Trends in Funding for Research on Pain: A Report on the National Institutes of Health Grant Awards Over the Years 2003 to 2007

Article Outline

• Abstract
• Materials and Methods
  • Search Procedures
  • Classification of Records
  • Review Process
  • Funding Data
  • Procedural Issues
  • Data Analysis
• Results
  • Organ System Review
  • Nausea and Dyspnea Research
• Discussion
  • Limitations
• Conclusions
• Acknowledgements
• Appendix A
• References
• Copyright

Abstract 

In recent years, the National Institutes of Health (NIH) has experienced unprecedented reductions in its customary annual budget increases. Consequently, researchers, health care policy planners and others have a pressing need for accurate information on NIH funding patterns. We created a unique and objective system for compiling, classifying, and analyzing data on NIH grant awards and funding for research on pain, nausea, and dyspnea using naïve observers, cross-validation by multiple raters, and face validation by experts. We present results of our method and analyses for the period from 2003 to 2007. Following a 12% increase from 2003 to 2004, funding for pain research fell by 9.4% per year on average over the next 3 years. The percent of the total NIH budget going to support pain research increased to 0.78% in 2004 but fell to 0.61% in 2007. A piecewise regression model confirmed the declining trend represented a significant fit to the data (R²=0.98, p=0.024). Separate breakdowns by Institutes showed similar patterns. Analyses of nausea and dyspnea research support revealed small but steady increases over the same period. Declining support for pain research disproportionate to decreases in the NIH budget signals a need for measures to promote funding for meritorious applications.

Perspective

Results of 5 year trends in numbers of grants and funding for research in pain, nausea, and dyspnea by the NIH show overall declines for pain but slight increases for nausea and dyspnea. Declining support for pain research that exceeds the reductions in the total NIH budget signals a need for measures to increase pain research funding.

Key words: Pain, research funding, nausea, dyspnea, federal grants, health policy

The number of grants and the amount of funding awarded by the National Institutes of Health (NIH) for research in a particular area gauges the success of initiatives advocating investigation in that area. As efforts to make well-informed decisions about health care resource allocation and public policy gain momentum, the need for reliable data resources intensifies. Clinicians, researchers, professional societies, patient advocacy groups, and public policy experts have pressing needs for information that accurately tracks NIH research funding patterns.

The NIH provides about one third of all funding for biomedical research in the United States.¹⁰ Unfortunately, no reliable and objective means for tracking NIH research funding patterns in specific disease areas exist. Although the NIH releases summary information about funding allocations to different disease entities and conditions, it does not disclose the process it uses to categorize grant activities according to these distinctions.

The level of funding in 2003 has special significance as it represents the end of a 5-year congressionally mandated initiative to double NIH funding. The 5-year period beginning in 1999 and ending in 2003 saw annual funding increases of between 14 and 16%, approaching this target.⁶ Annual appropriations have slowed in subsequent years and in fact have failed to keep pace with the Biomedical Research and Development Price Index (BRDPI), the estimated rate of inflation in science-related research costs.⁶, ⁷ The NIH budget for 2008 was projected at 12% lower than its peak in 2004, when adjusted for inflation using the BRDPI.¹ A flat budget is proposed for 2009, giving the NIH the same amount for research and development as it received in 2008.² The present funding cycle marks the end of a 22-year period of average annual increases in research funding of about 3.34% per year, and represents an unprecedented recession in funding for biomedical research.⁵

With decreased funding for research and continuing pressures on resources to support national security and military efforts, major natural disasters, and uncertain economic status, competition for limited research funds will intensify.⁸ The ability to track funding patterns becomes increasingly important for policy-making decisions. Many NIH Institutes have committed program experts whose task it is to encourage pain research but they have lacked a rigorously designed, sensitive tool to help them evaluate funding patterns.

In an earlier report,³ we described the development of a tool for tracking patterns in NIH awards for research in pain, and reported results obtained using this tool for the year 2003. A key finding of that report was that less than 1% of all NIH funding in 2003 was allocated to research having a primary emphasis on pain.

