakfast cereals on diet, calcium intake and breakfast patterns, portion sizes, eating patterns for children, seniors, and other demographic groups etc
“The techniques described herein can also be used for product development and product reformulation (e.g. by identifying the nutrient requirements in a population, such as calcium fortification, folate enrichment and enrichment etc
“The information in the illustrative and exemplary embodiments may be helpful in a regulatory environment to assist with claims documentation, policy development, and data for regulatory comments and fortification review.
“The information could also be used to prepare abstracts and manuscripts for scientific journals, supplement internal and external laboratory research projects, or for any other scientific purpose.”
“Additionally, information from an illustrative or exemplary embodiment could be useful for data for speeches, presentations, public relations facts and advertising copy as well as consumer information.”
“According to a more specific aspect of an illustrative preferred embodiment, we use a food description reduction algorithm that reduces large amounts of food item data from a 14-day diet database into smaller amounts of useful data for identifying nutrients in foods actually eaten by participants in dietary intake studies. A particular embodiment uses a subset of data fields to identify food items on the order over 5,000 from more than a billion possible. For example, this data field subset could include 8-dimensional coordinates that represent food item identification (e.g. type, form and characteristic, flavor, class, preparation method, packaging type and special label code). In the preferred embodiment, many codes are combined and then grouped according to nutritional factors. These codes, when combined, are mapped to nutrient values that are based on food nutrient profiles and portion sizes.
“In the example and the illustrative embodiment, groupings are performed on the basis of a lookup table that uses four keys:
“In the illustrated and illustrative embodiment, food descriptor mapping proceeds by scanning a table that contains data to determine if the food item within the dietary intake data is defined and has a combination key. Multiple iterative scans may yield additional combination keys, which can be combined to create a combination code for the specific food item in the food intake data. The combination code food descriptor generated is stored in a food-portion data file. This data file contains foods that have been previously described by combining a portion data set with a nutrition information data set. The combination code food descriptor is located within the food portion file. In this example embodiment, it is mapped into a simpler unique food designationator (for storage space considerations). If the code is not located (indicating, for instance, that a new food item has been reported in the dietary intake dataset), an exception is created so that a dietary intake specialist can dynamically update the lookup tables to include this new item. This can be done iteratively to allow for interactively defining new foods items as they are introduced and appear in the dietary intake data.
“In accordance to another aspect of the illustrative embodiment, a household Master Analysis is performed to track individuals consumers?even through multiple intake survey from different times periods. For many other research purposes, it is necessary to have individual data about food intake and nutrition. This invention’s preferred embodiment is exemplary in that it can track individual person’s dietary intake using dietary intake data sets. These data sets are usually designed at the household level, but include enough data to allow individual tracking, if handled properly.
“For example, it is possible to get more precise dietary intake results by using data sets containing dietary intake data from different surveys. These surveys often survey the eating habits of the same households as the same people within the households. But, household compositions can change (e.g. when college students move on to the next stage of their lives), and different households may have different reporters/diarists during different periods. In the exemplary embodiment, it is possible to determine whether a household or individual is included in multiple food intake survey. The preferred embodiment assigns each individual a unique individualID. This individual ID is different from the one used to code them within the food intake data sets. To allow individual tracking of dietary intake, unique individual IDs can be assigned to individuals who have reported on different dietary intake surveys. The exemplary embodiment can analyze the food intake survey results using individual data. This allows for more precise results as the eating habits and dietary intakes of individuals can be weighted to make them one person. The ability to track individual consumption over a longer time span, such as years, can provide significant flexibility and advantages. You can find valuable information by tracking how an individual’s eating habits change over time, such as the age of a person.
“According to a further aspect, an exemplary and ilustrative embodiment, data are combined to create demographic-based (e.g. age and sex), categories for portion sizes determinations.
“In accordance to yet another aspect of an exemplary embodiment, recipe file are used for extracting nutrient information form food descriptors. It is possible to identify a food item in the food intake data and determine the nutrients that are present in that particular food. The example embodiment’s food intake survey data does not include information about an individual’s actual portion size. Instead, the portion size information is obtained using a different set of data based on the age, gender, and other demographics of those who ate the food item. The preferred embodiment uses recipes to estimate the nutrition that the consumer received from the food product once the food item and the portion size have been established.
The nutrient information within the nutrient set may not provide a complete nutrient profile for every one of the thousands of food products. The nutrient set may not include information about the nutrition obtained from eating a combination, but the nutrient set might contain complete information about the ingredients of such foods (e.g. flour, butter, milk, oil, and other components of a pancake batter recipe). This aspect of the preferred and illustrated illustrative embodiment is preserved so that recipe files can be used to separate particular food items into their constituent parts. To dynamically determine the overall nutritional content of the food item, the nutrient data can be used to identify the nutrients in each component.
“FIG. “FIG. System 100 uses a unique methodology that utilizes 14-day food diaries and additional data sets to provide nutrient portion information.
“As shown at FIG. 1 is the 14-day food intake data collection 110. This input is used in the overall dietary intake analysis system 100 in the preferred exemplary embodiment. The 14-day food intake data 110 could be, for instance, the NPD/NET dataset (see FIG. 1A) is a result of extensive data collection from many households, who record exactly what they eat over a 14-day period.
