Abstract
Background:
Chemotherapy dose reductions are associated with poorer survival. To better understand the role of clinician- and facility-level factors in chemotherapy dosing, we conducted an analysis within a large, real-world cohort of women with stages I-IIIA breast cancer.
Methods:
Our cohort included 8,540 breast cancer patients receiving chemotherapy at Kaiser Permanente Northern California between 2006 and 2019. Patients were treated across 22 facilities by 198 clinicians. We evaluated associations between clinician- and facility-level factors related to dose reductions at the start of chemotherapy (first cycle dose proportion, FCDP, <90%) and throughout treatment (average relative dose intensity, ARDI, <90%). Prevalence ratios (PR) and corresponding 95% confidence intervals (CI) were estimated for the clinician and facility factors in relation to chemotherapy dose reductions.
Results:
Factors associated with an increased likelihood of dose reduction were increased clinician years since medical school (FCDP<90%: PR≥30 vs <10 years:1.78, p-trend=0.03; ARDI<90%: PR≥30 vs <10 years:1.29, p-trend=0.03) and treatment at less urban facilities (ARDI<90%: PR<100% vs 100% urban:1.38, p-trend=0.002). Factors associated with a decreased likelihood of dose reduction were higher annual clinician volume of stages I-IIIA breast cancer patients (FCDP<90%: PR≥30 vs ≤15 patients:0.64, p-trend=0.03; ARDI<90%: PR≥30 vs ≤15 patients:0.76, p-trend=0.01), higher treatment facility annual volume of stages I-IIIA breast cancer patients (ARDI<90%: PR≥200 vs <75 patients:0.85, p-trend=0.03), and a larger practice size (FCDP<90%: PR≥10 vs ≤5 oncologists:0.53, p-trend=0.02).
Conclusions:
Clinician- and facility-level factors were associated with chemotherapy dose reductions. Practice-level changes, such as increasing breast cancer patient volumes and practice size, may support optimal dosing practices.
Keywords: first cycle dose reduction, average relative dose intensity, chemotherapy, breast cancer, clinician factors, facility factors
Introduction
Chemotherapy has been used for treating breast cancer (BC) since the mid-twentieth century.1 Regimens were initially used in the metastatic setting, then adopted for use in locally advanced and early-stage disease.2 During our study period, approximately one-third of patients with stage I–IIIA breast cancer received curative-intent chemotherapy, which is the focus of this analysis.3 BC chemotherapy is typically dosed based on body surface area.4 Relative dose intensity (RDI) is the ratio of the amount of chemotherapy delivered to the standard dose intensity for a chemotherapy regimen per unit time.5 Although studies have shown that RDI reductions may correlate with poorer survival outcomes,5-8 dose reductions may be clinically appropriate in certain cases (e.g., in response to toxicity or intolerance of side effects).9 However, there are also circumstances in which a dose reduction may have occurred prematurely or due to non-clinical factors.
Previous studies have noted that patient-level factors such as older age, obesity, Black race, low educational attainment, and presence of comorbidities are associated with chemotherapy dose reductions in early-stage disease.10-14 Although studied less frequently, facility-level factors such as hospital type,15 smaller practice size, and rural location16 have also been associated with chemotherapy dose reductions. There is limited literature, however, regarding the influence of clinician-level factors on dose reductions, which is particularly salient as clinician age,17-19 years of experience,19-21 race,18,22 gender,18,20,23,24 and patient volume25 can influence clinical decision-making and patient outcomes. Given the evolving landscape of breast cancer chemotherapy for early-stage disease including changes in dosing guidelines,4,26 broader use of gene recurrence score testing for decision-making on when to use chemotherapy,27-29 and the adoption of novel dosing strategies (e.g., dose-dense regimens),30 there are numerous contributing factors to a clinicians’ treatment decisions.
To better understand the role of clinician- and facility-level factors in chemotherapy dosing, we conducted an analysis within a large, real-world cohort of women diagnosed with stages I-IIIA BC. In our primary analyses, we examined the associations between clinician- and- facility-level factors and chemotherapy dose reduction, including both first cycle dose proportion (FCDP) <90% and average relative dose intensity (ARDI) <90%. In secondary analysis, we evaluated how those associations varied across patient age, body mass index (BMI), and calendar time.
Methods
Study Population and Data Sources
The Optimal Breast Cancer Chemotherapy Dosing (OBCD) study includes 34,109 women aged ≥18 years at diagnosis with stages I-IIIA BC enrolled at Kaiser Permanente Northern California (KPNC) from 2006-2019 or Kaiser Permanente Washington (KPWA) from 2004-2015. Institutional Review Board approval was obtained from all sites (Memorial Sloan Kettering Cancer Center, Rutgers University, KPNC, and KPWA), with a waiver of consent to use patient data from KPNC and KPWA.
Eligible participants were women enrolled in KPNC/KPWA diagnosed with primary unilateral BC and no history or same-day diagnosis of any cancer (except non-melanoma skin cancer) who did not opt out of research and had available medical records. This analysis was limited to women diagnosed between 2006-2019 at KPNC who received chemotherapy for BC (n=12,302). We then restricted to those who underwent surgery and received only adjuvant chemotherapy (n=10,969), were not known to be part of a randomized controlled trial (n=10,404), had data to ascertain FCDP/ARDI (n=9,550), and data for relevant patient-level covariates (n=9,136) and clinician/facility-level factors, resulting in a final analytic cohort of 8,450 women. Clinicians and facilities were included based on patient eligibility.
Cases and relevant tumor data were obtained from the KPNC Cancer Registry and Virtual Data Warehouse (VDW).31 Data on patient-level and treatment variables (drug name, dosage, date, and administration route) were collected from the VDW tumor registry and medical records, as described previously.32 Clinician/facility-level factors were collected from the VDW and internal KPNC data. American Community Survey (ACS) data was used to determine facility urbanicity.
Clinician and Facility Factors
Clinician factors included gender, years since medical school, patient-clinician racial/ethnic concordance, and average annual stages I-IIIA BC patient volume. Years since medical school was calculated at the time of first encounter with the patient and categorized into four groupings to represent different career stages. Thus, the same clinician may appear in more than one category. The clinician’s average patient volume represents the average number of stages I-IIIA BC patients seen annually during the years in which they saw these patients within KPNC.
Facility factors included practice size, average annual stages I-IIIA BC patient volume, and treatment facility urbanicity. Within KPNC, each clinician has a home treatment facility; treatment facilities are grouped into larger geographic “service areas.” Practice size reflects the number of oncologists working within the clinician’s service area during the study period. Treatment facility annual patient volume was the total number of stages I-IIIA BC patients seen at that facility divided by the number of years in the study period. Facility urbanicity was determined using 5-year estimates from ACS 2008-2012 data (representing the study midpoint). The percentage reflects the proportion of residents within the corresponding census tract residing in a rural block. Due to limited heterogeneity in urbanicity, the variable was dichotomized (100% urban; <100% urban).
Dosing Variables
Intended regimens and administration schedules were identified, as detailed elsewhere,33 using algorithms followed by chart abstraction in cases of ambiguity. The primary outcome of interest was a chemotherapy dose reduction, assessed two ways: (1) FCDP<90% and (2) ARDI<90%. Both measures are based on a comparison of the observed to expected dosing for the received cycles of the specific intended regimen.
FCDP was calculated by dividing the observed dose by expected dose at first chemotherapy administration. We dichotomized to FCDP <90% or ≥90%, with values <90% being considered a first cycle dose reduction. The 90% cutoff was selected, based on previous studies,13,14 to reflect a clinically meaningful dose reduction while balancing statistical power.14
Each drug’s RDI11,14 was calculated by dividing the cumulative dose received across all cycles received, and dividing this by the cumulative planned dose for the number of cycles received (dose per unit time). This value was then averaged across drugs to calculate the ARDI, which was then dichotomized to <90% or ≥90%.
Statistical Analyses
We conducted generalized linear models of the Poisson family with a log-link function and robust standard errors to estimate prevalence ratios (PR) and corresponding 95% confidence intervals (CI), accounting for clustering at the clinician level, for each factor of interest in relation to the primary dose-related outcome. For example, for FDCP<90%, minimally-adjusted generalized linear models include age, year of diagnosis, and cancer stage (model A). Additional patient-level factors were included in subsequent models to account for confounding, including race, ethnicity, neighborhood income, body mass index (BMI) at diagnosis, Charlson Comorbidity Index (CCI), and human epidermal growth factor-2 (HER2) status) (model B). The final, fully-adjusted multivariable models for FCDP<90% adjusted for any statistically significant clinician and facility factors from the minimally adjusted models (model C). A similar model-building approach was used for ARDI<90% (models D, E, and F), with the exception that unlike model B, model E includes two unique patient-level variables (estrogen receptor/ progesterone receptor status and cancer grade) in the ARDI analyses. All models account for clustering of patients treated by the same clinician. For exposures modelled as grouped linear variables, p-trends were calculated using a Wald test.
In secondary analyses, we examined stratification by BMI (<30kg/m2; ≥30kg/m2), age at diagnosis (<65; ≥65 years), and time (2006-2014; 2015-2019) in fully-adjusted models including patient-level factors and significant clinician/facility-level factors. These stratifications were informed by prior studies that have identified patients with high BMI or older age may be particularly vulnerable to chemotherapy dose reduction.10,14,34-38,39-41 The time point for the time-stratified analysis was selected based on the 2015 implementation of a subspecialized model of oncology care at KPNC. We therefore examined the associations before/after 2015 to determine if drivers of chemotherapy dose reduction differed as the makeup of a clinicians’ patient panels became more homogenous in cancer type. Results are presented separately for each group with the p-interaction determined using a likelihood ratio test.
All analyses were conducted using Stata 15.1 (StataCorp, College Station, TX).
