Determinants of Enrollment in Cancer Clinical Trials: The Relationship Between the Current State of Knowledge, Societal Disease Burden, and Randomized Clinical Trial Enrollment

Whether clinical cancer research currently focuses on gaps in the evidentiary basis for clinical guidelines and/or on cancers that impose greater societal burden is unclear. This study assessed the relationship between cancer research efforts in terms of planned randomized controlled trial (RCT) enrollment, objective measures of evidence quality, and a cancer’s burden on society. The authors calculated the planned RCT enrollment listed on ClinicalTrials.gov for the 17 most prevalent solid cancers. Using cancer type as the unit of analysis, linear regression was used to examine the association between planned enrollment in RCTs and 1) evidence quality, as measured by the absolute number and percent of highest quality category (category 1 [C1]) recommendations in the NCCN Clinical Practice Guidelines in Oncology for each cancer, and 2) measures of burden on society, including prevalence, incidence, person-years of life lost (PYLL), and disability-adjusted life years (DALY). Non-normal distributions were log transformed when appropriate. Overall, 15% of the NCCN recommendations were based on the highest quality evidence. Results produced 1260 RCTs. Planned RCT enrollment ranged from 2270 (testis) to 492,876 (breast) and was correlated neither with absolute number nor percent of C1 recommendations for that cancer. Planned RCT enrollment was positively correlated with a cancer’s prevalence (P=.01), incidence (P<.01), PYLL (P<.01), and DALY (P<0.01). In multivariate analysis, prevalence (P<.01) and PYLL (P<.01) had the strongest association with planned RCT enrollment. Findings showed, therefore, that planned cancer RCT enrollment is associated with higher societal disease burden, not the quality of a cancer’s clinical guidelines.

Abstract

Whether clinical cancer research currently focuses on gaps in the evidentiary basis for clinical guidelines and/or on cancers that impose greater societal burden is unclear. This study assessed the relationship between cancer research efforts in terms of planned randomized controlled trial (RCT) enrollment, objective measures of evidence quality, and a cancer’s burden on society. The authors calculated the planned RCT enrollment listed on ClinicalTrials.gov for the 17 most prevalent solid cancers. Using cancer type as the unit of analysis, linear regression was used to examine the association between planned enrollment in RCTs and 1) evidence quality, as measured by the absolute number and percent of highest quality category (category 1 [C1]) recommendations in the NCCN Clinical Practice Guidelines in Oncology for each cancer, and 2) measures of burden on society, including prevalence, incidence, person-years of life lost (PYLL), and disability-adjusted life years (DALY). Non-normal distributions were log transformed when appropriate. Overall, 15% of the NCCN recommendations were based on the highest quality evidence. Results produced 1260 RCTs. Planned RCT enrollment ranged from 2270 (testis) to 492,876 (breast) and was correlated neither with absolute number nor percent of C1 recommendations for that cancer. Planned RCT enrollment was positively correlated with a cancer’s prevalence (P=.01), incidence (P<.01), PYLL (P<.01), and DALY (P<0.01). In multivariate analysis, prevalence (P<.01) and PYLL (P<.01) had the strongest association with planned RCT enrollment. Findings showed, therefore, that planned cancer RCT enrollment is associated with higher societal disease burden, not the quality of a cancer’s clinical guidelines.

Given the finite resources for cancer research, understanding the focus of current clinical research efforts and the factors influencing these efforts is important.1 Ideally, research efforts should be focused on maximizing the potential for public good.2 Understanding whether research resources are applied in relation to the quality of current clinical evidence or to societal disease burden is a key step toward their rational distribution.

