Relative Survival Regression Research Articles

Nationwide treatment patterns and survival of older patients with prostate cancer.

To examine the clinical features, applied treatments, and survival of older patients with prostate cancer compared with younger patients. All patients diagnosed with prostate cancer between 2005 and 2015 in the Netherlands were identified from the nationwide population-based Netherlands Cancer Registry (NCR). Patient and tumour characteristics, as well as applied treatments, were described by age groups and prognostic risk groups. Relative survival rates were determined, including multivariable relative survival regression analyses. Additionally, to assess if age was associated with receiving curative treatment, a multivariable logistic regression analysis was performed. In total, 48% of all patients were 70 years of age or older. Older patients had a higher prostate specific antigen (PSA) level, a higher Gleason score, as well as a higher disease stage at diagnosis. The 10-year relative survival decreased with increasing age, and after adjustment for disease stage, Gleason score, PSA level, and comorbidities, older patients had a worse survival rate. Older patients with intermediate- or high-risk disease appeared to be treated less often with curative intent compared with younger patients after adjustment for tumour stage, Gleason score, PSA level, and comorbidities. Older patients with intermediate/high risk prostate cancer treated with curative intent showed a 10-year relative survival rate similar to younger patients. The survival of older patients was worse than younger patients. This might be due to suboptimal treatment, as older patients were less often treated with curative intent. Although the increased risk of treatment complications should be considered, age alone should not be a decisive factor when offering a treatment.

Can cancer registries show whether treatment is contributing to survival increases for melanoma of the skin at a population level?

It is uncertain whether survival increases from melanoma recorded by some population registries include a treatment effect. The US Surveillance, Epidemiology and End Results (SEER) programme has good data quality control, large numbers of cases enabling high statistical precision and summary stage plus thickness, which we consider to be a best-case population registry scenario to investigate potential for a treatment effect. We have investigated SEER data to indicate whether survivals increases are fully attributable to earlier diagnosis and other non-treatment factors. Through relative survival regression, the effects of diagnostic period on 5-year excess mortality were investigated, adjusting for socio-demographic factors, lesion sub-site, histology, thickness and stage at diagnosis in 1990-2009 (n = 99 690 cases). The reduction in excess mortality (95% confidence interval) between 1990-1999 and 2000-2009 was 31 (20-41)% for localised melanoma, 18 (12-22)% for regional melanoma and 3 (-5-10)% for melanomas with distant spread. Younger age was predictive of a greater percentage reduction. Treatment benefits are inferred from the higher survivals in 2000-2009 but uncertainty remains due to incomplete data to adjust for non-treatment factors and a lack of treatment data. Registries should use new information systems to collect more complete data on stage, other prognostic indicators, co-morbidities and treatment, to provide more definitive and detailed information on population effects of cancer control.

Avoidable deaths and random variation in patients' survival

Background:Random error in the numbers of avoidable deaths among cancer patients has not been considered in earlier studies.Methods:Methods to obtain valid confidence intervals (CIs) for numbers of avoidable deaths were developed. The excess mortality rates were estimated for patients diagnosed with colon cancer in five cancer control regions in Finland during 2000–2007 using a relative survival regression model. Numbers of avoidable deaths due to colon cancer and other causes, respectively, were estimated in different scenarios.Results:Altogether, 4139 and 1335 out of 10 772 patients under 90 years at diagnosis were estimated to have died due to colon cancer and other causes, respectively, during the first 5 years after diagnosis. If all the patients had shared the relative survival of the largest cancer control region to which the country capital belongs, the estimated number of avoidable deaths would have been 146 (95% CI 3–290).Conclusion:Random error in numbers of avoidable deaths, often substantial, can be quantified by realistic error margins, based on appropriate statistical methods.

Conditional relative survival of cancer patients and conditional probability of death

Little information is available on the conditional probabilities of death among patients who survive for >5 years after a diagnosis with cancer. The objective of this study was to estimate the conditional probabilities of death for breast cancer, prostate cancer, colorectal cancer, and lung cancer in France. The study included data from the French Network of Cancer Registries from 205,562 patients aged <75 years who were diagnosed with cancer between 1989 and 1997. The conditional probabilities of death were calculated by using a relative survival regression model in which age was included as a covariate. After the first year and until 10 years after diagnosis, the annual probability of death decreased dramatically for colorectal cancer: It was the same in all age groups after 3 years, and it was approximately 1% at 10 years. For prostate cancer, the decrease was not as great, and the conditional probability of death remained higher among younger patients at >4% at 10 years. During the 3 years after diagnosis, the probability of death was greater for older patients with breast cancer; then, it decreased less for younger patients compared with older patients, leading to a greater conditional probability of death among younger patients at 4 years and up to 10 years. The annual probability of death in patients with lung cancer decreased for both sexes but remained substantially higher for men than for women, reaching approximately 8% and 5%, respectively, at 10 years. Further studies would facilitate a better understanding of the observed differences in relative survival within European countries.

An approach to estimation in relative survival regression

The goal of relative survival methodology is to compare the survival experience of a cohort with that of the background population. Most often an additive excess hazard model is employed, which assumes that each person's hazard is a sum of 2 components--the population hazard obtained from life tables and an excess hazard attributable to the specific condition. Usually covariate effects on the excess hazard are assumed to have a proportional hazards structure with parametrically modelled baseline. In this paper, we introduce a new fitting procedure using the expectation-maximization algorithm, treating the cause of death as missing data. The method requires no assumptions about the baseline excess hazard thus reducing the risk of bias through misspecification. It accommodates the possibility of knowledge of cause of death for some patients, and as a side effect, the method yields an estimate of the ratio between the excess and the population hazard for each subject. More importantly, it estimates the baseline excess hazard flexibly with no additional degrees of freedom spent. Finally, it is a generalization of the Cox model, meaning that all the wealth of options in existing software for the Cox model can be used in relative survival. The method is applied to a data set on survival after myocardial infarction, where it shows how a particular form of the hazard function could be missed using the existing methods.