Results
Among 8,450 women in the analytic cohort, 46.7% were aged 50-64 years at diagnosis and 55.9% were non-Hispanic White (Table 1). The majority (74.7%) had no medical comorbidities. Stage II disease was the most common (52.6%), followed by I (36.7%), then IIIA (10.7%). Of 198 clinicians, over half (53.5%) were male, 67.2% were a different race/ethnicity than their patient, and 55.1% saw 15 or fewer stage I-IIIA BC patients yearly (Table S1). Of the 22 treatment facilities, 90.9% were urban and 50% saw between 75-199 stage I-IIIA BC patients yearly (Table S2).
Table 1.
OBCD Study stages I – IIIa breast cancer patient sociodemographic characteristics
| Characteristic | Overall N = 8,450 (100%)1 |
|---|---|
| Age at diagnosis | |
| 18-39 | 669 (7.9%) |
| 40-49 | 2,075 (24.6%) |
| 50-64 | 3,949 (46.7%) |
| 65-74 | 1,513 (17.9%) |
| 75+ | 244 (2.9%) |
| Race/ethnicity | |
| Hispanic | 1,328 (15.7%) |
| Non-Hispanic American Indian or Alaskan Native, more than one race, or other | 24 (0.3%) |
| Non-Hispanic Asian | 1,584 (18.7%) |
| Non-Hispanic Black | 717 (8.5%) |
| Non-Hispanic Native Hawaiian or Other Pacific Islander | 74 (0.9%) |
| Non-Hispanic White | 4,723 (55.9%) |
| Median neighborhood household income | |
| Q1: <66,392 | 1,962 (23.2%) |
| Q2: $66,392-<$89,107 | 2,064 (24.4%) |
| Q3: $89,107-<$117,278 | 2,254 (26.7%) |
| Q4: >=$117,278 | 2,170 (25.7%) |
| Year of diagnosis | |
| Jan 2006 - Dec 2007 | 1,063 (12.6%) |
| Jan 2008 - April 2012 | 2,629 (31.1%) |
| May 2012 - Dec 2014 | 1,649 (19.5%) |
| Jan 2015 - Dec 2019 | 3,109 (36.8%) |
| Charlson comorbidity index | |
| 0 | 6,308 (74.7%) |
| 1 | 1,416 (16.8%) |
| 2 | 409 (4.8%) |
| 3+ | 317 (3.8%) |
| Stage | |
| Stage I | 3,100 (36.7%) |
| Stage II | 4,443 (52.6%) |
| Stage IIIA | 907 (10.7%) |
n (%)
In the fully-adjusted models accounting for patient/clinician/facility-level factors (models C and F; Table 2), patients whose clinicians saw a larger annual volume of stages I-IIIA BC were less likely to receive FCDP<90% (PR≥30 vs ≤15 patients:0.64; 95% CI:0.41-0.99; p-trend=0.03) and have an ARDI<90% (PR≥30 vs ≤15 patients:0.76; 95% CI:0.61-0.94; p-trend=0.01). Secondary analyses showed no significant differences across BMI-stratified (Table 3), age-stratified (Table 4), or time-stratified (Table 5) subgroups in relation to FCDP<90% (p-interactions >0.05 for all clinician/facility factors). However, the inverse association between a clinician’s patient volume and ARDI<90% was largely driven by the treatment of patients with a BMI ≥30kg/m2 (i.e., patients with obesity). Patients with obesity treated by high-volume clinicians were less likely to have an ARDI<90% (PR≥30 vs ≤15 patients:0.68; 95% CI:0.54-0.85; p-trend=0.001; Table 6), whereas no significant association was observed in non-obese patients (PR≥30 vs ≤15 patients:0.86; 95% CI:0.67-1.12; p-trend=0.26) (p-interaction=0.01). No significant interactions were observed across age-stratified (Table 7) or time-stratified subgroups (Table 8) for annual clinician patient volume with ARDI<90%.
Table 2.
Clinician/facility level factors associated with FCDP <90% and ARDI <90% of chemotherapy regimen
| FCDP <90% | ARDI <90% | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Characteristic | No n (%) |
Yes n (%) |
Model A PR (95% CI) |
Model B PR (95% CI) |
Model C PR (95% CI) |
No n (%) |
Yes n (%) |
Model D PR (95% CI) |
Model E PR (95% CI) |
Model F PR (95% CI) |
| TOTAL | 7,871 (93%) | 579 (6.9%) | 6,804 (81%) | 1,646 (19%) | ||||||
| Clinician Gender | ||||||||||
| Male | 2,969 (37.7%) |
302 (52.2%) |
Ref | Ref | Ref | 2,546 (37.4%) |
725 (44.0%) |
Ref | Ref | Ref |
| Female | 4,902 (62.3%) |
277 (47.8%) |
0.75 (0.52-1.09) |
0.77 (0.55-1.09) |
0.77 (0.55-1.09) |
4,258 (62.6%) |
921 (56.0%) |
0.87 (0.72-1.06) |
0.88 (0.74-1.06) |
0.86 (0.72-1.03) |
| P = 0.13 | P = 0.14 | P = 0.14 | P = 0.17 | P = 0.18 | P = 0.10 | |||||
| Clinician Years Since Medical School | ||||||||||
| <10 years | 1,462 (18.6%) |
71 (12.3%) |
Ref | Ref | Ref | 1,278 (18.8%) |
255 (15.5%) |
Ref | Ref | Ref |
| 10-20 Years | 3,355 (42.6%) |
231 (39.9%) |
1.53 (1.04-2.25) |
1.57 (1.08-2.30) |
1.57 (1.08-2.30) |
2,905 (42.7%) |
681 (41.4%) |
1.17 (0.97-1.40) |
1.19 (0.99-1.43) |
1.15 (0.96-1.39) |
| 21-30 Years | 1,824 (23.2%) |
176 (30.4%) |
1.76 (1.13-2.75) |
1.86 (1.20-2.89) |
1.86 (1.20-2.89) |
1,566 (23.0%) |
434 (26.4%) |
1.26 (1.02-1.55) |
1.29 (1.06-1.57) |
1.28 (1.06-1.55) |
| ≥30 Years | 1,230 (15.6%) |
101 (17.4%) |
1.67 (0.96-2.92) |
1.78 (1.07-2.96) |
1.78 (1.07-2.96) |
1,055 (15.5%) |
276 (16.8%) |
1.25 (0.96-1.63) |
1.27 (1.00-1.62) |
1.29 (1.02-1.65) |
| p-trend = 0.08 | p-trend = 0.03 | p-trend = 0.03 | p-trend = 0.10 | p-trend = 0.05 | p-trend = 0.03 | |||||
| Patient-Clinician Racial/Ethnic Concordance | ||||||||||
| Discordant | 5,129 (65.2%) |
394 (68.0%) |
Ref | Ref | Ref | 4,419 (65.0%) |
1,104 (67.1%) |
Ref | Ref | Ref |
| Concordant | 2,742 (34.8%) |
185 (32.0%) |
0.86 (0.66-1.12) |
0.99 (0.75-1.32) |
0.99 (0.75-1.32) |
2,385 (35.0%) |
542 (32.9%) |
0.92 (0.81-1.04) |
0.96 (0.84-1.09) |
0.99 (0.86-1.13) |
| P = 0.27 | P = 0.95 | P = 0.95 | P = 0.17 | P = 0.52 | P = 0.84 | |||||
| Average Clinician Patient Volume Per Year | ||||||||||
| ≤15 patients | 2,142 (27.2%) |
243 (42.0%) |
Ref | Ref | Ref | 1,820 (26.7%) |
565 (34.3%) |
Ref | Ref | Ref |
| 16-29 patients | 3,457 (43.9%) |
220 (38.0%) |
0.70 (0.47-1.05) |
0.68 (0.47-0.99) |
0.68 (0.47-0.99) |
2,998 (44.1%) |
679 (41.3%) |
0.83 (0.67-1.02) |
0.83 (0.68-1.00) |
0.84 (0.70-1.01) |
| ≥30 patients | 2,272 (28.9%) |
116 (20.0%) |
0.68 (0.42-1.11) |
0.64 (0.41-0.99) |
0.64 (0.41-0.99) |
1,986 (29.2%) |
402 (24.4%) |
0.79 (0.61-1.02) |
0.76 (0.60-0.96) |
0.76 (0.61-0.94) |
| p-trend = 0.09 | p-trend = 0.03 | p-trend = 0.03 | p-trend = 0.06 | p-trend = 0.02 | p-trend = 0.01 | |||||
| Practice Size | ||||||||||
| ≤ 5 oncologists | 685 (8.7%) |
88 (15.2%) |
Ref | Ref | Ref | 604 (8.9%) |
169 (10.3%) |
Ref | Ref | Ref |
| >5-10 oncologists | 3,866 (49.1%) |
285 (49.2%) |
0.70 (0.41-1.17) |
0.63 (0.42-0.97) |
0.63 (0.42-0.97) |
3,333 (49.0%) |
818 (49.7%) |
0.95 (0.71-1.28) |
0.89 (0.70-1.14) |
0.86 (0.67-1.11) |
| >10 oncologists | 3,320 (42.2%) |
206 (35.6%) |
0.58 (0.34-0.99) |
0.53 (0.34-0.83) |
0.53 (0.34-0.83) |
2,867 (42.1%) |
659 (40.0%) |
0.90 (0.67-1.19) |
0.85 (0.68-1.08) |
0.84 (0.67-1.07) |
| p-trend = 0.06 | p-trend = 0.02 | p-trend = 0.02 | p-trend = 0.39 | p-trend = 0.27 | p-trend = 0.29 | |||||
| Treatment Facility Breast Cancer Patient Volume Per Year | ||||||||||
| <75 patients | 1,848 (23.5%) |
183 (31.6%) |
Ref | Ref | Ref | 1,539 (22.6%) |
492 (29.9%) |
Ref | Ref | Ref |
| 75-<200 patients | 5,194 (66.0%) |
344 (59.4%) |
0.66 (0.45-0.97) |
0.77 (0.53-1.11) |
0.77 (0.53-1.11) |
4,559 (67.0%) |
979 (59.5%) |
0.71 (0.59-0.86) |
0.77 (0.65-0.93) |
0.77 (0.65-0.93) |
| ≥200 patients | 829 (10.5%) |
52 (9.0%) |
0.68 (0.41-1.14) |
0.67 (0.39-1.14) |
0.67 (0.39-1.14) |
706 (10.4%) |
175 (10.6%) |
0.82 (0.63-1.07) |
0.85 (0.66-1.08) |
0.85 (0.66-1.08) |
| p-trend = 0.06 | p-trend = 0.10 | p-trend = 0.10 | p-trend = 0.02 | p-trend = 0.03 | p-trend = 0.03 | |||||
| Treatment Facility Urbanicity | ||||||||||
| 100% urban | 7,391 (93.9%) |
553 (95.5%) |
Ref | Ref | Ref | 6,442 (94.7%) |
1,502 (91.3%) |
Ref | Ref | Ref |
| <100% urban | 480 (6.1%) |
26 (4.5%) |
0.84 (0.37-1.90) |
0.62 (0.30-1.29) |
0.62 (0.30-1.29) |
362 (5.3%) |
144 (8.7%) |
1.59 (1.29-1.95) |
1.38 (1.12-1.69) |
1.38 (1.12-1.69) |
| P = 0.68 | P = 0.21 | P = 0.21 | P < 0.001 | P = 0.002 | P = 0.002 | |||||
Model A adjusts for age at diagnosis (categorical), cancer stage (grouped linear), and year of diagnosis (categorical) with clustering.
