Comparison of nonbreast cancer incidence, survival and mortality between breast screening program participants and nonparticipants

Comparisons of cancer mortality between users and nonusers of screening are potentially biased because of the effects of self-selection. Previous studies of breast screening have found that individuals likely to participate have lower breast cancer mortality than those unlikely to participate. This study compares the incidence, survival and mortality for all cancer types other than breast between participants and nonparticipants in a service screening mammography program. British Columbian females having their first mammogram between the ages of 40 and 79 and the years 1988 and 2004 were identified as a cohort of "participants". Person-years of follow-up of participants were aggregated by age and year. Nonparticipant person-years were obtained by subtraction from the total female population. Cancer diagnoses other than breast were identified for participants and nonparticipants. Age, calendar year, and income adjusted relative risks of cancer incidence were estimated from generalized additive models with Poisson errors. Hazard ratios were estimated by Cox regression. Observed cancer mortality in participants was compared with expected mortality generated from nonparticipant incidence and survival rates. Incidence rates of cancer showed a mixed relationship with some elevated, some decreased and others similar to nonparticipant rates. Cancer survival was higher among participants for most cancer types, with an overall hazard ratio of 0.76 (0.73-0.79). Observed mortality in participants was less than expected for most cancers, with an overall mortality ratio of 0.60 (0.58-0.62). The general cancer experience of screening program participants is different from that of the general population.     

Patients with previous cancer should not be excluded in international comparative cancer survival studies

In international comparisons or time trend analyses of cancer survival, it is common practice to restrict analyses to patients with a first cancer, i.e., to exclude patients with previous cancer diagnoses. However, the proportion of cancer patients with known previous cancer depends on the time cancer registries have been running, which results in varying proportions of excluded patients across registries. If prognosis of patients with second cancers differs from prognosis of patients with first cancers, varying exclusions may bias survival comparisons. We empirically evaluate the dependence of proportions of patients recorded as having a first cancer on time since initiation of cancer registration and the impact of excluding patients with known previous cancer on cancer survival estimates using the data of the nationwide Finnish Cancer Registry. Among 20 common cancer sites investigated, the proportion of "first cancers" varied between 97.4 and 99.7% in 1953-1957, the first 5-years of cancer registration, and decreased continuously to levels between 83.9 and 92.7% in 1993-1997. Excluding patients with a previous cancer diagnosis had little impact on estimates of survival of cancer patients diagnosed in 1953-1957, but increased 5-year relative survival estimates among patients diagnosed in 1993-1997 for each of the 20 cancers. The extent of the increase varied by cancer site and age. The increase ranged up to 2.9% points for crude and up to 1.7% points for age adjusted 5-year relative survival. These results caution against exclusion of patients with previous cancer diagnosis in comparative analyses of cancer survival.     

Should relative survival be used with lung cancer data?

Background:

Under certain assumptions, relative survival is a measure of net survival based on estimating the excess mortality in a study population when compared with the general population. Background mortality estimates are usually taken from national life tables that are broken down by age, sex and calendar year. A fundamental assumption of relative survival methods is that if a patient did not have the disease of interest then their probability of survival would be comparable to that of the general population. It is argued, as most lung cancer patients are smokers and therefore carry a higher risk of smoking-related mortalities, that they are not comparable to a population where the majority are likely to be non-smokers.

Methods:

We use data from the Finnish Cancer Registry to assess the impact that the non-comparability assumption has on the estimates of relative survival through the use of a sensitivity analysis.

Results:

Under realistic estimates of increased all-cause mortality for smokers compared with non-smokers, the bias in the estimates of relative survival caused by the non-comparability assumption is negligible.

Conclusion:

Although the assumption of comparability underlying the relative survival method may not be reasonable, it does not have a concerning impact on the estimates of relative survival, as most lung cancer patients die within the first 2 years following diagnosis. This should serve to reassure critics of the use of relative survival when applied to lung cancer data.     

