A simple non-linear model in incidence prediction

Describing and Modeling the Burden of Hospitalization of Patients With Neoplasms in Ghana Using Routine Health Data for 2012-2017

The increasing cancer burden calls for reliable data on current and future associated hospitalizations to enable health care resource planning, especially in low- and middle-income countries. We provide nationwide estimates of the current and future burden of hospitalization because of neoplasms in Ghana.

Other data sets are useful to estimate epidemiologic trends in settings where cancer registries are not available.

A nationwide study of the incidence and trends of first and multiple basal cell carcinoma in the Netherlands and prediction of future incidence

BACKGROUND

Basal cell carcinoma (BCC) is the most frequently diagnosed malignancy worldwide and an ever increasing annual incidence is observed. However, nationwide registries of BCCs are very rare and often extrapolation of the data was necessary to estimate the absolute number of diagnoses. Since September 2016, all histopathologically confirmed BCCs are registered in the Netherlands, due to developments in automatic notification and import in the Netherlands cancer registry. This offers the unique possibility to assess the nationwide population-based incidence of first and multiple BCC.

OBJECTIVES

The current study aims to assess nationwide incidence and trends of first and multiple BCC in the Netherlands and to predict incidence rates up to 2029.

METHODS

All patients with histopathologically confirmed BCC between 2001 and 2019 were selected from the population-based Netherlands Cancer Registry. Age-standardized incidence rates were calculated and trends were analyzed with use of the estimated annual percentage change. Prediction of BCC incidence rates up to 2029 was based on a regression model.

RESULTS

In total, 601,806 patients were diagnosed with a first BCC over the period 2001-2019. The age-standardized incidence rates for both men and women with a first BCC increased over these years from 158 to 304 and 124 to 274 per 100,000 person-years, respectively. For male and female patients aged between 30-39 years, decreases in annual incidences of -3.6% and -3.0% were found in recent years, respectively. For patients aged 50 years or older an ever increasing trend was found. A quarter of the patients with a first primary BCC developed one or more subsequent BCCs within three years. Increases in incidence of 30.4% (male) and 25.3% (female) is expected in the next 10 years.

CONCLUSIONS

BCC incidence doubled over the past decades. Trends seemed to stabilize in recent years for patients aged below 50 years. This might be a first sign of a decreasing trend. The incidence keeps rising in patients aged 50 years and older. In the next decade a further increase in BCC incidence is expected.

Planning for tomorrow: global cancer incidence and the role of prevention 2020-2070

Cancer is currently the first or second most common contributor to premature mortality in most countries of the world. The global number of patients with cancer is expected to rise over the next 50 years owing to the strong influence of demographic changes, such as population ageing and growth, on the diverging trends in cancer incidence in different regions. Assuming that the latest incidence trends continue for the major cancer types, we predict a doubling of the incidence of all cancers combined by 2070 relative to 2020. The greatest increases are predicted in lower-resource settings, in countries currently assigned a low Human Development Index (HDI), whereas the predicted increases in national burden diminish with increasing levels of national HDI. Herein, we assess studies modelling the future burden of cancer that underscore how comprehensive cancer prevention strategies can markedly reduce the prevalence of major risk factors and, in so doing, the number of future cancer cases. Focusing on an in-depth assessment of prevention strategies that target tobacco smoking, overweight and obesity, and human papillomavirus infection, we discuss how stepwise, population-level approaches with amenable goals can avert millions of future cancer diagnoses worldwide. In the absence of a step-change in cancer prevention delivery, tobacco smoking will remain the leading preventable cause of cancer, and overweight and obesity might well present a comparable opportunity for prevention, given its increasing prevalence globally in the past few decades. Countries must therefore instigate national cancer control programmes aimed at preventing cancer, and with some urgency, if such programmes are to yield the desired public health and economic benefits in this century.

Incidence of Multiple vs First Cutaneous Squamous Cell Carcinoma on a Nationwide Scale and Estimation of Future Incidences of Cutaneous Squamous Cell Carcinoma

Importance

Until now, most studies on cutaneous squamous cell carcinoma (cSCC) incidence rates concerned only the first cSCC per patient. Given the increase in incidence rates and the frequent occurrence of subsequent cSCCs per patient, population-based data on the incidence rates of both first and multiple cSCCs are needed.

Objectives

To calculate annual age-standardized incidence rates for histopathologically confirmed first and multiple cSCCs per patient and to estimate future cSCC incidence rates up to 2027.

Design, Setting, and Participants

A nationwide population-based epidemiologic cohort study used cancer registry data on 145 618 patients with a first histopathologically confirmed cSCC diagnosed between January 1, 1989, and December 31, 2017, from the Netherlands Cancer Registry and all patients with multiple cSCCs diagnosed in 2017.