One measure of the quality of current clinical evidence for a particular cancer type is the quality of the evidentiary base of clinical practice guidelines promulgated by professional societies such as NCCN. A recent analysis found substantial gaps in the quality of evidence guiding clinical decisions for numerous cancer types, because most recommendations in the NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines) for the 10 most common cancers are based on evidence rated lower than category 1 (C1).3 Furthermore, quality of evidence, as determined by the proportion of C1 recommendations, varies widely by cancer type. It is unclear whether cancer research efforts are focused in areas of lower evidence quality, in which greater opportunity may exist for discovery, or in areas of greater evidence quality, in which past successes raise additional research questions that are ripe for further study.

In addition to current evidence quality, societal disease burden should ideally guide research efforts.4 Although NIH funding seems to be allocated to diseases that pose a greater burden,5 the relationship between this burden and projected enrollment in cancer randomized clinical trials (RCTs) has never been formally studied. How RCT enrollment from all funders is distributed with respect to societal disease burden and the current quality of evidence is unknown. Therefore, to assess how current cancer research efforts are prioritized, the authors investigated whether planned enrollment in current clinical trials in oncology is distributed among cancer types proportionate to evidence quality and societal burden. Assessment of the current state of clinical trial efforts with regard to these 2 factors is the first step toward optimal allocation of the finite resources available for cancer research.

Materials and Methods

Measurement of Current Research Effort

RCTs are widely seen as the gold standard for determining the comparative effectiveness of cancer treatments.6 One measure of research effort is the planned enrollment of active phase III RCTs. The authors measured planned RCT enrollment rather than actual enrollment of completed RCTs to approximate current clinical research effort. To determine the number of patients planned to enroll in all RCTs, they queried ClinicalTrials.gov on May 1, 2011, for active, interventional, phase III RCTs for each of the 17 most prevalent solid tumors. ClinicalTrials.gov is a repository of clinical trials mandated by the FDA Modernization Act of 1997 as a registry of clinical trials investigating experimental treatments of serious health conditions.7 Both federal and private organizations that sponsor and implement clinical studies are responsible for ensuring that accurate information about their studies is submitted to ClinicalTrials.gov.8

For each cancer type, the authors included RCTs of therapies focused on one tumor type exclusively and RCTs of therapies focused on the cancer type of interest when included in a group of closely related malignancies. Because they focused on treatment, trials of diagnosis, screening, prevention, and health services research were excluded. They limited the group to only those trials focused on biological end points (tumor size, recurrence, and survival), excluding those focusing on mood, sexual function, or communication. The total number of planned enrollees were tallied for the RCTs of each cancer. If the total planned enrollment was not listed, as was the case for 110 of 1260 (8.7%) of the trials listed, the average enrollment for trials of that cancer type was imputed. Per cancer type, total planned enrollment was missing most often for pancreas cancer trials (13 of 40; 33%), and least often for testicular cancer trials (0 of 6; 0%). Although imputing the average values introduces a level of uncertainty, leaving these trials out could have falsely underestimated the amount of research occurring in this cancer type.

Assessment of Current Clinical Evidence Quality

The NCCN Guidelines provide a set of recommendations for how best to treat various cancers based on the most current research. Each specific recommendation in the NCCN Guidelines is ascribed a category indicating the quality of evidence available. C1 recommendations are based on high-level evidence with uniform panel consensus that the intervention is appropriate. Category 2a (C2a) recommendations are based on lower-level evidence with uniform panel consensus that the intervention is appropriate, whereas category 2b (C2b) recommendations are based on lower-level evidence with nonuniform panel consensus that the intervention is appropriate. Category 3 (C3) recommendations are based on any level of evidence and about which major panel disagreement exists that the intervention is appropriate.9 The authors examined the NCCN Guidelines for the 17 most prevalent solid tumors (bladder, breast, cervix, central nervous system, colorectal, endometrial, esophagus, hepatobiliary, kidney, lung, melanoma, oral cavity and pharynx, ovarian, pancreas, prostate, stomach, and testis cancers). When multiple options were available in a single clinical scenario, only the recommendation in the highest category of evidence and consensus was counted.

