Defining the Components of the Research Pipeline

Abstract

The National Institutes of Health (NIH) has proposed that research moves from “bench to bedside” through a Pipeline consisting of distinct research categories bridged by bidirectional translation. The original NIH Pipeline has now been expanded to include practice‐ and community‐based research. We developed operational definitions for the elements along the expanded NIH Pipeline. The proposed definitions were pilot‐tested and refined using the Computer Retrieval of Information on Scientific Projects (CRISP) and the Secure Internet‐Linked Web Technologies (SILK) database containing funding information for individual NIH awards. Using our definitions, two reviewers independently agreed on the categorization of 94% of 866 National Heart, Lung, and Blood Institute (NHLBI)‐funded projects, and 99% of projects could be placed into a single research category.

Introduction

Much has been written about the gap between research and practice. Legislators and their constituents have expressed concern that the large amounts of money appropriated for medical research have not consistently resulted in a commensurate impact on the health of the public. Researchers and policymakers have become increasingly frustrated that clinicians seem slow in implementing the findings of research, while clinicians argue that much of the research that is being conducted is not relevant to their practices or is impossible to implement. Many have attributed at least some of the problem to a lack of investment in clinical research. To address all of these concerns, in 2002, the National Institutes of Health (NIH) undertook a strategic planning process guided by NIH Director, Elias Zerhouni, resulting in what has been called the NIH Roadmap for Medical Research. ¹

One of the products of the Roadmap Project was the reconceptualization of a Research Pipeline connecting basic research to practice. This Pipeline includes two distinct types of research, basic and human, and translation is required to connect each to the other and to practice, and the process is, in theory at least, bidirectional. An expanded version of this Pipeline was articulated by Westfall et al. ( Figure 1 ). ² They argued that practice‐ and community‐based investigations should also be distinguished from human research. Their conceptualization is then identical to the NIH model except that it includes three research categories and three translational steps. This broader conceptualization has gained acceptance from the Agency for Healthcare Research and Quality ³ and has been incorporated into recent Clinical and Translational Science Award (CTSA) applications.

Figure 1.

The Westfall expanded Research Pipeline.

The purpose of this project was to develop and test operational definitions of the components of the Westfall Pipeline that could be used to evaluate current and future trends in the NIH and other research and research funding.

Methods

We first looked at the NIH Web sites in an attempt to find definitions for the components of the original Pipeline. Information was available that could be used to define “human” research but no definitions were available for the other categories. ⁴ Definitions were available in the medical literature for the added components in the expanded Pipeline. ⁵ We also consulted with researchers involved in CTSA applications. We then established proposed definitions and modified them following each of two pilot studies. The first pilot study was conducted on grants awarded in 2006 by the National Eye Institute and the second on grants awarded in the same year by the National Institute on Deafness and Other Communication Disorders. The changes made following each of the pilot studies were relatively minor and were prompted by the grants that were the most challenging for the authors to categorize, such as training grants, technology development projects, and disease‐specific clinics located in academic centers.

The authors then attempted to categorize all NHLBI external grants awarded in 2006 using the refined definitions. The grants were placed in categories after reading the abstract for each grant. The abstracts were found on the Computer Retrieval of Information on Scientific Projects (CRISP) database, which also provided the grant number, the investigator's name, the investigator's institution, the project title, and, in some cases, a list of key terms. ⁶ Each of the authors independently reviewed and categorized the abstracts and then met to discuss the differences and arrive at a joint conclusion. The monetary value of each grant was obtained from the SILK database by searching for the grant number taken from the CRISP database. ⁷ We included the total amounts of the awards, regardless of the proposed duration of the projects. Funds for projects that crossed research categories were divided equally among those categories.

Results

Our final operational definitions are shown in Table 1 . The “basic research” category was ultimately defined as studies involving nonhuman animals and human molecules, cells, and tissues that could not be tracked back to the individuals from which they were obtained. The “human research” category was defined as research conducted in carefully controlled settings on intact, living humans or on human parts that could be linked directly to the individuals from which they were obtained (for which informed consent would have been obtained). For example, genetic studies in which participants and their genetic material could be linked were considered to be human research, but when genetic material and its source could not be linked, it was considered to be basic research. Technology development projects were likewise categorized based upon the subjects used in the research (nonhuman vs. human subjects).

Table 1.

Definitions of funding categories.

