A review of policy dissemination and implementation research funded by the National Institutes of Health, 2007–2014

Background Policy has a tremendous potential to improve population health when informed by research evidence. Such evidence, however, typically plays a suboptimal role in policymaking processes. The field of policy dissemination and implementation research (policy D&I) exists to address this challenge. The purpose of this study was to: (1) determine the extent to which policy D&I was funded by the National Institutes of Health (NIH), (2) identify trends in NIH-funded policy D&I, and (3) describe characteristics of NIH-funded policy D&I projects. Methods The NIH Research Portfolio Online Reporting Tool was used to identify all projects funded through D&I-focused funding announcements. We screened for policy D&I projects by searching project title, abstract, and term fields for mentions of “policy,” “policies,” “law,” “legal,” “legislation,” “ordinance,” “statute,” “regulation,” “regulatory,” “code,” or “rule.” A project was classified as policy D&I if it explicitly proposed to conduct research about the content of a policy, the process through which it was developed, or outcomes it produced. A coding guide was iteratively developed, and all projects were independently coded by two researchers. ClinicalTrials.gov and PubMed were used to obtain additional project information and validate coding decisions. Descriptive statistics—stratified by funding mechanism, Institute, and project characteristics—were produced. Results Between 2007 and 2014, 146 projects were funded through the D&I funding announcements, 12 (8.2 %) of which were policy D&I. Policy D&I funding totaled $16,177,250, equivalent to 10.5 % of all funding through the D&I funding announcements. The proportion of funding for policy D&I projects ranged from 14.6 % in 2007 to 8.0 % in 2012. Policy D&I projects were primarily focused on policy outcomes (66.7 %), implementation (41.7 %), state-level policies (41.7 %), and policies within the USA (83.3 %). Tobacco (33.3 %) and cancer (25.0 %) control were the primary topics of focus. Many projects combined survey (58.3 %) and interview (33.3 %) methods with analysis of archival data sources. Conclusions NIH has made an initial investment in policy D&I research, but the level of support has varied between Institutes. Policy D&I researchers have utilized a variety of designs, methods, and data sources to investigate the development processes, content, and outcomes of public and private policies. Electronic supplementary material The online version of this article (doi:10.1186/s13012-015-0367-1) contains supplementary material, which is available to authorized users.

The field of health policy dissemination and implementation research (policy D&I) exists to address these challenges. Policy D&I is focused on generating knowledge to effectively spread research evidence among policymakers and integrate evidence-based interventions into policy designs [5]. Policy D&I, known as policy "knowledge translation and exchange" outside of the USA, has evolved over the past 40 years. The 1970s and 1980s were marked by interest in how social science research was used in public policymaking and saw the development of key contributions to the field-such as Weiss' typology of research evidence [7] and Caplan's two communities theory of knowledge utilization [17]. In the 1990s and 2000s, frameworks and instruments for policy D&I emerged [21][22][23][24][25] and the field was embraced within the discipline of public health [9,26,27]. Today, policy D&I is a priority of the World Health Organization [28,29] and a major focus of journals such as Health Research Policy and Systems and Evidence & Policy.
The history of policy D&I has demonstrated that government funding is essential for a county to develop and maintain a robust policy D&I research infrastructure [22,30] [6]. This is potentially changing, however, as the National Institutes of Health (NIH)-the largest government funder of health research in the USA-has recently invested in the broader field of D&I research and identified policy D&I as a priority [31].
In 2002, the National Institute of Mental Health issued a Program Announcement (PAR) entitled "Dissemination and Implementation Research in Mental Health" that highlighted the importance of developing effective strategies to disseminate research findings to the policymakers [32]. In 2005, the PAR was re-issued under the broader name "Dissemination and Implementation Research in Health" (D&IRH) with more participating Institutes (e.g., the National Cancer Institute, National Institute on Drug Abuse) [33]. In 2007, the NIH Center for Scientific Review created a permanent D&IRH study section to evaluate D&I proposals, re-affirming NIH's investment in the field [34]. In 2013, the "D&IRH" PAR was re-issued, this time with a greater emphasis on policy D&I research [35]. For example, the PAR encouraged applications that "address[ed] the complexity of bridging research, policy and practice." NIH has expressed interest in policy D&I research, but the extent to which it has funded research in this area is unclear. Two recent studies reviewed NIH D&I funding related to specific topics; one study focused on nursing  [36] and the other focused on cancer control [37]. A similar review of policy D&I research is needed to assess growth of the field and identify funding gaps. Such a study would support growth in the field of policy D&I in the USA by providing investigators with a menu of approaches to policy D&I research that can be adopted and adapted when developing proposals. This article reports the results of a study that was conducted to address these knowledge gaps. The aims of the study were to: (1) determine the extent to which policy D&I research has been funded by NIH between 2007 and 2014, (2) identify trends in NIH-funded policy D&I research, and (3) describe the characteristics of NIH-funded policy D&I research projects.

