An assessment of the collective efforts of clinical trials to provide evidence-based practice guidelines in cancer care.

Abstract

6019 Background: It is unclear how many of the recommendations in the National Comprehensive Cancer Network (NCCN) guidelines are based Category 1 (C1: “high level evidence”), or whether current cancer research efforts are focusing on the knowledge gap between guidelines and supporting evidence. We defined the knowledge gap as the percent of NCCN recommendations based on lower levels of evidence, and assessed the relation between the knowledge gap and cancer research efforts in terms of clinical trial enrollment. Methods: We examined the active, phase 3 randomized controlled trials (RCTs) listed on clinicaltrials.gov for the 17 most prevalent solid tumors. We used the NCCN guidelines to tally the percentage of C1 recommendations. We used the Pearson coefficient and linear regression to examine whether enrollment in active phase 3 RCTs is associated with 1) the knowledge gap and 2) measures of burden to society including prevalence, incidence, person years life lost (PYLL) and disability adjusted life years (DALY). Results: We identified 834 treatment recommendations for the 17 included cancer types. Overall, 15% of the NCCN recommendations were rated as C1, varying from 40% for hepatobiliary to 0% for endometrial cancer. There was no association between RCT enrollment and the proportion of C1 recommendations in univariate or multivariate analyses. RCT enrollment positively correlated with measures of burden to society including prevalence (p-value < 0.001), incidence (p-value 0.001), and DALY (p-value 0.038). Breast and prostate cancers have a large amount of RCT enrollment per PYLL and a relatively small knowledge gap. Melanoma, ovarian, and endometrial cancers have a moderate amount of RCT enrollment per PYLL and a relatively large knowledge gap. Lung, esophagus, bladder and CNS cancers have a low amount of RCT enrollment per PYLL and a moderate to large knowledge gap. Conclusions: RCT enrollment is not correlated to the knowledge gap. Current planned enrollment in RCTs is partially predicted by burden to society, with the strongest correlation to incidence and prevalence. Most recommendations in the NCCN guidelines are not based on C1 evidence.