br In our practice many patients undergo external beam
In our practice, many patients undergo external beam DHR 123 (EBRT) and brachytherapy (BT) at different locations. Our objective was to compare rates of completion of radiation within 60 days (the suggested metric of the Commission on Cancer, ) among patients re-ceiving treatment at one institution to those receiving treatment at mul-tiple institutions. Secondary outcomes included compliance with
sensitizing chemotherapy, completion of brachytherapy, total radiation dose, recurrence rate, and overall survival.
2. Materials and methods
A retrospective chart review was performed that included patients diagnosed with cervical cancer between January of 2000 to December of 2016 who were planned to undergo primary treatment with sensitiz-ing chemotherapy, EBRT and BT at the primary institution (PI). Patients with locally advanced stage IB1-IVA cervical cancer who had established care with a gynecologic oncologist at the PI were included in this study. While the patients were seen at the PI they were often given the choice of several locations to undergo their radiation therapy and several pa-tients opted to receive a portion of their RT at an outlying facility (OLF).
Eligible patients were identified by using the Oncology Registry Da-tabase. This database identifies reportable cancers using both national and state standards. Patients with cervical cancer were identified by using ICD-10 codes as well as pathology reports. Each case was then in-dividually reviewed by certified tumor registrars to determine if these were new diagnoses. Patients who met criteria were then added to the database and followed for the duration of their life. Follow-up infor-mation was obtained from medical record review, coordination with the Columbus oncology associates database, and letters sent to primary physicians when necessary. Information regarding death was obtained from review of monthly death certificates for the state of Ohio. If infor-mation was not available, local obituary searches were performed as needed.
For patients who met inclusion criteria, a thorough chart review of histology, imaging, clinic notes, chemotherapy summaries, and radia-tion therapy summaries was performed. The distance to treatment facil-ity was calculated using Google maps  with the patient's home address listed on their clinic chart at initial visit and the respective treat-ment facility. The distance in miles was recorded for the route with the shortest duration, regardless of the mileage.
Exclusion criteria included an incomplete record of radiation ther-apy, no available histologic diagnosis, histology other than squamous or adenocarcinoma, transfers of care, use of palliative radiation therapy rather than curative, previous hysterectomy or hysterectomy for initial cancer treatment with use of adjuvant therapy, and patients who did not complete or initiate RT due to medical complications or limitations.
Overall 209 charts were reviewed. Of these patients, 99 were ex-cluded based on exclusion criteria. An additional ten patients were ex-cluded as BT was not completed due to medical complications or limitations, indicating that completion of therapy was not related to
compliance. BT was not completed in these ten patients for various rea-sons including difficulty with smit sleeve placement (n = 3), necrotiz-ing fasciitis of the perineum (n = 1), debilitating stroke (n = 2), declining overall clinical status (n = 3), and sudden death (n = 1). The 100 patients that met inclusion criteria were divided into two groups for data analysis (Fig. 1).
The two groups consisted of those patients that had all of their radi-ation therapy (EBRT and BT) at the PI (n = 75) and those who had all or part of their radiation therapy (RT) at another institution (n = 25). All of the patients who underwent a portion of their RT outside of the PI com-pleted their EBRT and BT at separate facilities. Descriptive information on the study sample was tabulated using means, medians, standard deviations and ranges for numeric variables and percentages for nominal (categorical) variables. Univariate compar-isons of demographic and clinical factors for the two groups (patients who received EBRT and BT at the PI and patients who received EBRT and/or BT at an OLF) were made using independent samples t-tests for continuous variables and chi-square tests for categorical data. For data that was not normally distributed (based on the Levene's test), in-dependent samples t-tests not assuming equal variances or Wilcoxon rank sum tests were conducted. Statistical significance was based on traditional two-sided tests with the alpha error set at 5%. Statistical anal-yses were conducted using IBM SPSS Statistics version 19.0 (Armonk, NY).