The preferred embodiment system 100, shown in FIG. 1. includes a data integration mechanism 200, which links information from the food consumption data set 110 with information in two other data sets, namely, a food nutrition data set 120 and a portion data set 130 in the exemplary embodiment.
“Food nutrient set 120″ is a collection of detailed nutrition data, which are generated by food researchers. It includes information about thousands (e.g. 120 nutrients) of uniquely identified foods that have been assigned to specific 200 (e.g. 179) food groups. The University of Minnesota’s Nutrition Data System for Research is an example of a food nutrient data collection 120. The preferred embodiment 100 uses the University of Minnesota Nutrition Coordinating Center’s NDS-R software to assign complete nutritional profiles to all foods in the 14-day food intake dataset 110. FIG. 1A.”
“As mentioned above, the food intake data 110 used in the preferred embodiment 110 does not include information about the individual’s portion sizes. Instead, portions are kept on a household basis and not for individual consumption. It is important to know the amount of each nutrient that has been consumed by each person. Not just each household. Many households often consist of several individuals from different demographic groups (e.g., older Americans and younger Americans; children of different genders; etc.). Marketing and health research can benefit greatly from individual-based data.
“To address the issue with the exemplary data set 110 on food intake, we link an additional data set 130 on food portion sizes obtained from a different source, e.g. the USDA’s Continuing Survey of Individual Food Intakes (CSFII). FIG. 1A. 1A. The 24-hour recall method of collecting data on diet does not always reflect the ‘usual’ pattern. Food intake. The USDA data set is more comprehensive than the 140-day food intake data set 110 in some respects. The USDA survey requires that individuals record how many calories they ate during 24-hour periods. The USDA data set 130 provides detailed demographic information, such as household size, income, age, and sex. The preferred exemplary and illustrative embodiment 100 uses this USDA data set 130 to obtain demographically-based food portion size data. It is possible to calculate statistically from the USDA data set 130 that, for example, a 34-year-old male eating a steak dinner with french fries and steak will likely eat approximately xxxg of steak, and yyyg of french fries. A 12-year-old female eating the same meal would eat zzzg of steak, and aaag grams of French fries. System 100 uses this demographically-based portion size information to estimate individual portion size of food intakes recorded in the dietary intake data set 110. System 100 could also contain the Pyramid Servings data from USDA CSFII using the CSFII Food codes.
“Briefly the food coding structure used in data integration operation 200 is based upon 53 basic food categories as shown in the exemplary embodiment. Each food category has a unique and progressive level of detail that describes the food in that category. For example, food coding could include the following individual codes:
“Example System Implementation.”
“FIG. FIG. 4 illustrates an example block diagram for an illustrative and non-limiting implementation system 100. 1. FIG. Referring to FIG. FIG. FIG. 5 shows how the mass storage devices 152 (which contain the various data sets 110-120, 130) can be input in a variety ways. For example, the portable mass storage device (154), other optical magnetic (155), or another mass storage device (152) may be supplied. The data sets may also be transmitted over a computer network (158) such as the Internet.
“In the example embodiment 100 mass storage device 152 is shown in FIG. 4, which may contain one or more magnetic disk drives, can also store additional data structures such as a recipe file, a food portion link table 150, a household mastertable 160, and other data structures (e.g. 135). These data structures (145, 150 and 160 in the exemplary embodiment) are created and/or updated automatically by system 100. They are used in data integration process 200.
“As also illustrated in FIG. “As also shown in FIG.
“These diverse processes, 400, 500 and 600, are controlled interactively by one or more users via display workstations 162 to generate a statistical data collection called ‘food base? 140 is the combined data from the three input data sets 110 to 120 and 130. This statistical data set 140 can be stored on the same mass storage device 152 or a different one by the computation block 156. This statistical data set 140 can be transferred to another location using optical disk 154 or 155, network 150, or any other means for further manipulation and analysis.
“In the example embodiment, the exact same or different computation block is 156. A personal computer, for example, performs statistical, trend, and other analysis 700 on a data set 140 in response user commands. To generate reports 310, the data set 140 is input via one or more workstations 162. These reports 310 can be printed, displayed on a computer 162, or transmitted over the network 158 (e.g. via email, web pages, etc.).
“As shown in FIG. 4. The exemplary system 100 also includes an additional Health Focus Data set 135 that is directly linked to households in the food intake data sets 110 (and so to household master tables 160). NPD Group may supply the Health Focus data set 135. The Health Focus data set 135 allows system 100 users to examine attitudes towards health and how they relate to dietary intake at a household-level. In the exemplary embodiment, data set 110 on food intake preferably includes results from multiple surveys (e.g., a survey that is about ten years old and a more recent survey), in order to provide long-term trends information.
Referring to FIG. 5. In one specific example, computer 156 executes several processing routines to implement data integration process 200. Computer 200 might execute the following example:
“As also illustrated in FIG. 5 illustrates how exemplary system 100 could include a modeling component in the form a modeler 1060, which can be used to model nutrient data, portion data, and recommended daily allowances. This modeling capability can improve the operation of exemplary 100.