Implications of incomplete registration of deaths on long‐term survival estimates from population‐based cancer registries

International comparison of population‐based cancer survival is a key component of monitoring progress against cancer. Its validity depends to an unknown degree on completeness of ascertainment of deaths in the cancer registries involved which may vary according to legal and administrative circumstances. The aim of this study was to assess the impact of incomplete registration of deaths through various mechanisms on the validity of long‐term absolute and relative survival estimates. For that purpose, we simulated underascertainment of deaths through linkage failure of registry data with death certificates with probabilities between 0.1 and 5%, and underascertainment of deaths by unregistered annual emigration with probabilities between 0.05 and 2%, using data from the Finnish Cancer Registry. The expected impact on estimates of 5‐, 10‐ and 15‐year absolute and relative survival was assessed. We demonstrate that even modest levels of under‐registration of deaths may lead to severe overestimation of long‐term survival estimates, ranging from 0 to 31 percent units in the scenarios assessed. In general, relative survival is much more affected than absolute survival, and potential problems are much larger for relative survival estimates in older compared with younger patients. Potential overestimation strongly increases with length of follow‐up, and this increase is particularly pronounced for under‐registration of deaths because of unrecorded emigration. Every effort should be made in cancer registry based survival analyses to ascertain deaths with close to 100% completeness. When such completeness cannot be achieved, long‐term relative survival estimates and their comparison across populations must be interpreted with much caution. © 2009 UICC     

Bias in relative survival methods when using incorrect life-tables: lung and bladder cancer by smoking status and ethnicity in New Zealand

Relative survival and excess mortality approaches are commonly used to estimate and compare net survival from cancer. These approaches are based on the assumption that the underlying (non-cancer) mortality rate of cancer patients is the same as that of the general population. This assumption is likely to be violated particularly in the context of smoking-related cancers. The magnitude of this bias has not been estimated. The objective of this article is to estimate the bias in relative survival ratios (RSRs) and excess mortality rate ratios (EMRRs) from using total population compared to correct subpopulation specific life-tables. Analyses were conducted on 1996-2001 linked census-cancer data (including smoking status) for people with lung and bladder cancer, using sex-specific (standard practice), sex- and ethnic-specific, sex- and smoking-specific and sex-, ethnic- and smoking-specific life-tables. Five-year RSRs using sex-specific life-tables, compared to fully stratified life-tables, were underestimated by 10-25% for current smoking and Maori populations. For example, the current smoker male bladder cancer RSR was 0.700 for sex-specific life-tables, compared to 0.838 for fully stratified life-tables. Similarly, EMRRs comparing current to never smokers and Maori to non-Maori were overestimated using sex-specific life-tables only: modestly only for lung cancer, but markedly for bladder cancer. For example, the EMRR comparing current to never smokers with bladder cancer in a fully adjusted regression model was 1.475 when using sex-specific life-tables only, but reduced to 1.098 when using fully stratified life-tables. Substantial bias can occur when estimating relative cancer survival across subpopulations if non-matching life-tables are used.     

Recent cancer survival in Germany: An analysis of common and less common cancers

The monitoring of cancer survival by population‐based cancer registries is a prerequisite to evaluate the current quality of cancer care. Our study provides 1‐, 5‐ and 10‐year relative survival as well as 5‐year relative survival conditional on 1‐year survival estimates and recent survival trends for Germany using data from 11 population‐based cancer registries, covering around one‐third of the German population. Period analysis was used to estimate relative survival for 24 common and 11 less common cancer sites for the period 2007–2010. The German and the United States survival estimates were compared using the Surveillance, Epidemiology and End Results 13 database. Trends in cancer survival in Germany between 2002–2004 and 2008–2010 were described. Five‐year relative survival increased in Germany from 2002–2004 to 2008–2010 for most cancer sites. Among the 24 most common cancers, largest improvements were seen for multiple myeloma (8.0% units), non‐Hodgkin lymphoma (6.2% units), prostate cancer (5.2% units) and colorectal cancer (4.6% units). In 2007–2010, the survival disadvantage in Germany compared to the United States was largest for cancers of the mouth/pharynx (?11.0% units), thyroid (?6.8% units) and prostate (?7.5% units). Although survival estimates were much lower for elderly patients in both countries, differences in age patterns were observed for some cancer sites. The reported improvements in cancer survival might reflect advances in the quality of cancer care on the population level as well as increased use of screening in Germany. The survival differences across countries and the survival disadvantage in the elderly require further investigation.     

Estimating relative survival among people registered with cancer in England and Wales.