We now report the results of an analysis of new data obtained for the years 2004 through 2007 and identify trends in NIH grant awards and funding for pain over the 5-year period from 2003 to 2007.

Back to Article Outline

Materials and Methods 

We developed a systematic method for obtaining data on grant awards made by the NIH and evaluating their relevance to advancing the field of pain research. In 2005 we published the methods used to develop these procedures and reported on findings of an application of this approach in an analysis of pain research funding in 2003.³ As a comparison and demonstration of how these methods might be extended to other domains, we also applied this approach to the areas of nausea and dyspnea research. The goal of the method was to provide an objective and verifiable classification tool for measuring grant awards and funding patterns. Using the Computer Retrieval of Information on Scientific Projects (CRISP) developed and maintained by the NIH as our primary data source for grant information,¹⁴ we developed a method to extract records having potential relevance to the research areas of interest and compiled a database containing detailed information on each grant, including project title, abstract and keywords, that could then be queried and reviewed record by record. The method comprised a system of search terms, classification rules, and review procedures that we developed under consultation with experts in the fields of basic and clinical pain research, practice and policy. After developing the initial set of search terms and classification rules, we tested the method on the 2003 grant portfolio and submitted the results to examination by the experts. The recommended revisions were implemented and submitted to 2 more rounds of revision to ensure that the process was sufficiently inclusive of all grants considered by the experts to be “pain research.”

Search Procedures 

We executed an inclusive search query to CRISP specifying fiscal year without restriction as to grant mechanism. Drawing on a list of pain terminology provided by the International Association for the Study of Pain (IASP) and experts in pain research and pain management, we developed an extensive list of 53 search terms for querying CRISP to cast as wide a net as possible and thereby minimize errors of omission. Search terms included pain-specific words (e.g., pain, nociceptor, hyperalgesia), terms describing disease conditions having pain as a significant or identifying feature (e.g., fibromyalgia, migraine), and conditions having pain as 1 of several important symptoms (e.g., osteoarthritis, angina). A consequence of using such a comprehensive search list was that we netted numerous false positive records found by review to be irrelevant unless co-appearing with another term (e.g., arthritis pain). A careful inspection of these records revealed over 300 false positives could be avoided by eliminating some terms. Our review also revealed that interstitial cystitis frequently failed appropriate classification as relevant to pain. Adding interstitial cystitis to the list of pain condition terms rectified this omission. Table 1 details the revised search term list used for the analyses described in this report.

Table 1. Pain Terms and Conditions Used to Query the CRISP Database

Classification of Records 

The review process consisted of reading the project name, abstract, and keywords (defined as the Text Object) and classifying each record according to a set of decision criteria (Appendix A). We constructed these criteria to allow the description of the project's specific aims, goals, or hypotheses to stand as the true measure of the relevance of the grant to the area of research interest. If the abstract for the project used 1 of the search terms specific to pain or used a pain condition term specifically to describe the purpose of the project, then the grant qualified as relevant and was classified “primary pain.” If the abstract applied these terms only generally or in combination with other symptom or condition terms, then the grant was considered marginally relevant and we labeled it “secondary pain.” If the grant failed both conditions, it was classified not relevant, or “non-pain.” For example, the statement “This study will determine: (1) whether pre-operative psychophysical testing to evaluate pain responses is useful in predicting acute and longer-term pain outcomes after lumpectomy” would be classified primary pain. The statement “Neurogenic inflammation is a well established principle in some inflammatory conditions where neurotransmitters such as substance P are key mediators of pain sensation, neutrophil infiltration, edema, and local proteolysis” would be considered secondary pain. The statement “Patient distress associated with symptoms such as pain, nausea, dyspnea, and even to a degree depression and anxiety, have been the focus of some targeted research. Less is known about how to address sources of distress outside of these traditional domains” would receive a non-pain classification. We also identified each record as basic science or clinical research according to the NIH definition¹³ and categorized the research according to organ system or disease entity. Since classification depends on the review of the abstract, if an abstract was unavailable or was not unique to the record (i.e., duplicates another record as in the case of supplemental grants), then we considered the record unclassifiable.

