We found a match
Your institution may have rights to this item. Sign in to continue.
- Title
Direct modelling of age standardized marginal relative survival through incorporation of time-dependent weights.
- Authors
Lambert, Paul C.; Syriopoulou, Elisavet; Rutherford, Mark R.
- Abstract
<bold>Background: </bold>When quantifying the probability of survival in cancer patients using cancer registration data, it is common to estimate marginal relative survival, which under assumptions can be interpreted as marginal net survival. Net survival is a hypothetical construct giving the probability of being alive if it was only possible to die of the cancer under study, enabling comparisons between populations with differential mortality rates due to causes other the cancer under study. Marginal relative survival can be estimated non-parametrically (Pohar Perme estimator) or in a modeling framework. In a modeling framework, even when just interested in marginal relative survival it is necessary to model covariates that affect the expected mortality rates (e.g. age, sex and calendar year). The marginal relative survival function is then obtained through regression standardization. Given that these covariates will generally have non-proportional effects, the model can become complex before other exposure variables are even considered.<bold>Methods: </bold>We propose a flexible parametric model incorporating restricted cubic splines that directly estimates marginal relative survival and thus removes the need to model covariates that affect the expected mortality rates. In order to do this the likelihood needs to incorporate the marginal expected mortality rates at each event time taking account of informative censoring. In addition time-dependent weights are incorporated into the likelihood. An approximation is proposed through splitting the time scale into intervals, which enables the marginal relative survival model to be fitted using standard software. Additional weights can be incorporated when standardizing to an external reference population.<bold>Results: </bold>The methods are illustrated using national cancer registry data. In addition, a simulation study is performed to compare different estimators; a non-parametric approach, regression-standardization and the new marginal relative model. The simulations study shows the new approach is unbiased and has good relative precision compared to the non-parametric estimator.<bold>Conclusion: </bold>The approach enables estimation of standardized marginal relative survival without the need to model covariates that affect expected mortality rates and thus reduces the chance of model misspecification.
- Subjects
LATENT variables; SURVIVAL analysis (Biometry); DEATH rate; PARAMETRIC modeling; COMPUTER simulation; RESEARCH; RESEARCH methodology; MEDICAL cooperation; EVALUATION research; COMPARATIVE studies; TUMORS; PROBABILITY theory
- Publication
BMC Medical Research Methodology, 2021, Vol 21, Issue 1, p1
- ISSN
1471-2288
- Publication type
journal article
- DOI
10.1186/s12874-021-01266-1