•  
  •  
 

Article Title

Applying a Neutropenia Risk Model to Cancer Patients Using VDW Data: A CRN Pilot Study

Publication Date

4-30-2015

Keywords

neutropenia, risk model

Abstract

Background/Aims: Neutropenia, an extreme reduction in infection-fighting white blood cells, is a dangerous side effect of cytotoxic chemotherapy for cancer treatment, especially when accompanied by fever. Neutropenia and febrile neutropenia are associated with treatment delays, chemotherapy dose reductions, hospitalizations, antibiotic use, death and high costs. The granulocyte-colony stimulating factors (G-CSFs), filgrastim, pegfilgrastim and tbo-filgrastim, reduce the risk of neutropenia and febrile neutropenia; however, they are also associated with additional patient visits, side effects and high treatment-related costs. G-CSF use was found to be inconsistent with recommended clinical guidelines (overuse among those at low risk and underuse among those at high risk for neutropenia), thus making it part of the Choosing Wisely campaign. Individual patient factors are associated with increased risk for neutropenia and febrile neutropenia beyond that of the prescribed chemotherapy. No study has investigated the effect of providing medical oncologists with neutropenia risk predictions in real time to optimize the use of G-CSF and reduce neutropenia/febrile neutropenia outcomes among patients. The ongoing work will assess the feasibility of applying an existing patient neutropenia risk model developed by Lyman et al. to existing electronic data files. This work builds upon feasibility work currently being conducted at HealthPartners Institute, Minneapolis, MN, and adds two more CRN sites (Group Health Research Institute, Seattle, WA, and Henry Ford Health System, Detroit, MI).

Methods: We will investigate the applicability of an adapted risk model to readily available electronic data (excluding clinical oncology treatment modules). We will adapt Lyman’s model to work with available data, provide external validation and calibrate the adapted model for our patient population.

Results: To date, we have obtained IRB approval to conduct the study and are developing a distributed modular program for extraction of clinical and administrative data variables available in the HMORN’s virtual data warehouse.

Discussion: This presentation will report preliminary data on neutropenia and febrile neutropenia rates, chemotherapy treatment variations and G-CSF use among racially and geographically diverse sites. The current CRN study will provide a necessary foundation for a planned external proposal designed to test and implement a real-world risk assessment tool intended for use by clinicians at the point of care.

Share

COinS