Main Outcomes and Measures

Age-standardized incidence rates for cSCC-standardized to the European Standard Population 2013 and United States Standard Population 2000-were calculated per sex, age group, body site, and disease stage. A regression model with positive slope was fitted to estimate cSCC incidence rates up to 2027.

Results

A total of 145 618 patients in the Dutch population (84 572 male patients [58.1%]; mean [SD] age, 74.5 [11.5] years) received a diagnosis of a first cSCC between 1989 and 2017. Based on incident data, European Standardized Rates (ESRs) increased substantially, with the highest increase found among female patients from 2002 to 2017, at 8.2% (95% CI, 7.6%-8.8%) per year. The ESRs for first cSCC per patient in 2017 were 107.6 per 100 000 person-years (PY) for male patients, an increase from 40.0 per 100 000 PY in 1989, and 68.7 per 100 000 PY for female patients, an increase from 13.9 per 100 000 PY in 1989, which corresponds with a US Standardized Rate of 71.4 per 100 000 PY in 2017 for men and 46.4 per 100 000 PY in 2017 for women. Considering multiple cSCCs per patient, ESRs increased by 58.4% for men (from 107.6 per 100 000 PY to 170.4 per 100 000 PY) and 34.8% for women (from 68.7 per 100 000 PY to 92.6 per 100 000 PY). Estimation of ESRs for the next decade show a further increase of 23.0% for male patients (ESR up to 132.4 per 100 000 PY [95% prediction interval, 125.8-139.0 per 100 000 PY]) and 29.4% for female patients (ESR up to 88.9 per 100 000 PY [95% prediction interval, 84.3-93.5 per 100 000 PY]).

Conclusions and Relevance

This nationwide epidemiologic cohort study suggests that incidence rates of cSCC keep increasing, especially among female patients, and that the occurrence of multiple cSCCs per patient significantly adds to the current and future burden on dermatologic health care. Revision of skin cancer policies are needed to halt this increasing trend.

Predicting future cancer burden in the United States by artificial neural networks

Aims: To capture the complex relationships between risk factors and cancer incidences in the US and predict future cancer burden. Materials & methods: Two artificial neural network (ANN) algorithms were adopted: a multilayer feed-forward network (MLFFNN) and a nonlinear autoregressive network with eXogenous inputs (NARX). Data on the incidence of the four most common tumors (breast, colorectal, lung and prostate) from 1992 to 2016 (available from National Cancer Institute online datasets) were used for training and validation, and data until 2050 were predicted. Results: The rapid decreasing trend of prostate cancer incidence started in 2010 will continue until 2018-2019; it will then slow down and reach a plateau after 2050, with several differences among ethnicities. The incidence of breast cancer will reach a plateau in 2030, whereas colorectal cancer incidence will reach a minimum value of 35 per 100,000 in 2030. As for lung cancer, the incidence will decrease from 50 per 100,000 (2017) to 31 per 100,000 in 2030 and 26 per 100,000 in 2050. Conclusion: This up-to-date prediction of cancer burden in the US could be a crucial resource for planning and evaluation of cancer-control programs.

An R package for an integrated evaluation of statistical approaches to cancer incidence projection

Background

Projection of future cancer incidence is an important task in cancer epidemiology. The results are of interest also for biomedical research and public health policy. Age-Period-Cohort (APC) models, usually based on long-term cancer registry data (> 20 yrs), are established for such projections. In many countries (including Germany), however, nationwide long-term data are not yet available. General guidance on statistical approaches for projections using rather short-term data is challenging and software to enable researchers to easily compare approaches is lacking.

Methods

To enable a comparative analysis of the performance of statistical approaches to cancer incidence projection, we developed an R package (incAnalysis), supporting in particular Bayesian models fitted by Integrated Nested Laplace Approximations (INLA). Its use is demonstrated by an extensive empirical evaluation of operating characteristics (bias, coverage and precision) of potentially applicable models differing by complexity. Observed long-term data from three cancer registries (SEER-9, NORDCAN, Saarland) was used for benchmarking.

Results

Overall, coverage was high (mostly > 90%) for Bayesian APC models (BAPC), whereas less complex models showed differences in coverage dependent on projection-period. Intercept-only models yielded values below 20% for coverage. Bias increased and precision decreased for longer projection periods (> 15 years) for all except intercept-only models. Precision was lowest for complex models such as BAPC models, generalized additive models with multivariate smoothers and generalized linear models with age x period interaction effects.

Conclusion

The incAnalysis R package allows a straightforward comparison of cancer incidence rate projection approaches. Further detailed and targeted investigations into model performance in addition to the presented empirical results are recommended to derive guidance on appropriate statistical projection methods in a given setting.

Statistical projection methods for lung cancer incidence and mortality: a systematic review

OBJECTIVES

To identify and summarise all studies using statistical methods to project lung cancer incidence or mortality rates more than 5 years into the future.

