Regional differences in diagnostic stages and survival in colorectal and lung cancer: Estimating costs and health gains of earlier diagnosis using Norwegian register data

Abstract

Background: An estimated one out of three individuals in Norway will at some point receive a cancer diagnosis. Previous research has identified a geographical difference in incidence of cancer across the different Norwegian counties. Earlier analysis has also reviled that a higher degree of cancer severity is usually associated with increased treatment costs. In addition, healthcare spending on cancer in Norway has increased substantially over the last decade. Objective: The main objective of this thesis was to investigate stage-specific costs and survival for colorectal and lung cancer in Norway, in order to investigate potential health and financial gains by earlier diagnosis, specifically by regional assessment of best practice. Methods: Data from the Cancer Registry of Norway was used to statistically test for significant difference between several prognostic factors and their effect on survival time. Variations related to any significant difference in survival, were then assessed by both parametric and non-parametric estimations. Using the dataset from the CRN, regional variations was investigated to assess best practice and incremental health gains were calculated using predictions from two different survival estimation methods. Analyzing potential cost saving scenarios using stage-specific costs for the two cancers were possible through a pathway model exploring the cost of lung cancer treatment, and a previously published decision-analytic model for colorectal cancer, both using a Norwegian perspective. Results: The main non-parametric estimation reveled a regional variation in median survival time of between 6.19-9.23 and 48.24-66.59 months for lung and colorectal cancer patients, respectfully. The incremental health gain between the highest and lowest grade of cancer severity was 44.78 and 168.8 months for lung and colorectal cancer, respectfully. By applying a scenario of best regional practice, the results yielded a potential per patient and yearly cost saving for lung and colorectal cancer of NOK 12 269 and NOK 11 202, and approximately NOK 37 mill and NOK 50 million, respectfully. Conclusion: There exists regional difference in the prognostic staging distribution and survival, which could indicate a disruption within the Norwegian principle of equal access and services. There is also substantial health and financial gains to be achieved from earlier diagnosis. Additional research should focus on gathering more detailed documentation and reporting on cancer treatment, socioeconomic variables, and supplier-demand variations. Furthermore, additional probabilistic decision-analytic modeling could investigate the uncertainty of cancer related costs, and also explore other financial scenarios for different sub-groups.