“Example of Software Architecture”
“FIG. “FIG. The FIG. 6 illustrates a sample software architecture diagram that shows how the FIG. The example embodiment’s household master routine 400 processes the dietary information set 110 to determine individual dietary consumption records. (i.e. by resolving households into individuals and matching up individuals in different surveys within the 110 dietary intake dataset 110 from different times periods to unique individual IDs assigned to system 100 so that an individual is not counted twice, but instead has all of his/her dietary intake details from different surveys considered to be associated with the same person?regardless of whether the individual was a reporter on some or all of the surveys).
“In this example, the dietary input data set 110 is processed using the food descriptor reduce algorithm 500. This coded each food in the intake set 100 that a given person eats based on specific fields within the diet data set. The specific exemplary, illustrative, but not-limiting embodiment defines an 8-dimensional coordinate system that specifies the following food item parameters.
“The food descriptor algorithm 500 generates a combination number (?combo_CD?) This is used as input to the food portion link table 150. In the exemplary embodiment, the food-portion link table 150 links specific food items with the recipe information from recipe file 140 and the food portion sizes information from the portion data set 130.
“The exemplary data integration process 200 moves through the dietary input data set 110 dynamically solving individual identifiers using the household master routine 400. Demographic information (primarily gender in the illustrated embodiment) is extracted from the 110 dietary intake data sets 110 and applied to 130 portion size databases. A portion size data set 130 produces a portion code that, statistically, indicates how much of a particular food item the individual is likely to have eaten (based on their demographic information). In the two examples, this portion size information can be used in two different ways. The portion size information is applied in the food nutrients data set 120 to allow for eventual calculation of the amount consumed of nutrients. (Note that the total amount consumed of nutrient depends on the amount of food eaten and the amount consumed). The portion size information from the portion data set 130 is used as an input to food-portion link 150. In the example embodiment, the food portion link table uses the recipe information from recipe file 140 and the combination code to generate the food code to be output to the output data sets 140.
“In this example embodiment, the recipe codes obtained from recipe files145 are used to determine the specific constituent ingredients within each food item. It also indicates the amount of each ingredient within each food item. Recipe files 145 have been created in the preferred illustrative and exemplary embodiments to accurately reflect every nutritionally important ingredient within every one of many thousand foods people eat. The recipe files 145 can be used to very precisely specify the ingredients of food items and the proportions. Other cases, such as homemade food like pancakes or soup, the recipe files might be less precise and may instead use the more detailed ingredient breakdowns from the dietary intake database 110. In any event, the purpose of recipe files 145 is to as accurately as possible resolve individual food items identified in the dietary intake data set 110 into their respective individual nutritionally-significant constituents and corresponding relative amounts of each.”
“The exemplary embodiment uses this recipe information in conjunction with portion size information (e.g., in a straightforward mathematical multiplication or other scaling) to obtain the amount (e.g., in grams or other convenient quantity units) of each nutritionally-significant constituent within the food item that was consumed. This information, along with the constituent identification and quantity information, is used to index nutrition data set 120. This generates a list of nutrients in each constituent component in the same amount. This nutrient list/quantity is output by the nutrition data set 120 to the output data sets 140.
Data set 140 also receives in the exemplary embodiment the person identifier of the dietary intake data sets 110. This allows for individual tracking of consumption trends. To facilitate demographic-based analysis, the output data set 140 also contains demographic information, such as age, gender and other factors. The output data set 140 may store the food code, which can be used to identify the specific food. It also has the ability to store information about serving sizes, obtained from the portion data set 130.
“FIG. “FIG.7” is an example diagram that illustrates a specific, non-limiting, but exemplary data set linkage. This FIG. This FIG. 7 diagram shows exemplary links between an example household Master Table 160, the dietary intake information set 110, portion size data set 130, and the food portion table 150. The illustrative and exemplary embodiment shows a human dietary researcher adding the portion code to food portion link 150 after a specific recipe has been finalized. When the food descriptor algorithm 500 routines are completed, the portion code (portion_cd), is added to the diary 110. After this, the serving sizes of the person are added to the diary 110.
“As illustrated in FIG. “As shown in FIG. Once assigned, the person ID from the household master table 160 is included in the individual dietary intake records of dietary database 110. This allows system 100 to associate data set 110’s dietary intake information with the person identification assigned by exemplary systems 100. System 100 can personalize the dietary intake data 110 to reflect individual information. It is also able to assign the exact same person ID to multiple participants in multiple dietary intake survey.
The exemplary system 100 focuses primarily on the eight dimensions of food coordinate data as illustrated in this exemplary portion of a dietary intake dataset 110. The exemplary system 100 converts the information in the eight fields of dietary intake data sets 110 into a 32-digit combination codes that are then used to index into the food-portion-link table. This unique code uniquely defines thousands upon thousands of foods and associates them with the recipe file 145 as well as the portion information in the portion data set 130. The food-portion table 150 links the dietary intake data sets 110 and 130. It is possible to calculate the statistically correct portion size using the portion data set 130 once the individual’s demographics and the specific food item have been determined. This information is combined with the recipe information used by System 100 to calculate the nutrient levels from the nutrient set 120.