Review Process 

We conducted separate searches and reviews of grant award activity for each year of interest and compiled a dataset consisting of all unique records obtained for the year under review. The authors of the 2005 article conducted the review of 2003 data. For each subsequent review, we selected 2 independent reviewers having no prior familiarity with the objectives of the review. After receiving training on the review process and the classification criteria, the reviewers separately examined each record in the dataset. On review completion, the first author conducted a second review to identify and correct any obvious rule misapplications. Reviewers discussed and resolved any remaining classification disputes for 100% agreement.

Funding Data 

To evaluate funding patterns, we obtained funding data on individual grants, total budget information for each Institute, and NIH total budget for fiscal years 2003 through 2007 from multiple publicly accessible NIH web resources. The NIH Office of Extramural Research provides data on individual grant funding in separate files organized by state, US possessions and territories, and foreign sites.¹⁶ The NIH Office of Budget website provides data on total appropriations for each Institute and NIH total actual obligations for each year.¹¹ To provide year to year comparisons, amounts were corrected to constant dollars using the Office of Management and Budget Gross Domestic Product Price Index.¹² This index provides a more conservative adjustment for inflation than the BRDPI.⁸ Because the funding data obtained to prepare funding analyses for the 2005 report were incomplete, we reacquired these data and performed a new analysis for the present report.

Procedural Issues 

Our initial review of 2003 data was completed in 2004. Subsequently, we performed annual updates, adding consecutive years' data 1 year at a time. While preparing data for this report, we discovered a complication with our method for querying CRISP and how CRISP performs its record keeping. CRISP provides a data field labeled Fiscal Year that corresponds to the fiscal year (defined as beginning October 1 and ending September 30) in which a grant is awarded. When a new grant record enters CRISP, the fiscal year field is assigned the value of the fiscal year it was awarded. For multi-year grants, continuation awards trigger the addition of a new record to CRISP with the new fiscal year value assigned. By performing a set of queries to CRISP in which we included the entire 5-year period of interest as well as the prior year (fiscal years 2002 to 2007), we discovered many records having grant start and end dates in our period of interest but that escaped detection in our previous CRISP searches. This misstep appeared to result from a failure to update these records with new fiscal year assignments. As a result, records obtained for the 2005 report under-represented the actual numbers. The values for 2003 reported here use these corrected data.

Data Analysis 

We present descriptive data and 5-year trend analyses on numbers of grants and award amounts. As these data represent exact values, population based statistical analysis methods do not apply and can not be used to evaluate such things as random versus systematic variation. However, procedures to model the data can be useful for identifying potentially meaningful and informative trends. Such methods do allow for statistical tests to interpret how well a model fits the data (e.g., likelihood ratios and coefficients of variation), and tests of whether the model fits significantly better than chance (i.e., p values). Post hoc inspection of descriptive counts and funding data revealed an apparent increase from 2003 to 2004 and a consistent decline thereafter. A piecewise regression model, sometimes called a “hockey stick model”,⁷ with a pivot point at 2004 can identify trends in data having characteristics of this sort. The model provides 1 intercept and slope for initial change, here termed the gain, and another for subsequent increases or decreases, here designated the declination. We used linear regression maximum likelihood estimates¹⁵ to provide separate tests of goodness of fit and significance for the overall model and separately for gain and declination. For counts data we also applied generalized linear models with Poisson distributions that provide Chi-square based tests for goodness of fit for the overall model and gain and declination parameters.

