Endoscopic colorectal cancer screening: a cost-saving analysis.

BACKGROUND: Comprehensive analyses have shown that screening for cancer usually induces net costs. In this study, the possible costs and savings of endoscopic colorectal cancer screening are explored to investigate whether the induced savings may compensate for the costs of screening. METHODS: A simulation model for evaluation of colorectal cancer screening, MISCAN-COLON, is used to predict costs and savings for the U.S. population, assuming that screening is performed during a period of 30 years. Plausible baseline parameter values of epidemiology, natural history, screening test characteristics, and unit costs are based on available data and expert opinion. Important parameters are varied to extreme but plausible values. RESULTS: Given the expert opinion-based assumptions, a program based on every 5-year sigmoidoscopy screenings could result in a net savings of direct health care costs due to prevention of cancer treatment costs that compensate for the costs of screening, diagnostic follow-up, and surveillance. This result persists when costs and health effects are discounted at 3%. The "break-even" point, the time required before savings exceed costs, is 35 years for a screening program that terminates after 30 years and 44 years for a screening program that continues on indefinitely. However, net savings increase or turn into net costs when alternative assumptions about natural history of colorectal cancer, costs of screening, surveillance, and diagnostics are considered. CONCLUSIONS: Given the present, limited knowledge of the disease process of colorectal cancer, test characteristics, and costs, it may well be that the induced savings by endoscopic colorectal cancer screening completely compensate for the costs.

Age-specific reduction in breast cancer mortality by screening: an analysis of the results of the Health Insurance Plan of Greater New York study.

The 14-year follow-up mortality results from the randomized breast cancer screening trial of the Health Insurance Plan of Greater New York (HIP) were analyzed with respect to the problem of age-specific screening effects. Mortality reduction was measured in three different ways and appears to be homogeneous across age groups. This finding challenges the widespread opinion that the results of the HIP study support the conclusion that breast cancer screening is not effective below age 50.

Care and costs for advanced cervical cancer.

The types, amounts and costs of hospital and home care in patients who died from cervical cancer are investigated, using both national data sources and hospital files. Our goal has been assessment of the savings on treatment and care of advanced cervical cancer resulting from cervical cancer screening. Hospital costs account for 70% of the total cost per patient of Dfl 29,200. The amount of hospital care decreases significantly with increasing age. The average number of days of hospitalisation per patient with advanced disease decreases from 62 days below age 50 to less than 10 days at age 70 and older. In-hospital medical procedures, home care and nursing home care account for 24, 22 and 8% of the costs, respectively. Mass screening programmes for cervical cancer will result in a reduction in both advanced disease and mortality. The potential savings compensate approximately 10% of the costs of screening.

In search of the best upper age limit for breast cancer screening.

The aim of this study was to determine the best upper age limit for a breast cancer screening programme. We used a model-based study using optimistic and pessimistic assumptions, concerning improvement of prognosis due to screen-detection and duration of the period of mammographic detectability, resulting in upper and lower limits for favourable and unfavourable effects. Under pessimistic assumptions, the balance between positive and negative effects of screening remains favourable up to an age of around 80 years. Under optimistic assumptions, this balance never becomes clearly negative with increase of the upper age limit of a screening programme. When including the costs in the analysis, the balance between effects and costs of increasing the upper age limit from 69 to 75 years is likely to be at least as favourable as intensifying a screening programme within the age group 50-69 years. A further increase leads to a markedly less favourable balance. Competing causes of death do not lead to missing net benefit for women up to at least age 80 years, but the disproportional rise of negative effects of screening with age in older women leads to a lower cost-effectiveness ratio than intensifying screening at ages 50-69 years.

Impact of systematic false-negative test results on the performance of faecal occult blood screening.

The impact of systematic false-negative test results on mortality reduction and on programme sensitivity of annual faecal occult blood testing in ages 50-84 years is explored using a microsimulation model. We made calculations for test sensitivities of 80, 50 and 30%. In order to reproduce a cancer detection rate of 2.2 per 1000 at the first screening, the corresponding mean preclinical sojourn times had to be 1.42, 2.30 and 3.84 years, respectively. The fraction systematic results among the false-negative results is varied between 0 and 100%. With 80% test sensitivity, the reduction in mortality due to screening decreases from 25% without systematic results to 23% when all false-negative results are systematic and the programme sensitivity decreases from 63 to 58%. With 30% test sensitivity, mortality reduction decreases from 21 to 11% and programme sensitivity decreases from 52 to 27%. The impact of systematic false-negative test results is important if annual FOBT screening is considered.

