Socioeconomic gaps over time in colorectal cancer survival in England: flexible parametric survival analysis

Background Despite persistent reports of socioeconomic inequalities in colorectal cancer survival in England, the magnitude of survival differences has not been fully evaluated. Methods Patients diagnosed with colon cancer (n=68 169) and rectal cancer (n=38 267) in England (diagnosed between January 2010 and March 2013) were analysed as a retrospective cohort study using the National Cancer Registry data linked with other population-based healthcare records. The flexible parametric model incorporating time-varying covariates was used to assess the difference in excess hazard of death and in net survival between the most affluent and the most deprived groups over time. Results Survival analyses showed a clear pattern by deprivation. Hazard ratio of death was consistently higher in the most deprived group than the least deprived for both colon and rectal cancer, ranging from 1.08 to 1.17 depending on the model. On the net survival scale, the socioeconomic gap between the most and the least deprived groups reached approximately −4% at the maximum (−3.7%, 95% CI −1.6 to −5.7% in men, −3.6%, 95% CI −1.6 to −5.7% in women) in stages III for colon and approximately −2% (−2.3%, 95% CI −0.2 to −4.5% in men, −2.3%, 95% CI −0.2 to −4.3% in women) in stage II for rectal cancer at 3 years from diagnosis, after controlling for age, emergency presentation, receipt of resection and comorbidities. The gap was smaller in other stages and sites. For both cancers, patients with emergency presentation persistently had a higher excess hazard of death than those without emergency presentation. Conclusion Survival disparities were profound particularly among patients in the stages, which benefit from appropriate and timely treatment. For the patients with emergency presentation, excess hazard of death remained high throughout three years from the diagnosis. Public health measures should be taken to reduce access inequalities to improve survival disparities.


Code
Operation Name for Colon Cancer H04.1 Proctocolectomy NEC, Panproctocolectomy and Ileostomy H04.2 Panproctocolectomy and anastomosis of ileum to anus and creation of pouch HFQ H04.3 Panproctocolectomy and anastomosis of ileum to anus NEC H04.8 Other specified total excision of colon and rectum H04.9 Panproctocolectomy NEC, Total excision of colon and rectum, unspecified-H05.1 Total colectomy and anastomosis of ileum to rectum H05.2 Total colectomy and ileostomy and creation of rectal fistula HFQ H05.3 Total colectomy and ileostomy NEC H05.8 Total excision of colon, other specified H05.9 Total excision of colon, Unspecified H06.1 Extended right hemicolectomy and end to end anastomosis H06.2 Extended right hemicolectomy and anastomosis of ileum to colon H06.3 Extended right hemicolectomy and anastomosis NEC H06.4 Extended right hemicolectomy and ileostomy HFQ H06.8 Other specified extended excision of right hemicolon H06.9 Extended excision of Right hemicolon, unspecified, excision of Right colon and surrounding tissue H07.1 Right hemicolectomy and end to end anastomosis of ileum to colon, Ileocaecal resection H07.2 Right hemicolectomy and side to side anastomosis of ileum to transverse colon, H07.3 Right hemicolectomy and anastomosis NEC H07.4 Right hemicolectomy and ileostomy HFQ H07.8 Other specified other excision of right hemicolon H07.9 Other excision of right hemicolon, unspecified; Right hemicolectomy NEC H08.1 Transverse colectomy and end to end anastomosis H08.2 Transverse colectomy and anastomosis of ileum to colon H08.3 Transverse colectomy and anastomosis NEC H08.4 Transverse colectomy and ileostomy HFQ H08.5 Transverse colectomy and exteriorisation of bowel NEC CODE COLOSTOMY SPERATELY H08.8 Other specified excision of transverse colon H08.9 Excision of transverse colon, unspecified H09.1 Left hemicolectomy and end to end anastomosis of colon to rectum H09.2 Left hemicolectomy and end to end anastomosis of colon to colon H09.3 Left hemicolectomy and anastomosis NEC H09.4 Left hemicolectomy and ileostomy HFQ H09.5 Left hemicolectomy and exteriorisation of bowel NEC CODE COLOSTOMY SEPERATELY H09.8 Excision of left hemicolon, Other specified H09.9 Left hemicolectomy NEC, Excision of left hemicolon, Unspecified H10.1 