Model B includes additionally adjustments to Model A for significant patient factors in patient-level analyses: race/ethnicity, income, BMI, CCI, and HER2.
Model C includes no additional adjustment to Model B because no statistically significant clinician or facility factors were identified in Model A. Hence, Model B and Model C show the same results.
Model D adjusts for age at diagnosis (categorical), cancer stage (grouped linear), and year of diagnosis (categorical) with clustering
Model E includes additionally adjustments to Model D for significant patient factors in patient-level analyses: race/ethnicity, income, BMI, CCI, HER2, grade, and ER/PR receptor status.
Model F includes additionally adjustments to Model E for significant clinician or facility factors from Model D: treatment facility urbanicity.
Abbreviations: Ref: Reference; BMI: body mass index; ER/PR: estrogen receptor/progesterone receptor; CCI: Charlson Comorbidity Index; HER2: human epidermal growth factor receptor 2
Note: Although treatment facility volume was significant in model D, it was excluded from model F due to collinearity with treatment facility urbanicity. Treatment facility urbanicity was also excluded from model F for treatment facility volume.
Table 3.
Clinician/facility level factors associated with FCDP <90%, stratified by BMI
| First Cycle Dose Proportion <90% | |||||||
|---|---|---|---|---|---|---|---|
| BMI <30kg/m2 | BMI ≥30kg/m2 | ||||||
| Characteristic | No n (%) |
Yes n (%) |
Adjusted for patient-level + clinician/facility with clusteringa PR (95% CI) |
No n (%) |
Yes n (%) |
Adjusted for patient-level + clinician/facility with clusteringa PR (95% CI) |
p-interaction |
| TOTAL | 5,150 (97%) | 149 (2.8%) | 2,632 (86%) | 424 (14%) | |||
| Clinician Gender | |||||||
| Male | 1,956 (38.0%) | 76 (51.0%) | Ref | 977 (37.1%) | 223 (52.6%) | Ref | P = 0.81 |
| Female | 3,194 (62.0%) | 73 (49.0%) | 0.65 (0.40-1.05) | 1,655 (62.9%) | 201 (47.4%) | 0.82 (0.57-1.18) | |
| P = 0.08 | P = 0.29 | ||||||
| Clinician Years Since Medical School | |||||||
| <10 years | 908 (17.6%) | 20 (13.4%) | Ref | 540 (20.5%) | 51 (12.0%) | Ref | P = 0.65 |
| 10-30 Years | 2,200 (42.7%) | 59 (39.6%) | 1.21 (0.69-2.11) | 1,120 (42.6%) | 169 (39.9%) | 1.71 (1.12-2.63) | |
| 21-30 Years | 1,235 (24.0%) | 41 (27.5%) | 1.35 (0.71-2.56) | 567 (21.5%) | 133 (31.4%) | 1.98(1.23-3.19) | |
| ≥30 Years | 807 (15.7%) | 29 (19.5%) | 1.45 (0.65-3.22) | 405 (15.4%) | 71 (16.7%) | 1.89 (1.10-3.24) | |
| P-trend = 0.38 | p-trend = 0.02 | ||||||
| Patient-Clinician Racial/Ethnic Concordance | |||||||
| Discordant | 3,149 (61.1%) | 94 (63.1%) | Ref | 1,931 (73.4%) | 296 (69.8%) | Ref | P = 0.63 |
| Concordant | 2,001 (38.9%) | 55 (36.9%) | 0.95 (0.65-1.37) | 701 (26.6%) | 128 (30.2%) | 1.01 (0.71-1.43) | |
| P = 0.77 | P = 0.95 | ||||||
| Average Clinician Patient Volume Per Year | |||||||
| ≤15 patients | 1,464 (28.4%) | 55 (36.9%) | Ref | 657 (25.0%) | 185 (43.6%) | Ref | P = 0.06 |
| 16-29 patients | 2,265 (44.0%) | 53 (35.6%) | 0.71 (0.42-1.19) | 1,146 (43.5%) | 165 (38.9%) | 0.70 (0.46-1.05) | |
| ≥30 patients | 1,421 (27.6%) | 41 (27.5%) | 0.89 (0.46-1.73) | 829 (31.5%) | 74 (17.5%) | 0.57 (0.37-0.88) | |
| p-trend = 0.69 | p-trend = 0.01 | ||||||
| Practice Size | |||||||
| ≤ 5 Oncologists | 489 (9.5%) | 20 (13.4%) | Ref | 184 (7.0%) | 67 (15.8%) | Ref | P = 0.51 |
| >5-10 Oncologists | 2,549 (49.5%) | 81 (54.4%) | 0.81 (0.42-1.56) | 1,270 (48.3%) | 201 (47.4%) | 0.57 (0.37-0.89) | |
| >10 Oncologists | 2,112 (41.0%) | 48 (32.2%) | 0.56 (0.29-1.10) | 1,178 (44.8%) | 156 (36.8%) | 0.50 (0.31-0.81) | |
| p-trend = 0.05 | p-trend = 0.04 | ||||||
| Treatment Facility Breast Cancer Patient Volume Per Year | |||||||
| <75 patients | 1,168 (22.7%) | 55 (36.9%) | Ref | 650 (24.7%) | 127 (30.0%) | Ref | P = 0.26 |
| 75-<200 patients | 3,473 (67.4%) | 85 (57.1%) | 0.59 (0.35-1.00) | 1,668 (63.4%) | 255 (60.1%) | 0.82 (0.55-1.23) | |
| ≥200 patients | 509 (9.9%) | 9 (6.0%) | 0.41 (0.17-0.96) | 314 (11.9%) | 42 (9.9%) | 0.70 (0.36-1.37) | |
| p-trend = 0.02 | p-trend = 0.23 | ||||||
| Treatment Facility Urbanicity | |||||||
| 100% urban | 4,879 (94.7%) | 144 (96.6%) | Ref | 2,427 (92.2%) | 403 (95.0%) | Ref | P = 0.89 |
| <100% urban | 271 (5.3%) | 5 (3.4%) | 0.55 (0.17-1.79) | 205 (7.8%) | 21 (5.0%) | 0.66 (0.28-1.54) | |
| P = 0.32 | P = 0.34 | ||||||
Model adjusted for age at diagnosis (categorical), cancer stage (grouped linear), year of diagnosis (categorical), race/ethnicity, income, Charlson score, and HER2 status.
Abbreviations: Ref: Reference; HER2: human epidermal growth factor receptor 2
Table 4.