Construction of Other Variables

The authors quantified a cancer type’s societal burden using the most current estimates of its incidence, prevalence, person-years of life lost (PYLL), and disability-adjusted life years (DALY). They obtained incidence estimates from the American Cancer Society,10 prevalence estimates from the 2008 SEER database,11 data for the 2008 PYLL from the NCI’s Cancer Trends Progress Report,12 and data for 2004 age-adjusted DALY per 100,000 people from the WHO’s disease and injury estimates for the United States.13

Statistical Method

The authors conducted bivariate analyses, determining the Pearson correlation between the total planned enrollment in phase III RCTs and 1) measures of a cancer’s evidence quality (number and percent of recommendations that are C1), and 2) measures of societal burden (prevalence, incidence, PYLL, and DALY). They also used multiple linear regressions with a forward inclusion alpha level of 0.05 and a removal alpha level of 0.1 to construct a model that best explained the variance in planned RCT enrollment. Variables with non-normal distributions were log-transformed for analysis. As a sensitivity analysis, they repeated these bivariate and multivariate analyses using the number of active RCTs instead of total planned enrollment as the dependent variable of interest.

To examine graphically the relationship between evidence quality and societal burden-adjusted clinical trial enrollment, the authors compared each cancer’s proportion of C1 recommendations with its total planned RCT enrollment divided by DALY. DALY was chosen over other measures of societal burden because it is the most comprehensive measure, including information about the societal disease cost such as the number of persons affected and the number of years affected individuals lose to disability and death. Analyses of DALY do not include central nervous system, renal, or testicular cancers, because DALY estimates were not available for these cancers.

All statistical analysis was performed with STATA version 12. Graphs were created in Microsoft Excel (version 14.0). This study was exempt from Human Investigational Committee review.

Results

A total of 1260 active phase III RCTs met the inclusion criteria, with an overall planned enrollment of 1,085,913 participants (Table 1). Planned completion time for all clinical trials ranged from 0.3 to 26.0 years, with a median of 5.0 years. The average planned time for trial completion per cancer type generally ranged from 4 to 9 years. Testicular cancer was an outlier, with an average trial completion time of 14 years.

Table 1

Active Cancer Clinical Trials and Patient Enrollment Data

Table 1

The authors identified 834 treatment recommendations in the NCCN Guidelines for the 17 included cancer types (Table 2). The number of recommendations varied from 9 for kidney to 113 for colorectal cancer. Overall, 15% of these were C1 recommendations, 79% were C2a recommendations, 5% were C2b recommendations, and 1% were C3 recommendations. A large degree of variation was seen in the distribution of evidence quality and consensus across cancer types (Figure 1). Hepatobiliary cancer had the highest proportion of C1 recommendations (40%), and endometrial cancer had the lowest (0%). Breast cancer had the highest number of C1 recommendations (N=32), twice as many as the cancer types with the next highest number (oral cavity/pharynx and lung cancers had the same number, with 16 for each).

On bivariate analysis, neither the number nor the percent of C1 recommendations was correlated with planned RCT enrollment, suggesting cancer types with lower evidence quality do not receive greater clinical research resources (Table 3). Measures of societal burden, such as prevalence (P=.01; Figure 2), incidence (P<.01), PYLL (P<.01; Figure 3), and DALY (P<.01), positively correlated with planned RCT enrollment.

Table 2

Categories of Evidence Quality for NCCN Cancer Treatment Recommendations

Table 2

On multivariate regression, PYLL (P<.01) and prevalence (P<.01) were independently associated with RCT enrollment. The other covariates tested (proportion of C1 recommendations, number of C1 recommendations, incidence, and DALY) did not meet inclusion criteria. The R2 of the model was 0.83. As a sensitivity analysis, the authors used the number of active RCTs instead of total planned enrollment as the dependent variable. Using this method, the results were similar to those listed earlier: PYLL (P<.01) and prevalence (P<.01) were both independently associated with the number of RCTs per cancer type.