Because routinely collected survival data for cancer patients in England and Wales do not typically specify cause of death, conventional estimates of survival in cancer patients based on such data are a measure of their mortality from all causes rather than their mortality due to cancer. As a result, trends in survival over time are difficult to interpret because changes in overall survival may well reflect changes in the risk of death from other causes, rather than from the cancer of interest. One way of overcoming this problem is to use some form of 'relative survival' defined as a measure of survival corrected for the effect of other independent causes of death. Since this concept was first introduced, various methods for calculating relative survival have been proposed and this had led to some confusion as to the most appropriate choice of estimate. This paper aims to provide an introduction to the concept of relative survival and reviews some of the suggested methods of estimation. In addition, a particularly simple, but robust approach, is highlighted based on expected and observed mortality. This method is illustrated using preliminary data from the Office for National Statistics on cancer survival in patients born after 1939 and diagnosed with cancer during 1972-84. The examples presented, although limited to analyses on a small number of selected sites, highlight some encouraging trends in survival in people aged under 35 diagnosed with leukaemia, Hodgkin's disease and testicular cancer during this period.     

Explaining gastric cancer survival differences among European countries

Wide geographic variability in incidence and mortality rates for gastric cancer exists throughout the world despite persistent decreases over several decades. Variability in survival from gastric cancer is also evident and countries with higher incidence rates of gastric cancer show better survival rates than countries with lower incidence. The aim of this study was to identify reasons for the association between incidence and survival and to obtain survival estimates and differences corrected for this variation, thus facilitating further interpretation by clinical factors such as stage and treatment. Relative survival rates for gastric cancer derived from the EUROCARE-2 database for 47 cancer registries in 17 European countries were analyzed with regression methods to adjust differences by age, sex, period of diagnosis, subsite of the stomach, histologic type and stage at diagnosis. Overall, nearly 60% of the variability in gastric cancer relative survival was explained by differences in these variables. Factors are related to treatment and general management of patients is expected to explain the residual variability in gastric cancer survival between European countries. There is a need to improve completeness and standardization of detailed information collected on gastric cancer patients to allow detailed comparative analyses and interpretation.     

The impact of life tables adjusted for smoking on the socio-economic difference in net survival for laryngeal and lung cancer

Background:

Net survival is a key measure in cancer control, but estimates for cancers that are strongly associated with smoking may be biased. General population life tables represent background mortality in net survival, but may not adequately reflect the higher mortality experienced by smokers.

Methods:

Life tables adjusted for smoking were developed, and their impact on net survival and inequalities in net survival for laryngeal and lung cancers was examined.

Results:

The 5-year net survival estimated with smoking-adjusted life tables was consistently higher than the survival estimated with unadjusted life tables: 7% higher for laryngeal cancer and 1.5% higher for lung cancer. The impact of using smoking-adjusted life tables was more pronounced in affluent patients; the deprivation gap in 5-year net survival for laryngeal cancer widened by 3%, from 11% to 14%.

Conclusions:

Using smoking-adjusted life tables to estimate net survival has only a small impact on the deprivation gap in survival, even when inequalities are substantial. Adjusting for the higher, smoking-related background mortality did increase the estimates of net survival for all deprivation groups, and may be more important when measuring the public health impact of differences or changes in survival, such as avoidable deaths or crude probabilities of death.     

Spatial variation in prostate cancer survival in the Northern and Yorkshire region of England using Bayesian relative survival smoothing

Primary Care Trust (PCT) estimates of survival lack robustness as there are small numbers of deaths per year in each area, even when incidence is high. We assess PCT-level spatial variation in prostate cancer survival using Bayesian spatial models of excess mortality. We extracted data on men diagnosed with prostate cancer between 1990 and 1999 from the Northern and Yorkshire Cancer Registry and Information Service database. Models were adjusted for age at diagnosis, period of diagnosis and deprivation. All covariates had a significant association with excess mortality; men from more deprived areas, older age at diagnosis and diagnosed in 1990-1994 had higher excess mortality. The unadjusted relative excess risks (RER) of death by PCT ranged from 0.75 to 1.66. After adjustment, areas of high and low excess mortality were smoothed towards the mean, and the RERs ranged from 0.74 to 1.49. Using Bayesian smoothing techniques to model cancer survival by geographic area offers many advantages over traditional methods; estimates in areas with small populations or low incidence rates are stabilised and shrunk towards local and global risk estimates improving reliability and precision, complex models are easily handled and adjustment for covariates can be made.     