Clinician/facility level factors associated with FCDP <90%, stratified by patient age at diagnosis
| First Cycle Dose Proportion <90% | |||||||
|---|---|---|---|---|---|---|---|
| <65 years | ≥65 years | ||||||
| Characteristic | No n (%) |
Yes n (%) |
Adjusted for patient-level + Clinician/facility with clusteringa PR (95% CI) |
No n (%) |
Yes n (%) |
Adjusted for patient-level + Clinician/facility with clusteringa PR (95% CI) |
p-interaction |
| TOTAL | 6,214 (94%) | 405 (6.1%) | 1,568 (90.3%) | 168 (9.7%) | |||
| Clinician Gender | |||||||
| Male | 2,380 (38.3%) | 212 (52.3%) | Ref | 553 (35.3%) | 87 (51.8%) | Ref | P = 0.98 |
| Female | 3,834 (61.7%) | 193 (47.7%) | 0.80 (0.54-1.17) | 1,015 (64.7%) | 81 (48.2%) | 0.73 (0.51-1.04) | |
| P = 0.25 | P = 0.08 | ||||||
| Clinician Years Since Medical School | |||||||
| <10 years | 1,169 (18.8%) | 50 (12.3%) | Ref | 279 (17.8%) | 21 (12.5%) | Ref | P = 0.82 |
| 10-30 Years | 2,622 (42.2%) | 154 (38.0%) | 1.60 (1.02-2.51) | 698 (44.5%) | 74 (44.0%) | 1.58 (0.91-2.75) | |
| 21-30 Years | 1,455 (23.4%) | 129 (31.9%) | 1.91 (1.19-3.08) | 347 (22.1%) | 45 (26.8%) | 1.60 (0.86-2.96) | |
| ≥30 Years | 968 (15.6%) | 72 (17.8%) | 1.89 (1.06-3.38) | 244 (15.6%) | 28 (16.7%) | 1.62 (0.81-3.21) | |
| P-trend = 0.03 | p-trend = 0.23 | ||||||
| Patient-Clinician Racial/Ethnic Concordance | |||||||
| Discordant | 4,057 (65.3%) | 271 (66.9%) | Ref | 1,023 (65.2%) | 119 (70.8%) | Ref | P = 0.28 |
| Concordant | 2,157 (34.7%) | 134 (33.1%) | 1.09 (0.80-1.47) | 545 (34.8%) | 49 (29.2%) | 0.84 (0.53-1.32) | |
| P = 0.59 | P = 0.44 | ||||||
| Average Clinician Patient Volume Per Year | |||||||
| ≤15 patients | 1,710 (27.5%) | 172 (42.5%) | Ref | 411 (26.2%) | 68 (40.5%) | Ref | P = 0.85 |
| 16-29 patients | 2,726 (43.9%) | 153 (37.8%) | 0.67 (0.45-1.02) | 685 (43.7%) | 65 (38.7%) | 0.68 (0.45-1.02) | |
| ≥30 patients | 1,778 (28.6%) | 80 (19.8%) | 0.65 (0.38-1.10) | 472 (30.1%) | 35 (20.8%) | 0.59 (0.39-0.90) | |
| p-trend = 0.07 | p-trend = 0.01 | ||||||
| Practice Size | |||||||
| ≤ 5 Oncologists | 516 (8.3%) | 60 (14.8%) | Ref | 157 (10.0%) | 27 (16.1%) | Ref | P = 0.69 |
| >5-10 Oncologists | 3,060 (49.2%) | 194 (47.9%) | 0.62 (0.39-0.99) | 759 (48.4%) | 88 (52.4%) | 0.64 (0.38-1.09) | |
| >10 Oncologists | 2,638 (42.5%) | 151 (37.3%) | 0.55 (0.34-0.90) | 652 (41.6%) | 53 (31.5%) | 0.44 (0.26-0.76) | |
| p-trend = 0.07 | p-trend = 0.003 | ||||||
| Treatment Facility Breast Cancer Patient Volume Per Year | |||||||
| <75 patients | 1,461 (23.5%) | 131 (32.4%) | Ref | 357 (22.8%) | 51 (30.4%) | Ref | P = 0.63 |
| 75-<200 patients | 4,104 (66.0%) | 241 (59.5%) | 0.76 (0.49-1.17) | 1,037 (66.1%) | 99 (58.9%) | 0.73 (0.48-1.11) | |
| ≥200 patients | 649 (10.4%) | 33 (8.1%) | 0.58 (0.31-1.11) | 174 (11.1%) | 18 (10.7%) | 0.70 (0.39-1.25) | |
| p-trend = 0.09 | p-trend = 0.15 | ||||||
| Treatment Facility Urbanicity | |||||||
| 100% urban | 5,831 (93.8%) | 387 (95.6%) | Ref | 1,475 (94.1%) | 160 (95.2%) | Ref | P = 0.86 |
| <100% urban | 383 (6.2%) | 18 (4.4%) | 0.62 (0.32-1.23) | 93 (5.9%) | 8 (4.8%) | 0.62 (0.23-1.67) | |
| P = 0.17 | P = 0.35 | ||||||
Model adjusted for year of diagnosis (categorical), cancer stage (grouped linear), race/ethnicity, income, BMI, Charlson score, and HER2 status.
Abbreviations: Ref: Reference; BMI: body mass index; HER2: human epidermal growth factor receptor 2
Table 5.
Clinician/facility level factors associated with FCDP <90%, stratified by calendar time
| First Cycle Dose Proportion <90% | |||||||
|---|---|---|---|---|---|---|---|
| 2006-2014 | 2015-2019 | ||||||
| Characteristic | No n (%) |
Yes n (%) |
Adjusted for patient-level + clinician/facility with clusteringa PR (95% CI) |
No n (%) |
Yes n (%) |
Adjusted for patient-level + clinician/facility with clusteringa PR (95% CI) |
p-interaction |
| TOTAL | 4,799 (91.9%) | 481 (9.1%) | 2,983 (97.0%) | 92 (3.0%) | |||
| Patient-Clinician Gender Concordance | |||||||
| Male | 2,263 (47.2%) | 273 (56.8%) | Ref | 670 (22.5%) | 26 (28.3%) | Ref | P = 0.98 |
| Female | 2,536 (52.8%) | 208 (43.2%) | 0.74 (0.52-1.07) | 2,313 (77.5%) | 66 (71.7%) | 0.76 (0.34-1.74) | |
| P = 0.11 | P = 0.52 | ||||||
| Clinician Years Since Medical School | |||||||
| <10 years | 1,028 (21.4%) | 59 (12.3%) | Ref | 420 (14.1%) | 12 (13.0%) | Ref | P = 0.26 |
| 10-30 Years | 1,826 (38.0%) | 183 (38.0%) | 1.77 (1.13-2.79) | 1,494 (50.1%) | 45 (48.9%) | 1.04 (0.48-2.26) | |
| 21-30 Years | 1,190 (24.8%) | 161 (33.5%) | 2.26 (1.35-3.80) | 612 (20.5%) | 13 (14.1%) | 0.87 (0.37-2.07) | |
| ≥30 Years | 755 (15.7%) | 78 (16.2%) | 1.75 (0.99-3.10) | 457 (15.3%) | 22 (23.9%) | 1.74 (0.47-6.41) | |
| p-trend = 0.02 | p-trend = 0.49 | ||||||
| Patient-Clinician Racial/Ethnic Concordance | |||||||
| Discordant | 3,090 (64.4%) | 326 (67.8%) | Ref | 1,990 (66.7%) | 64 (69.6%) | Ref | P = 0.90 |
| Concordant | 1,709 (35.6%) | 155 (32.2%) | 1.03 (0.76-1.41) | 993 (33.3%) | 28 (30.4%) | 0.91 (0.47-1.76) | |
| P = 0.84 | P = 0.79 | ||||||
| Average Clinician Patient Volume Per Year | |||||||
| ≤15 patients | 1,667 (34.7%) | 225 (46.8%) | Ref | 454 (15.2%) | 15 (16.3%) | Ref | P = 0.59 |
| 16-29 patients | 2,113 (44.0%) | 178 (37.0%) | 0.63 (0.42-0.93) | 1,298 (43.5%) | 40 (43.5%) | 1.09 (0.48-2.47) | |
| ≥30 patients | 1,019 (21.2%) | 78 (16.2%) | 0.58 (0.35-0.96) | 1,231 (41.3%) | 37 (40.2%) | 0.98 (0.36-2.67) | |
| p-trend = 0.02 | p-trend = 0.92 | ||||||
| Practice Size | |||||||
| ≤ 5 Oncologists | 464 (9.7%) | 76 (15.8%) | Ref | 209 (7.0%) | 11 (12.0%) | Ref | P = 0.21 |
| >5-10 Oncologists | 2,352 (49.0%) | 225 (46.8%) | 0.56 (0.36-0.87) | 1,467 (49.2%) | 57 (62.0%) | 0.80 (0.27-2.38) | |
| >10 Oncologists | 1,983 (41.3%) | 180 (37.4%) | 0.53 (0.33-0.85) | 1,307 (43.8%) | 24 (26.0%) | 0.38 (0.13-1.12) | |
| p-trend = 0.07 | p-trend = 0.01 | ||||||
| Treatment Facility Breast Cancer Patient Volume Per Year | |||||||
| <75 patients | 1,075 (22.4%) | 137 (28.5%) | Ref | 743 (24.9%) | 45 (48.9%) | Ref | P = 0.24 |
| 75-<200 patients | 3,261 (68.0%) | 301 (62.6%) | 0.89 (0.59-1.33) | 1,880 (63.0%) | 39 (42.4%) | 0.41 (0.19-0.88) | |
| ≥200 patients | 463 (9.6%) | 43 (8.9%) | 0.72 (0.38-1.35) | 360 (12.1%) | 8 (8.7%) | 0.39 (0.18-0.86) | |
| p-trend = 0.30 | p-trend = 0.03 | ||||||
Model adjusted for age at diagnosis (categorical), cancer stage (grouped linear), race/ethnicity, income, BMI, Charlson score, and HER2.
Abbreviations: Ref: Reference; BMI: body mass index; HER2: human epidermal growth factor receptor 2
Note: Treatment practice urbanicity was redacted from this table due to small cell counts. No significant associations were observed between treatment facility urbanicity and FCDR in either time period.
Table 6.