“Exemplary & Illustrative Process Flow”
“FIGS. 8A?8E shows an example process flowchart showing the steps taken by the preferred embodiment. The process flow shown in the example begins at a “start?” FIG. 8A (block 1100). In the exemplary embodiment, the first step is to read in a record of dietary intake from data set 110. Then add a unique person identifier to an incoming dietary intake household filing 1102 for a specific year (block 1104). The dietary intake data set 110, as mentioned above, is organized by household but also contains individual demographic information for each household member.
“FIG. “FIG. 8A Block 1104. 8A block 1104. This exemplary embodiment creates a unique identifier for each person in order to better track them. The 110 exemplary dietary intake data sets are more concerned at the household level. Different household members could act as reporters at different times within the exemplary diet intake data sets. This can be problematic if one wants to determine individual food consumption. FIG. 100 is an example of the system 100. 8B routine 1104 is used to identify individuals who participated in the dietary intake study in previous years and to assign the same ID (e.g. a sequential or another value) to them. The example embodiment will match individual consumption data from data set 110 that correspond to different survey times periods and record them in output data set 140 with the same unique individual ID to enable individual tracking of intake and nutrition. Routine 1104 adds household master files 160 to accommodate new people, and generates new identifiers for them on an as-needed base.
Referring to FIG. “Referring to FIG. If there is a match (?”yes? Exit to decision block 1108), then it is not necessary to update the master family file 160 with a new ID. If there is not a preexisting identifier in the master household 160 that corresponds to this individual (?no?) Exit to decision block 1108, the preferred embodiment adds a record to the master file 160 and increments the?last name identifier? Counter to generate a unique person identifier. The example embodiment shows routine 1104 updating the dietary intake file household file 1102 for the year with the new/existing individual ID (block 1112) and returns (block 11114).
Referring again to FIG. 8A. Once routine 1100 has determined that the incoming diet intake household file 1102 contains a unique ID for each person associated to a dietary intake record (FIG. ?away from your home? diaries 1116 and 1118 to create an individual-specific common diary (block 1120). This is the exemplary and illustrated embodiment 100. It uses these diaries 1116 for identifying and filtering out those who have not reported correctly eating food items. This information is not included in the dietary intake data 110. The example embodiment 100 describes the meals that have been reported as eaten, but contains some fields that are null (i.e. it fails to indicate what was eaten at each meal). The example embodiment codes non-reported home meals as one value (e.g.?10?). Away from home, non-reported meals will be coded with a different value (e.g.?10?). These records will be filtered later. Non-reported meals can be distinguished from valid skipped meals. The preferred embodiment 100 tracks for certain analysis.
“FIG. 8C is a flowchart showing illustrative and exemplary steps taken by the diary creation program 1120 in FIG. 8A. Referring to FIG. Referring to FIG. 8C, the example embodiment first verifies, for each in-home record with a particular value indicating away home eating that there are at least one or more food records in the associated away home file 1118 (block 1102). If this is not true, the example embodiment 100 codes each in-home record with a specific coding that indicates a non-reported dinner (block 1122). Block 1124 merges the away from home and in-home files 1116, 1118 to create a diary for each time period. This example uses the diary file 1126 as a tool to identify each person (here identified with household ID, member number) and determine the day and meal. If there is no valid helping code for a food record, each food item in the record that corresponds to that person, day, and meal will be coded to indicate an unreported in-home meal (block 1128). This data processing eliminates non-reported meal records to improve the accuracy of the resulting data set 140.
“Referring again to FIG. 8A. Once the food intake diary 1126 is created, the exemplary embodiment matches household records with diary records to obtain the personal identifier and gender from the updated household tableau 1102 (block 1102). Next, the process 1100 exemplar performs the food descriptor removal algorithm 500. This is the process of generating a code value for each record in the diary file 1126 that either identifies an item from the food portion table 150 or a new item (i.e. an item not yet identified within the food-portion table 150).
“In the example embodiment, reduction routine 500 allows preferred embodiment to tie together data sets 110 and 130 regarding dietary intake. The reduction algorithm 500, in the exemplary embodiment, is not able to generate all possible combinations (eight factororial), but it can find an existing food code and create a place for a food code using only few thousand records (e.g. on the order 2000). The reduction algorithm 500 significantly reduces the number of food items that the dietary intake information set 110 can theoretically identify into a smaller number of food items most people eat.
“FIG. 8D illustrates steps that the food descriptor algorithm 500 of the illustrated embodiment performs. This particular example uses eight fields from the dietary intake database 110 (i.e. Type, Form and Characteristic, Flavors, Classification, Preparation Type, Packaging Type, Special label Code). The exemplary embodiment uniquely identifies over 5000 food items (block 112). Each of the fields in the example embodiment contains a numeric number between 1 and 3 characters. In the example embodiment, many of these codes are combined and grouped according to dietary intake factors that relate to the nutritional makeup of the food. These groups are stored in the food-portion link lookup table 150.