Sigmoid colectomy and end to end anastomosis of ileum to rectum H10.2 Sigmoid colectomy and anastomosis of colon to rectum H10.3 Sigmoid colectomy and anastomosis NEC H10.4 Sigmoid colectomy and ileostomy HFQ H10.5 Sigmoid colectomy and exteriorisation of bowel NEC H10.8 Other specified excision of sigmoid colon H10.9 Unspecified excision of sigmoid colon H11. Panproctocolectomy and anastomosis of ileum to anus and creation of pouch HFQ H04.3 Panproctocolectomy and anastomosis of ileum to anus NEC H04.8 Other specified total excision of colon and rectum H04.9 Panproctocolectomy NEC, Total excision of colon and rectum, unspecified-H05.1 Total colectomy and anastomosis of ileum to rectum H05.2 Total colectomy and ileostomy and creation of rectal fistula HFQ H05.3 Total colectomy and ileostomy NEC H05.8 Total excision of colon, other specified H05.9 Total excision of colon, Unspecified H06.1 Extended right hemicolectomy and end to end anastomosis H06.2 Extended right hemicolectomy and anastomosis of ileum to colon H06.3 Extended right hemicolectomy and anastomosis NEC H06.4 Extended right hemicolectomy and ileostomy HFQ H06.9 Extended excision of Right hemicolon, unspecified, excision of Right colon and surrounding tissue H07.1 Right hemicolectomy and end to end anastomosis of ileum to colon, Ileocaecal resection H07.2 Right hemicolectomy and side to side anastomosis of ileum to transverse colon, H07.3 Right hemicolectomy and anastomosis NEC H07.4 Right hemicolectomy and ileostomy HFQ H07.8 Other specified other excision of right hemicolon H07.9 Other excision of right hemicolon, unspecified; Right hemicolectomy NEC H08.1 Transverse colectomy and end to end anastomosis H08.3 Transverse colectomy and anastomosis NEC H08.4 Transverse colectomy and ileostomy HFQ H08.5 Transverse colectomy and exteriorisation of bowel NEC CODE COLOSTOMY SPERATELY H08.8 Other specified excision of transverse colon H09.1 Left hemicolectomy and end to end anastomosis of colon to rectum H09.2 Left hemicolectomy and end to end anastomosis of colon to colon H09.3 Left hemicolectomy and anastomosis NEC H09.4 Left hemicolectomy and ileostomy HFQ H09.5 Left hemicolectomy and exteriorisation of bowel NEC CODE COLOSTOMY SEPERATELY H09.8 Excision of left hemicolon, Other specified H09.9 Left hemicolectomy NEC, Excision of left hemicolon, Unspecified H10.1 Sigmoid colectomy and end to end anastomosis of ileum to rectum H10.2 Sigmoid colectomy and anastomosis of colon to rectum H10.3 Sigmoid colectomy and anastomosis NEC H10.4 Sigmoid colectomy and ileostomy HFQ H10.5 Sigmoid colectomy and exteriorisation of bowel NEC H10.8 Other specified excision of sigmoid colon H10.9 Unspecified excision of sigmoid colon H11.1 Colectomy and end to end anastomosis of colon to colon NEC H11.2 Colectomy and side to side anastomosis of ileum to colon NEC H11.3 Colectomy and anastomosis NEC H11.4 Colectomy and ileostomy NEC H11.5 Colectomy and exteriorisation of bowel CODE COLOSTOMY SEPERATELY H11.8 Other excision of colon, other specified H11.9 Hemicolectomy NEC; Colectomy NEC, Other excision of colon, unspecified; BMJ Publishing Group Limited (BMJ) disclaims all liability and responsibility arising from any reliance Supplemental material placed on this supplemental material which has been supplied by the author(s)

Code
Operation Name for Rectal Cancer H29.1 Subtotal excision of colon and rectum and creation of colonic pouch and anastomosis of colon to anus H29.2 Subtotal excision of colon and rectum and creation of colonic pouch NEC H29.3 Subtotal excision of colon and creation of colonic pouch and anastomosis of colon to rectum H29.4 Subtotal excision of colon and creation of colonic pouch NEC H29.8 Subtotal excision of colon, Other specified H29.9 Subtotal excision of colon, Unspecified H33.1 Abdominoperineal excision of rectum and end colostomy; APR; SCAPER H33.2 Proctectomy and anastomosis of colon to anus H33.3 Anterior resection of rectum and anastomosis of colon to rectum using staples H33.4 Anterior resection of rectum and anastomosis NEC H33.5 Hartmann procedure, Rectosigmoidectomy and closure of rectal stump and exteriorisation of bowel CODE COLOSTOMY SEPERATELY H33.6 Anterior resection of rectum and exteriorisation, CODE COLOSTOMY SEPARATELY H33.7 Perineal resection of rectum HFQ H33.8 Anterior Resection of Rectum