Clinician/facility level factors associated with ARDI <90%, stratified by BMI
| Average Relative Dose Intensity <90% | |||||||
|---|---|---|---|---|---|---|---|
| BMI <30 | BMI ≥30 | ||||||
| Characteristic | No n (%) |
Yes n (%) |
Adjusted for patient-level + clinician/facility with clusteringa PR (95% CI) |
No n (%) |
Yes n (%) |
Adjusted for patient-level + clinician/facility with clusteringa PR (95% CI) |
p-interaction |
| TOTAL | 4,528 (85%) | 771 (15%) | 2,199 (72%) | 857 (28%) | |||
| Clinician Gender | |||||||
| Male | 1,716 (37.9%) | 316 (41.0%) | Ref | 801 (36.4%) | 399 (46.6%) | Ref | P = 0.06 |
| Female | 2,812 (62.1%) | 455 (59.0%) | 0.88 (0.71-1.09) | 1,398 (63.6%) | 458 (53.4%) | 0.84 (0.69-1.03) | |
| P = 0.24 | P = 0.09 | ||||||
| Clinician Years Since Medical School | |||||||
| <10 years | 809 (17.9%) | 119 (15.4%) | Ref | 456 (20.7%) | 135 (15.8%) | Ref | P = 0.63 |
| 10-30 Years | 1,935 (42.7%) | 324 (42.0%) | 1.08 (0.85-1.36) | 943 (42.9%) | 346 (40.4%) | 1.20 (0.96-1.51) | |
| 21-30 Years | 1,080 (23.9%) | 196 (25.4%) | 1.15 (0.89-1.49) | 465 (21.2%) | 235 (27.4%) | 1.38 (1.10-1.73) | |
| ≥30 Years | 704 (15.5%) | 132 (17.1%) | 1.22 (0.92-1.63) | 335 (15.2%) | 141 (16.4%) | 1.36 (1.03-1.80) | |
| p-trend = 0.16 | p-trend = 0.02 | ||||||
| Patient-Clinician Racial/Ethnic Concordance | |||||||
| Discordant | 2,759 (60.9%) | 484 (62.8%) | Ref | 1,617 (73.5%) | 610 (71.2%) | Ref | P = 0.28 |
| Concordant | 1,769 (39.1%) | 287 (37.2%) | 0.95 (0.81-1.12) | 582 (26.5%) | 247 (28.8%) | 1.01 (0.83-1.22) | |
| P = 0.56 | P = 0.95 | ||||||
| Average Clinician Patient Volume Per Year | |||||||
| ≤15 patients | 1,269 (28.0%) | 250 (32.4%) | Ref | 537 (24.4%) | 305 (35.6%) | Ref | P = 0.01 |
| 16-29 patients | 2,002 (44.2%) | 316 (41.0%) | 0.88 (0.70-1.12) | 953 (43.3%) | 358 (41.8%) | 0.82 (0.66-1.02) | |
| ≥30 patients | 1,257 (27.8%) | 205 (26.6%) | 0.86 (0.67-1.12) | 709 (32.2%) | 194 (22.6%) | 0.68 (0.54-0.85) | |
| p-trend = 0.26 | p-trend = 0.001 | ||||||
| Practice Size | |||||||
| ≤ 5 Oncologists | 432 (9.5%) | 77 (10.0%) | Ref | 160 (7.3%) | 91 (10.6%) | Ref | P = 0.44 |
| >5-10 Oncologists | 2,233 (49.3%) | 397 (51.5%) | 0.96 (0.68-1.35) | 1,061 (48.2%) | 410 (47.8%) | 0.76 (0.58-1.01) | |
| >10 Oncologists | 1,863 (41.1%) | 297 (38.5%) | 0.90 (0.64-1.25) | 978 (44.5%) | 356 (41.5%) | 0.78 (0.60-1.02) | |
| p-trend = 0.44 | p-trend = 0.34 | ||||||
| Treatment Facility Breast Cancer Patient Volume Per Year | |||||||
| <75 patients | 998 (22.0%) | 225 (29.2%) | Ref | 513 (23.3%) | 264 (30.8%) | Ref | P = 0.76 |
| 75-<200 patients | 3,090 (68.2%) | 468 (60.7%) | 0.74 (0.59-0.92) | 1,424 (64.8%) | 499 (58.2%) | 0.79 (0.65-0.96) | |
| ≥200 patients | 440 (9.7%) | 78 (10.1%) | 0.86 (0.60-1.24) | 262 (11.9%) | 94 (11.0%) | 0.79 (0.57-1.09) | |
| p-trend = 0.09 | p-trend = 0.05 | ||||||
| Treatment Facility Urbanicity | |||||||
| 100% urban | 4,319 (95.4%) | 704 (91.3%) | Ref | 2,050 (93.2%) | 780 (91.0%) | Ref | P = 0.02 |
| <100% urban | 209 (4.6%) | 67 (8.7%) | 1.67 (1.24-2.26) | 149 (6.8%) | 77 (9.0%) | 1.21 (1.02-1.45) | |
| P = 0.001 | P = 0.03 | ||||||
Model adjusted for age at diagnosis (categorical), cancer stage (grouped linear), year of diagnosis (categorical), race/ethnicity, income, Charlson score, HER2 status, grade, ER/PR status, and facility urbanicity.
Abbreviations: Ref: Reference; ER/PR: estrogen receptor/progesterone receptor; HER2: human epidermal growth factor receptor 2. Note: Treatment facility volume was excluded from all adjusted models due to collinearity with practice urbanicity. Additionally, treatment facility volume adjusted models excluded practice urbanicity.
Table 7.
Clinician/facility level factors associated with ARDI <90%, stratified by patient age at diagnosis
| Average Relative Dose Intensity <90% | |||||||
|---|---|---|---|---|---|---|---|
| <65 years | ≥65 years | ||||||
| Characteristic | No n (%) |
Yes n (%) |
Adjusted for patient-level + clinician/facility with clusteringa PR (95% CI) |
No n (%) |
Yes n (%) |
Adjusted for patient-level + clinician/facility with clusteringa PR (95% CI) |
p-interaction |
| TOTAL | 5,408 (82%) | 1,211 (18%) | 1,319 (76%) | 417 (24%) | |||
| Clinician Gender | |||||||
| Male | 2,065 (38.2%) | 527 (43.5%) | Ref | 452 (34.3%) | 188 (45.1%) | Ref | P = 0.26 |
| Female | 3,343 (61.8%) | 684 (56.5%) | 0.87 (0.72-1.06) | 867 (65.7%) | 229 (54.9%) | 0.81 (0.64-1.01) | |
| P = 0.16 | P = 0.07 | ||||||
| Clinician Years Since Medical School | |||||||
| <10 years | 1,026 (19.0%) | 193 (15.9%) | Ref | 239 (18.1%) | 61 (14.6%) | Ref | P = 0.76 |
| 10-30 Years | 2,296 (42.5%) | 480 (39.6%) | 1.12 (0.90-1.38) | 582 (44.1%) | 190 (45.6%) | 1.25 (0.96-1.62) | |
| 21-30 Years | 1,253 (23.2%) | 331 (27.3%) | 1.30 (1.06-1.60) | 292 (22.1%) | 100 (24.0%) | 1.20 (0.88-1.63) | |
| ≥30 Years | 833 (15.4%) | 207 (17.1%) | 1.32 (1.03-1.68) | 206 (15.6%) | 66 (15.8%) | 1.24 (0.86-1.80) | |
| p-trend = 0.01 | p-trend = 0.39 | ||||||
| Patient-Clinician Racial/Ethnic Concordance | |||||||
| Discordant | 3,518 (65.1%) | 810 (66.9%) | Ref | 858 (65.0%) | 284 (68.1%) | Ref | P = 0.58 |
| Concordant | 1,890 (34.9%) | 401 (33.1%) | 1.02 (0.87-1.19) | 461 (35.0%) | 133 (31.9%) | 0.93 (0.74-1.17) | |
| P = 0.85 | P = 0.55 | ||||||
| Average Clinician Patient Volume Per Year | |||||||
| ≤15 patients | 1,485 (27.5%) | 397 (32.8%) | Ref | 321 (24.3%) | 158 (37.9%) | Ref | P = 0.06 |
| 16-29 patients | 2,369 (43.8%) | 510 (42.1%) | 0.89 (0.73-1.09) | 586 (44.4%) | 164 (39.3%) | 0.73 (0.57-0.92) | |
| ≥30 patients | 1,554 (28.7%) | 304 (25.1%) | 0.80 (0.63-1.02) | 412 (31.2%) | 95 (22.8%) | 0.64 (0.48-0.84) | |
| p-trend = 0.07 | p-trend = 0.001 | ||||||
| Practice Size | |||||||
| ≤ 5 Oncologists | 462 (8.5%) | 114 (9.4%) | Ref | 130 (9.9%) | 54 (12.9%) | Ref | P = 0.17 |
| >5-10 Oncologists | 2,655 (49.1%) | 599 (49.5%) | 0.89 (0.69-1.15) | 639 (48.4%) | 208 (49.9%) | 0.76 (0.53-1.07) | |
| >10 Oncologists | 2,291 (42.4%) | 498 (41.1%) | 0.90 (0.71-1.13) | 550 (41.7%) | 155 (37.2%) | 0.67 (0.48-0.94) | |
| p-trend = 0.64 | p-trend = 0.03 | ||||||
| Treatment Facility Breast Cancer Patient Volume Per Year | |||||||
| <75 patients | 1,232 (22.8%) | 360 (29.7%) | Ref | 279 (21.2%) | 129 (30.9%) | Ref | P = 0.68 |
| 75-<200 patients | 3,622 (67.0%) | 723 (59.7%) | 0.79 (0.66-0.96) | 892 (67.6%) | 244 (58.5%) | 0.69 (0.53-0.88) | |
| 200+ patients | 554 (10.2%) | 128 (10.6%) | 0.86 (0.65-1.14) | 148 (11.2%) | 44 (10.6%) | 0.73 (0.53-1.00) | |
| p-trend = 0.08 | p-trend = 0.02 | ||||||
| Treatment Facility Urbanicity | |||||||
| 100% urban | 5,115 (94.6%) | 1,103 (91.1%) | Ref | 1,254 (95.1%) | 381 (91.4%) | Ref | P = 0.92 |
| <100% urban | 293 (5.4%) | 108 (8.9%) | 1.36 (1.05-1.78) | 65 (4.9%) | 36 (8.6%) | 1.42 (0.87-2.33) | |
| P = 0.02 | P = 0.16 | ||||||
Model adjusted for cancer stage (grouped linear), year of diagnosis (categorical), race/ethnicity, income, BMI, Charlson score, HER2 status, grade, ER/PR status, and facility urbanicity
Abbreviations: Ref: Reference; BMI: body mass index; ER/PR: estrogen receptor/progesterone receptor; HER2: human epidermal growth factor receptor 2. Note: Treatment facility volume was excluded from all adjusted models due to collinearity with practice urbanicity. Additionally, treatment facility volume adjusted models excluded practice urbanicity.
Table 8.