Back to Article Outline

Results 

We downloaded all records resulting from querying CRISP and performed separate reviews for each additional year of data (Table 2). Records requiring review varied from a high of 2705 for 2003 data, the first year of review in which all records were new, to a low of 354 for 2006 data. Reviewer agreement varied from a high of 93.8% to a low of 85.8%. After correcting for rule misapplication, reviewer agreement increased to between 91.4% and 99.8%. A second review that included 2002 records with fiscal year verified by grant start and end dates netted an additional 306 grants not included in the 2005 report. These records were added to the database and subjected to ad hoc review.

Table 2. Results of the Review Process for NIH Grants from 2003 Through 2007.

We reviewed data for total pain research grants as well as comparisons by research type (basic science and clinical), Institute, and as a percentage of the total NIH budget. We present descriptive information on numbers of pain grants and dollar amounts awarded for each of the 5 consecutive years from 2003 to 2007. We also present trends in the numbers of grants and dollar amounts as predicted by the piecewise regression model. The descriptive data almost uniformly indicated that numbers of grants and dollar amounts increased from 2003 to 2004 and decreased from 2004 to 2007 for all comparisons. Statistical tests evaluated how well this model fit the actual data for each comparison of interest. To simplify this report, we only report data on research having a primary focus on mechanisms and treatment of pain, i.e., records classified primary pain.

The number of grants awarded for primary pain research varied from a high of 642 in 2004 to a low of 586 in 2007 (Table 3). Clinical research grants outnumbered basic science in all years but 2003. Piecewise regression models for the numbers of basic science, clinical, and total pain grants provided significant fits to the data (Table 4). Gain and declination parameters for basic science, clinical, and total pain research were all significant with generally larger standard errors of estimate (S.E.) for gain than for declination. Fig 1 displays the trends predicted by the model as well as the actual number of grants for basic science, clinical, and total pain research. Trend lines depict significant increases from 2003 to 2004 and decreases from 2004 to 2007 for basic science, clinical research, and the combination. Standardized coefficients (β) for the gain and declination for basic science, clinical research, and total pain were all significant and in the expected direction.

Table 3. Number of Grants for Primary Pain Research by Year and Research Type

Year	Basic Science	Clinical	Total
2003	325(50.6%)	317(49.4%)	642
2004	358(49.0%)	373(51.0%)	731
2005	353(49.7%)	357(50.3%)	710
2006	280(45.8%)	331(54.2%)	611
2007	283(48.3%)	303(51.7%)	586

Table 4. Test of the Piecewise Regression Model for Change in Numbers of Grants and Dollar Amounts

Modeled Change in Numbers of Grants
Research Type	Model Fit (Likelihood Ratio)	Model Fit (p value)	Gain β	Gainp (S.E.)	Declination β	Declinationp (S.E.)
Basic science	8.29	0.016	0.46	0.043(18.92)	−1.07	0.000(6.49)
Clinical research	22.14	0.000	−1.05	0.000(3.65)	0.70	0.000(1.25)
Total pain research	13.02	0.001	0.93	0.000(19.85)	−1.11	0.000(6.81)

Modeled Change in Amount of Funding
Research Type	Model Fit (R2)	Model Fit (pvalue)	Gain β	Gainp (S.E.)	Declination β	Declinationp (S.E.)
Basic science	0.92	0.077	0.27	0.364(7.52)	−1.07	0.043(2.58)
Clinical research	0.99	0.011	0.80	0.011(2.2)	−1.13	0.006(0.75)
Total pain research	0.98	0.025	0.53	0.057(7.4)	−1.15	0.013(2.52)

• 
  • View Large Image
  • Download to PowerPoint

• Figure 1. 

  Number of grants for pain research by research type.

In the 2005 report matches between grants and funding data were achieved for only 87% of the 2003 records. For the present report, grant records and funding data matched on average 98% over the 5 years, varying from 96% in 2003 to 99% in 2007 (Table 5). Dollar amounts awarded for pain research showed patterns parallel to that for numbers of grants, with increases from 2003 to 2004 and declines from 2004 to 2007 (Fig 2). Piecewise regression models for the annual change in dollar amounts fit well for clinical and total pain grants but not for basic science (Table 4). Gain parameters were in the expected direction but did not reach significance for basic science. However, declination parameters were significant and in the expected direction. For clinical and total pain research, declination coefficients were slightly larger for dollar amounts than for numbers of grants, suggesting that from 2004 to 2007 dollar amounts declined faster than the number of grants awarded.