STUDY TYPE

Systematic review.

METHODS

We performed a systematic literature search in multiple electronic databases to identify studies published from 1 January 1988 to 14 August 2018, which used statistical methods to project lung cancer incidence and/or mortality rates. Reference lists of relevant articles were checked for additional potentially relevant articles. We developed an organisational framework to classify methods into groups according to the type of data and the statistical models used. Included studies were critically appraised using prespecified criteria.

RESULTS

One hundred and one studies met the inclusion criteria; six studies used more than one statistical method. The number of studies reporting statistical projections for lung cancer increased substantially over time. Eighty-eight studies used projection methods, which did not incorporate data on smoking in the population, and 16 studies used a method which did incorporate data on smoking. Age-period-cohort models (44 studies) were the most commonly used methods, followed by other generalised linear models (35 studies). The majority of models were developed using observed rates for more than 10 years and used data that were considered to be good quality. A quarter of studies provided comparisons of fitted and observed rates. While validation by withholding the most recent observed data from the model and then comparing the projected and observed rates for the most recent period provides important information on the model's performance, only 12 studies reported doing this.

CONCLUSION

This systematic review provides an up-to-date summary of the statistical methods used in published lung cancer incidence or mortality projections. The assessment of the strengths of existing methods will help researchers to better apply and develop statistical methods for projecting lung cancer rates. Some of the common methods described in this review can be applied to the projection of rates for other cancer types or other non-infectious diseases.

The future burden of cancer in Canada: Long-term cancer incidence projections 2013-2042

BACKGROUND

Cancer is the leading cause of death in Canada and the estimated annual spending associated with cancer is approximately $7.5 billion. Projecting the future burden of cancer in Canada is essential for health planning and evaluation. We aimed to estimate the future incidence of cancer in Canada to 2042.

METHODS

Age-sex-region-specific cancer incidence data were obtained for the years 1983-2012 and cancer incidence was projected from 2013 to 2042 for the top five cancer sites. The modelling algorithm combined a mixture of cancer projection methods to select the best-fitted model. When the chosen model produced by the modelling algorithm resulted in estimates that were not consistent with expert opinion, an alternate model was selected that took into consideration historical changes in policy, screening and lifestyle behaviours. Incidence projections were made for Canada and its provinces.

RESULTS

Lung cancer incidence is estimated to rise to 14,866 cases in men and 19,162 in women in 2042. Colorectal cancer incidence is estimated to rise to 28,146 in men and 21,102 in women. Cases of bladder cancer are projected to rise to 10,708 and 3,364 in men and women, respectively. Breast cancer incidence is predicted to rise to 40,712 and prostate cancer incidence is projected to rise to 92,949.

CONCLUSION

These cancer incidence projections up to 2042 can be used for planning cancer control strategies and prevention programs. Given the ongoing changes in the prevalence of risk factors and in cancer prevention policies, these estimates should be interpreted with caution.

Estimating the current and future cancer burden in Canada: methodological framework of the Canadian population attributable risk of cancer (ComPARe) study

INTRODUCTION

The Canadian Population Attributable Risk of Cancer project aims to quantify the number and proportion of cancer cases incident in Canada, now and projected to 2042, that could be prevented through changes in the prevalence of modifiable exposures associated with cancer. The broad risk factor categories of interest include tobacco, diet, energy imbalance, infectious diseases, hormonal therapies and environmental factors such as air pollution and residential radon.

METHODS AND ANALYSIS

Using a national network, we will use population-attributable risks (PAR) and potential impact fractions (PIF) to model both attributable (current) and avoidable (future) cancers. The latency periods and the temporal relationships between exposures and cancer diagnoses will be accounted for in the analyses. For PAR estimates, historical exposure prevalence data and the most recent provincial and national cancer incidence data will be used. For PIF estimates, we will model alternative or 'counterfactual' distributions of cancer risk factor exposures to assess how cancer incidence could be reduced under different scenarios of population exposure, projecting incidence to 2042.

DISSEMINATION

The framework provided can be readily extended and applied to other populations or jurisdictions outside of Canada. An embedded knowledge translation and exchange component of this study with our Canadian Cancer Society partners will ensure that these findings are translated to cancer programmes and policies aimed at population-based cancer risk reduction strategies.

Cancer burden in slovenia with the time trends analysis

BACKGROUND

The aim of our study was to describe cancer burden and time trends of all cancers combined, the most frequent as well as the rare cancers in Slovenia.

PATIENTS AND METHODS

The principal data source was the population-based Cancer Registry of Republic of Slovenia. The cancer burden is presented by incidence and prevalence for the period 1950-2013 and by mortality for years 1985-2013. The time trends were characterized in terms of an average annual percent change estimated by the log-linear joinpoint regression. The Dyba-Hakulinen method was used for estimation of incidence in 2016 and the projections of cancer incidence for the year 2025 were calculated applying the Globocan projection software.

