Data sources:
We obtained linked administrative data through IC/ES, formerly known as the Institute of Clinical and Evaluative Sciences of Ontario. Demographic data was available from the Ontario Registered Person Database. The modified Charlson-Deyo comorbidity score was calculated based on diagnoses registered for patients at hospital admissions in the year preceding surgery 33. Severe obesity (class 3 obesity, BMI>40 according to the CDC classification) was collected from unique surgical and anesthesia billing codes. Procedures and treatments were extracted from the Ontario Health Insurance Plan databases, used for documenting and billing all healthcare interactions by providers, and from the Discharge Abstract Database maintained by the Canadian Institute of Health Information (CIHI). Cancer diagnoses and histologies, as well as stage information, is collected by Cancer Care Ontario (CCO) in the Ontario Cancer Registry. CCO collects stage data based on the staging criteria of the American Joint Committee on Cancer or the Collaborative Stage initiative. For cases with more than one valid stage value, a resolved “best stage” is derived based on a pre-specified algorithm. Overall survival, from diagnosis to death, was defined as the primary outcome for this analysis.
Exposure variables, including marginalization scores, income quintiles and rural vs. urban residency, are neighborhood-based and assessed using conversion software from Statistics Canada to match individuals’ postal codes to small geographical units (Census Tracts and Dissemination Area)34. The Marginalization Score is an Ontario adaptation of the Canadian Marginalization Index previously validated for health research in Ontario and includes domains of material deprivation, residential instability, dependency and ethnic concentration35. The residential instability index reflects housing instability, number of residents per dwelling, family unit size and composition. The material deprivation index includes information on education, income, government support and unemployment. The ethnic concentration index reflects the proportion of new immigrants (< 5 years) in the community and those who self-identify as a minority. The dependency index reflects the proportion of seniors and dependent minors in the community; since uterine cancer is prevalent in women in their 60s and 70s and is uncommon in young women, the dependency scale was not considered relevant in this population and was therefore not used in this analysis. Each of the three domains in the index were evaluated separately and as a combined summary marginalization score, as previously reported 35. Marginalization indices are reported in quintiles, where quintile 5 reflects the highest degree of marginalization.