Medical device tracking: results of a case study of the implantable cardioverter defibrillator.

Case reports were received of a fatal tachycardia caused by a malfunction of an implantable cardioverter defibrillator (ICD), a device that is subject to the tracking regulations of the Food and Drug Administration's Center for Devices and Radiological Health. The case reports led to a decision to notify 5,604 patients of the need for reprogramming of their ICDs to prevent the tachycordia. In the first 60 days, a total of 98.7% of the patients were successfully located and their devices reprogrammed. Multiple logistic regression analysis was conducted to examine an extensive array of factors that might have been related to the time to reprogramming. Patient-specific factors such as age, sex, and ejection fraction did not serve as a barrier to reprogramming in the first week (p = NS). Patients whose regular physician had >5 patients with the ICD subject to the recall were significantly more likely to have their ICDs reprogrammed in the first week (odds ratio [OR] 2.11, 95% confidence interval [CI] 1.85 to 2.43, p<0.001). Patients who changed physicians were significantly less likely to undergo reprogramming in the first week (OR 0.73, 95% CI 0.63 to 0.86, p<0.001). The experience of the recall of this tracked device is highly encouraging because it demonstrates that most tracked device recipients can be successfully located and receive medical intervention. Although tracking devices is a manufacturer's responsibility, the clinical community plays a critical role in its success. This report highlights the importance of understanding that role among physicians.

Study design in the evaluation of breast cancer imaging technologies.

RATIONALE AND OBJECTIVES: Bringing a new imaging technology to market is a complex process. Beyond conceptualization and proof of concept, obtaining U.S. Food and Drug Administration (FDA) approval for clinical use depends on the documented experimental establishment of safety and efficacy. In turn, safety and efficacy are evaluated in the context of the intended use of the technology. The purpose of this study was to examine a conceptual framework for technology development and evaluation, focusing on new breast imaging technologies as a highly visible and current case in point. MATERIALS AND METHODS: The FDA views technology development in terms of a preclinical and four clinical phases of assessment. With a concept of research and development as a learning model, this phased-assessment concept of regulatory review against intended use was integrated with a five-level version of a hierarchy-of-efficacy framework for evaluating imaging technologies. Study design and analysis issues are presented in this context, as are approaches to supporting expanded clinical indications and new intended uses after a new technology is marketed. CONCLUSION: Breast imaging technologies may be intended for use as replacements for standard-of-care technologies, as adjuncts, or as complementary technologies. Study designs must be appropriate to establish claims of superiority or equivalence to the standard for the intended use. Screening technologies are ultimately judged on their demonstrated effectiveness in decreasing cause-specific mortality through early detection, but they may be brought to market for other uses on the basis of lesser standards of efficacy (eg, sensitivity, specificity, positive and negative predictive value, and stage of disease detected).

Comparison of risk estimates for selected diseases and causes of death.

BACKGROUND: Lifetime risk estimates of disease are limited by long-term data extrapolations and are less relevant to individuals who have already lived a period of time without the disease, but are approaching the age at which the disease risk becomes common. In contrast, short-term age-conditional risk estimates, such as the risk of developing a disease in the next 10 years among those alive and free of the disease at a given age, are less restricted by long-term extrapolation of current rates and can present patients with risk information tailored to their age. This study focuses on short-term age-conditional risk estimates for a broad set of important chronic diseases and nondisease causes of death among white and black men and women. METHODS: The Feuer et al. (1993, Journal of the National Cancer Institute) [15] method was applied to data from a variety of sources to obtain risk estimates for select cancers, myocardial infarction, diabetes mellitus, multiple sclerosis, Alzheimer's, and death from motor vehicle accidents, homicide or legal intervention, and suicide. RESULTS: Acute deaths from suicide, homicide or legal intervention, and fatal motor vehicle accidents dominate the risk picture for persons in their 20s, with only diabetes mellitus and end-stage renal disease therapy (for blacks only) having similar levels of risk in this age range. Late in life, cancer, acute myocardial infarction, Alzheimer's, and stroke become most common. The chronic diseases affecting the population later in life present the most likely diseases someone will face. Several interesting differences in disease and death risks were derived and reported among age-specific race and gender subgroups of the population. CONCLUSION: Presentation of risk estimates for a broad set of chronic diseases and nondisease causes of death within short-term age ranges among population subgroups provides tailored information that may lead to better educated prevention, screening, and control behaviors and more efficient allocation of health resources.

The partial testing design: a less costly way to test equivalence for sensitivity and specificity.

We propose a new, less costly, design to test the equivalence of digital versus analogue mammography in terms of sensitivity and specificity. Because breast cancer is a rare event among asymptomatic women, the sample size for testing equivalence of sensitivity is larger than that for testing equivalence of specificity. Hence calculations of sample size are based on sensitivity. With the proposed design it is possible to achieve the same power as a completely paired design by increasing the number of less costly analogue mammograms and not giving the more expensive digital mammograms to some randomly selected subjects who are negative on the analogue mammogram. The key idea is that subjects who are negative on the analogue mammogram are unlikely to have cancer and hence contribute less information for estimating sensitivity than subjects who are positive on the analogue mammogram. To ascertain disease state among subjects not biopsied, we propose another analogue mammogram at a later time determined by a natural history model. The design differs from a double sampling design because it compares two imperfect tests instead of combining information from a perfect and imperfect test.

Americans' knowledge of cancer risk and survival.

