Reasons the Reported Funding Levels for Categories using the Research, Condition and Categorization (RCDC) Process Might Change as Compared With Previously Reported Years

1. Uniform Definitions
   
2. Computer Sorting
   
3. Category Scope
4. Uniform Scope of Search
   
5. 100-Percent Reporting Rule
   
6. Accounting at the Subproject Level
   
7. Business Rules and Organizational Reporting
   
8. Projects Reported in Different Areas

	1. Uniform Definitions
	Scientific experts across the NIH worked together for 2 years to develop a single, NIH-wide list of scientific terms and concepts for each of the 215 reported categories. These terms and concepts, along with a few standard business rules, are referred to as "definitions" for the various categories. The definitions are not stated in sentences; rather they take the form of a list of scientific terms and concepts. In addition, each term or concept is given a positive or negative weight relative to its relationship to the category. The new process is not simply a keyword search strategy. Before RCDC, each IC sorted, categorized, and reported its own research portfolio, using approaches often unique to the individual IC. In many instances, the ICs used similar definitions but IC interpreted the definitions based on its mission, focus, and portfolio. The new RCDC process not only uses a universal definition for each reporting category, but also applies it uniformly to projects across all of the ICs. The result is a more consistent report of funding across the NIH. For more information about how the NIH developed RCDC definitions, please see http://report.nih.gov/rcdc/category_process.
	2. Computer Sorting
	RCDC uses a text-mining computerized application to categorize and assign NIH-funded grants, contracts, and in-house research (NIH intramural research). Text mining is a process that categorizes and clusters projects using words and multiword phrases from a project description to generate an outline of each project's content based on the meaning of those words and phrases. Before RCDC, ICs applied scientific judgment to assign projects to categories based on their IC's research mission. Sometimes the IC program staff and/or analysts read and categorized each project using a manual sorting process. Some ICs used the Computer Retrieval of Information on Scientific Projects (CRISP) system and supplemented it with staff review. For more information about CRISP-on-the-Web, please see http://report.nih.gov/crisp/index.aspx. For decision making, the computerized RCDC process relies on the clarity and specificity of the text provided in the researcher-generated grant, contract, or intramural project descriptions. For example, a principal investigator writing about plans to develop a new material for tooth restoration might use the terms "stress response" and "fatigue." Text mining might initially select that project as relevant to Behavioral and Social Sciences (BSSR) because those terms are included in the BSSR category. RCDC excludes such misreads by adding negative weights to some category terms and concepts within a specific category
	3. Category Scope
	RCDC categories might be more restrictive or inclusive than previous IC category definitions. A category scope might now be defined more narrowly, for example, the project list may now exclude some basic research or research that is related but not directly relevant. Basic research—research that focuses on discovery of fundamental, underlying processes—often has a direct impact on many research areas and diseases. This fact makes basic research more difficult to categorize, but its contribution is significant. For example, research on osteogenesis (bone growth) might provide information that can be used to treat osteoporosis (bone loss), but the basic knowledge is not directly related to osteoporosis. Another example involves findings in one area of E. coli investigation that had an unexpected impact on possible breast cancer treatments. Since these findings were not expected, it is unlikely they would be categorized in Breast Cancer. Instead, the findings would be included only in the categories relevant to the ideas proposed in the grant information. In the future, if this scientific relationship holds true, the category definition for breast cancer might change to include E. coli. In other cases, the RCDC category may be broader in scope than a particular IC might have previously reported. For example, due to the new agreed-upon definitions established by scientific experts, RCDC selects more research on the trigeminal (facial) nerve for the category Temporomandibular Joint and Muscle Disorders (pain in jaw muscles and joint) than ICs had previously included in this category. By their nature, some categories are more likely to have been broadly defined through the RCDC process. For example, the concepts included in the category Genetics are intended to capture a wide range of projects and types of research. In contrast, the scientific experts interpreted the category Biotechnology to have a specific and direct meaning, resulting in a project listing that is narrower in scope.
	4. Uniform Scope of Search
	In most instances, the RCDC process searches the text of the same four submitted sections (title, abstract, specific aims, and public health relevance) for each funded grant, contract, or intramural project to find terms that match the scientific concepts contained in the category definition. Previously, some NIH ICs used additional sections of the grants and contracts, such as the research methods and significance sections. They also might have used other sources of information, such as progress reports or referenced articles from intramural investigators, to help categorize intramural research. In some cases, this strategy might have been more precise, but the overall effect was uneven. RCDC applies a more consistent approach across all IC reporting.
	5. 100-Percent Reporting Rule
	While some NIH ICs previously prorated (divided proportionately) project dollars based on the perceived relevance for different categories, the RCDC process allocates 100 percent of the project dollars to each of the categories to which the project is assigned. Projects that are only slightly related to a category do not get categorized. RCDC categorizes each subproject from NIH multicomponent grants independently. This practice results in a method of proration for larger grants. NIH chose the 100-percent allocation method because descriptive information about the various components of a study (that is, each hypothesis or specific aim) does not necessarily equate to how dollars are spent for a category by the investigator. Without detailed accounting of how dollars were spent by category, proration is methodologically very challenging. To further explain the RCDC method, here is an invented example. A molecular genetics study of craniofacial development with three specific aims, two of which involve mouse stem cells and one that involves a biotechnology, might have been categorized by the IC as 100-percent Genetics, 66-percent Stem Cell Research, and 33-percent Biotechnology. RCDC would categorize this project as 100-percent Genetics and 100-percent Stem Cell Research, but it would not categorize the project in Biotechnology. This 100-percent reporting rule will influence the total funding reported for any category, and the rule will be applied consistently. In addition, all of the categories added together will not equal the total NIH appropriation because work conducted in many research projects has relevance to more than one category. But if all projects from all categories are combined, duplicate projects are removed, and total dollars are added, the sum will be less than the NIH budget because the categories do not encompass all types of biomedical research. As a result, some NIH-funded projects might not fall into any of the 215 reported categories.
	6. Accounting at the Subproject Level
	RCDC categorizes grants at the subproject level while some NIH ICs previously categorized grants only at the primary or parent grant level. For example, a large grant might include a data coordinating center and multiple related trials that are broken out as subprojects. Using RCDC, each of these individual subprojects would be assigned to categories separately. As a result, an individually calculated dollar amount for each subproject would be counted at 100 percent in those categories to which it is assigned. For this reason, accounting at the subproject level can easily influence the total funding reported for any category.
	7. Business Rules and Organizational Reporting
	For some categories, the RCDC process presets parameters, or "business rules," to determine whether a project should be included in a category. For example, all projects categorized as Autism will automatically be included in the Mental Health category as part of the RCDC business rules agreed to by the NIH IC scientific experts. In the past, some ICs also used business rules to categorize projects. In the historic coding processes, these IC-specific business rules were sometimes associated with the organizational structure of the IC. For instance, all projects administered by a specific center or division in an IC might be automatically assigned to a certain reporting category. For RCDC, only in situations in which the category matches an IC name is such a business rule in place.
	8. Projects Reported in Different Areas
	In the RCDC process, a project may be reported in areas in which an NIH IC has never reported before. In the past, some ICs might have reported only on categories in which estimated funding for that category exceeded a certain dollar threshold. The RCDC process scans and assigns all research projects and dollars attributable to relevant categories whether or not a project was previously reported in that area. This difference does not represent any change in research priorities. It only reflects changes in the way the new RCDC process categorizes research projects. For example, an IC might not have reported on the broad research category Lung. However, the IC might have reported on lung conditions or diseases specific to that IC’s mission, such as Asthma, Lung Cancer, or Tuberculosis.