RESULTS

In recent years, near 14,000 Slovenes were diagnosed with cancer per year and just over 6,000 died; more than 94,000 people who were ever diagnosed with cancer are currently living among us. The total burden of cancer is dominated by five most common cancer sites: skin (non-melanoma), colon and rectum, lung, breast and prostate, together representing almost 60% of all new cancer cases. On average the incidence of common cancers in Slovenia is increasing for 3.0% per year in last decade, but the incidence of rare cancers is stable.

CONCLUSIONS

Because cancer occurs more among the elderly, and additionally more numerous post-war generation is entering this age group, it is expected that the burden of this disease will be growing further, even if the level of risk factors remains the same as today.

Unemployment and prostate cancer mortality in the OECD, 1990-2009

The global economic downturn has been associated with increased unemployment in many countries. Insights into the impact of unemployment on specific health conditions remain limited. We determined the association between unemployment and prostate cancer mortality in members of the Organisation for Economic Co-operation and Development (OECD). We used multivariate regression analysis to assess the association between changes in unemployment and prostate cancer mortality in OECD member states between 1990 and 2009. Country-specific differences in healthcare infrastructure, population structure, and population size were controlled for and lag analyses conducted. Several robustness checks were also performed. Time trend analyses were used to predict the number of excess deaths from prostate cancer following the 2008 global recession. Between 1990 and 2009, a 1% rise in unemployment was associated with an increase in prostate cancer mortality. Lag analysis showed a continued increase in mortality years after unemployment rises. The association between unemployment and prostate cancer mortality remained significant in robustness checks with 46 controls. Eight of the 21 OECD countries for which a time trend analysis was conducted, exhibited an estimated excess of prostate cancer deaths in at least one of 2008, 2009, or 2010, based on 2000-2007 trends. Rises in unemployment are associated with significant increases in prostate cancer mortality. Initiatives that bolster employment may help to minimise prostate cancer mortality during times of economic hardship.

Selecting the minimum prediction base of historical data to perform 5-year predictions of the cancer burden: The GoF-optimal method

BACKGROUND

Predicting the future burden of cancer is a key issue for health services planning, where a method for selecting the predictive model and the prediction base is a challenge. A method, named here Goodness-of-Fit optimal (GoF-optimal), is presented to determine the minimum prediction base of historical data to perform 5-year predictions of the number of new cancer cases or deaths.

METHODS

An empirical ex-post evaluation exercise for cancer mortality data in Spain and cancer incidence in Finland using simple linear and log-linear Poisson models was performed. Prediction bases were considered within the time periods 1951-2006 in Spain and 1975-2007 in Finland, and then predictions were made for 37 and 33 single years in these periods, respectively. The performance of three fixed different prediction bases (last 5, 10, and 20 years of historical data) was compared to that of the prediction base determined by the GoF-optimal method. The coverage (COV) of the 95% prediction interval and the discrepancy ratio (DR) were calculated to assess the success of the prediction.

RESULTS

The results showed that (i) models using the prediction base selected through GoF-optimal method reached the highest COV and the lowest DR and (ii) the best alternative strategy to GoF-optimal was the one using the base of prediction of 5-years.

CONCLUSIONS

The GoF-optimal approach can be used as a selection criterion in order to find an adequate base of prediction.

Estimating cancer incidence using a Bayesian back-calculation approach

We propose a Bayesian hierarchical model for the calculation of incidence counts from mortality data by a convolution equation that expresses mortality through its relationship with incidence and the survival probability density. The basic idea is to use mortality data together with an estimate of the survival distribution from cancer incidence to cancer mortality to reconstruct the numbers of individuals who constitute previously incident cases that give rise to the observed pattern of cancer mortality. This model is novel because it takes into account the uncertainty from the survival distribution; thus, a Bayesian-mixture cure model for survival is introduced. Furthermore, projections are obtained starting from a Bayesian age-period-cohort model. The main advantage of the proposed approach is its consideration of the three components of the model: the convolution equation, the survival mixture cure model and the age-period-cohort projection within a directed acyclic graph model. Furthermore, the estimation are obtained through the Gibbs sampler. We applied the model to cases of women with stomach cancer using six age classes [15-45], [45-55], [55-65], [65-75], [75-85] and [85-95] and validated it by using data from the Tuscany Cancer Registry. The model proposed and the program implemented are convenient because they allow different cancer disease to be analysed because the survival time is modelled by flexible distributions that are able to describe different trends.