The authors further examined the relationship between evidence quality and DALY-adjusted clinical trial enrollment to determine the effect of evidence quality while adjusting for societal burden. This was done graphically by comparing the proportion of C1 recommendations with the total planned RCT enrollment divided by DALY for each cancer type (Figure 4). Cancer types are arranged in order of decreasing evidence quality, and the bars show each cancer’s total planned RCT enrollment adjusted by DALY. If the hypothesis that cancers with lower evidence quality are receiving more attention in clinical research were true, the total planned RCT enrollment adjusted for DALY (bar height) would increase as research quality (line height) decreased going left to right. This hypothesis is not supported by the data as presented in Figure 4. Dividing the cancer types into groups based on their total planned RCT enrollment (high: >1500 patients per DALY; moderate: 750-1500 patients per DALY; low: <750 patients per DALY) and evidence quality (high: >15% C1 recommendations; low: <15% C1 recommendations) reveals the following observations. Adjusted for DALY, breast and prostate cancer fall into a category of cancers associated with high planned RCT enrollment and relatively high evidence quality. Melanoma and ovarian and colorectal cancers fall into a category of cancers associated with moderate planned RCT enrollment and relatively low evidence quality. Hepatobiliary, stomach, pancreas, and cervix cancers have low planned RCT enrollment and relatively low evidence quality; and lung, esophagus, bladder, and endometrial cancers have low planned RCT enrollment and moderate to low evidence quality.

Figure 1
Figure 1

The distribution of the categories of evidence supporting the NCCN Clinical Practice Guidelines in Oncology. Cancer types are arranged from highest to lowest percentage of category 1 recommendations.

Abbreviations: CNS, central nervous system; OC, oral cavity.

Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 11, 8; 10.6004/jnccn.2013.0114

Discussion

The authors found that measures of a cancer’s societal burden were positively associated with research effort. This finding was consistent with that of a prior study showing an association between measures of societal burden and the amount of money allocated to cancer types by the NIH.5 They also found that evidence quality, which was generally low but varied widely across cancer types, was not associated with research effort. These results suggest that disparities in evidence quality between cancer types are likely to persist based on the status quo.

Current research efforts, as measured by planned RCT enrollment in currently active RCTs, is likely related to a complex set of factors, including the magnitude of past successes, investigator interest, existing clinical trial infrastructure, availability and motivation of patients to be enrolled, and differences in funding opportunities by cancer type. Although many RCTs do not meet accrual goals,14 the authors propose that planned RCT enrollment can quantify the magnitude of clinical research for a particular cancer. As these RCTs mature and generate results, cancer types with relatively high currently planned RCT enrollment are more likely to gain even higher quality recommendations than those with less research effort. Therefore, although cancer types with at least a moderate amount of planned RCT enrollment and low evidence quality, such as melanoma, ovarian cancer, and colorectal cancer (Figure 4), may see improvement in evidence quality in the coming years as current RCTs mature, deficiencies in evidence quality for cancer types with lower RCT enrollment will likely persist into the future.

Table 3

Bivariate Analyses of Correlates to Total Planned Enrollment in Clinical Trials

Table 3
Figure 2
Figure 2

The relationship between the prevalence and planned randomized controlled trial (RCT) enrollment In the United States. Cancer types are represented as points on the scatterplot to show the relationship between prevalence and total planned enrollment in phase III randomized clinical trials.

Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 11, 8; 10.6004/jnccn.2013.0114

Figure 3
Figure 3

The relationship between the person-years of life lost (PYLL) and randomized controlled trial (RCT) enrollment in the United States. Cancer types are represented as points on the scatterplot to show the relationship between PYLL and total planned enrollment in phase III randomized clinical trials.

Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 11, 8; 10.6004/jnccn.2013.0114

According to this analysis, breast and prostate cancers have the highest planned RCT enrollment when adjusted for societal burden. These cancer types are extremely prevalent, and patient outcomes have already benefited from a large body of research relative to most other cancers. Previous successes in clinical trials or a relatively high number of survivors may have led to more attention and funding for further RCTs. Prior clinical trial success and large numbers of disease survivors may be unmeasured confounding factors that could mask the true associations between RCT enrollment and the variables examined in the present study. The presence of these and other unknown confounding variables highlight the complexity of factors influencing RCT enrollment. In contrast to breast and prostate cancers, lung cancer has a relatively low planned RCT enrollment when adjusted for its burden on society. Other cancers for which a relative lack of research exists when adjusting for their burden on society include esophagus, bladder, cervix, and pancreas cancers. Well-designed clinical trials are needed to improve evidence quality in the setting of limited resources.

Figure 4
Figure 4

Total planned enrollment in currently active clinical trials controlled for disability-adjusted life years (DALY). The bars and left side y-axis values represent the number of people enrolled in currently active clinical trials adjusted for age-adjusted DALY per 100,000 people. The line graph and right side y-axis values represent the proportion of recommendations that are category 1.

Abbreviations: OC, oral cavity; RCT, randomized controlled trial.

Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 11, 8; 10.6004/jnccn.2013.0114

The NIH, the largest funder of clinical trials, considers public health needs, the scientific quality of research, the probability of success, the maintenance of a diverse portfolio, and the maintenance of an adequate scientific infrastructure in choosing which studies to fund.4 In using the probability of success as a criterion, the NIH may focus on RCTs that further refine previously proven treatments. Using the NIH’s last stated criterion—the maintenance of an adequate scientific infrastructure—may induce a push for better evidence in cancers that have low evidence quality. For cancers about which more is known regarding ideal treatment, planned RCT enrollment may be large as the differences between treatment groups and expected gains become incrementally smaller. Using the total number of trials, instead of the total planned enrollment, the authors repeated the bivariate and multivariate analyses above and found no differences (data not shown).

This study is limited by the challenge of assigning a level of knowledge to a complex treatment recommendation. When 2 or more competing options were listed with differing categories of evidence and consensus for a clinical scenario, the authors counted only the highest quality recommendation among the group. However, lower-quality recommendations may sometimes be more appropriate than C1 recommendations for certain categories of patients, based on clinical factors such as comorbidity, side-effect profile, age, preference, and treatment availability. Furthermore, although a uniform definition of C1 recommendations exists, the number of recommendations of any type varied widely among cancer types. In light of this complexity, efforts are currently underway to better summarize the comparative effectiveness research available to guide physicians in treatment selection.15-18

This study should be interpreted with the understanding that using the number or percentage of C1 recommendations cannot fully characterize the level of evidence that exists for a particular disease. Well-studied diseases may have more total recommendations, thus falsely diluting the percentage of C1 recommendations for well-studied cancers and falsely elevating the percentage of C1 recommendations for less well-studied cancers.

Also potentially limiting this study is the fact that measures of cancer burden and clinical trial enrollment do not come from the same years. PYLL and DALY for cancer types were based on data from 2008 and 2004, respectively, which were the most recent years available for this analysis, whereas clinical trials active as of May 2011 were included. Therefore, differing timeframes for analysis may have skewed the results, in particular for diseases that have seen a more recent change in the therapeutic landscape. However, because clinical trials active in 2011 would have been developed in the preceding years, their design and allocation would have been informed by the state of cancer burden in those earlier years.

The authors used phase III RCTs to measure research effort, because they are the gold standard in comparing competing treatment options and determining safety and efficacy. Limitations exist in the external validity and applicability of RCTs, because they often focus on a narrowly selected population in a controlled setting.19,20 Another limitation to this study is the assumption that RCTs always lead to higher evidence quality; RCTs also may be poorly designed, generating confusing results with poor external validity, or new RCTs may generate findings contradictory to previous RCTs. At an even more fundamental level, whether higher evidence quality leads to meaningful improvements in patient care and outcomes is unclear.

