Validation of the National Cancer Registration and Analysis Service prostate cancer registry with data from the CAP study

• Published in: The Lancet (British edition) Vol. 388; p. S77
• Main Authors: Merriel, Samuel W D, Dr, Turner, Emma L, PhD, Walsh, Eleanor, MSc, Young, Grace, MSc, Metcalfe, Chris, PhD, Hounsome, Luke, PhD, Tudge, Isobel, Donovan, Jenny, PhD, Hamdy, Freddie, Prof, MD, Neal, David, Prof, FRCS, Martin, Richard M, Prof, PhD
• Format: Journal Article
• Language: English
• Published: London Elsevier Ltd 01.11.2016; Elsevier Limited
• Subjects: Cancer; Health risk assessment; Internal Medicine; Medical research; Prostate cancer
• ISSN: 0140-6736; 1474-547X
• DOI: 10.1016/S0140-6736(16)32313-3

Abstract Background Prostate cancer is the second most common cancer in the UK, with 39 741 cases diagnosed in 2014. The National Cancer Registration and Analysis Service (NCRAS) collects data about cancer (including prostate cancer) in England. The CAP study is a cluster-randomised controlled trial investigating the effectiveness and cost-effectiveness of prostate-specific antigen (PSA) testing. The value of cancer registries is high if they receive and process quality data, which can be evaluated in various domains, including completeness and validity. This study aimed to compare the completeness and accuracy of the NCRAS registry with data from the CAP study. Methods We compared the percentage completeness and agreement of prostate cancer diagnosis by Union Internationale Contre le Cancer (UICC) TNM stage and Gleason grade in the NCRAS cancer registry with information collected via independent medical record review on participants in CAP. CAP study participants were matched to NCRAS registry entries by their National Health Service number. Agreement was assessed with Cohen's κ. Findings 1356 men were included in this study (mean age 75·1 years, SD 5·1). Both the NCRAS registry (97·9%) and CAP study (98·5%) had high levels of completeness for date of diagnosis. Agreement was high for combined Gleason score (κ=0·90, 95% CI 0·89–092) and low for American Joint Committee on Cancer (AJCC) group (0·48, 0·43–0·53) overall. TNM staging agreement was 0·35 (0·31–0·37) for T, 0·51 (0·45–0·57) for N, and 0·58 (0·51–0·66) for M stage. Agreement was moderate when considering local (T1–3N0M0) versus metastatic disease (T4NxMx/TxN1Mx/TxNxM1) (0·54, 0·44–0·64). Interpretation The NCRAS prostate cancer registry has a high level of completeness for case registration, and strong accuracy for Gleason grade. Agreement of exact TNM staging and AJCC group seems low, which could be explained if staging data were collected from different sources (ie, pathological vs imaging staging methods), and needs to be explored further. Prostate cancer stage and accuracy of grade data for the NCRAS registry need repeated assessment to drive improvements in data quality. Funding None.