Clinician/facility level factors associated with ARDI <90%, stratified by calendar time
| Average Relative Dose Intensity <90% | |||||||
|---|---|---|---|---|---|---|---|
| 2006-2014 | 2015-2019 | ||||||
| Characteristic | No n (%) |
Yes n (%) |
Adjusted for patient-level +clinician/facility with clusteringa PR (95% CI) |
No n (%) |
Yes n (%) |
Adjusted for patient-level + clinician/facility with clusteringa PR (95% CI) |
p-interaction |
| TOTAL | 4,146 (79%) | 1,134 (21%) | 2,581 (84%) | 494 (16%) | |||
| Clinician Gender | |||||||
| Male | 1,940 (46.8%) | 596 (52.6%) | Ref | 577 (22.4%) | 119 (24.1%) | Ref | P = 0.62 |
| Female | 2,206 (53.2%) | 538 (47.4%) | 0.83 (0.68-1.03) | 2,004 (77.6%) | 375 (75.9%) | 0.88 (0.68-1.15) | |
| P = 0.09 | P = 0.36 | ||||||
| Clinician Years Since Medical School | |||||||
| <10 years | 910 (22.0%) | 177 (15.6%) | Ref | 355 (13.8%) | 77 (15.6%) | Ref | P = 0.01 |
| 10-30 Years | 1,577 (38.0%) | 432 (38.1%) | 1.32 (1.03-1.68) | 1,301 (50.4%) | 238 (48.2%) | 0.85 (0.66-1.09) | |
| 21-30 Years | 1,017 (24.5%) | 334 (29.5%) | 1.56 (1.24-1.98) | 528 (20.5%) | 97 (19.6%) | 0.85 (0.61-1.20) | |
| ≥30 Years | 642 (15.5%) | 191 (16.8%) | 1.45 (1.05-1.98) | 397 (15.4%) | 82 (16.6%) | 0.96 (0.67-1.36) | |
| p-trend = 0.01 | p-trend = 0.98 | ||||||
| Patient-Clinician Racial/Ethnic Concordance | |||||||
| Discordant | 2,650 (63.9%) | 766 (67.6%) | Ref | 1,726 (66.9%) | 328 (66.4%) | Ref | P = 0.11 |
| Concordant | 1,496 (36.1%) | 368 (32.4%) | 0.94 (0.79-1.12) | 855 (33.1%) | 166 (33.6%) | 1.18 (0.95-1.47) | |
| P = 0.50 | P = 0.13 | ||||||
| Average Clinician Patient Volume Per Year | |||||||
| ≤15 patients | 1,415 (34.1%) | 477 (42.1%) | Ref | 391 (15.1%) | 78 (15.8%) | Ref | P = 0.13 |
| 16-29 patients | 1,846 (44.5%) | 445 (39.2%) | 0.77 (0.62-0.97) | 1,109 (43.0%) | 229 (46.4%) | 1.02 (0.79-1.32) | |
| ≥30 patients | 885 (21.4%) | 212 (18.7%) | 0.74 (0.56-0.97) | 1,081 (41.9%) | 187 (37.9%) | 0.85 (0.63-1.14) | |
| p-trend = 0.02 | p-trend = 0.17 | ||||||
| Practice Size | |||||||
| ≤ 5 Oncologists | 408 (9.8%) | 132 (11.6%) | Ref | 184 (7.1%) | 36 (7.3%) | Ref | P = 0.28 |
| >5-10 Oncologists | 2,031 (49.0%) | 546 (48.2%) | 0.79 (0.59-1.06) | 1,263 (48.9%) | 261 (52.8%) | 0.95 (0.68-1.34) | |
| >10 Oncologists | 1,707 (41.2%) | 456 (40.2%) | 0.82 (0.62-1.08) | 1,134 (43.9%) | 197 (39.9%) | 0.85 (0.62-1.16) | |
| p-trend = 0.45 | p-trend = 0.19 | ||||||
| Treatment Facility Breast Cancer Patient Volume Per Year | |||||||
| <75 patients | 888 (21.4%) | 324 (28.6%) | Ref | 623 (24.1%) | 165 (33.4%) | Ref | P = 0.66 |
| 75-<200 patients | 2,871 (69.3%) | 691 (60.9%) | 0.79 (0.64-0.99) | 1,643 (63.7%) | 276 (55.9%) | 0.72 (0.58-0.90) | |
| ≥200 patients | 387 (9.3%) | 119 (10.5%) | 0.88 (0.61-1.28) | 315 (12.2%) | 53 (10.7%) | 0.72 (0.53-0.97) | |
| p-trend = 0.19 | p-trend = 0.01 | ||||||
| Treatment Facility Urbanicity | |||||||
| 100% urban | 3,962 (95.6%) | 1,043 (92.0%) | Ref | 2,407 (93.3%) | 441 (89.3%) | Ref | P = 0.86 |
| <100% urban | 184 (4.4%) | 91 (8.0%) | 1.36 (1.11-1.68) | 174 (6.7%) | 53 (10.7%) | 1.34 (0.93-1.91) | |
| P = 0.004 | P = 0.11 | ||||||
Model adjusted for age at diagnosis (categorical), cancer stage (grouped linear), race/ethnicity, income, BMI, Charlson score, HER2 status, grade, ER/PR status, and facility urbanicity
Abbreviations: Ref: Reference; BMI: body mass index; ER/PR: estrogen receptor/progesterone receptor; HER2: human epidermal growth factor receptor 2. Note: Treatment facility volume was excluded from all adjusted models due to collinearity with practice urbanicity. Additionally, treatment facility volume adjusted models excluded practice urbanicity.
Patients whose clinicians had greater years since medical school were more likely to receive a FCDP<90% (PR≥30 vs <10 years:1.78; 95% CI:1.07-2.96; p-trend=0.03) and an ARDI<90% (PR≥30 vs <10 years:1.29; 95% CI:1.02-1.65; p-trend=0.03). In the time-stratified secondary analyses, among patients diagnosed before 2015 (i.e., before KPNC’s transition toward a subspecialized model of oncology care), clinician years since medical school was significantly associated with ARDI<90% (PR≥30 vs <10 years:1.45; 95% CI:1.05-1.98; p-trend=0.01), whereas this was no longer seen for patients diagnosed in 2015 onward (PR≥30 vs <10 years:0.96; 95% CI:0.67-1.36, p-trend=0.98) (p-interaction=0.01; Table 8). No significant interactions were observed across BMI- or age-stratification for years since medical school with ARDI<90%. Clinician gender and patient-clinician racial/ethnic concordance were not significantly associated with FCDP<90% or ARDI<90%.
When examining facility-level factors, patients treated at less urban facilities were more likely to have an ARDI<90% (PR<100% vs 100% urban:1.38; 95% CI:1.12-1.69; p-trend=0.002). The association between urbanicity and ARDI<90% was reflected in both BMI groups, but primarily driven by the treatment of non-obese patients (BMI <30: PR<100% vs 100% urban:1.67; 95% CI:1.24-2.26; p-trend=0.001 versus BMI ≥30: PR<100% vs 100% urban: 1.21; 95% CI:1.02-1.45; p-trend=0.03) (p-interaction=0.02). No significant interactions were observed across age- or time-stratification for urbanicity with ARDI<90%.
Patients treated within larger oncology practices were less likely to receive FCDP<90% (PR≥10 vs ≤5 oncologists:0.53; 95% CI:0.34-0.83; p-trend=0.02) whereas patients treated at higher-volume treatment facilities were less likely to have an ARDI<90% (PR≥200 vs <75 patients:0.85; 95% CI:0.66-1.08; p-trend=0.03). These associations did not significantly differ across subgroups in BMI-, age-, or time-stratified analyses.
Given heterogeneity in carboplatin dosing over the study period and our use of actual body weight for dosing calculations, we did a sensitivity analysis on the fully-adjusted models for FCDP/ARDI<90% excluding carboplatin. Results were very comparable, with only one association changing statistical significance, although effect estimates remained similar: greater years since medical school was no longer significantly associated with FCDP <90% (PR≥30 vs ≤10 years:1.73; 95% CI:0.98-3.05; p-trend=0.06).
Discussion
In this paper, we show that clinician- and facility-level factors were associated with chemotherapy dose reductions in women with stages I-IIIA BC. We evaluated dose reductions through both FCDP and ARDI to ascertain how prescribing behavior changes throughout the course of treatment. If a dose reduction occurs at the start of chemotherapy, reflected through FCDP, then it could be due to a clinician’s concern for the patient’s ability to tolerate treatment or the use of adjusted ideal body weight instead of actual body weight. Whereas dose reductions occurring throughout the course of chemotherapy, reflected in the RDI, could additionally reflect responsive prescribing to address toxicities and symptoms. We found that greater number of clinician years since medical school and treatment at less urban treatment facilities were associated with an increased likelihood of dose reduction. Conversely, higher average annual volume of similarly-staged BC patients for both the clinician and treatment facility, and larger practice size were associated with a lower likelihood of dose reduction. These findings underscore the importance of both clinician experience and practice setting in influencing chemotherapy dosing decisions.
Patients treated by clinicians with ≥30 years since medical school were more likely to receive dose reductions at the outset and over the course of treatment, as compared to those treated by less experienced clinicians. Clinicians with greater years in practice may have more anecdotal evidence of poor outcomes with full dose chemotherapy and hence prefer to start patients on a lower chemotherapy dose to balance efficacy and toxicity. Additionally, these clinicians may see patients with more complex medical histories who may be more likely to receive dose modifications at the start of chemotherapy to avoid toxicities as well as more likely to have responsive dose modifications throughout the course of chemotherapy given their increased risk of experiencing toxicities. Despite careful adjustment for patient factors such as BMI and comorbidity, we cannot dismiss the possibility of some residual confounding by patient complexity and cannot infer the clinical appropriateness of these dose reductions. In the time-stratified analyses, years since medical school was no longer significantly associated with ARDI<90% post-2015. In subspecialization, fewer clinicians provide care for a given cancer type, and clinicians who provide care for that cancer are more likely to have larger patient panels with that cancer type. Perhaps as clinicians began to have increasingly homogenous patient panels, there was an increased level of familiarity with managing toxicities and dosing guidelines, contributing towards fewer reductions.