Despite these limitations, this study is an important document of the current level of knowledge for different cancer types and how the state of the art affects research priorities. A periodic reassessment of this relationship is critical to determine the efficacy and validity of current prioritization strategies for improving patient outcomes. Highlighting areas of greatest need may help policy-makers target cancer types for increased resource allocation. If desirable, changing the current pattern of clinical trial efforts to take into account the need for high-quality evidence and a cancer’s burden on society will require concerted and targeted efforts to overcome existing research inertia.

Conclusions

In conclusion, evidence quality is not associated with the number of patients being enrolled to cancer RCTs. Current planned enrollment, with certain exceptions such as lung cancer, is associated with societal disease burden as measured by cancer prevalence, incidence, DALY, and PYLL. Prioritization of research questions is necessary to guide the direction of future resource allocation in cancer research, so that progress can be made in the areas of greatest need.

Drs. Lloyd, Buscariollo, Makarov, and Yu have disclosed that they have no financial interests, arrangements, affiliations, or commercial interests with the manufacturers of any products discussed in this article or their competitors. Dr. Gross has disclosed that he receives a research grant from Medtronic and is on the advisory board for Fair Health Inc.

References

  • 1.

    GreenwaldPCullenJWMcKennaJW. Cancer prevention and control: from research through applications. J Natl Cancer Inst1987;79:389400.

  • 2.

    BurnetNGJefferiesSJBensonRJ. Years of life lost (YLL) from cancer is an important measure of population burden—and should be considered when allocating research funds. Br J Cancer2005;92:241245.

    • Search Google Scholar
    • Export Citation
  • 3.

    PoonachaTKGoRS. Level of scientific evidence underlying recommendations arising from the National Comprehensive Cancer Network Clinical Practice Guidelines. J Clin Oncol2011;29:186191.

    • Search Google Scholar
    • Export Citation
  • 4.

    Committee on NIH Research Priority-Setting Process. Scientific opportunities and public needs: improving priority setting and public input at the National Institutes of Health. Washington, DC: National Academy Press; 1998.

    • Search Google Scholar
    • Export Citation
  • 5.

    GrossCPAndersonGFPoweNR. The relation between funding by the National Institutes of Health and the burden of disease. N Engl J Med1999;340:18811887.

    • Search Google Scholar
    • Export Citation
  • 6.

    HarrisRPHelfandMWoolfSH. Current methods of the US Preventive Services Task Force: a review of the process. Am J Prev Med2001;20(3 Suppl):2135.

    • Search Google Scholar
    • Export Citation
  • 7.

    McCrayATIdeNC. Design and implementation of a national clinical trials registry. J Am Med Inform Assoc2000;7:313323.

  • 8.

    Food and Drug Administration Modernization Act of 1997. Pub L No. 105-115 §113 111 Stat 2296.

  • 9.

    NCCN Categories of Evidence and Consensus. National Comprehensive Cancer Network Web site. Available at: http://www.nccn.org/professionals/physician_gls/categories_of_consensus.asp. Accessed January 29 2012.

    • Search Google Scholar
    • Export Citation
  • 10.

    Estimated New Cancer Cases and Deaths by Sex US 2012. National Cancer Institute. Available at: http://www.cancer.org/acs/groups/content/@epidemiologysurveilance/documents/document/acspc-032009.pdf. Accessed January 27 2012.

    • Search Google Scholar
    • Export Citation
  • 11.

    HowladerNNooneAMKrapchoM eds. SEER Cancer Statistics Review 1975-2008National Cancer Institute. Bethesda, MD based on November2011SEER data submission. Available at: http://seer.cancer.gov/csr/1975_2008/. Accessed January 27 2012.

    • Search Google Scholar
    • Export Citation
  • 12.

    Cancer Trends Progress Report 2009/2010 Update. National Cancer Institute. Available at: http://progressreport.cancer.gov/doc_detail.asp?pid=1&did=2009&chid=96&coid=930&mid#life. Accessed January 12 2012.