Women who participate in spontaneous screening are not at higher risk for cervical cancer than women who attend programme screening.

Up to 1995, programme screening for cervical cancer in The Netherlands was targeted at women between 35 and 54 years of age at 3-yearly intervals. Spontaneous screening in addition to programme screening was common practice. Our aim was to compare the underlying risk for cervical neoplasia for women involved in both types of screening. From the national pathological database, we retrieved all primary smears (n=693318) taken in 1994 in The Netherlands. Among the smears registered for screening purposes (39%), 79% was taken within the mass screening programme and 21% was taken for spontaneous screening. The underlying risk was studied from the detection rates of histologically confirmed severe dysplasia or worse, using a multivariate loglinear model, including age and screening history. The detection rate of at least severe dysplasia, adjusted for age and screening history, was equal for women who had a spontaneous smear and for those who had a programme smear (odds ratio (OR): 0.97; 95% Confidence Interval (CI): 0.84-1.14). In our data, women participating in spontaneous screening were not at a higher risk for cervical cancer than women who used programme screening. Therefore, all asymptomatic women in the Netherlands should follow the general guidelines for age-range and screening-interval.

A model-based prediction of the impact on reduction in mortality by a breast cancer screening programme in the city of Florence, Italy.

The efficacy of breast cancer screening for women older than 50 years has been shown in several studies. Service screening is now ongoing or planned in several countries in Europe. MISCAN, a computer simulation programme, has been used to analyse data from the Florence District Programme (FDP) breast cancer experience. First, the model was fitted to the screening results for the period 1975-1986. A good correspondence between the model outcomes and the FDP results was achieved. It was then used to predict the impact on mortality of the new starting programme of the city of Florence (63,000 women, 50-69 years old). Assuming a 70% attendance rate, then for the city of Florence, 2563 screen-detected breast cancers are predicted for the period 1991-2020 out of the total number of 9095 breast cancers for all ages (28%). A total of 3720 deaths for breast cancer are expected without screening. An absolute reduction of 472 deaths (13%) is predicted for the whole population. The estimated number of years of life gained by screening until 2020 is 4354. Simulation by MISCAN has previously been a useful support tool for decision-making about screening. The present paper is the first based on a southern European experience. The possibility of applying MISCAN to predict the impact of a national programme in Italy is discussed.

How cost-effective is breast cancer screening in different EC countries?

Should the decision to start breast cancer screening in the Netherlands and in the U.K. be followed by other EC countries? This question has been addressed in an exploratory analysis of the differences in cost-effectiveness of breast cancer screening in Spain, France, the U.K. and the Netherlands. A detailed cost-effectiveness analysis of breast cancer screening in the Netherlands has been used as the starting point. Country specific data on incidence, mortality, demography, screening organisation and price levels in health care have been used to predict the costs and effects of nationwide screening programmes, in which women aged 50-70 are invited for 2-yearly mammographic screening. The relative effect of screening is highest in the U.K. (16.55 life-years gained per 1000 screens) and lowest in Spain (8.23 life-years gained per 1000 screens). The cost per screen is highest in Spain (38 pounds) and lowest in the U.K. (18 pounds). In comparison with the yearly health expenditures per capita, the cost per life-year gained is 2.8 times higher in the Netherlands, 3.1 times higher in the U.K., 6.5 times higher in France and 20.6 times higher in Spain. These marked differences show that no uniform policy recommendations for breast cancer screening can be made for all countries of the EC.

Diagnostic and treatment procedures induced by cervical cancer screening.