“A COMBO KEY file 1134, which correlates combination key information and column name/column value information, is kept. The COMBO KEY in the example embodiment may be a four-digit sequential number that identifies one eighth of a 32-byte character cod. The column name field can be used to identify the column in the dietary intake data 110. One or more values may be provided in the column value field that can apply to this specific column and its associated food category. You may also want to keep a category code that specifies a certain number of unique codes (e.g. 53) that identify a specific type of food group (e.g. 10=cereals and 3=milk, 53=baby foods, etc ).
“As an illustration, consider the following values:
“Type: 0\nForm: 12\nCharacteristic: 0\nFlavor: 249\nClassification: 2\nPrep Method: 0\nPackage Type: 3\nSpecial Label: 34”
“Assume that the following rows are found in the COMBO KEY 1134 file:”
“COMBO? “COMBO?” 34nLABEL CODE CD
“In the above example, the reduction algorithm 500 would take every dimension and scale cmbntn_val columns for the values given for each column name and category code. The COMBO KEY will be converted into a four-character string, and then inserted in the appropriate section of a 32 character code.
“000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000” pkg specialnmthd. type label”
“Not all dimensions are necessary to identify a food. The algorithm can be flexible enough to assign column names 0 (null), which are not found in the scan of the preferred exemplary embodiment.
“Thus, in the above example combo key file, 134 fragment, one scans for cmbntn_val=249 and col_nm=flvr_Cd. The algorithm will then find the corresponding unique combo_key equal or greater than 299. In the exemplary embodiment, that value is converted to?0299 by the algorithm. The algorithm in the exemplary embodiment converts that value to?0299?. Next, SPC_LBL_CD=34 for category=10 is found. The unique combination key for that record’s record is then converted to string?3243? and then inserted into the eighth of the octet. These are the only dimensions required for this particular food item. There are no other column names (not illustrated) for this category code. The 32-character code corresponds to this:
FIGS. FIGS. 9A?9L illustrate an example of the progression of data structures as described above. FIG. FIG. 9A illustrates an example excerpt from a food diary 1126. Not all fields are shown due to space limitations. The exemplary system 100 will begin to process the diary 1126 information. A first record will be pointed to by the FIG. 9B). The first record is read in, and then the steps 1132 and 1136 of FIG. 8D is used to create a food code that can be rewritten to the journal 1126 (see FIG. 9C). FIG. FIG. 9D shows an excerpt of the exemplary diary 1126, taken after the FIG. 8D routine 500 is executed for all records in the extract.
“FIG. FIG. 9E illustrates an example combo key file 1134, which is used by FIG. 8D block 1132 is used to convert the eight dimensions of a food item in the dietary intake data 110 into a 32-character unique number that maps to the food portion table 150. FIG. 9E is an example of the records returned by the combo key file 1132. 9E shows an excerpt from the combo key file 1134 records for an exemplary diary entry. 9F containing a category code for a particular value (in our case,?27?). FIG. FIG. 9F highlights the relevant data fields. The key columns in this example are: TY_CD, FRM_CD(Form), FLVR_CD(Flavor), FLVR_CD [Flavor], PKG_TY_CD (?Packaging Type?), SPC_LBL_CD (?Special Label?). Other highlighted columns are not relevant for this particular category code.
“FIG. 9G illustrates an example in which the Type (TY_CD=1) was included in a CMBNTN_VAL within a specific record number 1017. This is an example where the type column name can be parsed to create a combination key that is based on the?Type? information. FIG. FIG. 9H illustrates a similar example, where another combination key (??1020?) is used. 9H shows a similar example where another combination key (?1020? in this instance) is created using the?Form. information in the diary 1126. FIGS. FIGS. 9I, 9J, and 9K provide additional examples of combination codes that can be used for flavor, package type, and special label coding. These are just examples. A different diary entry can generate different combination codes depending on the specific values of the eight food descriptor dimensions.
“In this case, the 32-digit food code obtained by adding the different combination codes in a specific order is defined as:
FIG. 9L highlights the matching records that make the combination key. 9L. FIG. FIG. 9M shows the corresponding diary record 1126, which is returned with the 32-digit combination codes obtained from the 500 food descriptor reduction process.
“Once the data-reduction algorithm has determined the 32-character key using the record number of found items in the combo key (block 1136), then the reduction algorithm searches for the 32 character code in the food portion line file 150 (block 1138). The data reduction algorithm 500 scans the file 150 containing food-portion links. This file contains information about foods that have been previously defined (i.e. by previous operations of System 100). It combines the portion size data set 120 and the nutrition data sets 130 to generate this 32-character code. If found (?yes? If found (?yes?), exit to decision block 1140. The food diary record 1126 will be updated with the current food key, portion code, and recipe code (block 1142), which were obtained from the food-portion table 150 (see FIG. 7). 7). Storage space consideration. Routine 1100 increments a counter that tracks the number of times this food item has been used in the diary 1126. The diary 1126 record also includes the food_cd value as well as the portion code from the portion data set 120. This data can be later merged into the final output data sets 140.