Basic research	Research involving animals or animal‐derived biochemicals, cells, or tissues, or in vitro human studies in which the human participants were not identifiable.
Phase I translation	Projects intended to bridge the gap between basic and human research.
Phase I translational research	Studies of the efficacy or effectiveness of various phase I translational strategies.
Human research	Research involving intact, living humans in carefully controlled settings.
Phase II translation	Projects intended to bridge the gap between human research and practice‐ and community‐based research.
Phase II translational research	Studies of the efficacy or effectiveness of phase II translational strategies.
Practice‐ and community‐based research	Research conducted in representative practice and/or community settings, which is grounded in, informed by, and intended to directly improve clinical or community health practice.
Phase III translation	Dissemination and implementation of the findings of practice‐ and community‐based research. (This has traditionally been called quality improvement.)
Phase III translational research	Dissemination and implementation research.
Training grants	Projects designed to develop research curricula for groups of learners whose future research paths have yet to be determined.

The category of practice‐ and community‐based research was defined as research conducted in practice and/or community settings that is grounded in, informed by, and intended to directly improve clinical or community health practices. ⁵ On the basis of this definition, we placed phase I, II, and III clinical trials (safety and efficacy studies) in the “human research” category and phase IV trials (effectiveness studies) in the “practice‐ and community‐based research” category. Studies involving patients cared for in special disease‐specific clinics (eg, sickle‐cell disease clinics and congestive heart failure clinics), clinics not likely to differ much when in an academic center or in the community if they met all other definitional criteria, were assigned to the practice‐based/community‐based category.

The process of “doing” translation (translation) was distinguished from the process of “studying how to do it” (translational research). For example, in the “phase I translation (T1)” category, we included projects intended to bridge the gap between basic and human research, and “phase I translational research” was defined as studies of the efficacy or effectiveness of various phase I translational strategies. Training grants for individual researchers were categorized based upon the specific research project proposed for the training period. General training grants were put into a separate category separate from the Pipeline. This resulted in the 10 possible categories.

The NHLBI awarded 866 external grants and contracts in 2006. The authors independently agreed on the categorization of 814 (94%) of them. Initial disagreement occurred most often for projects ultimately categorized as human research (eg, epidemiological studies and some health services research). Greater than 99% of the projects could be comfortably placed into a single category. For those that could not, the funding amounts were split evenly between the categories.

Table 2 lists the numbers of external grants and contracts and the amounts of funding awarded in each of the 10 categories. Far more grants and grant money were awarded in the “basic research” category (526 grants for $191,149,404) than in all the other categories combined (343 grants and $119,789,554).

Table 2.

Number of awards and funding amounts by category.

Research category	Number of awards	Funding amounts
Basic research	523	$191,149,404
Phase I translation	21	$2,915,620
Phase I translational research	0	$0
Human research	218	$85,197,890
Phase II translation	18	$2,730,458
Phase II translational research	0	$0
Practice‐ or community‐based research	55	$23,865,539
Phase III translation	7	$1,212,654
Phase III translational research	10	$1,510,630
Training (general)	14	$2,356,763

Funded phase I translational projects consisted primarily of conferences involving basic and human researchers and computer database development. Phase II translational projects included conferences and systematic reviews. Of particular note is that no funding was allocated to T1 or T2 translational research. In contrast, most of the funding for T3 was for research.

Discussion

It was possible to create a set of definitions that corresponds to the categories of research and translation along the expanded NIH Research Pipeline and to the publicly available data about the NIH externally funded research projects. This should make it possible to more clearly communicate about the categories and calculate and track the distribution of research funds that are expended for research and translation at the various points along the Pipeline.

We welcome further discussion on these definitions since these terms are now being used commonly, particularly in connection with the NIH‐sponsored CTSAs. In particular, the early translational steps may require further definition. With clearer definitions, the kind of analysis conducted for the 2006 NHLBI funding could be easily repeated for any of the Institutes and for any year in which funding has been posted. This could then be used to track the changes in funding allocation from institute to institute, or from year to year, which we believe would be an asset to future decisionmakers by providing a way to monitor the progress of the NIH Roadmap.

The expanded conceptual model chosen has not yet been universally accepted. However, those who favor the original NIH model can simply combine the amounts allocated to human research and practice‐ and community‐based research into a single human research category and the amounts allocated to T2 and T3 translation into a single T2 translational category. However, both versions may be oversimplifications of how research actually makes its way into practice, and large areas such as health services research, clinical epidemiology, clinical decision making, medical humanities, and social sciences may not be represented well within either model.

Conclusion

It is possible, using the definitions developed, to categorize a great majority of the research projects funded by the NIH and track changes in this funding over time.