    • Search Google Scholar
    • Export Citation
  • 13.

    The global burden of disease: 2004 update. The World Health Organization. Available at: http://www.who.int/healthinfo/global_burden_disease/2004_report_update/en/index.html. Accessed January 29 2remo012.

    • Search Google Scholar
    • Export Citation
  • 14.

    ChengSKDietrichMSDiltsDM. A sense of urgency: evaluating the link between clinical trial development time and the accrual performance of cancer therapy evaluation program (NCI-CTEP) sponsored studies. Clin Cancer Res2010;16:55575563.

    • Search Google Scholar
    • Export Citation
  • 15.

    LiECDeMartinoJ. Preliminary report: the development of the NCCN Comparative Therapeutic Index as a clinical evaluative process for existing data in oncology. J Natl Compr Canc Netw2010;5(Suppl 5):S19.

    • Search Google Scholar
    • Export Citation
  • 16.

    WilenskyGR. Developing a center for comparative effectiveness information. Health Aff (Millwood)2006;25:w572585.

  • 17.

    IglehartJK. Prioritizing comparative-effectiveness research—IOM recommendations. N Engl J Med2009;361:325327.

  • 18.

    SoxHCGreenfieldS. Comparative effectiveness research: a report from the Institute of Medicine. Ann Intern Med2009;151:203205.

  • 19.

    SteinbergM. The overthrow of the (evidence) hierarchy. Pract Radiat Oncol2011;1:8182.

  • 20.

    GrossCPMalloryRHeiatAKrumholzHM. Reporting the recruitment process in clinical trials: who are these patients and how did they get there?Ann Intern Med2002;137:1016.

    • Search Google Scholar
    • Export Citation

If the inline PDF is not rendering correctly, you can download the PDF file here.

Correspondence: James B. Yu, MD, Yale University School of Medicine, Department of Therapeutic Radiology, LL-516 Smilow, PO Box 208040, New Haven, CT 06510-3221. E-mail: james.b.yu@yale.edu

Article Sections

Figures

  • View in gallery

    The distribution of the categories of evidence supporting the NCCN Clinical Practice Guidelines in Oncology. Cancer types are arranged from highest to lowest percentage of category 1 recommendations.

    Abbreviations: CNS, central nervous system; OC, oral cavity.

  • View in gallery

    The relationship between the prevalence and planned randomized controlled trial (RCT) enrollment In the United States. Cancer types are represented as points on the scatterplot to show the relationship between prevalence and total planned enrollment in phase III randomized clinical trials.

  • View in gallery

    The relationship between the person-years of life lost (PYLL) and randomized controlled trial (RCT) enrollment in the United States. Cancer types are represented as points on the scatterplot to show the relationship between PYLL and total planned enrollment in phase III randomized clinical trials.

  • View in gallery

    Total planned enrollment in currently active clinical trials controlled for disability-adjusted life years (DALY). The bars and left side y-axis values represent the number of people enrolled in currently active clinical trials adjusted for age-adjusted DALY per 100,000 people. The line graph and right side y-axis values represent the proportion of recommendations that are category 1.

    Abbreviations: OC, oral cavity; RCT, randomized controlled trial.

References

  • 1.

    GreenwaldPCullenJWMcKennaJW. Cancer prevention and control: from research through applications. J Natl Cancer Inst1987;79:389400.

  • 2.

    BurnetNGJefferiesSJBensonRJ. Years of life lost (YLL) from cancer is an important measure of population burden—and should be considered when allocating research funds. Br J Cancer2005;92:241245.

    • Search Google Scholar
    • Export Citation
  • 3.

    PoonachaTKGoRS. Level of scientific evidence underlying recommendations arising from the National Comprehensive Cancer Network Clinical Practice Guidelines. J Clin Oncol2011;29:186191.

    • Search Google Scholar
    • Export Citation
  • 4.