Table 5. Funding for Pain Research by Year and Research Type¹

• 
  • View Large Image
  • Download to PowerPoint

• Figure 2. 

  Dollar amounts for pain research by research type.

We obtained NIH data on actual obligations for the years 2003 to 2007 that include funding for each Institute or Center (IC), bridge funds, the Roadmap initiative, and building and facilities.¹¹ In 2008 constant dollars,¹⁷ NIH total funding increased from $30.6 billion in 2003 to $31.4 billion in 2004, an increase of 2.4%, but decreased each year thereafter by an average of 1.9% per year. In contrast, awards for pain research increased from 2003 to 2004 by 12.2% but decreased thereafter by 9.4% per year on average (Fig 3).

• 
  • View Large Image
  • Download to PowerPoint

• Figure 3. 

  Annual percent change in dollar amounts for pain research and for the total NIH budget.

An examination of the percent of total funding going to pain research provides a view of funding patterns independent of the effects of fluctuations in total budget appropriations. Funding for pain research represented 0.71% of the total NIH budget in 2003, increasing to 0.78% in 2004, but decreased steadily thereafter to a low of 0.61% in 2007 (Fig 4). The piecewise regression model having an initial increase followed by a steady decline thereafter provided a significant fit to these data (R²=0.98, p=0.024), whereas a model having an increase followed by a flat trajectory (i.e., no downward trend) did not (R²=0.004, p>0.05).

• 
  • View Large Image
  • Download to PowerPoint

• Figure 4. 

  Dollar amounts for pain research as a percent of the total NIH budget.

In the 2005 report we presented details on number of grants and funding amounts for pain research by each IC. We provide these data in Supplemental Material A, Supplemental Material B. Here we present predicted trends in funding by IC. Applying a piecewise regression model with parameters for gain and declination, we modeled annual change in funding for each IC. Fig 5 displays modeled trends in funding for pain research for the 12 ICs having the highest award amounts for pain in 2003. The overall model fit 8 of these ICs with marginal or better significance (R²=0.96 to 0.924; p=0.039 to 0.076) attributed entirely to fits of the declination parameters (β=−1.134 to −1.031; p=0.020 to 0.042). An additional 2 ICs had marginally significant declination parameters (β=−1.067 and −1.008; p=0.075 and 0.078). Estimates of the gain parameters failed to fit the data for any of the ICs. The results of the model fits indicate pain research funding by IC did not increase significantly from 2003 to 2004, although the combined pattern shows a significant increase. Significant or marginal decreases in awards for pain research from 2004 to 2007 were found for 10 of the top 12 ICs funding pain research, with the National Institute of Dental and Craniofacial Research (NIDCR) and the National Institute of Child Health and Human Development (NICHD) the exceptions.

Supplemental Material A. Numbers of Grants by Institute and Research Type

Supplemental Material B. Dollar Amounts and Percent of Total Budget by Institute and Research Type

• 
  • View Large Image
  • Download to PowerPoint

• Figure 5. 

  Trends in dollar amounts for pain research by institute.

Fig 6 shows the percent of total funding by individual ICs going to pain research. The pattern of predicted change in percent of funding closely matches that predicted for dollar amounts. The piecewise regression model fit the same 8 top ICs funding pain research, with declination accounting for model fits while none of the gain parameters fit the data. Significant or marginal declination β coefficients were found for the same 10 ICs, ranging from −1.136 (p=0.025) to −1.000 (p=0.092) with NIDCR and NICHD again the exceptions.

• 
  • View Large Image
  • Download to PowerPoint

• Figure 6. 

  Trends in dollar amounts for pain research as a percent of institute budget by institute. Percents are displayed on a log scale due to the wide range in values.

