The National Cancer Registration and Analysis Service (NCRAS), part of Public Health England (PHE), is the population-based cancer registry for England. It registers all cases of cancer diagnosed in England, and the data is used to support the NHS and researchers. The data is received from various sources within the NHS, for example from lab tests and clinical decisions, but also from the cause of death recorded on the death certificate. There are a small proportion of cancers where the only information received is from the death certificate, these are called death certificate only cases. When this occurs, GPs are contacted to find out additional clinical information on the cancer, however there is a high rate of non-response. These cases are excluded from certain calculations, e.g., survival estimates therefore, they can influence how the effectiveness of a health system is assessed and the accuracy of official public health statistics.
This project will use CPRD data to further understand the DCO cases. The DCO cases will be combined with CPRD data, to investigate whether any information on the diagnosis of cancer or symptomatic presentation relating to an undiagnosed cancer was recorded by the GP. Tables will be produced which summarise whether cancer or plausible cancer-relevant consultations have been recorded, and their timing pre-death. If there are DCO cases with adequate information on the cancer, tables will be produced which summarise the demographics of the individuals with cancer, i.e. sex and age group at diagnosis.
Cancer registration data requires complete population coverage to provide unbiased, precise estimates for all cancer types, including rare cancers such as eye. The proportion of cancer registrations that are only identified using information on the death certification (DCO cases) indicate a failure to capture relevant health services information, and a high percentage indicates incomplete population coverage. The presence of DCO cases has consistently been an important marker of quality, and an increase in the availability of source documents to cancer registration staff has reduced the proportion from over 8% in the 1980s to less than 1%. However, they have not been eradicated, and a priority for cancer registration is to understand why. DCOs are routinely excluded from survival calculations. As survival estimates attempt to assess the effectiveness of the health system in treating patients with cancer, patients in which cancer is only found after death cannot contribute to this assessment. Therefore, the presence of these could bias findings, particularly in international comparisons, or regarding the examination of disparities in cancer outcomes within the UK.
During the registration process, Cancer Registration Officers (CROs) write to a GP to ask for more information when a DCO is found. Not all GPs respond. The availability of primary care information in CPRD permits investigation into these cases for the first time.
The DCO cases will be linked using the established CPRD-NCRAS linker file using the tumour identifier. Descriptive statistics (counts, proportions) will be calculated which describe: the level of ascertainment of CPRD data for DCOs; the completeness of required information to register a case; any socio-demographic bias present in matched DCO-CPRD cases. The required information to register a case is: confirmation of the cancer diagnosis; investigations undertaken; dates of diagnosis and investigations.
Primary outcome: number of DCO patients with CPRD data; difference in days between the date of death as recorded by NCRAS and CPRD data; number of CPRD records with cancer coded as the medical term selected; range of consultation and observation date; number of days between consultation and observation date and date of death from the death certification; completeness of diagnosis, referral and event fields.
Secondary outcomes: number of DCO patients actively under care at the time of death; number of DCOs with a matched site between the death certification and CPRD data; proportion of match DCO-CPRD records which are referred to oncology, or to a relevant diagnostic test (as defined by Be Clear on Cancer evaluations).
Lucy Ellis-Brookes - Chief Investigator - National Cancer Intelligence Network - NCIN
Katherine Henson - Corresponding Applicant - Public Health England
Brian Rous - Collaborator - Public Health England
Carolynn Gildea - Collaborator - Public Health England
Catherine Welham - Collaborator - Public Health England
Chloe Bright - Collaborator - Public Health England
Georgios Lyratzopoulos - Collaborator - University College London ( UCL )
Jack Anderson - Collaborator - Public Health England
Lucy Irvine - Collaborator - Public Health England
Sam Winters - Collaborator - Public Health England
Victoria Coupland - Collaborator - Public Health England
Roger Hill - Collaborator - Public Health England
NCRAS Cancer Registration Data