Making relative survival analysis relatively easy

In survival analysis we are interested in time from the beginning of an observation until certain event (death, relapse, etc.). We assume that the final event is well defined, so that we are never in doubt whether the final event has occurred or not. In practice this is not always true. If we are interested in cause-specific deaths, then it may sometimes be difficult or even impossible to establish the cause of death, or there may be different causes of death, making it impossible to assign death to just one cause. Suicides of terminal cancer patients are a typical example. In such cases, standard survival techniques cannot be used for estimation of mortality due to a certain cause. The cure to the problem are relative survival techniques which compare the survival experience in a study cohort to the one expected should they follow the background population mortality rates. This enables the estimation of the proportion of deaths due to a certain cause. In this paper, we briefly review some of the techniques to model relative survival, and outline a new fitting method for the additive model, which solves the problem of dependency of the parameter estimation on the assumption about the baseline excess hazard. We then direct the reader's attention to our R package relsurv that provides functions for easy and flexible fitting of all the commonly used relative survival regression models. The basic features of the package have been described in detail elsewhere, but here we additionally explain the usage of the new fitting method and the interface for using population mortality data freely available on the Internet. The combination of the package and the data sets provides a powerful informational tool in the hands of a skilled statistician/informatician.

Quantitative exploration of possible reasons for the recent improvement in breast cancer survival

Breast cancer mortality has been declining in many countries including Canada because of improvements in survival. This study attempts to explain observed trends in breast cancer survival with special attention given to the role of improvements in early detection and treatment. This study is based on 4,312 women diagnosed with primary invasive breast carcinoma treated in a Canadian breast center between 1976 and 2000 and followed to the end of 2001. Observed and relative survival rates were calculated. Multivariate relative survival regression models were used to assess trends in breast cancer survival over the study period. The proportion of women with small tumors (< or = 10 mm) was higher in late 1990s, while that of women with regional involvement was lower compared to earlier periods. Adjuvant chemotherapy or endocrine therapy use increased steadily from 6.6% to 84.0% during the study period. Five-year relative survival rates ranged between 82.1% and 83.7% between 1976 and 1990, and increased thereafter to reach 87.6% in 1991-95, and 92.1% in 1996-2000. During the first five years after diagnosis, women diagnosed in 1991-95 and 1996-2000 experienced a reduction in breast cancer mortality of 28% (Relative Risk (RR)= 0.72; 95% CI: 0.59-0.89) and 49% (RR = 0.51; 95% CI: 0.39-0.68) respectively compared to women diagnosed in 1976-90. Improvement in breast cancer survival in 1990's could not be explained by characteristics of women, biology of the tumor, advancements in early detection and type of initial treatments. A substantial increase in breast cancer survival was observed in the 1990s but the reasons for this improvement remain elusive. Better knowledge of these reasons could help not only to further reduce the burden related to breast cancer but also the burden related to other major cancer sites.

Relative survival analysis in R

Relative survival techniques are used to compare the survival experience in a study cohort with the one expected should they follow the background population mortality rates. The techniques are especially useful when the cause-specific death information is not accurate or not available since they provide a measure of excess mortality in a group of patients with a certain disease. There are several approaches to modeling relative survival, but there is no widely used statistical package that would incorporate the relevant techniques. The existing software was mostly written by the authors of different methods, in different computer languages and with different requirements for the data input, which makes it almost impossible for a user to choose between available models. We describe our R package relsurv that provides functions for easy and flexible fitting of several relative survival regression models.

A relative survival regression model using B-spline functions to model non-proportional hazards.

Relative survival, a method for assessing prognostic factors for disease-specific mortality in unselected populations, is frequently used in population-based studies. However, most relative survival models assume that the effects of covariates on disease-specific mortality conform with the proportional hazards hypothesis, which may not hold in some long-term studies. To accommodate variation over time of a predictor's effect on disease-specific mortality, we developed a new relative survival regression model using B-splines to model the hazard ratio as a flexible function of time, without having to specify a particular functional form. Our method also allows for testing the hypotheses of hazards proportionality and no association on disease-specific hazard. Accuracy of estimation and inference were evaluated in simulations. The method is illustrated by an analysis of a population-based study of colon cancer.

Parametric Relative Survival Regression Using Generalized Linear Models with Application to Hodgkin’s Lymphoma

SUMMARY Changes in survival rates during 1940–1992 for patients with Hodgkin’s disease are studied by using population-based data. The aim of the analysis is to identify when the breakthrough in clinical trials of chemotherapy treatments started to increase population survival rates, and to find how long it took for the increase to level off, indicating that the full population effect of the breakthrough had been realized. A Weibull relative survival model is used because the model parameters are easily interpretable when assessing the effect of advances in clinical trials. However, the methods apply to any relative survival model that falls within the generalized linear models framework. The model is fitted by using modifications of existing software (SAS, GLIM) and profile likelihood methods. The results are similar to those from a cause-specific analysis of the data by Feuer and co-workers. Survival started to improve around the time that a major chemotherapy breakthrough (nitrogen mustard, Oncovin, prednisone and procarbazine) was publicized in the mid 1960s but did not level off for 11 years. For the analysis of data where the cause of death is obtained from death certificates, the relative survival approach has the advantage of providing the necessary adjustment for expected mortality from causes other than the disease without requiring information on the causes of death.