Vismodegib: the proof of concept in Basal cell carcinoma

Although basal cell carcinoma (BCC) is the most common cancer worldwide, its metastatic dissemination is exceptional. Before 2012, we had a few treatment options available for metastatic or locally advanced cases. Management of these patients was complicated due to the lack of scientific data, the deterioration of a patient’s general status, the patient’s advanced age, and the presence of multiple comorbidities. The hedgehog signaling pathway is dysregulated in BCC. The exploration of this signaling pathway yielded to a major milestone in the treatment of advanced BCC. Vismodegib (GDC-0449), an oral small-molecule agent that targets the Hedgehog signaling pathway, demonstrates high levels of activity in clinical trials. It was approved in January 2012 for the treatment of locally advanced or metastatic BCC. Vismodegib confirms, once again, the interest in exploring the signal transduction pathways in cancers.

Trends in incidence and predictions of cutaneous melanoma across Europe up to 2015

BACKGROUND

Melanoma is a significant health problem in Caucasian populations. The most recently available data from cancer registries often have a delay of several months up to a few years and they are generally not easily accessible.

OBJECTIVES

To assess recent age- and sex-specific trends in melanoma incidence and make predictions for 2010 and 2015.

METHODS

A retrospective registry-based analysis was performed with data from 29 European cancer registries. Most of them had data available from 1990 up to 2006/7. World-standardized incidence rates (WSR) and the estimated annual percentage change (EAPC) were computed. Predictions were based on linear projection models.

RESULTS

Overall the incidence of melanoma is rapidly rising and will continue to do so. The incidence among women in Europe was generally higher than in men. The highest incidence rates were seen for Northern and north-western countries like the UK, Ireland and the Netherlands. The lowest incidence rates were observed in Portugal and Spain. The incidence overall remained stable in Norway, where, amongst young (25-49 years) Norwegian males rates significantly decreased (EAPC -2.8, 95% CI -3.6; -2.0). Despite a low melanoma incidence among persons above the age of 70, this age group experienced the greatest increase in risk during the study period.

CONCLUSIONS

Incidence rates of melanoma are expected to continue rising. These trends are worrying in terms of disease burden, particularly in eastern European countries.

Predictions of cancer incidence in Wielkopolska in 2018

Aim of the study

The cancer incidence in Wielkopolska in 2008 was one of the highest in the country and was higher than in Poland by 21% in men and by 14% in women. We can quantify the future burden of cancer from two different perspectives: the number of new cancer cases due to population change and new cases due to risk change. Making predictions of number of new cancer cases in Wielkopolska in 2018.

Material and methods

These projections of number of cancer cases, age specific rates and age-standardized rates for 2018 (all cancers and the most frequent cancers for men and women) has been based on the historical trends of cancer incidence in Wielkopolska in 1999-2008 and demographical prognosis of Central Statistical Office using the method of Hakulinen and Dyba.

Results

There will be over 8000 new cancer cases in men in Wielkopolska in 2018 and over 7000 in women. Compare to the period 2004-2008 the number of cancer cases will increase by 45% for men and by nearly 30% for women. About 2/3 of the increase in Wielkopolska is predicted to be connected with demography change, 1/3 with risk change.

Conclusions

The predicted increase of number of cancer cases in Wielkopolska in 2018 will be the result of: changes in the population (bigger impact of the older age groups), an influence of the risk factors (mainly smoking) and a participation in the screening programs.

Bayesian approach to predicting cancer incidence for an area without cancer registration by using cancer incidence data from nearby areas

This paper compares three different methods for performing cancer incidence prediction in an area without a cancer registry under a Bayesian framework, using linear and log-linear age-period models with either age-specific slopes or a common slope across age groups. The three methods assume that a nearby area with a cancer registration has similar incidence and mortality patterns as the area of interest without a cancer registry where the cancer incidence prediction is carried out. The three methods differ in modeling strategies: (i) modeling the incidence rate directly; (ii) modeling the ratio of the number of incident cases to that of mortality cases; and (iii) modeling the difference between the incidence rate and the mortality rate. Strategy (iii) is a new approach in this type of projection. Empirical assessment is made using real data from the cancer registry of Tarragona, Spain, to predict cancer incidence in Girona, Spain, and vice versa. Predictions of short-term (3-4 years) incidence were made for 2001 in Tarragona using observed cancer incidence and mortality data for 1994-1998 from Girona. Short-term predictions were made for 2002 in Girona using Tarragona's 1994-1998 data. Additionally, long-term (10 years) incidence rate predictions were made for 2002 in Girona using data from Tarragona for the period 1985-1992. Our results suggest that extrapolating time-trends of incidence rates minus mortality rates may have the best predictive performance overall. These methods of population-level disease-incidence prediction are highly relevant to health care planning and policy decisions.

A statistical model of the international spread of wild poliovirus in Africa used to predict and prevent outbreaks

BACKGROUND

Outbreaks of poliomyelitis in African countries that were previously free of wild-type poliovirus cost the Global Polio Eradication Initiative US$850 million during 2003-2009, and have limited the ability of the program to focus on endemic countries. A quantitative understanding of the factors that predict the distribution and timing of outbreaks will enable their prevention and facilitate the completion of global eradication.