The amount of diagnostic and treatment procedures induced by cervical cancer screening has been assessed prospectively and related to mortality reduction. Assumptions are based on data from Dutch screening programmes and on a scenario for future developments. With 5 invitations for screening, between ages 37-70 every eight years, 13 deaths are avoided per million women per screening year. Each death avoided is balanced by 2800 preventive smears, 9 women referred to a gynaecology department and 4 minor treatment procedures (conserving treatment or exconisation). 25 invitations in a life-time avoids 27 deaths per million women per screening year but with 7300 preventive smears, 22 referrals and 8 small treatment procedures. Thus intensifying screening will not only result in diminishing returns of extra screening efforts, but also in increasing risk for women to undergo unnecessary (no invasive disease or death avoided) diagnostic and treatment procedures. The balance between beneficial and adverse effects deteriorates strongly when hysterectomies play an important part in the management of cervical intraepithelial neoplasia.

A measure of survival time for long-term follow-up studies of the elderly.

Studies of determinants of longevity in the elderly by means of a follow-up study are usually based on samples with a wide age range and some selectivity. This poses the problem of constructing a measure of longevity that: (1) will make subjects in the sample comparable regarding survival time, independent of their age and sex; (2) is suitable for long-term follow-up studies; (3) can be used to check the absence or presence of selectivity of the sample with regard to longevity; (4) improves the comparability of studies with different designs and sampling schemes; (5) can take the mortality development over time into account. An individual measure of survival time is presented that satisfies these requirements: the realized probability of dying (RPD). Its construction is described. The RPD is derived from population life tables based on age, year of birth, and sex. For each subject, the relative position on the survival curve within the birth cohort is determined. In an illustration, the RPD is applied to data from a 28-year follow-up study of the elderly in the Netherlands. A comparison is made with other survival measures commonly used in this type of study. It is concluded that the RPD is a powerful and valid measure of longevity in elderly subjects, and that it can be useful in the study of determinants of longevity.

Duration of preclinical cervical cancer and reduction in incidence of invasive cancer following negative pap smears.

BACKGROUND: An investigation has been made into the differences between estimates for the duration of preclinical cervical cancer resulting from two types of studies. A median duration of 5-10 years was suggested by the observed build-up of incidence of invasive cervical cancer after one or more negative smears. Much longer median durations of more than 15 years have been reported from fitting statistical models to screening data. METHODS: We developed one of these statistical models and fitted it to clinical incidence and screening data from British Columbia, which resulted in estimated mean durations of 12 years for pre-invasive stages, and 4-5 years for screen-detectable stages. The model is used to predict the build-up of the incidence of invasive cancer after one and after two negative smears. RESULTS: The model predictions appear to correspond closely to the observed incidence trends following negative smears. The apparent contradiction between model estimates and observed data is explained by recognizing that many of the women who have had negative smears will have further Pap smears, resulting in earlier diagnosis of invasive cervical cancers and thus an apparent faster build-up of the incidence. CONCLUSIONS: When the impact of further Pap smears is neglected, the data suggest that the risk of invasive cancer following one or more negative smears returns to close to prescreening levels within 6-10 years. This is an overestimation of the risk of clinical invasive cancer. In the case of cessation of screening it will take longer before the incidence of clinical cancer will increase. Where there is continuous screening the screen-detected cancers have a relatively favourable prognosis, thus contributing less to the serious morbidity and mortality risks associated with invasive cancer. This should be taken into account in making comparisons with the prescreening situation.

Screening for congenital heart malformation in child health centres.

BACKGROUND: Although screening for congenital heart malformations is part of the child health care programme in several countries, there are very few published evaluations of these activities. This report is concerned with the evaluation of this screening at the Dutch Child Health Centres (CHC). METHODS: All consecutive patients, aged between 32 days and 4 years, presented at the Sophia Children's Hospital Rotterdam throughout a period of 2 years, with a congenital heart malformation were included in this study. Paediatric cardiologists established whether or not these patients were diagnosed after haemodynamic complications had already developed (diagnosed 'too late'). Parents and CHC-physicians were interviewed in order to establish the screening and detection history. Test properties were established for all patients with a congenital heart malformation (n = 290), intended effects of screening were established in patients with clinically significant malformations (n = 82). RESULTS: The sensitivity of the actual screening programme was 0.57 (95% CI : 0.51-0.62), the specificity 0.985 (95% CI : 0.981-0.990) and the predictive value of a positive test result 0.13 (95% CI: 0.10-0.19). Sensitivity in a subpopulation of patients adequately screened was 0.89 (95% CI: 0.74-0.96). Adequately screened patients were less likely to be diagnosed 'too late' than inadequately screened patients (odds ratio [OR] = 0.20, 95% CI: 0.04-1.05). The actual risk of being diagnosed 'too late' in the study-population (48%) was only slightly less than the estimated risk for patients not exposed to CHC-screening (58%, 95% CI: 43%-72%). Adequately screened patients however were at considerably less risk (17%, 95% CI: 4%-48%). CONCLUSION: Screening for congenital heart malformations in CHC contributes to the timely detection of these disorders. The actual yield, however, is far from optimal, and the screening programme should be improved.

