Clinical reporting for personalized cancer genomics requires extensive access to subscription-only literature

Abstract

Objective: Medical care for cancer is increasingly directed by genomic laboratory testing for alterations in the tumor genome that are significant for diagnosis, prognosis and therapy. Uniquely in medicine, providers must search the biomedical literature for each patient to determine the clinical significance of these alterations. Access to published scientific literature is frequently subject to high fees, with access limited to institutional subscriptions. We sought to investigate the degree to which the scientific literature is accessible to clinical cancer genomics providers, and the potential role of university and hospital system libraries in information access for cancer care.

Methods: We identified 265 journals that were accessed during the interpretation and reporting of clinical test results from 1,842 cancer patients at the University Health Network (Toronto, Canada). We determined the degree of open access for this set of clinically important literature, and for any journals not available through open access we surveyed subscription access at seven academic hospital systems and at their affiliated universities.

Results: This study found that nearly half (116/265) of journals have open access mandates that make articles freely available within one year of release. For the remaining subscription access journals, universities provided a uniformly high level of access, but access available through hospital system collections varied widely.

Conclusion: This study highlights the importance of different modes of access to the use of the scientific literature in clinical practice and points to challenges that must be overcome as genomic medicine grows in scale and complexity.

Keywords: Cancer; genetics; health sciences library collections; oncology; open access; pathology; precision medicine; subscription access; university collections.

Figure 1

Journals required for interpretation of clinical cancer genomic testing results for a large patient series.

Figure 2

Survey of university and hospital system subscription coverage for literature used in clinical cancer genomics. (A) Proportion of subscription-access journals in the leukemia and lung cancer clinical interpretation datasets that were covered by the holdings of university and health science center libraries surveyed. (B) Proportion of journals and journal articles from subscription-access journals in the combined clinical interpretation dataset covered by these collections. (C) Ratio of the number of instances in the combined clinical interpretation dataset to the number of journals covered by each institutional collection (red line). (D) The average impact factor of journals in each institutional collection (purple line). Proportion of the total number of journals covered by subscription access is shown for comparison (gray bars).

Figure 3

Overlap between university and hospital system library collections. (A) Holdings of surveyed hospital system libraries, grouped by Web of Science category. (B) Web of Science categories for subscription-access journals that were not covered by university holdings, by institution. (C) Additive coverage for university collections, combined with affiliated hospital system collections. Percent coverage is shown for the university system alone (gray), and for the combined university and hospital collections (green).