METHODS AND FINDINGS

Children with poliomyelitis in Africa from 1 January 2003 to 31 December 2010 were identified through routine surveillance of cases of acute flaccid paralysis, and separate outbreaks associated with importation of wild-type poliovirus were defined using the genetic relatedness of these viruses in the VP1/2A region. Potential explanatory variables were examined for their association with the number, size, and duration of poliomyelitis outbreaks in 6-mo periods using multivariable regression analysis. The predictive ability of 6-mo-ahead forecasts of poliomyelitis outbreaks in each country based on the regression model was assessed. A total of 142 genetically distinct outbreaks of poliomyelitis were recorded in 25 African countries, resulting in 1-228 cases (median of two cases). The estimated number of people arriving from infected countries and <5-y childhood mortality were independently associated with the number of outbreaks. Immunisation coverage based on the reported vaccination history of children with non-polio acute flaccid paralysis was associated with the duration and size of each outbreak, as well as the number of outbreaks. Six-month-ahead forecasts of the number of outbreaks in a country or region changed over time and had a predictive ability of 82%.

CONCLUSIONS

Outbreaks of poliomyelitis resulted primarily from continued transmission in Nigeria and the poor immunisation status of populations in neighbouring countries. From 1 January 2010 to 30 June 2011, reduced transmission in Nigeria and increased incidence in reinfected countries in west and central Africa have changed the geographical risk of polio outbreaks, and will require careful immunisation planning to limit onward spread. Please see later in the article for the Editors' Summary.

[Statistical Analysis of Rates and Trends (SART): a web-based tool for statistical calculation of population indicators]

We propose a web-based tool (SART: http://regstattools.net/sart.html) that automates calculations to obtain various population indicators that can be used for the control of diseases or health events. SART has four modules: a) a descriptive module that allows calculation of the number of cases and their percentage, the crude rate, the adjusted rate, the truncated rate and the cumulative rate; b) the estimated annual percentage change of rates; c) calculation of expected cases; and d) the standardized incidence of mortality ratio. SART requests a base file and input parameters from the user before processing the data. The data and the results obtained are processed and then sent by email to the user. The results are provided by sex and for each of the study variables (diseases, ethnic groups, geographic areas...) introduced into the base file.

NORDCAN--a Nordic tool for cancer information, planning, quality control and research

The NORDCAN database and program ( www.ancr.nu ) include detailed information and results on cancer incidence, mortality and prevalence in each of the Nordic countries over five decades and has lately been supplemented with predictions of cancer incidence and mortality; future extensions include the incorporation of cancer survival estimates.

MATERIAL AND METHODS

The data originates from the national cancer registries and causes of death registries in Denmark, Finland, Iceland, Norway, Sweden, and Faroe Islands and is regularly updated. Presently 41 cancer entities are included in the common dataset, and conversions of the original national data according to international rules ensure comparability.

RESULTS

With 25 million inhabitants in the Nordic countries, 130 000 incident cancers are reported yearly, alongside nearly 60 000 cancer deaths, with almost a million persons living with a cancer diagnosis. This web-based application is available in English and in each of the five Nordic national languages. It includes comprehensive and easy-to-use descriptive epidemiology tools that provide tabulations and graphs, with further user-specified options available.

DISCUSSION

The NORDCAN database aims to provide comparable and timely data to serve the varying needs of policy makers, cancer societies, the public, and journalists, as well as the clinical and research community.

Analysis of nonlinear patterns of change with random coefficient models

Nonlinear patterns of change arise frequently in the analysis of repeated measures from longitudinal studies in psychology. The main feature of nonlinear development is that change is more rapid in some periods than in others. There generally also are strong individual differences, so although there is a general similarity of patterns for different persons over time, individuals exhibit substantial heterogeneity in their particular response. To describe data of this kind, researchers have extended the random coefficient model to accommodate nonlinear trajectories of change. It can often produce a statistically satisfying account of subject-specific development. In this review we describe and illustrate the main ideas of the nonlinear random coefficient model with concrete examples.