NEC, Rectosigmoidectomy and anastomosis of colon to rectum Excision of rectum, other specified H33.9 Rectosigmoidectomy NEC, Excision of rectum, unspecified; H34.1 Open excision of lesion of rectum: Open removal of polyp; Yorke Mason H34.2 Open cauterisation of lesion of rectum, Diathermy H34.5 Open destruction of lesion of rectum NEC H34.8 Open removal of lesion of rectum, other specified H40.1 Trans-sphincteric excision of mucosa of rectum H40.2 Trans-sphincteric excision of lesion of rectum H40.3 Trans-sphincteric destruction of lesion of rectum H40.8 Other specified operations on rectum through anal sphincter H40.9 Unspecified operations on rectum through anal sphincter X14.1 Total exenteration of pelvis X14.2 Anterior exenteration of pelvis X14.3 Posterior exenteration of pelvis X14.8 Other specified clearance of pelvis X14.9 Clearance of pelvis, unspecified BMJ Publishing Group Limited (BMJ) disclaims all liability and responsibility arising from any reliance Supplemental material placed on this supplemental material which has been supplied by the author(s)

Supplementary Material 2. Definition of Chronic and Acute Comorbidities.
Ten and 14 comorbidities of the 17 comorbidities in the Charlson Comorbidity Index were selected for chronic and acute comorbidities, respectively, on the basis of their clinical relevance to the selection and timeliness of treatment for CRC e.g. invasive or less invasive treatment, curative or palliative treatment.
Obesity defined here as body mass index: BMI > 30 was included independently in the acute phase only 0 to 0.5 years before CRC diagnosis because BMI is a reversible condition i.e. a timevarying variable and may confound with stage at diagnosis.
Unlike the Charlson Comorbidity Index, the comorbidities in our study were counted but not assigned a weight. Abbreviations: SES, socioeconomic status.
Four variables, namely, stage at diagnosis, histology, tumour grade and emergency presentation were missing in some patients.
The missingness was highest in stage with 30.6% and 27.3% for colon and rectal cancer, respectively.
Tumour grade was missing at 24.2% and 22.3%, emergency presentation at 9.9% and 6.7%, and histology at less than 3% for colon and rectal cancer, respectively.
Those variables were multiply imputed for 30 times under the missing at random assumption.
Following variables were used for the multiple imputation: sex, age group, cancer site, number of chronic and acute comorbidities, receipt of major surgery, vital status, Nelson-Aalen estimator and government office region.
The distributions of stage, histology, tumour grade and emergency presentation after the imputation were as follows.
When evaluating association between each variable and survival in the Cox regression, the Wald test, instead of likelihood ratio test, was used for both imputed and completed data to account for the uncertainty in imputed data.
For FPM, complete data only were used because imputed data are not technically supported to use in FPM.
Therefore, analyses using imputed data in the first analysis Cox regression were considered sensitivity analyses and results are shown in Appendix Table 3. Abbreviations: 95% CI, 95% confidence interval; HR, hazard ratio; NA, not applicable (not included in multivariable model); SES, socioeconomic status. a Rightsided colon includes ascending colon, hepatic flexure and caecum. Transverse colon includes transverse colon and splenic flexure. Left-sided colon includes descending colon and sigmoid colon. b Information did not match across different datasets in 4 patients for colon cancer therefore excluded. c All variables are mutually adjusted. For SES only, adjusted HRs are shown without interaction between SES and stage. For other variables, interaction between SES and stage is adjusted. Abbreviations: 95% CI, 95% confidence interval; HR, hazard ratio; SES, socioeconomic status. a Information did not match across different datasets in 4 patients for colon cancer therefore excluded. b Adjusted for sex, age, site, histology (imputed), tumour grade (imputed), emergency presentation (imputed), receipt of major resection, number of chronic and acute comorbidities. c Adjusted for sex, age, site, tumour grade (imputed), emergency presentation (imputed), receipt of major resection, number of chronic and acute comorbidities.