Summary: | Introduction: Lung cancer is the commonest cause of cancer related death in men and women in England. In 2004 the National Lung Cancer Audit (NLCA) was created, as a national non-mandatory contemporary dataset of clinical features of individuals with lung cancer in part to identify variations in clinical practice and outcomes. The main aims of this dissertation are to determine the validity and representativeness of this dataset and then to investigate what factors influence access to surgery and chemotherapy and subsequent survival. In addition I have taken the opportunity afforded by this large dataset to describe the natural history of lung cancer in young adults (20-40 years). Methods: In order to establish if the dataset was representative, I created a measure of case ascertainment at the level of an NHS Trust, and examined the distribution of patient features and outcomes for varying levels of case ascertainment. I have then quantified the impact of patient and NHS Trust level features on access to surgery in people with non-small cell lung cancer and access to chemotherapy in people with small cell lung cancer using multivariate logistic regression. I have also conducted a series of survival analyses using Cox regression. Results: I have found no evidence that patient features vary systematically according to levels of case ascertainment in the NLCA. Age, sex, performance status, stage and co-morbidity all influenced the likelihood of having surgery for people with non-small cell lung cancer. Those patients first seen in a thoracic surgical centre where more likely to receive surgery than patients seen at peripheral centres (adjusted OR 1.51, 95% CI 1.16, 1.97), and surgery had a significant benefit on mortality (adjusted HR 0.41, 95% CI 0.39, 0.44). Although the resection rate was higher for patients first seen at a surgical centre (17% v 12%) these patients did equally well after surgery suggesting they were not a higher risk group. Individuals with small cell lung cancer first seen in a hospital with a high participation in clinical trials, (>5% of expected lung cancer patients being entered into clinical trials), were more likely to receive chemotherapy (adjusted OR 1.42, 95% CI 1.06, 1.90). Chemotherapy was associated with an improvement in survival (adjusted HR 0.51, 95% CI 0.46, 0.56), and amongst those patients receiving chemotherapy, mortality was not affected by the trial status of the hospital where they were first seen. In limited stage small cell disease, those patients who had chemo-radiotherapy had an improved survival compared with those patients who received chemotherapy alone (adjusted HR 0.72, 95% CI 0.62, 0.84). This dataset of English patients with lung cancer contains one of the largest cohorts of young adults (20-40 years) with lung cancer (N=583). I have been able to demonstrate that the majority present with a good performance status (0 or 1 in 80% of those with PS recorded), but advanced (stage IV) disease at diagnosis (55% of those with stage recorded). Those who have surgery have a survival profile similar to their older counterparts. Conclusion: The National Lung Cancer Audit is a representative, contemporary cohort of people with lung cancer, which can provide valuable information for health service research in lung cancer. I have found evidence that there is variation in access to treatment based on the facilities or the performance of individual NHS Trusts. My results suggest that by improving access to thoracic surgery for those individuals with non-small cell lung cancer we may be able to raise the resection rate and improve five year survival. The pattern is similar for people with small cell lung cancer and access to chemotherapy. What this research cannot explain is the aetiology for this variation, and where in the diagnostic pathway changes need to be made to improve the active management and access to potentially curative regimes. As the audit matures with more detailed information on NHS Trust level care, further analyses will be possible to try and determine more clearly what explains these variations, and how we might intervene to reduce them.
|