Organ System Review 

The review of grant records included identification of the research's focus on specific diseases or organ systems. Multisomatoform pain conditions were defined as pain conditions having no identifiable or significant underlying physical cause and which include profound psychosocial impairment. Included among these conditions are fibromyalgia and irritable bowel syndromes. Neurological research remained the most researched pain area over all 5 years with the number of grants increasing from 70 in 2003 to 119 in 2007 and dollar amounts increasing from $21.1 to $30.5 million. Multisomatoform pain research also increased over the 5-year span while all others decreased or remained at the same level. Both had much larger increases in basic science than in clinical research (Table 6A and B). Neurological basic science research increased by 51% ($17 to $25.7 million) compared to a 17% increase in clinical research. Multisomatoform basic science research increased even more dramatically from $2.6 to $5.5 million, a 112% increase, compared to clinical research, a 26% increase. By contrast, award amounts for cancer pain research grants decreased by 38% over the 5-year period from $15.5 to $9.6 million (40 to 29 grants), with a much sharper decline in basic science ($4.2 to $0.9 million) than in clinical research ($11.3 to $8.8 million). Cardiac pain was the least researched over all 5 years, with the total number of grants decreasing from 7 to 2 and dollar amounts shrinking from $2.1 to $0.39 million.

Table 6A. Number of Grants for Pain Research by Disease or Organ System

Table 6B. Dollar Amounts for Pain Research by Organ System¹

Nausea and Dyspnea Research 

As with the review of pain research, we re-queried the CRISP database and included years 2002 to 2007. The new search did not uncover any records missed during earlier searches. Reviews of all records for primary research for nausea and dyspnea revealed that, unlike the pattern found for pain, research on both nausea and dyspnea increased steadily over the 5-year period from 2003 to 2007. Numbers of dyspnea grants increased from 13 to 21 with dollar amounts increasing from $3.19 to $6.67 (Table 7A). Nausea grants increased from 17 to 32 and dollar amounts increased from $3.55 to $7.57 million (Table 7B).

Table 7A. NIH Grants for Dyspnea Research from 2003 to 2007

Table 7B. NIH Grants for Nausea Research from 2003 to 2007

Back to Article Outline

Discussion 

This review revealed: 1) an apparent increase in NIH support for pain research from 2003 to 2004; and 2) a decline in support over the subsequent 3 years. Tests of the piecewise regression model confirmed both trends as significant. A further evaluation of this model comparing it with a model having no change after 2004 provided confirmation of significant declines in awards during this period. As Fig 1 and Fig 2Fig show, predicted trends follow actual values tightly, revealing the “hockey stick” characteristic of gains early followed by lengthy declines thereafter. Both basic science and clinical pain research displayed this pattern. The pattern also provided a good fit to the data for a majority of individual ICs (Figs 5 and Figs 6). The increase in 2004 follows an increase in the total budget for NIH even as the mandate to double NIH funding by 2003 ended. This may represent a carryover effect for pain research as the increase in grant numbers and awards outpaced the NIH overall. However, the declines in pain research thereafter far exceeded decreases in the total NIH budget (Fig 3). The finding of changes in the percent of the total NIH budget going to pain research cannot be explained by changes in NIH budget because we could assume such changes would affect all research areas equally, leaving the percent of the total budget unchanged. As Fig 4 reveals, this has not been the case for pain research. By contrast, nausea and dyspnea research saw slow but steady increases over the 5 years, confirming that the decline in pain awards is not an artifact of an overall decline in NIH funding affecting all research areas equally.