Forecasting age-specific breast cancer mortality using functional data models

Accurate estimates of future age-specific incidence and mortality are critical for allocation of resources to breast cancer control programmes and evaluation of screening programmes. The purpose of this study is to apply functional data analysis techniques to model age-specific breast cancer mortality time trends, and forecast entire age-specific mortality functions using a state-space approach. We use annual unadjusted breast cancer mortality rates in Australia, from 1921 to 2001 in 5 year age groups (45 to 85+). We use functional data analysis techniques where mortality and incidence are modelled as curves with age as a functional covariate varying by time. Data are smoothed using non-parametric smoothing methods then decomposed (using principal components analysis) to estimate basis functions that represent the functional curve. Period effects from the fitted coefficients are forecast then multiplied by the basis functions, resulting in a forecast mortality curve with prediction intervals. To forecast, we adopt a state-space approach and an automatic modelling framework for selecting among exponential smoothing methods.Overall, breast cancer mortality rates in Australia remained relatively stable from 1960 to the late 1990s, but have declined over the last few years. A set of four basis functions minimized the mean integrated squared forecasting error and account for 99.3 per cent of variation around the mean mortality curve. Twenty year forecasts suggest a continuing decline, but at a slower rate, and stabilizing beyond 2010. Forecasts show a decline in all age groups with the greatest decline in older women. The proposed methods have the potential to incorporate important covariates such as hormone replacement therapy and interventions to represent mammographic screening. This would be particularly useful for evaluating the impact of screening on mortality and incidence from breast cancer.

Melanoma incidence in central Italy will go on increasing also in the near future: A registry-based, age–period–cohort analysis

The aim of the study was to evaluate malignant melanoma incident trends in central Italy by means of an age-period-cohort approach. A total of 1977 malignant melanoma (15-84 years) incidents in the area of the Tuscany Cancer Registry between 1987 and 2001 were analysed. Poisson regression has been used to estimate age, cohort and period effect. A nonlinear regression model was used to estimate the expected number of new cases in the period 2002-2006. Incidence rates increased in all age, period and cohort groups. The model that best fitted the data included age and 'drift'. The linear effect ('drift') showed, in each age group, an increase of the risk of malignant melanoma diagnosis of about 36.6% every 5 years of period or cohort. For the period 2002-2006, 1112 new cases were predicted with a standardized rate (age 15-84 years) of 19.2x100.000. In the Tuscany Cancer Registry area, no clues for malignant melanoma incidence rates levelling off were documented. Growing rates and number of malignant melanoma are expected in the near future.

Age, period and cohort effects in Bayesian smoothing of spatial cancer survival with geoadditive models

Based on the example of data on breast cancer survival in a specific area in France, this paper describes a Bayesian approach to analysing individual continuous-time spatially referenced survival data. Starting from the well-known Cox model we develop a geoadditive survival model with a baseline effect, spatial effects, age, period and cohort effects as well as the effect of the number of metastases at the time of diagnosis. Furthermore, we also investigate temporal and spatial variations in the effect of the number of metastases. Our approach is particularly useful since we find clear hints for a violation of the proportional hazards assumption and the existence of different spatial patterns for patients with no, one and more than one metastasis. The reliability of our approach is attested by comparison with a parametric model and with several simulated data sets with known risk profiles.

Predicting the future burden of cancer

As observations in the past do not necessarily hold into the future, predicting future cancer occurrence is fraught with uncertainty. Nevertheless, predictions can aid health planners in allocating resources and allow scientists to explore the consequence of interventions aimed at reducing the impact of cancer. Simple statistical models have been refined over the past few decades and often provide reasonable predictions when applied to recent trends. Intrinsic to their interpretation, however, is an understanding of the forces that drive time trends. We explain how and why cancer predictions are made, with examples to illustrate the concepts in practice.

Predictions of skin cancer incidence in the Netherlands up to 2015

BACKGROUND

Skin cancer is an important, growing public health problem among white caucasians, causing a heavy burden on dermatologists and general practitioners.

OBJECTIVES

To predict the future incidence of skin cancer in the Netherlands up to 2015.

METHODS

Expected numbers of skin cancer cases in the Netherlands up to 2015 were calculated by trend modelling of observed rates for melanoma and squamous cell carcinoma (SCC) between 1989 and 2000 obtained from the Netherlands Cancer Registry and for basal cell carcinoma (BCC) obtained from the Eindhoven Cancer Registry; these rates were then multiplied by the predicted age distributions. Incidence rates were fitted to four different models, and predictions were based on the best fitting model.

RESULTS

An increase of 80% in the total number of skin cancer patients is expected in the Netherlands: from 20 654 in 2000 to 37 342 in 2015. The total number of melanoma cases is expected to increase by 99%, with the largest increase for males (males aged 35-64, 111%; males aged > or = 65, 139%). Numbers of patients with SCC will increase overall by 80%, mainly among older males and females (increase of 79%) and females aged 35-64 (increase of 93%). The number of cases of BCC will increase by 78%, with the largest increase for the combined groups, those aged 15-64 (males, 66% increase; females, 94% increase), especially for sites other than the head and neck. The contribution of demographic changes (ageing effect) was largest for males with BCC and SCC (35-44%).

