Cancer incidence estimation from mortality data: a validation study within a population-based cancer registry
MetadataShow full item record
AuthorRedondo Sánchez, Daniel; Rodríguez Barranco, Miguel; Ameijide, Alberto; Alonso, Francisco Javier; Fernández-Navarro, Pablo; Jiménez Moleón, José Juan; Sánchez, María José
Cancer incidenceEstimationGoodness-of-fit testMortality-to-incidence ratioValidation
Redondo-Sánchez, D., Rodríguez-Barranco, M., Ameijide, A. et al. Cancer incidence estimation from mortality data: a validation study within a population-based cancer registry. Popul Health Metrics 19, 18 (2021). https://doi.org/10.1186/s12963-021-00248-1
SponsorshipSubprogram "Cancer surveillance" of the CIBER of Epidemiology and Public Health (CIBERESP); MINECO/FEDER PGC2018-098860-B-I00; Andalusian Department of Health Research, Development and Innovation PI-0152/2017
Background: Population-based cancer registries are required to calculate cancer incidence in a geographical area, and several methods have been developed to obtain estimations of cancer incidence in areas not covered by a cancer registry. However, an extended analysis of those methods in order to confirm their validity is still needed. Methods: We assessed the validity of one of the most frequently used methods to estimate cancer incidence, on the basis of cancer mortality data and the incidence-to-mortality ratio (IMR), the IMR method. Using the previous 15-year cancer mortality time series, we derived the expected yearly number of cancer cases in the period 2004– 2013 for six cancer sites for each sex. Generalized linear mixed models, including a polynomial function for the year of death and smoothing splines for age, were adjusted. Models were fitted under a Bayesian framework based on Markov chain Monte Carlo methods. The IMR method was applied to five scenarios reflecting different assumptions regarding the behavior of the IMR. We compared incident cases estimated with the IMR method to observed cases diagnosed in 2004–2013 in Granada. A goodness-of-fit (GOF) indicator was formulated to determine the best estimation scenario. Results: A total of 39,848 cancer incidence cases and 43,884 deaths due to cancer were included. The relative differences between the observed and predicted numbers of cancer cases were less than 10% for most cancer sites. The constant assumption for the IMR trend provided the best GOF for colon, rectal, lung, bladder, and stomach cancers in men and colon, rectum, breast, and corpus uteri in women. The linear assumption was better for lung and ovarian cancers in women and prostate cancer in men. In the best scenario, the mean absolute percentage error was 6% in men and 4% in women for overall cancer. Female breast cancer and prostate cancer obtained the worst GOF results in all scenarios. Conclusion: A comparison with a historical time series of real data in a population-based cancer registry indicated that the IMR method is a valid tool for the estimation of cancer incidence. The goodness-of-fit indicator proposed can help select the best assumption for the IMR based on a statistical argument.