Cancer net survival on registry data: Use of the new unbiased Pohar‐Perme estimator and magnitude of the bias with the classical methods

Net survival, the survival which might occur if cancer was the only cause of death, is a major epidemiological indicator required for international or temporal comparisons. Recent findings have shown that all classical methods used for routine estimation of net survival from cancer‐registry data, sometimes called “relative‐survival methods,” provide biased estimates. Meanwhile, an unbiased estimator, the Pohar‐Perme estimator (PPE), was recently proposed. Using real data, we investigated the magnitude of the errors made by four “relative‐survival” methods (Ederer I, Hakulinen, Ederer II and a univariable regression model) vs. PPE as reference and examined the influence of time of follow‐up, cancer prognosis, and age on the errors made. The data concerned seven cancer sites (2,51,316 cases) collected by FRANCIM cancer registries. Net survivals were estimated at 5, 10 and 15 years postdiagnosis. At 5 years, the errors were generally small. At 10 years, in good‐prognosis cancers, the errors made in nonstandardized estimates with all classical methods were generally great (+2.7 to +9% points in prostate cancer) and increased in age‐class estimations (vs. 5‐year ones). At 15 years, in bad‐ or average‐prognosis cancers, the errors were often substantial whatever the nature of the estimation. In good‐prognosis cancers, the errors in nonstandardized estimates of all classical methods were great and sometimes very important. With all classical methods, great errors occurred in age‐class estimates resulting in errors in age‐standardized estimates (+0.4 to +3.2% points in breast cancer). In estimating net survival, cancer registries should abandon all classical methods and adopt the new Pohar‐Perme estimator.     

Multiple tumours in survival estimates

In international comparisons of cancer registry based survival it is common practice to restrict the analysis to first primary tumours and exclude multiple cancers. The probability of correctly detecting subsequent cancers depends on the registry’s running time, which results in different proportions of excluded patients and may lead to biased comparisons. We evaluated the impact on the age-standardised relative survival estimates of also including multiple primary tumours.Data from 2,919,023 malignant cancers from 69 European cancer registries participating in the EUROCARE-4 collaborative study were used. A total of 183,683 multiple primary tumours were found, with an overall proportion of 6.3% over all the considered cancers, ranging from 0.4% (Naples, Italy) to 12.9% (Iceland). The proportion of multiple tumours varied greatly by type of tumour, being higher for those with high incidence and long survival (breast, prostate and colon-rectum). Five-year relative survival was lower when including patients with multiple cancers. For all cancers combined the average difference was –0.4 percentage points in women and –0.7 percentage points in men, and was greater for older registries. Inclusion of multiple tumours led to lower survival in 44 out of 45 cancer sites analysed, with the greatest differences found for larynx (–1.9%), oropharynx (–1.5%), and penis (–1.3%).Including multiple primary tumours in survival estimates for international comparison is advisable because it reduces the bias due to different observation periods, age, registration quality and completeness of registration. The general effect of inclusion is to reduce survival estimates by a variable amount depending on the proportion of multiple primaries and cancer site.     

Differences in survival for patients with familial and sporadic cancer

Family history of cancer is a well‐known risk factor but the role of family history in survival is less clear. The aim of this study was to investigate the association between family history and cancer survival for the common cancers in Sweden. Using the Swedish population‐based registers, patients diagnosed with the most common cancers were followed for cancer‐specific death during 1991–2010. We used multivariate proportional hazards (Cox) regression models to contrast the survival of patients with a family history of cancer (individuals whose parent or sibling had a concordant cancer) to the survival of patients without a family history. Family history of cancer had a modest protective effect on survival for breast cancer (hazard ratio (HR) = 0.88, 95% confidence interval (95% CI) = 0.81 to 0.96) and prostate cancer (HR = 0.82, 95% CI = 0.75 to 0.90). In contrast, family history of cancer was associated with worse survival for nervous system cancers (HR = 1.24, 95% CI = 1.05 to 1.47) and ovarian cancer (HR = 1.20, 95% CI = 1.01 to 1.43). Furthermore, the poorer survival for ovarian cancer was consistent with a higher FIGO stage and a greater proportion of more aggressive tumors of the serous type. The better survival for patients with a family history of breast and prostate cancer may be due to medical surveillance of family members. The poor survival for ovarian cancer patients with an affected mother or sister is multifactorial, suggesting that these cancers are more aggressive than their sporadic counterparts.     