    Committee on NIH Research Priority-Setting Process. Scientific opportunities and public needs: improving priority setting and public input at the National Institutes of Health. Washington, DC: National Academy Press; 1998.

    • Search Google Scholar
    • Export Citation
  • 5.

    GrossCPAndersonGFPoweNR. The relation between funding by the National Institutes of Health and the burden of disease. N Engl J Med1999;340:18811887.

    • Search Google Scholar
    • Export Citation
  • 6.

    HarrisRPHelfandMWoolfSH. Current methods of the US Preventive Services Task Force: a review of the process. Am J Prev Med2001;20(3 Suppl):2135.

    • Search Google Scholar
    • Export Citation
  • 7.

    McCrayATIdeNC. Design and implementation of a national clinical trials registry. J Am Med Inform Assoc2000;7:313323.

  • 8.

    Food and Drug Administration Modernization Act of 1997. Pub L No. 105-115 §113 111 Stat 2296.

  • 9.

    NCCN Categories of Evidence and Consensus. National Comprehensive Cancer Network Web site. Available at: http://www.nccn.org/professionals/physician_gls/categories_of_consensus.asp. Accessed January 29 2012.

    • Search Google Scholar
    • Export Citation
  • 10.

    Estimated New Cancer Cases and Deaths by Sex US 2012. National Cancer Institute. Available at: http://www.cancer.org/acs/groups/content/@epidemiologysurveilance/documents/document/acspc-032009.pdf. Accessed January 27 2012.

    • Search Google Scholar
    • Export Citation
  • 11.

    HowladerNNooneAMKrapchoM eds. SEER Cancer Statistics Review 1975-2008National Cancer Institute. Bethesda, MD based on November2011SEER data submission. Available at: http://seer.cancer.gov/csr/1975_2008/. Accessed January 27 2012.

    • Search Google Scholar
    • Export Citation
  • 12.

    Cancer Trends Progress Report 2009/2010 Update. National Cancer Institute. Available at: http://progressreport.cancer.gov/doc_detail.asp?pid=1&did=2009&chid=96&coid=930&mid#life. Accessed January 12 2012.

    • Search Google Scholar
    • Export Citation
  • 13.

    The global burden of disease: 2004 update. The World Health Organization. Available at: http://www.who.int/healthinfo/global_burden_disease/2004_report_update/en/index.html. Accessed January 29 2remo012.

    • Search Google Scholar
    • Export Citation
  • 14.

    ChengSKDietrichMSDiltsDM. A sense of urgency: evaluating the link between clinical trial development time and the accrual performance of cancer therapy evaluation program (NCI-CTEP) sponsored studies. Clin Cancer Res2010;16:55575563.

    • Search Google Scholar
    • Export Citation
  • 15.

    LiECDeMartinoJ. Preliminary report: the development of the NCCN Comparative Therapeutic Index as a clinical evaluative process for existing data in oncology. J Natl Compr Canc Netw2010;5(Suppl 5):S19.

    • Search Google Scholar
    • Export Citation
  • 16.

    WilenskyGR. Developing a center for comparative effectiveness information. Health Aff (Millwood)2006;25:w572585.

  • 17.

    IglehartJK. Prioritizing comparative-effectiveness research—IOM recommendations. N Engl J Med2009;361:325327.

  • 18.

    SoxHCGreenfieldS. Comparative effectiveness research: a report from the Institute of Medicine. Ann Intern Med2009;151:203205.

  • 19.

    SteinbergM. The overthrow of the (evidence) hierarchy. Pract Radiat Oncol2011;1:8182.

  • 20.

    GrossCPMalloryRHeiatAKrumholzHM. Reporting the recruitment process in clinical trials: who are these patients and how did they get there?Ann Intern Med2002;137:1016.

    • Search Google Scholar
    • Export Citation

Article Information

PubMed

Google Scholar

Related Articles

Metrics

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 53 53 3
PDF Downloads 4 4 1
EPUB Downloads 0 0 0