On July 25, 2001, the Office of Dietary Supplements sponsored a workshop on the assessment of dietary supplement use and the potential need for a national database of dietary supplements. The database would be accessible to anyone and would be especially valuable to researchers interested in assessing dietary supplement intake. It would include vitamins, minerals, botanicals and a wide variety of other products, as established by the Dietary Supplement Health and Education Act of 1994. The charge for the day was to assess various models of food composition and dietary supplement ingredient databases and identify key components of database development. These components included intended uses; structural features; information content and maintenance, oversight and costs of a national dietary supplement database. The participants included domestic and foreign researchers, members of the dietary supplement industry, and government officials.
· Models of dietary supplement databases
A representative from the National Center for Health Statistics (NCHS) described the data collection methods used to capture information about dietary supplement intake in the National Health and Nutrition Examination Survey (NHANES). This nationally representative survey is designed for many research purposes such as estimating the prevalence and distribution of health conditions and related risk factors in the population. NHANES provides information on the relation of diet, nutritional status and health over time. It also serves as a tool for identifying emerging public health issues.
Data on use of dietary supplements in NHANES are collected during home interviews. Survey participants are asked about dietary supplement use during the past month. If they use supplements, the interviewer asks to see the product container. Informational items collected from the label include product name, manufacturer and address. The interviewer asks participants how long the product was taken, how often, and in what amounts.
Once the field-work is completed, NCHS staff attempt to obtain label information from the product manufacturers, either from the producer directly, through the Internet, or by using the Physicians’ Desk Reference. The current NCHS database on dietary supplement use contains about 1,600 products with 914 unique dietary supplement ingredients. The database includes the product name, serving size (e.g., 1 capsule), recommended dosage, source of product information, product type (e.g., for children), form of the product (e.g., caplet), and names of ingredients as listed in “Supplement Facts” panel and their quantity. The ingredients listed below the “Supplements Facts” panel on the label are not currently entered into the database.
The NCHS representative highlighted many challenges in creating and using a dietary supplement database. These include: identifying and classifying products (e.g., determining whether products such as teas should be classified as foods, dietary supplements or both); constructing a database that can be searched for various forms of an ingredient; and matching reported products with their designators in the database because many products have similar names.
Issues that must be addressed for any database of dietary supplements include: changes in labeling that are not accompanied by changes in product formulations and vice versa; inability to identify and quantify ingredients in products that are proprietary blends; inclusion of directions, instructions, claims or certifications; categorization schemes for ingredients; utility of product bar codes; and tracking changes in product formulations.
A representative from the Australian Government described a database called the Australian Register of Therapeutic Goods (ARTG). That database was established under the authority of therapeutic goods legislation and is primarily used for regulatory purposes. These include ensuring the safety of products on the market in Australia, as well as their quality and efficacy. The ARTG contains prescription and over-the-counter products (including products used in the practice of complementary and alternative medicine). Many of the products that are defined in the U.S. as dietary supplements are considered complementary medicines in Australia. The speaker mentioned that categorizing products is a challenge, but that a database that supports a comprehensive search strategy may alleviate that problem. The ARTG captures many features about the product by asking manufacturers questions that are recorded onto a computer disk. The information from manufacturers is downloaded to the database. Soon this system will function through the Internet in real time. The type of information contained in the current database includes the manufacturer’s contact information, product name, type and strength of product, dosage form, active ingredients and excipients, indications for use and claims (both health risk reduction and warnings). The ARTG maintains the confidentiality of proprietary blends.
· General discussion by meeting participants
There was general agreement that the main purpose of a national U.S. database is to support and enhance research activities. There are many types of research that could be facilitated by a national dietary supplement database. In addition to studies investigating possible relationships between dietary supplement use and several health outcomes, one participant indicated that a national dietary supplement database could be used in combination with other data sources to perform a cost-benefit analysis of the use of dietary supplements. Discussion centered around whether a national database would be designed for safety or efficacy purposes. A national dietary supplement database may be used for regulatory purposes as well.
A number of researchers described different databases that they have developed for other federal agencies and individual projects including: NHANES, the Women’s Healthy Eating and Living Study, the Women’s Intervention Nutrition Study, the Multi-Ethnic Cohort, the Nurses Health Study and Health Professionals Study, INTERMAP, and the Vitamins and Lifestyle Study. A variety of software packages were used to construct databases that contain anywhere from 51 to 5,400 dietary supplement products. The software packages employed include PowerBuilder and Sybase, DataEase, NetPlus, MicroSoft (MS) Access, Visual FoxPro, Sun-Unix Text Processing, MS Excel, Nutrition Data System for Research, and SAS. All of these projects received some, and in many cases most, of their funding from the National Institutes of Health.
Representatives from the dietary supplement trade associations indicated that they had databases of their products for internal use. One participant mentioned a database of food labels produced by the Food and Drug Administration (FDA) called Food Label and Package Survey (FLAPS). FLAPS data are publicly available. FLAPS includes up to 1,300 food labels of processed packaged foods that are regulated by FDA. FLAPS provides relevant labeling characteristics, such as health claims and nutrient content claims, on those packages.
The particular type of software and system architecture that would be most appropriate for constructing a national dietary supplement database was not decided and represents an area that requires additional consideration.
There was a great deal of discussion about the content of a national dietary supplement database. However, there was no agreement on the degree of specificity that a database should capture. There was also a general sense that ingredient categories were needed, but there was little agreement on what those categories should be. Speakers described broad ranges of information that they collect from dietary supplement labels. Some researchers are most interested in the information contained in the “Supplement Facts” panels, while others think it would be appropriate to include all ingredients and claims made on the label, directions for use, unique product codes (UPC) and the manufacturer. One industry speaker suggested using a unique code that is found on the products they manufacture in order to track the history of each reformulation of a product. An industry representative explained that at least one organization already has a system in place that stores all information that is contained on the product label.
Regardless of what information is ultimately determined to be most appropriate for the database, industry representatives expressed general support for such an effort. Mostly they offered to submit product labels to a central location, which would decrease the workload required to collect product information.
Maintenance, oversight and costs of a national dietary supplement database
There was a dynamic discussion of the issues surrounding maintenance of a national dietary supplement database. Many researchers indicated that they updated their databases daily, while others conducted annual updates or left the database static after its construction. Industry representatives explained that their products might be reformulated on average every 2 years with a range of 1-10 years. These changes in formulation make the question of updating the database a critical issue to address. There was no agreement on how often a database would be updated.
Another issue related to the maintenance of a national database is personnel. During the presentations and discussion, the participants indicated that personnel varied from less than one full time equivalent to 6 full time equivalents to maintain a dynamic database. It is not clear who would actually assume the responsibility of creating and maintaining a national dietary supplement database. NIH is a logical candidate given that 16 NIH Institutes and Centers currently support the USDA food composition databank and most dietary supplement databases used by US researchers have been developed with NIH funds.
Decisions regarding oversight for creating and maintaining a national dietary supplement database will require additional deliberation.
Additional information is needed to determine the cost of creating and maintaining a national dietary supplement database. Estimates could be drawn from the variety of databases that were described during the day-long discussion.