Endocervical status is not predictive of the incidence of cervical cancer in the years after negative smears.

The clinical relevance of the lack of endocervical cells was never well established in a longitudinal study with histologically proven cervical cancer as an end point. From the Dutch Network and National Database for Pathology, results for all negative smears obtained in 1990 and 1991 in the Netherlands were retrieved, as were data for all cytologic and histologic examinations performed after the negative smears before April 1998. There were no significant differences between the proportion of preinvasive lesions (cervical intraepithelial neoplasia 1, 2, and 3) detected after negative smears without endocervical cells compared with negative smears with endocervical cells. The proportion of women in whom invasive cancer developed was the same in both groups. These data suggest there is no reason to advise women with negative smears without endocervical cells to undergo an additional smear.

The microsimulation approach to epidemiologic modeling of helminthic infections, with special reference to schistosomiasis.

The microsimulation technique has been used since 1985 as a tool for epidemiologic modeling of helminthic infections. This technique is characterized by mimicking individual life histories, which makes it possible to include several relevant processes and mechanisms that have not so far been considered in applied modeling. Biological, epidemiologic, and social processes can be simulated in detail, which allows realistic prediction of the impact of control strategies. It is clear that careful quantification and validation of the many processes and parameters in the model requires close collaboration with experts working on control projects. In the development and application of a microsimulation model, we distinguish eight steps, ranging from the identification of questions the model will be designed to address, to the completion of a model that can be used as a routine decision-making tool in a control program.

SCHISTOSIM: a microsimulation model for the epidemiology and control of schistosomiasis.

A computer simulation model, SCHISTOSIM, has been developed for the epidemiology and control of schistosomiasis, based on the stochastic microsimulation technique. The eventual aim is to evaluate and predict the effects of different control strategies. In the current state of the model, human-, worm-, and infection-related aspects have been included. However, many others, including most transmission and transmission-related mechanisms, have yet to be modeled. By simulating a series of surveys and treatments in Burundi, short-term effects of this program were satisfactorily explained by the model. However, long-term predictions did not match the observed data. Possible extensions of the model to properly describe these effects are identified. The potential of SCHISTOSIM as a tool for the prediction of outcomes of alternative control strategies is illustrated and discussed.

The reproductive lifespan of Onchocerca volvulus in West African savanna.

The epidemiological model ONCHOSIM--a model and computer simulation program for the transmission and control of onchocerciasis--has been used to determine the range of plausible values for the reproductive lifespan of Onchocerca volvulus. Model predictions based on different lifespan quantifications were compared with the results of longitudinal skin-snip surveys undertaken in 4 reference villages during 13 to 14 years of successful vector control in the Onchocerciasis Control Programme in West Africa. Good fits between predicted and observed trends in skin microfilarial loads could be obtained for all villages. It is concluded that the reproductive lifespan of the savanna strain of O. volvulus lies between 9 and 11 years, and that 95% of the parasites reach the end of reproduction before the age of 13 to 14 years.

Economic aspects of cervical cancer screening.

The results of a cost-effectiveness analysis of cervical cancer screening in The Netherlands are reported, emphasizing the analysis of the costs of screening and consequent diagnosis and treatment. Many organized screening policies are evaluated, differing in age-range and interval between screens. The cost estimates are based on organization charts, file studies and tariffs. The costs of screening itself are by far the most important cost component. Screening increases the costs of diagnosis. Costs for primary treatment only rise for large screening policies. Screening causes savings in costs of terminal treatment, but these are small compared with the costs of screening. The costs per life-year gained for the most efficient policies amount to DFL 24,000 for the policy with 7 invitations per woman in a lifetime and rise considerably in case of more than 10 invitations. Cervical cancer screening appears to be less cost-effective than breast cancer screening, but compared with other services the results are comparatively good. Implementing one of the efficient organized screening policies and discouraging spontaneous screening beyond that schedule leads to considerable savings. Moreover, many organized policies which are not efficient are still superior to spontaneous screening.

