High-cost cancer imaging: Opportunities for utilization management. (June 2017)

• Record Type: Journal Article
• Authors: Copeland, Timothy P.; Franc, Benjamin L.
• Is Part Of: Journal of cancer policy. Volume 12(2017)
• Page Start: 16
• Page End: 20
• Publication Date: 2017-06
• Subjects: Diagnostic imaging -- Utilization -- Radiology -- Medical oncology -- Evidence-based practice; Cancer -- Government policy -- Periodicals; Cancer -- Patients -- Services for -- Periodicals; Medical Oncology -- Periodicals; Public Health -- Periodicals; Cancer; Periodicals; 362.196994
• DOI: 10.1016/j.jcpo.2016.12.004
• Languages: English
• ISSNs: 2213-5383

Abstract

Highlights: We examined the effect of patient and provider characteristics on 104 oncologists' imaging behaviors over 15 years at an academic cancer center. Imaging utilization variations were not based solely upon patient factors and treatment intensity, but also upon provider ordering preferences. Total imaging utilization was higher among patients with multiple ordering oncologists. Imaging utilization tended to increase as patients' temporal proximity to death decreased, declining slightly in the final 3 months of life. Our findings suggest improved coordination of care practices may reduce duplicative imaging from multiple ordering oncologists. Abstract: Objectives: To retrospectively evaluate utilization of high-cost cancer imaging to clarify the extent to which variations in provider preferences drive imaging utilization. Study design: At a United States academic cancer center, 4, 605 patients were identified with 29, 740 oncologist ordered tomographic imaging studies. Patients' dates of death ranged from January 2000 through December 2014. Imaging was restricted to CT, MR, and PET/PET-CT. Outcome variables were total imaging per patient and total imaging per patient by a single oncologist. The number of ordering oncologists per patient, patients receiving imaging in the final year of life, and patients receiving imaging ordered by a high-ordering oncologist were the predictors of interest. Methods: Zero-truncated negative binomial regressions were used to model collective and individual oncologist per patient imaging counts, with the exposure period defined as the number of days from diagnosis to death. Results: Patients with imaging ordered by one of the high-ordering medical oncologists predicted nearly a two-fold increase in total images from diagnosis to death (IRR, 1.93; 95% CI, 1.67–2.23). Increasing numbers of providers (2, 3, 4+ ordering oncologists) were associated with greater collective per patient imaging (IRRs 1.65, 2.19, 2.33). Mean imaging intensity increased in a linear manner as temporal proximity to death decreased, from 12 months pre-mortem to death, and imaging in the final year of life was associated with greater per patient imaging (IRR, 0.25; 95% CI, 0.23–0.27). Conclusion: These findings suggest heterogeneous provider ordering preferences and lapses in care coordination are drivers of high-cost cancer imaging utilization.