In this payer-based sample, we found that GEP testing was associated with less use of adjuvant chemotherapy. In analyses stratified by risk groups using traditional clinical prognostic criteria, GEP testing was associated with less chemotherapy use among women at high clinical risk of recurrence, but was associated with more chemotherapy use among women with low clinical risk of recurrence. These results suggest that GEP testing influences both the overall rate of chemotherapy use and the type of patients who receive chemotherapy.
Among the subgroup of women who received a GEP test, more women had their risk re-classified as lower than their clinical risk. GEP may result in more personalized care by providing additional reassurance to women at low clinical risk of recurrence who could forgo adjuvant chemotherapy, and identifying women who are at high recurrence risk who may benefit from more aggressive treatment [7
]. We did not see any association between GEP use and the occurrence of serious chemotherapy adverse-effects or charges during the 12 months following diagnosis. Together, these findings suggest that use of GEP may result in more “personalization” of chemotherapy use, but may not be associated with reduced serious chemotherapy-associated adverse events or cost savings, at least in the short term.
Prior studies provide limited evidence regarding the relationship between GEP testing and use of chemotherapy in routine practice settings. In one study of 269 women with non-metastatic breast cancer seen at a single cancer center, adjuvant chemotherapy was given to 7% of women with low recurrence score, compared to 42% and 86% of women with intermediate and high recurrence score, respectively [19
]. In a different study of women with OncotypeDX testing, an independent oncologist was asked to review each chart and make a recommendation about the need for adjuvant chemotherapy without access to this test result [31
]. OncotypeDX altered chemotherapy management for 38% of women for whom this independent assessment differed from that of the treating oncologist. In a third study, receipt of a GEP test was associated with a change in treatment recommendation, typically resulting in less use of chemotherapy, for one-third of patients, consistent with our findings [32
]. None of these prior studies examined the relationship between GEP use and serious chemotherapy-associated adverse events. We found that while GEP use was associated with less chemotherapy use, and changes in patterns of use within subgroups of women defined by conventional clinical characteristics, this did not appear to translate into fewer serious adverse events.
Several economic simulations, with a longer time horizon than our study, suggest that GEP-guided therapy is associated with modest cost-saving compared with treatment decisions based on traditional clinical risk factors [7
]. Our analysis, based on actual charges for care over a 12-month period, did not suggest a difference in costs associated with GEP test use. Based on the degree of precision of our estimates for GEP-related changes in total charges (), our sample size was sufficient to detect a relative savings as small as 16%; smaller differences may be meaningful from a societal perspective. The lack of association between GEP testing and charges may perhaps reflect that a decrease in the use of chemotherapy for some patients was in part offset by an increase in use for others.
The premise that a genomic test could facilitate more individually tailored decision-making about the use of adjuvant chemotherapy is appealing from the perspective of women, their providers, and health care systems. Ultimate assessment of the value of these tests will depend on the availability of data regarding long-term outcomes of treatment decisions and the ability to track changes in utilization and outcomes over time. Data from ongoing randomized controlled trials to test the impact of GEP testing on outcomes will be critical to the assessment of the value of these tests [33
Our study has several limitations. As our findings are based on observational data, we cannot conclude that GEP testing was the cause of the observed differences in the use of chemotherapy. We used propensity score adjustment to address the potential selection bias that is inherent in observational data [34
]. As noted above, we did not have data on long-term outcomes, which may be particularly important for women with non-metastatic disease. Finally, all women in our study were under age 65 and received coverage from a single, large health insurer. Thus, our findings may be less generalizable to the uninsured, older breast cancer patients covered by Medicare, or to women with Medicaid. The distribution of age, stage and nodal status in our sample was similar to those observed for women with breast cancer in the Breast Cancer Surveillance Consortium [35
]. Our sample does, however, include women who receive care across the US in diverse practice settings, including community-based practices.
In this study of breast cancer patients treated in routine practice settings, GEP testing was associated with an overall decline in the use of adjuvant chemotherapy. While GEP testing is being used to tailor decision-making about the use of adjuvant chemotherapy, we did not demonstrate differences in serious chemotherapy-associated adverse events or charges during the 12 months following diagnosis.