Colorectal cancer survival in the Nordic countries and the United Kingdom: excess mortality risk analysis of 5 year relative period survival in the period 1999 to 2000

A deficit in colorectal cancer survival in Denmark and in the UK compared to Sweden, Norway and Finland was found in the EUROCARE studies. We set out to explore if these differences still exist. Patients diagnosed with colorectal cancer as their first invasive cancer at age 15-89 in the period 1994-2000 were identified using data from 11 cancer registries in the UK and from four Nordic countries. Five-year relative period survival using deaths in 1999-2000 following cancers diagnosed in 1994-2000 was analysed with excess mortality risk modelling. Follow-up time since diagnosis with age as an effect-modifier in the first half year was the most important factor with the highest excess risk of death immediately after diagnosis and with higher age and decreasing with length of follow-up. Variations between countries were bigger in the first half year following diagnosis than in the interval 0.5-5 years with about 30% higher risk in UK and Denmark. The differences between countries are still substantial and the order has not changed, even if the five year relative survival has improved since the EUROCARE studies. Patient management, diagnostics, and comorbidity likely explain the excess deaths in UK and Denmark during the first 6 months. The effect of stage and quality of management and treatment should be examined in population based studies with detailed patient information. Use of more detailed age-intervals than conventionally applied in survival studies proved to be important in statistical modelling and is recommended for future studies.     

Loss-adjusted survival of cervix cancer in Khon Kaen, Northeast Thailand

For incident cancers of the cervix uteri (601 cases) registered in the population-based cancer registry of Khon Kaen province, Northeast Thailand, in 1985-1990 loss-adjusted survival probabilities were estimated by a logistic regression model with four prognostic factors (age at diagnosis, stage of disease, place of residence and treatment), and compared with observed survival, estimated by the actuarial method. All patients were followed up for a minimum of 5 years, using both passive and active methods. In all, 27.6% of patients were lost to follow-up within 5 years of the index date. The overall observed survival at 5 years was 56.8% and loss-adjusted survival was 54.7%. The difference between the loss-adjusted and observed survival at 5 years was small: 2.1% overall, varying between 0.8 and 3.5 percent units for any prognostic group. The assumption of independence of loss to follow-up and death in the calculation of survival by the actuarial method in this, and probably in other, population-based series, is reasonable and leads to no material bias in the estimates.     

Trends in survival after cancer diagnosis among HIV‐infected individuals between 1992 and 2009. Results from the FHDH‐ANRS CO4 cohort

Although the decline in cancer mortality rates with the advent of combination antiretroviral therapy (cART) in HIV‐infected individuals can be mostly explained by a decrease in cancers incidence, we looked here if improved survival after cancer diagnosis could also contribute to this decline. Survival trends were analyzed for most frequent cancers in the HIV‐infected population followed in the French Hospital Database on HIV: 979 and 2,760 cases of visceral and non‐visceral Kaposi's sarcoma (KS), 2,339 and 461 cases of non‐Hodgkin lymphoma (NHL) and Hodgkin's lymphoma (HL), 446 lung, 312 liver and 257 anal cancers. Five‐year Kaplan–Meier survival rates were estimated for four periods: 1992–1996, 1997–2000, 2001–2004 and 2005–2009. Cox proportional hazard models were used to compare survival across the periods, after adjustment for confounding factors. For 2001–2004, survival was compared to the general population after standardization on age and sex. Between the pre‐cART (1992–1996) and early‐cART (1997–2000) periods, survival improved after KS, NHL, HL and anal cancer and remained stable after lung and liver cancers. During the cART era, 5‐year survival improved after visceral and non‐visceral KS, NHL, HL and liver cancer, being 83, 92, 65, 87 and 19% in 2005–2009, respectively, and remained stable after lung and anal cancers, being 16 and 65%, respectively. Compared with the general population, survival in HIV‐infected individuals in 2001–2004 was poorer for hematological malignancies and similar for solid tumors. For hematological malignancies, survival continues to improve after 2004, suggesting that the gap between the HIV‐infected and general populations will close in the future.     