“If a 32-character code for food is not found in food-portion link file 150 (???), then the new code is added along with all eight dimensions (i.e., type, form and characteristics of the food item) to identify it. Exit to decision block 1140. The new 32-character food code is then added together with all eight dimensions (i.e. type, form and characteristics, flavor classification, packaging method, special label) to identify the food item. This creates a?new foods’. Table 1144 (block 1146). FIG. 8D continues until all entries in diary 1176 are processed (block 1148).
“In general, if the 32 character code is not found, an exception record will be created and one of two things will occur. If the 32-character code cannot be found because of a new or previously undefined value in the dietary intake database 110 (and assuming that the new value has validity), then the new value will go into the appropriate group according to a (human) dietetic intake scientist. If the new item is not for a food previously tracked, then a human dietary intake scientist will create a new recipe in file 140 based on nutrition data set 120, the values of the eight dimensions fields within new foods data file 144, and insert the new item into food-portion link 150. This allows later encounters with the same food item to get the appropriate nutrition information (FIG. 8A, blocks 1154 and 1152. FIG. 8A shows the exception handling that is possible after the data reduction algorithm 500 has been completed. It is based on the contents of the new foods files 1144. If there are any new foods in the new foods files 1144 after the data reduction algorithm 500 (?yes?) Exit to decision block 1150. Next, for each new food, a dietician creates or updates a recipe from the dietary intake database 110. Assigns a portion code and then adds the new (or updated record) to the food-portion link table 150 (block 1552). The data reduction algorithm 500 should be run again until there are no more exceptions. This is the end (?no). At this point (?no?), exit to decision block 1100. The preferred exemplary routine 1100 exports 2 deliverables (the diary files 1126 and 160, respectively) in the form an output data set 140. This is for final processing (block 1156).
“FIG. 8E is an illustration of a flowchart that illustrates the final merge procedure 1156 as shown in FIG. 8A. 8A. FIG. 8E: To create a temporary diary file 1126, we remove all records that indicate non-reported meals (block 1160). Is this temporary diary file 1126? This temporary diary file 1126 is combined with the nutrient information set 120 (or at most portions thereof) to create a food base data set 140. The specific embodiment is described for illustration purposes only. It initializes cereal flags, lunch, dinner, and breakfast flags. The routine 1156 calculates the correct amount for each nutrient based upon serving size (block 1164).
“GMI FOOD FOODnMEAL DAY HLP GRP MJRnperson ID NBR G FOOD CD YEAR FOOD NM CD FOOD NM, Plain Oats, NSF7800n108538 2 119 111106 Regular, Caffeinated NSF118? 1100\n108538 1 49805 2 1 19 31867 Skim milk ? ?24C 2200\n108538 1 49805 2 1 19 41375 Tomato Juice 62 1800\n108538 1 49805 2 1 19 101136 Raisin Bran ? ?77C 1900n108538? 49805? 2 19 101136 Raisin Bran 62 1800n108538? 49805? 2 19 51375 Tomato Juice 62 1800n108538? 49805? 2 19 101136 Raisin Bran 74 1900nNR Untoastedn108538? 49805? 2 19 5012361 Bagels Regular White or 74 1900nNR Regular Sugar Hyd Banana? ? ?63A 1800nor.NR, Fresh, Reg Sugarn108538 3 49805 2 19 441882 Peanut butter, Reduced Fat ?60B 1700\n108538 2 49805 2 1 19 5213731 Cracker, Regular Butter, 90 2300\nRegular\n108538 3 49805 2 1 19 253275 Vegetable, Plain, Black 70 2600\nBeans, Cooked\n108538 3 49805 2 1 19 253931 Combination Vegetable 69 2600\nDish, Plain Mixed Veg,\nCkd\n108538 3 49805 2 1 19 253310 Vegetable. Plain, Beets, 68 2600nRawn108538 2 49805 2 19 412062 Ketchup Regular 125? 2300\n108538 3 49805 2 1 19 331864 Apple Pie 102? 1500n108538 2 49805 3 1 19 331864 Apple Pie 102? ?75A 1900\nToasted\n108538 3 49805 2 1 19 284168 Ice Cream, 27 2200\nDiet/Lowcal/Sorbet, Van or\nNR, Reg Fa\n108538 4 49805 2 1 19 387384 popcorn light 84 2300”
“GminMeal Portion Cereal Sz Ind Grn Ind Nutrient Cd Srvng Sz Ind Grn Ind Grn Nutrient Cd Consumed Breakfast Lunch Dinner Snack”
“GminMeal TotnPerson Smry Cereal Grm Yogurt Rte Coffee Tea Presweet Energie Tot Carb Protein 137.43 2.30 00 5.82 n108538 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 138.86 9.35 10.13 7.09 8.87 1.59 n108538 0 0 0 0 0 0 0 0 0 0 0 0, 0 0 0 0 0 0 0 0 0 0 0 0 0 0 05 0 0 0 0 0 0 0 0 0 0 0 0 0 0 e 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1088538 0 0 0 0 0 0 0 0 0 0 0 0 0 001 0 0 0 0 0 0 0 0 0 0 0 0 0 0 sn108538 0 0 0 0 0 0 0 0 0 0 0 0? 0 0 0 0 0 0 0 0 0 0 0 0 0 0???
