Unwarranted variation in radiation therapy fractionation

Increased Variation in Esophageal Cancer Treatment and Geographic Healthcare Disparity in Michigan

BACKGROUND

Regional variation in complex healthcare is shown to negatively impact health outcomes. We sought to characterize geographic variance in esophageal cancer operation in Michigan.

STUDY DESIGN

Data for patients with locoregional esophageal cancer from the Michigan Cancer Surveillance Program from 2000 to 2013 was analyzed. We reviewed the incidence of esophageal cancer by county and region, and those with locoregional disease receiving an esophagectomy. Counties were aggregated into existing state-level "urban vs rural" designations, regions were aggregated using the Michigan Economic Recovery Council designations, and data was analyzed with ANOVA, F-test, and chi-square test.

RESULTS

Of the 8,664 patients with locoregional disease, 2,370 (27.4%) were treated with operation. Men were significantly more likely to receive esophagectomy than women (p < 0.001). Likewise, White, insured, and rural patients were more likely than non-White (p < 0.001), non-insured (p = 0.004), and urban patients (p < 0.001), respectively. There were 8 regions and 83 counties, with 61 considered rural and 22 urban. Region 1 (Detroit metro area, southeast) comprises the largest urban and suburban populations; with 4 major hospital systems it was considered the baseline standard for access to care. Regions 2 (west; p = 0.011), 3 (southwest; p = 0.024), 4 (east central; p = 0.012), 6 (northern Lower Peninsula; p = 0.008), and 8 (Upper Peninsula; p < 0.001) all had statistically significant greater variance in annual rates of operation compared with region 1. Region 8 had the largest variance and was the most rural and furthest from region 1. The variance in operation rate between urban and rural differed significantly (p = 0.005).

CONCLUSIONS

A significant increase in variation of care was found in rural vs urban counties, as well as in regions distant to larger hospital systems. Those of male sex, White race, rural residence, and those with health insurance were significantly more likely to receive operation.

Large variation in radiation therapy fractionation for multiple myeloma in Australia

AIM

To evaluate the patterns of use of different radiation therapy (RT) fractionation for multiple myeloma (MM) bone disease.

METHODS

This is a population-based cohort of patients with MM who had RT between 2012 and 2017 as captured in the statewide Victorian Radiotherapy Minimum Data Set in Australia. Data linkage was performed to identify subsets of RT delivered within 3 months of death. RT fractionation was classified into four groups: single-fraction (SFRT), 2-5, 6-10, and > 10 fractions. Changes in RT fractionation use over time were evaluated with the Cochran-Armitage test for trend. Factors associated with RT fractionation were evaluated using multivariate logistic regressions.

RESULTS

Nine hundred and sixty-seven courses of RT were delivered in 623 patients. The proportion of SFRT, 2-5, 6-10 and > 10 fractions RT was 18%, 47%, 28%, and 7%, respectively. There was an increase in the use of 2-5 fractions, from 48% in 2012 to 60% in 2017 (p-trend < .001), with corresponding decrease in the use of 6-10 fractions, from 26% in 2012 to 20% in 2017 (p-trend = .003). Nine percent (40/430) of RT courses at private institutions were SFRT, compared to 25% (135/537) in public institutions (p < .001). In multivariate analyses, treatment in private institution was the strongest predictor of multifraction RT use. SFRT use was more common closer to the end of life-18%, 14%, and 33% of RT within 2-3, 1-2, < 1 month of death, respectively.

CONCLUSION

There is increasing use of shorter course RT (2-5 fractions) for MM over time. SFRT use remains low, with large variation in practice.

Variations in whole brain radiation therapy fractionation for brain metastases in Victoria

INTRODUCTION

We aim to evaluate the use of different whole brain radiation therapy (WBRT) fractionation schedules for brain metastases (BM) in Victoria, and the factors associated with it.

METHODS

This is a population-based cohort of patients who received radiation therapy for BM between 2012 and 2017, as captured in the Victorian Radiotherapy Minimum Dataset. We excluded patients with primary brain tumour and those who had 'prophylactic' intent treatment. The Cochran-Armitage test was used to evaluate changing trend in WBRT fractionation. Multivariate multinomial logistic regressions were used to evaluate factors associated with WBRT fractionation.

RESULTS

Of the 3111 patients who had WBRT, 1048 (45%), 1291 (42%) and 312 (13%) had ≤5, 6-10 and >10 fractions WBRT respectively. There was progressive increase in ≤5 fractions WBRT use over time, from 37% in 2012 to 50% in 2017 (P-trend < 0.001). In multivariate analyses, increasing age, patients with gastrointestinal cancer, patients living in remote/regional areas and more recent treatment were associated with the use of shorter WBRT fractionation (≤5 fractions), while patients who had WBRT plus stereotactic radiosurgery, and those treated in private institutions were associated with the use of prolonged WBRT fractionation (≥6 fractions). Three hundred eighty-nine (13%) patients died within 30 days of WBRT, of which 241 (64%), 119 (32%) and 17 (5%) had ≤5, 6-10 and > 10 fractions WBRT respectively.

CONCLUSION

We observed large variations in WBRT fractionation that are associated with patient, tumour, treatment and institutional factors. It is important to continuously monitor and benchmark our practice in order to reduce potentially unwarranted variations.