A comparison of relative and cause‐specific survival by cancer site,age and time since diagnosis

Relative survival (RS) estimates are widely used by cancer registries, mainly because they do not rely on the well‐documented deficiencies of cause of death information. The aim of our study was to compare 5‐year cause‐specific survival (CSS) estimates and 5‐year RS estimates for different cancer sites by age and time since diagnosis, and discuss possible reasons for observed differences. Using data from the Cancer Registry of Norway, we identified 200,008 patients diagnosed with cancer at one of the 48 sites included in this analysis during the period 1996–2005, and followed them up until the end of 2010. CSS estimates were calculated (i) considering cause of death to be the cancer that was originally diagnosed and (ii) considering the cause of death to be a cancer within the same organ system. For most cancer sites the difference between CSS and RS estimates was small (<5%). The greatest differences were seen for rarer cancers such as mediastinum and Kaposi sarcoma. Including deaths from the same organ system in the calculation of CSS further reduced the differences for many sites. For younger age groups and shorter time since diagnosis, RS and CSS estimates tended to be similar, whereas CSS estimates tended to be lower than RS estimates with longer time since diagnosis in the oldest age groups. When compared to RS estimates CSS estimates were reliable for most of the cancer sites included in our analysis. There are, however, some exceptions where CSS estimates may not be recommended, including for rarer cancers and for patients aged 85 and above.     

Reduction of population-based cancer survival estimates by trace back of death certificate notifications: an empirical illustration

BackgroundSurvival studies using data from population-based cancer registries allow assessing effectiveness of cancer care on a population level. However, population-based cancer registries differ in the proportion of cases first notified by death certificate, as well as in the efforts to trace back such death certificate notifications (DCN). We aimed to assess the impact of such trace back on population-based cancer survival estimates.Materials and methodsIn this study from the population-based Saarland Cancer Registry (Germany) we investigated the survival experience of successfully traced back DCN cases from 1994 to 2003. Five-year relative survival of patients with DCN cancers and the effect of trace back on population-based 5-year relative survival estimates were analysed by age and tumour site.ResultsTwelve percent of all cancers were DCN and such cases occurred most often amongst sites with poor prognosis and amongst elderly patients. Approximately half of DCN cases could be successfully traced back. Five-year relative survival of patients with DCN cancers with trace back was 2%. The inclusion of DCN cancers with additional registrations reduced the 5-year relative survival estimate for all cancers combined by 4% points. Reductions were stronger for older patients and highly fatal cancers.ConclusionsTrace back results in increased inclusion of patients with very poor prognosis. Varying extent of trace back across registries may compromise comparability of cancer survival estimates and should be taken into account in comparative cancer survival studies.     

A wide difference in cancer survival between middle aged and elderly patients in Europe

Nowadays the burden of cancer in elderly people has reached an alarming extent. The purpose of this study is comparing cumulative and conditional relative survival in elderly patients between 65 and 84 years and younger adults aged from 55 to 64. Fifty-three cancer registries of 22 European countries, participating in the EUROCARE-3 programme, collected information on the cases diagnosed over the period 1990-1994. We computed cumulative and conditional relative survival for 16 cancer sites. Middle aged patients experienced a better prognosis than the elderly for all cancer sites, in both sexes and the differences were more marked at 1 than 5 years since diagnosis. The very large differences noted in the first period after cancer detection declined in the subsequent years and, when 5-years conditional survival was considered, for several cancers the elderly and younger adults had the same probabilities of surviving. The death relative excess risks (RERs) in the elderly with respect younger individuals were really very high and markedly larger at 1 than 5 years, and in women than men. Genitourinary and gynaecological cancers showed the highest RERs, around 2.0 and between 1.5 and 2.5 respectively. This very high early mortality could be due not only to clinical aspects: the barriers to health care access and a consequent late diagnosis might represent for elderly patients the main determinant of this very large prognostic disadvantage. In conclusion, clinical management of cancer in the elderly remains a major issue to be faced with complex social and health care policies.