“Gmi Tot Tot\nMeal Anml Veg Tot Sat Musat Pusat\nPerson Smry Protein Protein Alcohol Cholest Fat Acid Fat Acid Fat Acid Fructose Galactose Glucose Lactose\n108538 1 0.00 5.82 0.00 0.00 0.40 0.73 0.84 0.05 0.00 0.00 0.00\n108538 1 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00\n108538 1 3.27 0.00 0.00 1.73 0.12 0.05 0.01 0.00 0.00 0.00 4.24\n108538 1 0.00 0.00 0.00 0.00 0.02 0.02 0.04 3.99 0.00 2.94 0.00\n108538 1 0.00 5.05 0.00 0.00 0.13 0.22 0.71 2.96 0.00 3.45 0.00\n108538 1 0.00 5.89 0.00 0.00 0.12 0.07 0.39 0.11 0.00 0.22 0.00\n108538 2 0.00 0.88 0.00 0.00 0.16 0.04 0.08 2.38 0.00 3.70 0.00\n108538 2 0.00 7.01 0.00 0.00 1.47 4.28 3.03 0.00 0.00 0.06 0.00\n108538 2 0.00 1.00 0.00 0.00 0.99 2.30 0.53 0.03 0.00 0.06 0.00\n108538 3 0.00 9.09 0.00 0.00 0.16 0.05 0.26 0.84 0.00 0.63 0.00\n108538 3 0.00 2.79 0.00 0.00 0.02 0.01 0.11 0.56 0.00 0.62 0.00\n108538 3 0.00 0.82 0.00 0.00 0.02 0.02 0.03 0.05 0.00 0.10 0.00\n108538 3 0.00 0.19 0.00 0.00 0.01 0.01 0.02 0.46 0.00 0.96 0.00\n108538 3 0.00 3.78 0.00 0.00 5.03 8.79 5.29 6.84 0.00 2.41 0.00\n108538 3 1.52 2.67 0.00 31.25 3.36 2.88 0.81 0.16 0.00 0.37 1.08\n108538 3 0.00 0.40 0.00 0.00 0.00 0.01 0.05 0.56 0.00 0.62 0.00\n108538 4 0.00 8.69 0.00 0.00 3.17 8.14 2.76 0.07 0.00 0.07 0.00”
“Gmi Tot Sol\nMeal Diet Diet Insol Bc\nVperson Smry Maltose Sucrose Starch Fib Fib Diet Fib Pectins Tot Vitm A Equiv Retinol Vitm D\n108538 1 0.00 0.52 21.02 4.08 1.90 2.21 0.00 853.87 0.00 601.61 0.00\n108538 1 0.00 0.00 0.10 0.38 0.21 0.21 0.00 0.00 0.00 0.00 0.00\n108538 1 0.00 0.00 0.00 0.00 0.00 0.00 0.00 204.00 1.93 57.75 0.98\n108538 1 0.00 0.00 0.00 0.84 0.46 0.38 0.46 556.00 698.52 0.00 0.00\n108538 1 0.90 8.54 18.52 7.35 0.92 6.43 0.11 1364.00 0.40 224.43 1.24\n108538 1 0.00 0.15 25.73 1.29 0.50 0.79 0.00 0.00 0.00 0.00 0.00\n108538 2 0.00 5.73 3.52 2.11 0.53 1.59 0.53 81.00 43.18 0.00 0.00\n108538 2 0.19 2.34 6.40 0.68 0.20 0.48 0.22 0.00 0.00 0.00 0.00\n108538 2 0.00 0.75 6.75 0.26 0.15 0.11 0.00 0.00 0.00 0.00 0.00\n108538 3 0.00 1.57 9.72 6.69 2.51 4.19 0.00 2.00 1.26 0.00 0.00\n108538 3 0.00 0.65 0.82 3.61 1.41 2.20 0.91 3209.73 2389.84 0.00 0.00\n108538 3 0.00 4.18 0.34 0.98 0.59 0.39 0.22 33.60 10.30 0.00 0.00\n108538 3 0.35 1.42 0.00 0.17 0.04 0.13 0.00 1016.00 77.87 0.00 0.00\n108538 3 0.00 19.29 23.91 2.50 1.00 1.37 0.37 20.47 21.11 0.00 0.00\n108538 3 0.08 8.78 19.90 1.36 0.25 1.11 0.00 211.03 42.02 18.97 0.31\n108538 3 0.00 21.18 0.00 2.51 0.28 2.23 0.28 34.20 20.52 0.00 0.00\n108538 4 0.00 0.29 38.82 10.94 0.37 10.57 0.00 163.17 85.01 0.00 0.00”
“GminMeal Tot Alpha Beta Gamma Delta 1 0.00 0.00 0.04 0.13 0.09 1.41n108538 0.00 0.02 0.01 0.18n108538 0.00 0.09 1.69 0.00 0.04 0.08 0.37 0.41 1.47n108538 0.00 0.02 0.02 1.16n108538 0.00 0.59 1.62 1.47 0.00 2.64 0.00 0.00 1.07 0.00 9.67 19.92 1.51 0.28 2.25n108538 1.35 1.39 0.23 1.57n108538 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 5.00 0.35 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 n108538 n108538 1.25 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 1.25 66 0.00 0.00 0.00 0.00 0.00 0.00 5.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00???