The meager 0.6% investment in pain research by the NIH seems seriously out of scale with the impact of pain on the nation's health care burden. A new review by Max and Stewart shows that pain accounts for over 20% of all doctor visits, 10% of drug sales, and costs to developed countries of $1 trillion each year.⁹

The review process for this report differed from the 2005 report in 1 important respect: Reviewers in this report were naïve to the purposes and processes involved in developing the review protocol, thus providing a test of the objectivity of the review process. Agreement between reviewers indicates how sensitive the review process is to subjective interpretation. Compared to the data review described in the 2005 report, which achieved 93.8% agreement, subsequent reviews faired less well. This result seems unsurprising as reviewers were not as familiar with the criteria for decision making. Following the secondary review, which attempted to determine if disagreements resulted from misinterpretation or misapplication of rules, agreement increased substantially. Post hoc inspection of disagreements and interviews with reviewers indicated that most conflicts occurred as a result of making fine distinctions within categories (e.g., subcategories for primary pain) rather than large distinctions between categories (e.g., primary pain versus non-pain). Simplifying these distinctions will allow better agreement and improve the objectivity and efficiency of the process in future reviews.

Comparing basic science and clinical research activity over the 5-year period from 2003 to 2007 revealed a growing disparity in the number of basic science and clinical pain research grants (Table 3). In 2003 basic science grants outnumbered clinical slightly but began falling behind in 2004 and continued to trail in 2007. This disparity also appeared in dollar amounts (Table 4). The imbalance between clinical and basic pain research deviates significantly from the reported state of the NIH as a whole. According to NIH Director Zerhouni, in 2006 basic science comprised 55% of the NIH budget, compared to 40% for applied research including clinical.¹⁸ Observers of research funding have voiced concerns regarding NIH commitment to behavioral science research.⁴ Our findings suggest that basic science research is on the increase for neurological and multisomatoform pain but clinical pain research has not kept up. By contrast, cancer pain research has declined more rapidly for basic science than for clinical pain research.

Reports have suggested that the NIH has responded to the reductions in annual budget increases since 2003 by limiting the size of awards to both new and continuing projects in order to spread fewer dollars over more grants.⁸ Our findings support this view. Comparison of declination slopes revealed that for both basic science and clinical pain research dollar amounts have decreased more rapidly than have numbers of grants.

Analysis of trends in number of grants and awards by IC revealed declines for all but 2 of the top 12 ICs supporting pain research (Fig 5). The analysis of the percent of the total budget of each IC going to pain research provides an indication of the priority each IC has given to pain research compared to other areas of biomedical research. As shown in Fig 6, we found that only 2 of the top 12 ICs had maintained their support for pain research at a more or less constant proportion of their budget over the 5-year period.

Limitations 

Our finding of disparities in how records are maintained and updated in the CRISP database raises concerns regarding the reliability of basing studies such as ours on CRISP as the sole data source. We found that with each consecutive year of data obtained, fewer record errors occurred. Subsequent reviews should be able to avoid these issues as long as reviews are made of consecutive years. The 2005 report experienced issues with obtaining complete funding data. The present report faired much better, with from 96% to 99% complete data. The NIH has recently begun to roll out a new web-based initiative to provide reports on NIH budgets, grants, and awards in support of a wide range of research interests including pain.¹⁵ Perhaps this action represents a trend toward providing more transparency for NIH activities and allowing health care researchers and policy makers easier access to information on NIH funding patterns. Nevertheless, our experience suggests that a peer-reviewed algorithm for evaluating and classifying research according to specific criteria is necessary for providing an objective and verifiable means of tracking patterns in NIH funding and with sufficient sensitivity to detect meaningful variations in year-to-year changes.

Back to Article Outline

Conclusions 

The NIH has seen unprecedented decreases in funding in recent years. This report found that reductions in pain research outpaced that of the overall NIH budget. Our findings suggest that general changes to the budget of NIH or to individual Institutes will not in themselves lead to increased support for pain research. Additional measures need to be taken by NIH to improve the chances of funding for meritorious applications proposing research on pain. Our tool for evaluating trends in NIH funding was able to identify this concern and should help inform decisions by policy makers both within the NIH and those working to influence funding policy.

Back to Article Outline

Acknowledgements 

Support for this work was provided by a grant from the American Pain Society to authors Bradshaw and Chapman.

