Healthcare costs and employment outcomes associated with cancer in Scotland

Abstract

BACKGROUND: Cancer has substantial economic costs which have received little study in the Scottish population. Scotland holds rich public healthcare data that can be linked to enable detailed measurement of the healthcare use of people with cancer over many years. Analysis of these data can enhance understanding of cancer costs and how they change over time.

METHODS: I linked Scottish Morbidity Record (SMR) datasets and the Prescribing Information System (PIS) dataset to measure the healthcare use of people diagnosed with cancer and similar matched controls over eight years. Yearly and phase-of-care cost trajectories were charted. Generalised linear model (GLM) regression was used to analyse risk factors and estimate the excess costs of common cancers. I gained additional information by measuring costs for patients with and without pre-existing long-term conditions (LTCs). I also measured associations with employment outcomes in the United Kingdom Household Longitudinal Study (UKHLS) data using logistic regression and difference-in-differences (DiD) methods.

RESULTS: Costs varied considerably by cancer site, with the highest total and excess eight-year mean costs measured in people with non-Hodgkin lymphoma. Trajectories of costs showed rates of cost accumulation were highest in the treatment and end-of-life phases, but more healthcare was used in total during the intervening continuing phase. I also found that patients with pre-existing LTCs used considerable healthcare, but the magnitude of association was greatest in the subgroup with no pre-existing morbidity, both as a ratio of baseline costs and in absolute monetary units. Further analyses found that cancer was significantly associated with reduced odds of working at 3--5 years after diagnosis.

CONCLUSION: This thesis brings new insights into the long-term economic costs of cancer, which will help policymakers, health economists and clinicians allocate healthcare resources more efficiently and better understand the economic burden of cancer.