Inaccuracies in estimates of life expectancies of patients with bronchial cancer in clinical decision making.

Approximations of life expectancy in clinical decision making frequently assume constant disease-specific ("excess") mortality hazards over age at diagnosis and over time from diagnosis. This assumption is inconsistent with the longer relative survival of younger patients with bladder cancer and with the declines in mortality hazards from bladder and breast cancers over time from diagnosis. To estimate the error that may result from these assumptions, the authors derived excess mortality hazards from the Surveillance, Epidemiology and End Result (SEER) tumor registry for bronchial cancers stratified by age at diagnosis and time from diagnosis. They compared the life expectancies calculated by a model using an average constant annual cancer-specific mortality hazard over time from diagnosis with those calculated using data-derived cancer-specific annual mortality hazards that varied as a function of time from diagnosis. For younger patients with less advanced disease, the constant-average-mortality model underestimated life expectancies by up to 50% relative to those predicted by the time-variant model. For those over 75 years old at diagnosis, and for all patients with advanced disease, the constant-average-mortality model overestimated life expectancies by up to 65% relative to those predicted by the time-variant model. The authors conclude that predictions of life expectancy with bronchial cancer, and probably with other neoplasms, are limited by the widespread use of oversimplified methods of calculation and by the lack of data describing mortality hazards as a function of time from diagnosis.(ABSTRACT TRUNCATED AT 250 WORDS)

Model building on the basis of Dutch cervical cancer screening data.

A mass screening programme for cervical cancer is in progress in three pilot regions in The Netherlands. All women living in these regions aged 35-53 are invited to undergo screening at three-year intervals. The MISCAN simulation model was developed for the analysis and optimization of screening programmes. In this paper the model-based approach to evaluation is first outlined and then illustrated by analysing data from the first two screening rounds in the pilot regions. This analysis resulted in a rather restricted range of data-compatible assumptions for the mean duration of preclinical disease (14-19 yr) and the frequency of spontaneous regression of preinvasive lesions (45-65%), as well as a rather wide sensitivity range for the Pap smear (50-90%). These preliminary findings are compared with those of a previous MISCAN analysis of cervical cancer screening in British Columbia. On the basis of an assumed 18-yr duration, 50% regression and 70% sensitivity, a number of screening policies relating to the same age ranges but with different intervals are compared. Both the analysis and the policy comparisons are preliminary, but the findings are nevertheless reasonable and consistent with those of previous studies. A more complete MISCAN-based analysis of the Dutch screening programme and subsequent optimization of screening policies will be possible when further results become available and a cost-effectiveness analysis procedure has been incorporated into the MISCAN programme.

Modelling issues in cancer screening.

The two main goals of modelling cancer screening are data analysis and evaluation. In data analysis, analytical-numerical statistical models are used to test hypotheses about preclinical disease, the screening test, and the association between early detection and risk of dying from the cancer. Evaluation in cancer screening is supported by model-based prediction of screening effects and cost-effectiveness. Simulation models are suitable for these tasks, and can also be used to identify efficient age-ranges and intervals between screening tests. Striking differences exist between screening models for cervical cancer and breast cancer, which are the two cancer types for which screening is common practice. The two main problems in cervical cancer screening are the proportion of progressive and regressive among screen-detected lesions, and the impact of screening on incidence and mortality. In breast cancer, regression is not (yet) a big issue, and the relationship between screening and mortality reduction has been demonstrated in randomized controlled trials (at least for women older than 50 years). The weakest link in current breast cancer models is the association between earliness of detection and improvement in prognosis. The modelling outcomes and their usefulness are decisively influenced by the data sets that were used in quantifying the model, and the subclassifications of the data that were considered. New or pending modelling issues include HPV-based screening in cervical cancer, screening models for colorectal cancer, the use of surrogate outcome measures and model-based meta-analysis of screening trials.