The authors gratefully acknowledge the inspiration and guidance that the late Mitchell B. Max provided from the inception of this work to its completion. His enthusiasm and commitment to moving the pain field forward through research will be sorely missed.

Back to Article Outline

Appendix A 

Appendix A. Pain Research Classification Decision Rules

Back to Article Outline

References 

1. AAAS. AAAS R&D Funding Update on R&D in the FY 2008 NIH / HHS Budget. Accessed September 2, 2008 http://www.aaas.org/spp/rd/nih08p.htm2007;
   • View In Article
2. AAAS. AAAS Analysis of R&D in the FY 2009 Budget. Accessed September 2, 2008 http://www.aaas.org/spp/rd/prev09p.htm2008;
   • View In Article
3. Bradshaw DH, Nakamura Y, Chapman CR. National Institutes of Health grant awards for pain, nausea, and dyspnea research: An assessment of funding patterns in 2003. J Pain. 2005;6:277–293
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (168 KB)
   • CrossRef
4. Carstensen LL, Fiske ST, Greenough WT, Horowitz FD, McEwen BS, Takeuchi D, Waite LJ: An analysis of the NIH Basic Behavioral and Social Sciences Research Report submitted to the House and Senate committees on appropriations May, 2006.
   • View In Article
5. Heinig SJ, Krakower JY, Dickler HB, Korn D. Sustaining the engine of U.S. biomedical discovery. N Engl J Med. 2007;357:1042–1047
   • View In Article
   • CrossRef
6. Korn D, Rich RR, Garrison HH, Golub SH, Hendrix MJ, Heinig SJ, et al. Science policy (The NIH budget in the “postdoubling” era). Science. 2002;296:1401–1402
   • View In Article
   • CrossRef
7. Littell RC, Milliken GA, Stroup WW, Wolfinger RD. SAS System for Mixed Models. Cary, NC: SAS Institute; 1996;
   • View In Article
8. Loscalzo J. The NIH budget and the future of biomedical research. N Engl J Med. 2006;354:1665–1667
   • View In Article
   • CrossRef
9. Max MB, Stewart WF. The molecular epidemiology of pain: A new discipline for drug discovery. Nat Rev Drug Discov. 2008;7:647–658
   • View In Article
   • CrossRef
10. Moses H, Dorsey ER, Matheson DH, Thier SO. Financial anatomy of biomedical research. Jama. 2005;294:1333–1342
    • View In Article
    • CrossRef
11. NIH. History of total obligations by IC. Accessed July 15, 2008 http://officeofbudget.od.nih.gov/ui/SpendingHistory.htm
    • View In Article
12. NIH. National Institutes of Health Price Indexes. Accessed July 15, 2008 http://officeofbudget.od.nih.gov/UI/GDP_FromGenBudget.htm
    • View In Article
13. NIH. NIH Office of Extramural Research: Glossary and Acronym List. Accessed July 15, 2008 http://grants.nih.gov/Grants/glossary.htm
    • View In Article
14. NIH. CRISP: Computer Retrieval of Information on Scientific Projects. Accessed July 15, 2008 http://crisp.cit.nih.gov/
    • View In Article
15. NIH. Research Portfolio Online Reporting Tool (RePORT): Reports, Data and Analyses of NIH Research and Development Activities. Accessed August 8, 2008 http://report.nih.gov/award/award.cfm2008;
    • View In Article
16. Office of Extramural Research NIoH. NIH Extramural Awards by State and Foreign Site. Accessed October 28, 2007 http://grants.nih.gov/grants/award/awardtr.htm
    • View In Article
17. OMB. [cited 2008 7/15/2008] http://officeofbudget.od.nih.gov/UI/GDP_FromGenBudget.htmAccessed July 15, 2008
    • View In Article
18. Zerhouni EA. Research funding (NIH in the post-doubling era: realities and strategies). Science. 2006;314:1088–1090
    • View In Article
    • CrossRef