Methods
We used the NIH Research Portfolio Online Reporting Tool [38] to identify all projects that received funding during any US fiscal year between 2007 and 2014 through the D&I funding opportunity announcements (FOAs) Dissemination and Implementation Research in Health (PARs: 06-039, 06-071, 06-072, 06-520, 06-521, 07-086, 10-038, 10-039, 10-040, 13-054, 13-055,  or Dissemination and Implementation Research in Mental Health (PA 02-131). We selected 2007 as the starting point because it was the first year that projects supported by the Dissemination and Implementation Research in Health FOAs received funding. For each project, we extracted the title, abstract, project terms,  Percentages indicating the proportion of policy D&I within total D&I category (e.g., proportion of R01 projects funded through D&I FOAs that were policy D&I projects, 7/82 = 8.5 %) award amount per fiscal year, funding mechanism, and Institute.
Our review was guided by Bogenschneider's definition of policy as: "the development, enactment, and implementation of a plan or course of action carried out through law, rule, code, or other mechanism in the public or private sector" [5]. Accordingly, we identified potential policy D&I projects by searching the project title, abstract, and term fields for mentions of "policy," "policies," "law," "legal," "legislation," "ordinance," "statute," "regulation," "regulatory," "code," or "rule." Two coders then independently reviewed project abstracts and developed preliminary coding categories to capture project characteristics. These categories reflected themes in that data, categories used in previous NIH D&I funding reviews [36,37], and policy D&I scholarship [3-6, 8, 9]. The coders then jointly developed a coding guide and independently re-reviewed and coded the projects. Because many projects were not exclusively focused on policy D&I, projects were coded according to their policy D&I features. Incongruent coding decisions were identified in <10 % of projects and resolved through discussions.
We classified a project as policy D&I if it explicitly proposed to conduct empirical research about the "content" of a policy (e.g., analysis of the text of clean indoor air laws), the "process" through which it was developed (e.g., assessment of how state legislators use research evidence when developing clean indoor air laws), or the "outcomes" it produced (e.g., evaluation of the impacts of clean indoor air laws on cardiovascular health outcomes). These inclusion criteria were informed by Bogenschneider's definition of policy [6] and domains of health policy research proposed by Brownson and colleagues [3][4][5]. For projects classified as policy D&I, we searched for its NIH project number in ClinicalTrials.gov [39] and PubMed [40] to obtain additional information and validate coding decisions.
We calculated the total dollar amount awarded through all NIH research grants and through the D&I FOAs, stratified by funding mechanism and Institute. We also calculated the amount awarded for policy D&I projects, and the percentage of total D&I FOA funding they comprised, within strata. Data were managed and analyzed in Microsoft Excel.

Results and discussion
Between 2007 and 2014, 146 projects were funded through the D&I FOAs, 12 (8.2 %) of which were classified as policy D&I research (Fig. 1). A total of $16,177,250 was awarded for these projects, equivalent to 10.5 % of all funding through the D&I FOAs ( Table 1). The NIH Office of the Director allocated 69.7 % of its D&I FOA funding to a policy D&I project, indicating agency-level support for the field. The National Cancer Institute (NCI) was the primary funder of policy D&I research, supporting six projects which comprised 13.0 % of all NCI funding through the D&I FOAs. Some Institutes that were engaged in D&I funding-such as the National Institute of Nursing Research which funded seven D&I projects comprising over 1.0 % of the Institute's total research grant spending, a proportion larger than any other Institute-did not fund any policy D&I projects. Differences in policy D&I funding between Institutes could potentially be a reflection of varying levels of knowledge about policy D&I research among investigators within different health science disciplines or differences in the degree to which the disciplines are clinically, as opposed to policy, oriented.  [41,42], were the most common topics of focus. Many projects combined survey (58.3 %) and interview (33.3 %) methods with analysis of archival data sources, such as medical records (25.0 %) and policy documents (25.0 %). It was unclear; however, the extent to which projects used true mixed-method designs (e.g., the results of one method informed data collection of the other, or the results of both methods were presented together).

Limitations
We limited our review to projects funded through FOAs explicitly focused on D&I and may not have identified all NIH-funded policy D&I projects. Our review also did not capture policy D&I projects funded by other US government agencies or philanthropies. We did not attempt to obtain the full text of proposals and thus did not have complete information on project characteristics, such as information on how the projects planned to disseminate their findings. We also did not have data on the characteristics of policy D&I research proposals that were unsuccessful in obtaining funding through NIH. Thus, we were unable to identify specific characteristics that increased the likelihood of a policy D&I project being funded. We dichotomously

Conclusions
NIH has made an initial investment in policy D&I research, signaling that the field might be entering the mainstream US health research enterprise. The level of support for policy D&I research, however, has varied between Institutes and is probably not commensurate with the potential of evidence-based policy to positively impact human health. Policy D&I researchers have utilized a variety of designs, methods, and data sources to investigate the development processes, content, and outcomes of public and private policies. By mapping the characteristics of NIH funded policy D&I research projects, the current study provides