We observed that increasing volume of similarly-staged patients at both the clinician- and facility-level was associated with fewer dose reductions. Patients of clinicians who saw more stages I-IIIA BC patients annually were less likely to experience dose reductions. Notably, the treatment of patients with obesity largely drove the association seen between higher clinician volumes and decreased likelihood of an ARDI reduction. These high-volume clinicians may have more experience treating a diverse panel of BC patients and therefore feel more confident prescribing patients the full recommended dose, knowing that they can manage toxicities if they were to arise. Finally, increasing the annual volume of similarly-staged patients at treatment facilities was significantly associated with a lower likelihood of having an ARDI<90%. This may be due to higher-volume facilities having increased access to resources to assist with scheduling, thereby minimizing treatment interruptions, or established protocols to manage toxicities and reduce dose delays. Of note, larger practice size was also associated with a decreased likelihood of FCDP<90%, potentially reflecting the benefits of collective clinician experience or increased peer oversight within larger practices.
Finally, treatment facility urbanicity was inversely associated with ARDI<90%. This association was also reflected in the BMI-stratified analysis, primarily among the non-obese patients treated at less urban treatment facilities. Although the urbanicity of “less urban” facilities may vary, we generally viewed them to be in locations that had lower population density (compared to the 100% urban facilities), and potentially more difficult for patients to access through public transportation. Hence, these patients may have experienced dose reductions and delays contributing to a lower ARDI due to concerns over delayed access to treatment of toxicities. The generalizability of these findings to truly rural settings is limited. Future work that includes facilities across a more diverse spectrum of urbanicity and other aspects of geographic accessibility (e.g., patient distance from treatment facility, access to public transportation) would be helpful in better understanding these relationships.
This study has several strengths, including a large sample size of women with detailed chemotherapy information. We used these data to determine the intended regimen at the start of chemotherapy across all drug combinations described for stages I-IIIA BC in the NCCN guidelines, making this analysis unique among real-world data studies. Although this study was conducted within a single integrated healthcare system with all insured participants, prior studies have demonstrated that the patient population is demographically similar to the base population, and the use of this population uniquely reduces potential for confounding by insurance coverage.42 Even so, we recognize that generalizability may be limited to other health care systems and/or clinic models, given differences in degree of subspecialization among clinicians and other facility or systems-specific differences. Given the limited role of advanced practice providers (APPs) at KPNC oncology clinics during this time frame and setting, we have not accounted for this variable in our analyses. However, this may represent an important difference with other settings and may limit the generalizability of our findings to other settings where APPs play a role in chemotherapy management. In this study we have elected to focus on dosing, but future work may evaluate guideline concordance more broadly. A limitation of this analysis is the exclusion of patients who received neoadjuvant chemotherapy; however, their small numbers would have rendered the subgroup underpowered on its own and their inclusion risked inappropriate generalization of the findings. Additionally, our results capture dose reduction (and inherently delay in the ARDI calculations), but not other changes to chemotherapy delivery (e.g., early discontinuation). We are also unable to determine whether dose reductions were clinically appropriate without additional laboratory and patient reported outcomes data. This would be an important avenue for future research given that clinicians may utilize dose reductions doses or delay treatments as an alternative to early discontinuation, which is known to detrimentally impact survival.43,44 Furthermore, our data only captures patients with stages I-IIIA BC and does not reflect a clinician’s total patient caseload or a facility’s total patient volume. Facility-level factors such as practice size and treatment facility volume were calculated as static values for the study period and do not capture potential year-to-year variability due to new facility openings, renovations, or overall growth in KPNC membership. Although we accounted for clustering at the clinical-level, we cannot exclude the possibility of residual confounding from unmeasured patient-level characteristics that differ across locations or facilities.
In summary, this study underscores the complex interplay of factors influencing chemotherapy dosing decisions in stages I-IIIA BC with a focus on the role of clinician- and facility-level factors. Given the evolving evidence over the study period, there was likely heterogeneity in views on best practices regarding dose reductions. We observed that the associations between clinician- and facility-level factors and dose reductions may vary in certain subsets of patients, such as those with obesity. The results of our study have several potential implications related to practice structure and healthcare delivery. For example, system-level changes, such as adoption of a subspecialized model of oncology care, may mitigate the impact of some of these factors on chemotherapy dosing decisions. Furthermore, practice restructuring to increase volumes of similarly-staged BC patients and practice size may impact chemotherapy dosing. These findings can inform strategies to optimize delivery of appropriate chemotherapy and improve outcomes for BC patients.
Supplementary Material
Acknowledgements:
The funder did not play a role in the design of the study; the collection, analysis, and interpretation of the data; the writing of the manuscript; and the decision to submit the manuscript for publication.
Funding:
This work was supported by the National Cancer Institute at the National Institutes of Health (R37CA222793, U24CA171524, U01CA195565, P01CA154292, and P30CA008748; R50CA211115 to E.J.A.B.; and T32CA275764 to Y.S.); and the Geoffrey Beene Cancer Research Center at Memorial Sloan Kettering Cancer Center.
Footnotes
Conflict of Interest:
V.B., who is a JNCI Associate Editor and co-author on this paper, was not involved in the editorial review or decision to publish the manuscript. Dr. Bandera served as a member of Pfizer’s Advisory Board to enhance minority participation in clinical trials (8/2002-8/2023). Dr. Liu served as a member of Pfizer’s think tank on real-world evidence sponsored by Pfizer on 11/28/23. Dr. Liu’s research funding (institutional) unrelated to this current work: Genentech, AstraZeneca, Exact Sciences, Biotheranostics, BeiGene, and Bristol Myers Squibb. Dr. Wang is a former employee of Daiichi Sankyo, Inc (6/2022–4/2023). Other authors have no conflict of interest to declare.
Ethics statement: The procedures followed in this study were in accordance with the ethical standards of the Helsinki Declaration. Institutional Review Board approval was obtained from all sites (Memorial Sloan Kettering Cancer Center, Rutgers University, Kaiser Permanente Northern California, and Kaiser Permanente Washington) and a waiver of consent to use patient data was obtained from Kaiser Permanente Northern California and Kaiser Permanente Washington.
Data availability:
Data contain potentially identifiable information such as dates of events that cannot be shared openly without appropriate human subjects’ approval and data use agreements, but the data that support the findings of this study are available from the corresponding author upon reasonable request.
References
- 1.Verrill M. Chemotherapy for early-stage breast cancer: a brief history. Br J Cancer. Sep 2009;101 Suppl 1(Suppl 1):S2–5. doi: 10.1038/sj.bjc.6605268 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Hortobagyi GN. Developments in chemotherapy of breast cancer. Cancer. 2000;88(S12):3073–3079. doi: 10.1002/1097-0142(20000615)88:12+<3073::AID-CNCR26>3.0.CO;2-R [DOI] [PubMed] [Google Scholar]
- 3.Bhimani J, O'Connell K, Ergas IJ, et al. Trends in chemotherapy use for early-stage breast cancer from 2006 to 2019. Breast Cancer Res. Jun 13 2024;26(1):101. doi: 10.1186/s13058-024-01822-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.NCCN Clinical Practice Guidelines in Oncology: Breast Cancer. National Comprehensive Cancer Network. 2024. https://www.nccn.org/guidelines/guidelines-detail?category=1&id=1419
- 5.Lyman GH. Impact of chemotherapy dose intensity on cancer patient outcomes. J Natl Compr Canc Netw. Jan 2009;7(1):99–108. doi: 10.6004/jnccn.2009.0009 [DOI] [PubMed] [Google Scholar]
- 6.Nielson CM, Bylsma LC, Fryzek JP, Saad HA, Crawford J. Relative Dose Intensity of Chemotherapy and Survival in Patients with Advanced Stage Solid Tumor Cancer: A Systematic Review and Meta-Analysis. Oncologist. Sep 2021;26(9):e1609–e1618. doi: 10.1002/onco.13822 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Havrilesky LJ, Reiner M, Morrow PK, Watson H, Crawford J. A review of relative dose intensity and survival in patients with metastatic solid tumors. Crit Rev Oncol Hematol. Mar 2015;93(3):203–10. doi: 10.1016/j.critrevonc.2014.10.006 [DOI] [PubMed] [Google Scholar]