CONCLUSIONS

If incidence rates for skin cancers in the Netherlands continue to increase and population growth and ageing remain unabated, a rise in annual demand for care of more than 5% could occur, putting a heavy burden on general practitioners and dermatologists. In the absence of marked changes in current ultraviolet radiation exposure, these increases will probably continue after 2015.

Estimation and projection of the national profile of cancer mortality in China: 1991-2005

There are no national-level data on cancer mortality in China since two surveys in 1973–1975 and 1990–1992 (a 10% sample), but ongoing surveillance systems, based on nonrandom selected populations, give an indication as to the trends for major cancers. Based on a log-linear regression model with Poisson errors, the annual rates of change for 10 cancers and all other cancers combined, by age, sex and urban/rural residence were estimated from the data of the surveillance system of the Center for Health Information and Statistics, covering about 10% of the national population. These rates of change were applied to the survey data of 1990–1992 to estimate national mortality in the year 2000, and to make projections for 2005. Mortality rates for all cancers combined, adjusted for age, are predicted to change little between 1991 and 2005 (−0.8% in men and +2.5% in women), but population growth and ageing will result in an increasing number of deaths, from 1.2 to 1.8 million. The largest predicted increases are for the numbers of female breast (+155.4%) and lung cancers (+112.1% in men, +153.5% in women). For these two sites, mortality rates will almost double. Cancer will make an increasing contribution to the burden of diseases in China in the 21st century. The marked increases in risk of cancers of the lung, female breast and large bowel indicate priorities for prevention and control. The increasing trends in young age groups for cancers of the cervix, lung and female breast suggest that their predicted increases may be underestimated, and that more attention should be paid to strategies for their prevention and control.

Estimation and projection of the national profile of cancer mortality in China: 1991-2005

There are no national-level data on cancer mortality in China since two surveys in 1973-1975 and 1990-1992 (a 10% sample), but ongoing surveillance systems, based on nonrandom selected populations, give an indication as to the trends for major cancers. Based on a log-linear regression model with Poisson errors, the annual rates of change for 10 cancers and all other cancers combined, by age, sex and urban/rural residence were estimated from the data of the surveillance system of the Center for Health Information and Statistics, covering about 10% of the national population. These rates of change were applied to the survey data of 1990-1992 to estimate national mortality in the year 2000, and to make projections for 2005. Mortality rates for all cancers combined, adjusted for age, are predicted to change little between 1991 and 2005 (-0.8% in men and +2.5% in women), but population growth and ageing will result in an increasing number of deaths, from 1.2 to 1.8 million. The largest predicted increases are for the numbers of female breast (+155.4%) and lung cancers (+112.1% in men, +153.5% in women). For these two sites, mortality rates will almost double. Cancer will make an increasing contribution to the burden of diseases in China in the 21st century. The marked increases in risk of cancers of the lung, female breast and large bowel indicate priorities for prevention and control. The increasing trends in young age groups for cancers of the cervix, lung and female breast suggest that their predicted increases may be underestimated, and that more attention should be paid to strategies for their prevention and control.

Prediction of cancer incidence in the Nordic countries: empirical comparison of different approaches

Prediction of the future number of cancer cases is of great interest to society. The classical approach is to use the age-period-cohort model for making cancer incidence predictions. We made an empirical comparison of different versions of this model, using data from cancer registries in the Nordic countries for the period 1958-1997. We have applied 15 different methods to 20 sites for each sex in Denmark, Finland, Norway and Sweden. Median absolute value of the relative difference between observed and predicted numbers of cases for these 160 combinations of site, sex and country was calculated. The medians varied between 10.4 per cent and 15.3 per cent in predictions 10 years ahead, and between 15.1 per cent and 32.0 per cent for 20 year predictions. We have four main conclusions: (i) projecting current trends worked better than assuming that future rates are equal to present rates; (ii) the method based on the multiplicative APC model often overestimated the number of cancer cases due to its exponential growth over time, but using a power function to level off this growth improved the predictions; (iii) projecting only half of the trend after the first 10 years also gave better long-term predictions; (iv) methods that emphasize trends in the last decade seem to perform better than those that include earlier time trends.

Comparison of different approaches to incidence prediction based on simple interpolation techniques

The paper compares three different methods for performing disease incidence prediction based on simple interpolation techniques. The first method assumes that the age-period specific numbers of observed cases follow a Poisson distribution and the other two methods assume a normal distribution for the incidence rates. The main emphasis of the paper is on assessing the reliability of the three methods. For this purpose, ex post predictions produced by each method are checked for different cancer sites using data from the Cancer Control Region of Turku in Finland. In addition, the behaviour of the estimators of predicted expected values and prediction intervals, crucial for investigation of the reliability of prediction, are assessed using a simulation study. The prediction method making use of the Poisson assumption appeared to be the most reliable of the three approaches. The simulation study found that the estimator of the length of the prediction interval produced by this method has the smallest coverage error and is the most precise.