NBOCA Quarterly Report – Data Quality (England), Oct 2024

We have initially focused on data completeness as this aspect of data quality underpins what we can reliably and robustly report as an audit. In England, NATCAN receives information from the National Cancer Registration and Analysis Service (NCRAS), part of the National Disease Registration Service (NDRS), NHS England. NDRS collects patient-level data from all NHS acute providers on people with cancer using a range of national data-feeds. This includes the Cancer Registration datasets and the Cancer Outcomes and Services Dataset (COSD). COSD data are submitted to the NDRS on a monthly basis via Multidisciplinary Team electronic data collection systems.

NDRS has regional Data Liaison Managers who support NHS teams submitting cancer data to NDRS with advice and support on data improvement initiatives. We encourage trusts to get in touch with this team for support and advice. Contact details are found here.

For this quarterly report, the NBOCA utilised data from the Rapid Cancer Registration Dataset (RCRD). This dataset is compiled mainly from COSD records and is made available more quickly than the complete cancer registration dataset. Additional raw COSD items have been made available to NATCAN, however due to lack of processing their completeness/ accuracy is unknown. It also does not have complete coverage of all patients diagnosed with bowel cancer in England during the reporting period. For more information regarding the RCRD, including the data dictionary, please visit RCRD webpage.

Results are reported for a one year period to ensure sufficiently large numbers of records are included to produce reliable estimates at the trust level. The surgical reporting period is one quarter behind to allow time for procedures to occur and details to be submitted to NDRS.

This NBOCA report allocates diagnostic records to NHS organisations based on the “trust at diagnosis” recorded within the RCRD dataset.

Surgical records have been allocated based on the trust code recorded within the RCRD events dataset (based on data from HES). If more than one date for a major resection was recorded in either the RCRD events or COSD surgical dataset, the earliest date was used.

We encourage all provider teams to review their data completeness and make improvements as this will increase the number of patients who can be included in analyses and increase the range of analyses we can conduct. By focusing exclusively on data completeness for this report, we are aiming to spotlight areas where improvements are needed.

Overall, the RCRD has very good data completeness and many variables are 100% complete across diagnoses; however, some variables have less than 100% completeness or completeness varies widely between Trusts, and it is these variables that were selected for reporting. When analysing data quality, the NBOCA team recoded values as “missing” that would not be useful in practice and/or are not plausible values. For instance, values such as “not known”, “not specified”, or “not recorded” were considered to be missing, and implausible values such as performance status values other than 0-4 were considered to be missing.

The term “quarterly reporting” refers to the frequency of publication, with reports anticipated to be published in January, April, July and October each year. Data quality will be reported on a rolling basis with the starting period for each report moving three months forward.