Routes to Diagnosis data categorise the pathway a patient took to their diagnosis into one of eight Routes, assigned by a published methodological process and run routinely within CAS. The output is a single table with the Route and Route code that can be linked through tumourid.
The end of year AV20XX dataset in CASREF01 are used to produce these figures. Routes data used two standard datasets and one lookup table:
The formal publication explaining the Routes methodology: Routes to Diagnosis for cancer – Ellis-Brookes et al.pdf
The cohort was defined to align with the cohort specified in GDO_0025_tech_doc_inc, which in turn is generated from the methodology and standard restrictions in the CAS SOP #1 – Counting Cancer Cases. This SOP was followed, with the additional exclusions of testicular tumours in female patients and ovarian tumours in male patients for tumours whose ICD code starts with D. (The SOP handles C-coded tumours but does not yet perform any exclusions on D-coded tumours: sex-exclusions are performed on all tumours for our dataset). The cohort is created from AV2019.at_tumour_england. Cases not meeting the SOP criteria are excluded from the cohort. Once the restrictions have been applied the cohort is refined using diagnosis year and site codes to select the relevant cohort of interest.
The figures published in the GDO project will vary slightly from the official Routes to Diagnosis figures published by NCRAS. This is due to a difference in the cohort being used to create the denominator. For published Routes work this cohort only includes cases that were run through the Routes to Diagnosis algorithm and have been classified into a Route. This algorithm has some exclusion criteria around data quality for all datasets that may drop some cases but will also be subject to registry creep – with cases registered after the algorithm has been run not receiving a Route. The RTD2018 table was run on a cohort built from AV2018 for 2018. The 2013-2018 GDO cohort from AV2019 includes some cases registered late that were not included in the RTD2018 table. These cases are marked as “Not classified”, signifying that the Routes algorithm has not touched them. They are not assigned to the Unknown Route – this only happens for tumours that have been run through the algorithm. The addition of this classification was chosen so the cohort figures used here match with other GDO metrics.
The sites included here also vary from published Routes to Diagnosis data – so variation in results may also be down to the inclusion of different ICD10 codes, despite site names being identical in some cases. Age and geographic cuts may also differ.
Where no standard screening programme for the cancer sites in the initial data release exists, the screen detected Route is removed. This is indicated with ‘.c’. Some but not all sarcomas are eligible for screening, but some are screen detected and are counted as such.
Routes for 2019 diagnoses are not included here, as Routes to Diagnosis does not yet cover this year. This is indicated with ‘.a’.
Where a group size is very small, data are not available as a measure to protect patient confidentiality. This is indicated with ‘.k’.
The Routes to Diagnosis algorithm is run for tumours registered on the National Cancer Registration Database (ENCORE). When patients have multiple skin tumours, only the first BCC and cSCC of skin is registered on the database. This is standard cancer registry practice due to the high volume of these tumours diagnosed annually. Subsequent tumours included in the Get Data Out data are imputed from the existence of their path reports on the system (see the Skin SOP document for more details on this methodology). As these are imputed tumours, not registered tumours, they are not run through the Routes to Diagnosis Algorithm. Results for Routes to Diagnosis are presented for a Get Data Out group if 50% or more of tumours in the group have been run through the Routes algorithm. This is all groups in Get Data Out except for groups of ‘subsequent BCCs’ and ‘subsequent cSCCs’. For most groups the vast majority of tumours will have been run through the algorithm (>99%). The main exceptions are ‘all skin tumours’, ‘all skin cancers’ and ‘all keratinocyte cancers’, which contain significant numbers of subsequent tumours, but not more than 50%. This is indicated with ‘.b’. For any group, the column ‘Route not classified’ gives the count and percentage of tumours which were not run through the Routes algorithm for that group.
This is the first release that we are making data available for Kaposi’s sarcoma for Routes to Diagnosis. Previously, the only data published for that cancer site was Incidence. There were concerns that, as Kaposi’s Sarcoma can be treated in GUM/HIV clinics, which is not a typical Routes to Diagnosis pathway, the cancer registry may have missed clinically diagnosed cases. If those cases are being missed the data may have some bias. However, the percentage of Kaposi’s sarcoma with a confirmed route is as high as our other sites.
A full definition of these codes can be found in the GDO missing file.
The QA was done internally by the Get Data Out team.
Any questions, please contact charlotte.eversfield@nhs.net, sally.vernon1@nhs.net, hannah.maconochie1@nhs.net, sam.winters2@nhs.net.