“GminMeal Vitm Vitm & Person Smry Acid B6 Folate, B12 Calcium Phosphorus Magnesium Selenium and Iron”
“GminMeal Caproic Capric Lauric Palmitic Stearic Sodium Potassium Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid 1 380.08 159.39 0.02 0.00 0.00 0.01 0.00 0.02 0.00 0.00 n108538 131.65 0.00 0.00 0.00 0.00 0.00 0.00 0.11 0.00 0.53 n108538 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 5.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00??? 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 9.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00?? 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00?n108538???
“Gmi Meal Arachidic Behenic Myristoleic Palmitoleic Oleic Gadoleic Erucic Linoleic Linolenic Parinaric Arachidonic\nPerson Smry Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid Acid\n108538 1 0.00 0.00 0.00 0.00 0.73 0.00 0.00 0.80 0.05 0.00 0.00\n108538 1 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00\n108538 1 0.00 0.00 0.00 0.01 0.04 0.00 0.00 0.00 0.00 0.00 0.00\n108538 1 0.00 0.00 0.00 0.00 0.02 0.00 0.00 0.04 0.00 0.00 0.00\n108538 1 0.00 0.00 0.00 0.00 0.09 0.00 0.00 0.32 0.03 0.00 0.00\n108538 1 0.00 0.00 0.00 0.01 0.07 0.00 0.00 0.37 0.02 0.00 0.00\n108538 2 0.00 0.00 0.00 0.01 0.03 0.00 0.00 0.05 0.03 0.00 0.00\n108538 2 0.05 0.10 0.00 0.00 4.16 0.11 0.00 2.97 0.06 0.00 0.00\n108538 2 0.00 0.00 0.00 0.00 2.30 0.00 0.00 0.53 0.00 0.00 0.00\n108538 3 0.00 0.00 0.00 0.00 0.05 0.00 0.00 0.14 0.12 0.00 0.00\n108538 3 0.00 0.00 0.00 0.00 0.01 0.00 0.00 0.02 0.07 0.00 0.00\n108538 3 0.00 0.00 0.00 0.00 0.02 0.00 0.00 0.03 0.00 0.00 0.00\n108538 3 0.00 0.00 0.00 0.00 0.01 0.00 0.00 0.02 0.00 0.00 0.00\n108538 3 0.00 0.00 0.00 0.00 8.79 0.00 0.00 4.96 0.32 0.00 0.00\n108538 3 0.00 0.00 0.00 0.30 2.57 0.00 0.00 0.75 0.04 0.00 0.01\n108538 3 0.00 0.00 0.00 0.00 0.01 0.00 0.00 0.04 0.02 0.00 0.00\n108538 4 0.04 0.00 0.00 0.06 8.06 0.02 0.00 2.69 0.08 0.00 0.00”
“Docosp Docosh\nGmi Meal Eicos Enoic Enoic\nPerson Smry Acid Acid Acid Tryptophan Threonine Isoleucine Leucine Lysine Methionine Cystine Phenylalanine\n108538 1 0.00 0.00 0.00 0.07 0.19 0.23 0.45 0.23 0.12 0.14 0.31\n108538 1 0.00 0.00 0.00 0.00 0.00 0.00 0.03 0.00 0.00 0.00 0.00\n108538 1 0.00 0.00 0.00 0.05 0.14 0.20 0.32 0.26 0.09 0.03 0.15\n108538 1 0.00 0.00 0.00 0.00 0.04 0.04 0.04 0.04 0.00 0.00 0.04\n108538 1 0.00 0.00 0.00 0.08 0.15 0.18 0.32 0.15 0.07 0.11 0.22\n108538 1 0.00 0.00 0.00 0.07 0.17 0.22 0.42 0.14 0.11 0.12 0.29\n108538 2 0.00 0.00 0.00 0.01 0.03 0.03 0.06 0.04 0.01 0.02 0.04\n108538 2 0.00 0.00 0.00 0.09 0.25 0.31 0.51 0.37 0.09 0.09 0.37\n108538 2 0.00 0.00 0.00 0.01 0.03 0.03 0.07 0.02 0.02 0.02 0.05\n108538 3 0.00 0.00 0.00 0.10 0.39 0.40 0.73 0.63 0.14 0.10 0.49\n108538 3 0.00 0.00 0.00 0.02 0.10 0.10 0.12 0.14 0.04 0.02 0.09\n108538 3 0.00 0.00 0.00 0.01 0.03 0.03 0.04 0.03 0.01 0.01 0.03\n108538 3 0.00 0.00 0.00 0.00 0.01 0.00 0.01 0.01 0.00 0.00 0.00\n108538 3 0.00 0.00 0.00 0.05 0.11 0.14 0.25 0.09 0.05 0.07 0.18\n108538 3 0.00 0.00 0.00 0.05 0.15 0.18 0.39 0.18 0.09 0.08 0.21\n108538 3 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00\n108538 4 0.00 0.00 0.00 0.06 0.32 0.31 1.06 0.24 0.18 0.16 0.43”
Click here to view the patent on Google Patents.