- 8.Chang J Chemotherapy dose reduction and delay in clinical practice. evaluating the risk to patient outcome in adjuvant chemotherapy for breast cancer. Eur J Cancer. Apr 2000;36 Suppl 1:S11–4. doi: 10.1016/s0959-8049(99)00259-2 [DOI] [PubMed] [Google Scholar]
- 9.Denduluri N, Patt DA, Wang Y, et al. Dose Delays, Dose Reductions, and Relative Dose Intensity in Patients With Cancer Who Received Adjuvant or Neoadjuvant Chemotherapy in Community Oncology Practices. J Natl Compr Canc Netw. Nov 2015;13(11):1383–93. doi: 10.6004/jnccn.2015.0166 [DOI] [PubMed] [Google Scholar]
- 10.Lyman GH, Dale DC, Crawford J. Incidence and predictors of low dose-intensity in adjuvant breast cancer chemotherapy: a nationwide study of community practices. J Clin Oncol. Dec 15 2003;21(24):4524–31. doi: 10.1200/jco.2003.05.002 [DOI] [PubMed] [Google Scholar]
- 11.Griggs JJ, Culakova E, Sorbero ME, et al. Effect of patient socioeconomic status and body mass index on the quality of breast cancer adjuvant chemotherapy. J Clin Oncol. Jan 20 2007;25(3):277–84. doi: 10.1200/jco.2006.08.3063 [DOI] [PubMed] [Google Scholar]
- 12.Griggs JJ, Sorbero ME, Stark AT, Heininger SE, Dick AW. Racial disparity in the dose and dose intensity of breast cancer adjuvant chemotherapy. Breast Cancer Res Treat. Sep 2003;81(1):21–31. doi: 10.1023/a:1025481505537 [DOI] [PubMed] [Google Scholar]
- 13.Shayne M, Crawford J, Dale DC, Culakova E, Lyman GH. Predictors of reduced dose intensity in patients with early-stage breast cancer receiving adjuvant chemotherapy. Breast Cancer Res Treat. Dec 2006;100(3):255–62. doi: 10.1007/s10549-006-9254-4 [DOI] [PubMed] [Google Scholar]
- 14.Griggs JJ, Sorbero ME, Lyman GH. Undertreatment of obese women receiving breast cancer chemotherapy. Arch Intern Med. Jun 13 2005;165(11):1267–73. doi: 10.1001/archinte.165.11.1267 [DOI] [PubMed] [Google Scholar]
- 15.Bozcuk H, Yildiz M, Uçar S, Mutlu H, Coşkun HŞ. The Correlates of Dose Reduction in Chemotherapy for Patients with Common Cancers. a Prospective Study. Annals of Oncology. 2014;25:iv80. doi: 10.1093/annonc/mdu326.74 [DOI] [Google Scholar]
- 16.Lyman GH, Kuderer NM, Poniewierski MS, et al. Factors associated with reductions in chemotherapy dose intensity: Impact of clinical, sociodemographic and practice setting. Journal of Clinical Oncology. 2007;25(18_suppl):6550–6550. doi: 10.1200/jco.2007.25.18_suppl.6550 [DOI] [Google Scholar]
- 17.Becker D, Hershman D, Insel B, Neugut A. Relationship between physician characteristics and the use of newer chemotherapy for the treatment of colon cancer. Meeting Abstract. Journal of Clinical Oncology. 2011;29(15):6106. doi: 10.1200/jco.2011.29.15_suppl.6106 [DOI] [Google Scholar]
- 18.Hajjaj FM, Salek MS, Basra MK, Finlay AY. Non-clinical influences on clinical decision-making: a major challenge to evidence-based practice. J R Soc Med. May 2010;103(5):178–87. doi: 10.1258/jrsm.2010.100104 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Hamann J, Adjan S, Leucht S, Kissling W. Psychiatric decision making in the adoption of a new antipsychotic in Germany. Psychiatr Serv. May 2006;57(5):700–3. doi: 10.1176/ps.2006.57.5.700 [DOI] [PubMed] [Google Scholar]
- 20.Schwedhelm TM, Rees JR, Onega T, et al. Patient and physician factors associated with Oncotype DX and adjuvant chemotherapy utilization for breast cancer patients in New Hampshire, 2010-2016. BMC Cancer. Sep 3 2020;20(1):847. doi: 10.1186/s12885-020-07355-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.McKinlay JB, Burns RB, Durante R, et al. Patient, physician and presentational influences on clinical decision making for breast cancer: results from a factorial experiment. J Eval Clin Pract. Feb 1997;3(1):23–57. doi: 10.1111/j.1365-2753.1997.tb00067.x [DOI] [PubMed] [Google Scholar]
- 22.Modi SC, Whetstone LM, Cummings DM. Influence of patient and physician characteristics on percutaneous endoscopic gastrostomy tube decision-making. J Palliat Med. Apr 2007;10(2):359–66. doi: 10.1089/jpm.2006.0145 [DOI] [PubMed] [Google Scholar]
- 23.Bensing JM, van den Brink-Muinen A, de Bakker DH. Gender differences in practice style: a Dutch study of general practitioners. Med Care. Mar 1993;31(3):219–29. doi: 10.1097/00005650-199303000-00004 [DOI] [PubMed] [Google Scholar]
- 24.Franks P, Bertakis KD. Physician gender, patient gender, and primary care. J Womens Health (Larchmt). Jan–Feb 2003;12(1):73–80. doi: 10.1089/154099903321154167 [DOI] [PubMed] [Google Scholar]
- 25.Reeder-Hayes KE, Wheeler SB, Baggett CD, et al. Influence of provider factors and race on uptake of breast cancer gene expression profiling. Cancer. Apr 15 2018;124(8):1743–1751. doi: 10.1002/cncr.31222 [DOI] [PubMed] [Google Scholar]
- 26.Development and Update of Guidelines. National Comprehensive Cancer Network. 2024. https://www.nccn.org/guidelines/guidelines-process/development-and-update-of-guidelines
- 27.Paik S, Tang G, Shak S, et al. Gene Expression and Benefit of Chemotherapy in Women With Node-Negative, Estrogen Receptor–Positive Breast Cancer. Journal of Clinical Oncology. 2006;24(23):3726–3734. doi: 10.1200/jco.2005.04.7985 [DOI] [PubMed] [Google Scholar]
- 28.Sparano JA, Gray RJ, Makower DF, et al. Prospective Validation of a 21-Gene Expression Assay in Breast Cancer. New England Journal of Medicine. 2015;373(21):2005–2014. doi:doi: 10.1056/NEJMoa1510764 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Sparano JA, Gray RJ, Makower DF, et al. Adjuvant Chemotherapy Guided by a 21-Gene Expression Assay in Breast Cancer. N Engl J Med. Jul 12 2018;379(2):111–121. doi: 10.1056/NEJMoa1804710 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Gray R, Bradley R, Braybrooke J, et al. Increasing the dose intensity of chemotherapy by more frequent administration or sequential scheduling: a patient-level meta-analysis of 37 298 women with early breast cancer in 26 randomised trials. The Lancet. 2019;393(10179):1440–1452. doi: 10.1016/S0140-6736(18)33137-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Ross TR, Ng D, Brown JS, et al. The HMO Research Network Virtual Data Warehouse: A Public Data Model to Support Collaboration. EGEMS (Wash DC). 2014;2(1):1049. doi: 10.13063/2327-9214.1049 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Bhimani J, Wang P, Gallagher GB, et al. Patient factors and modifications to intended chemotherapy for women with Stages I-IIIA breast cancer. Int J Cancer. Oct 1 2025;157(7):1342–1353. doi: 10.1002/ijc.35494 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Bhimani J, O'Connell K, Ergas IJ, et al. Methodology for Using Real-World Data From Electronic Health Records to Assess Chemotherapy Administration in Women With Breast Cancer. JCO Clin Cancer Inform. Apr 2024;8:e2300209. doi: 10.1200/cci.23.00209 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Gouérant S, Leheurteur M, Chaker M, et al. A higher body mass index and fat mass are factors predictive of docetaxel dose intensity. Anticancer Res. Dec 2013;33(12):5655–62. [PubMed] [Google Scholar]
- 35.Carroll JP, Protani MM, Nguyen L, et al. Toxicity and tolerability of adjuvant breast cancer chemotherapy in obese women. Med Oncol. Apr 2014;31(4):881. doi: 10.1007/s12032-014-0881-z [DOI] [PubMed] [Google Scholar]
- 36.Madarnas Y, Sawka CA, Franssen E, Bjarnason GA. Are medical oncologists biased in their treatment of the large woman with breast cancer? Breast Cancer Res Treat. Mar 2001;66(2):123–33. doi: 10.1023/a:1010635328299 [DOI] [PubMed] [Google Scholar]
- 37.Kantor ED, O'Connell K, Ergas IJ, et al. Assessment of breast cancer chemotherapy dose reduction in an integrated healthcare delivery system. Breast Cancer Res Treat. Feb 2024;203(3):565–574. doi: 10.1007/s10549-023-07126-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Greenman CG, Jagielski CH, Griggs JJ. Breast cancer adjuvant chemotherapy dosing in obese patients: dissemination of information from clinical trials to clinical practice. Cancer. May 15 2008;112(10):2159–65. doi: 10.1002/cncr.23416 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Bowles EA, Bhimani J, O'Connell K, et al. Associations between age at diagnosis and chemotherapy dose reductions in women treated for stage I-IIIA breast cancer. JCO Oncology Practice. 2024;20(10_suppl):99–99. doi: 10.1200/OP.2024.20.10_suppl.99 [DOI] [Google Scholar]
- 40.Gajra A, Klepin HD, Feng T, et al. Predictors of chemotherapy dose reduction at first cycle in patients age 65 years and older with solid tumors. J Geriatr Oncol. Mar 2015;6(2):133–40. doi: 10.1016/j.jgo.2014.12.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Hwang IG, Kwon M, Kim JW, et al. Prevalence and Predictive Factors for Upfront Dose Reduction of the First Cycle of First-Line Chemotherapy in Older Adults with Metastatic Solid Cancer: Korean Cancer Study Group (KCSG) Multicenter Study. Cancers (Basel). Jan 18 2021;13(2)doi: 10.3390/cancers13020331 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Gomez SL, Shariff-Marco S, Von Behren J, et al. Representativeness of breast cancer cases in an integrated health care delivery system. BMC Cancer. Oct 14 2015;15:688. doi: 10.1186/s12885-015-1696-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Männle H, Münstedt K, Momm F, Siebers JW. On the Impact of Premature Termination of Chemotherapy on Relapse of Breast Cancer. International Journal of Oncology Research. 2021;4(1)doi: 10.23937/2643-4563/1710028 [DOI] [Google Scholar]
- 44.Hershman D, McBride R, Jacobson JS, et al. Racial disparities in treatment and survival among women with early-stage breast cancer. J Clin Oncol. Sep 20 2005;23(27):6639–46. doi: 10.1200/jco.2005.12.633 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data contain potentially identifiable information such as dates of events that cannot be shared openly without appropriate human subjects’ approval and data use agreements, but the data that support the findings of this study are available from the corresponding author upon reasonable request.