Risk and survival for colorectal cancer in northern Sweden : sociodemographic factors and surveillance programs

Abstract: BackgroundColorectal cancer (CRC) – i.e., cancer in the colon or rectum – is one of the most common cancers both globally and in Sweden. The risk for CRC is mainly related to age, heredity, and life-style risk factors. Previous studies have also demonstrated that individuals with lower socioeconomic status (SES), living alone, or far from care facilities may have a higher risk for CRC or a worse outcome.  In contrast to life-style or sociodemographic-associated risks, an inherited risk for CRC is difficult to modify. However, colonoscopic surveillance programs can be help prevent CRC in families with a known hereditary risk.The Northern Health Care Region (northern Sweden) is the most sparsely populated region in Sweden, and travel distances to care can be long. The population in Northern Sweden is on average older and has lower SES compared with the rest of the country. The impact of these sociodemographic differences on CRC in northern Sweden is not well known. AimThis thesis analyses CRC in a northern Sweden setting with regards to incidence, survival, and associated sociodemographic risk factors, including prevention for individuals with increased hereditary risk.MethodsPapers I and II, cohort studies from the Risk North database, link individual data from health care registers to other sociodemographic registers. In Paper I, the incidence, mortality, and survival for all CRC cases in northern Sweden were compared with the rest of Sweden for the period 2007-2013. Uni- and multivariable Cox regression analysis were used to assess the impact of sociodemographic factors and tumour stage on survival by calculating hazard ratios (HR). In Paper II, we analysed any association between travel time to care and CRC survival in northern Sweden during 2007-2013 using the same type of Cox regression analysis. Papers III and IV are based on a cohort of individuals with a family history of CRC, prospectively recorded from 1995 to 2012 in the colonoscopic surveillance register at the Cancer Prevention Clinic at Umeå University Hospital. In Paper III, we evaluated the cancer preventive effect of the performed colonoscopic surveillance. Observed cases of CRC were compared to a cohort estimate of cases without surveillance. Compliance with surveillance and colonoscopic quality was also analysed. In Paper IV, we examined the cost-effectiveness of the colonoscopic surveillance program in Paper III. A cost-utility analysis with a societal perspective was used and the stability of the results was tested in a sensitivity analysis.  ResultsThe age-adjusted incidence in colon cancer was 12.7% lower in northern compared to southern Sweden or 35.9/100 000 vs. 41.1/100 000 person years (p < 0.01). For rectal cancer, the incidence was 10.5% lower in the north (17.6 vs. 19.7 p <0.01). In subgroup analysis, the largest difference in incidence between northern and southern Sweden was found among individuals > 79 years age (colon - 190 vs. 237 ≈ 19.6%, rectal 72.4 vs. 88.0 ≈ 17.7%). For all of Sweden, the incidence in colorectal cancer was higher in males, individuals with lower SES, or individuals living alone. In univariable analyses of survival (all-cause and cause-specific) for colon and rectal cancer patients in all of Sweden, patients with high SES or co-habiting had a significantly better outcome compared to patients with low SES or living alone. HR for death ranged from 0.60 to 0.85 in the better-favoured risk group. No differences in colon or rectal cancer survival between northern and southern Sweden were demonstrated in the univariable analysis. However, in multivariable survival analysis, all-cause survival for colon cancer patients was better in southern Sweden (HR 0.92; 95% CI 0.86 – 0.97).  For cause-specific survival for colon cancer or in any analysis for rectal cancer, no differences between northern and southern Sweden were demonstrated. In analysis of travel time, no association between travel time and survival was found. In the evaluation of the colonoscopic surveillance programme, one case of CRC was observed, compared to 9.5-10.5 expected cases. Standardised Incidence Ratio (SIR) between observed and expected cases of CRC was 0.10 (CI 95% 0.0012–0.53) to 0.11 (CI 95% 0.0014–0.59. The compliance to the surveillance program was 90%. The adenoma detection rate was 14%, and 10% of the examinations were incomplete. In the cost-utility analysis, the net cost for surveillance was 233 038 €, while saving 64.8 Quality Adjusted Life Years (QALYs) compared to non-surveillance. The resulting Incremental Cost-Effectiveness Ratio (ICER) was 3596 €/QALY, ranging from -4620 €/QALY in the best-case scenario to 33 779 € /QALY in the worst-case scenario.ConclusionThe incidence of CRC was lower in northern Sweden and most evident in the elderly, raising questions on differences in life-style between northern and southern Sweden in the past. There were considerable sociodemographic disparities in CRC survival in Sweden, including a lower all-cause survival for colon cancer patients in the north. In this study, travel time to care in northern Sweden did not affect survival and the lower all-cause survival in northern Sweden cannot be fully explained. The colonoscopic surveillance of families in northern Sweden with inherited risk for CRC had a good cancer preventive effect, including a high cost-effectiveness. The reasons for the good effect may be high compliance, since the quality of the colonoscopies was moderate.