BACKGROUND: There are more than 500,000 deaths from cancer each year in the United States. This study examines Americans' knowledge of risk factors for breast, cervical, colon, and prostate cancers, which account for over 130,000 of these deaths, and their knowledge of the prospects of surviving these cancers following early detection. METHODS: Data were obtained from 12,035 subjects who completed the 1992 National Health Interview Survey Cancer Control Supplement which includes questions about cancer risk factors and survival. RESULTS: The majority of respondents were unable to identify major cancer risk factors when prompted with a list. Approximately two-thirds did not recognize that age increased the risk for breast and colon cancer, that diet increased the risk for colon cancer, or that multiple sex partners increased the risk for cervical cancer. Knowledge about survival was also poor. Only about half thought they had a good chance of survival following early detection of colon and cervical cancers, for which 5-year relative survival exceeds 90%. CONCLUSIONS: Americans lack knowledge about major risk factors for common cancers and about the prospects of survival following early detection. Knowledge about risk factors and about survival from cervical and colon cancers was poor at all ages, among all races, at all income levels, and at all educational levels. It was poorest among blacks and Hispanics and among those with the lowest income and least education. Americans need education about cancer risk factors and survival.

Cigarette smoking among U.S. adults by state and region: estimates from the current population survey.

BACKGROUND: Cigarette smoking is responsible for at least one third of all cancer deaths annually in the United States. Few sources exist in the peer-reviewed literature documenting state and regional differences in smoking behavior, despite the fact that cancer prevention and control efforts are increasingly being implemented below the national level. PURPOSE: Our goals were to determine smoking prevalence rates among men and women, by region, and for each of the 50 states and the District of Columbia from census survey data collected in 1992 and 1993 and to compare these rates with rates determined in 1985. METHODS: Every month, the U.S. Bureau of the Census collects labor force statistics on more than 100000 individuals on its Current Population Survey (CPS). For the September 1992, January 1993, and May 1993 CPS, the National Cancer Institute sponsored a 40-item Tobacco Use Supplement. The definition of a current smoker changed slightly between 1985 and 1992-1993. For the 1985 CPS, individuals were considered current smokers if they had smoked 100 cigarettes in their lifetime and were smoking at the time of interview; for the 1992-1993 CPS, current smokers included anyone who had smoked 100 cigarettes and was currently smoking every day or just on some days. We calculated current smoking rates (every day and some days combined) based on more than a quarter million adults (n = 266988) interviewed in 1992-1993. RESULTS: Substantial geographic variation exists in rates of current cigarette use among adults within the United States. In general, adults in the southern United States have higher rates of smoking and adults in the western states have lower rates of smoking and adults in the rest of the country, although differences in smoking behavior between men and women and among various racial and ethnic populations strongly influence these patterns. Only two states, Kentucky and West Virginia, exhibited adult smoking rates (men and women combined) of 30% or higher in 1992-1993; in contrast, in 1985, such rates were reported from 20 states. The only states in which the prevalence was below 20% in 1992-1993 were Utah (17.1%) and California (19.5%). Rates approaching 20% were reported from New Jersey (20.7%), Massachusetts (21.5%), and Nebraska, New York, and Hawaii (22.0% each) in 1992-1993. Rhode Island experienced the greatest relative decline in smoking prevalence from 1985 to 1992-1993, with a calculated relative change of -30.7% (based on a change in rate from 33.5% to 23.2%), followed by Delaware (-25.9%) the District of Columbia and New Jersey (-23.9% each), Connecticut (-23.2%), California (-22.9%), Alaska (-22.8%), Georgia (-22.6%), Massachusetts (-22.1%), and New York (-22.0%). CONCLUSIONS: Smoking rates are not uniform in the United States but vary considerably from state to state, even within the same region of the country. The CPS is the only mechanism currently capable of simultaneously monitoring smoking trends nationally, regionally, and on a state-by-state basis.

Recent trends in U.S. breast cancer incidence, survival, and mortality rates.

BACKGROUND: Clinical trials have demonstrated that use of mammographic screening and advances in therapy can improve prognosis for women with breast cancer. PURPOSE: We determined the trends in breast cancer mortality rates, as well as incidence and survival rates by extent of disease at diagnosis, for white women in the United States and considered whether these trends are consistent with widespread use of such beneficial medical interventions. METHODS: We examined mortality data from the National Center for Health Statistics and incidence and survival data by extent of disease from the Surveillance, Epidemiology, and End Results Program of the National Cancer Institute, all stratified by patient age, using statistical-regression techniques to determine changes in the slope of trends over time. RESULTS: The age-adjusted breast cancer mortality rate for U.S. white females dropped 6.8% from 1989 through 1993. A significant decrease in the slope of the mortality trend of approximately 2% per year was observed in every decade of age from 40 to 79 years of age. Trends in incidence rates were also similar among these age groups: localized disease rates increased rapidly from 1982 through 1987 and stabilized or increased more slowly thereafter; regional disease rates decreased after 1987; and distant disease rates have remained level over the past 20 years. Three-year relative survival rates increased steadily and significantly for both localized and regional disease from 1980 through 1989 in all ages, with no evidence of an increase in slope in the late 1980s. IMPLICATIONS: The decrease in the diagnosis of regional disease in the late 1980s in women over the age of 40 years likely reflects the increased use of mammography earlier in the 1980s. The increase in survival rates, particularly for regional disease, likely reflects improvements in systemic adjuvant therapy. Statistical modeling indicates that the recent drop in breast cancer mortality is too rapid to be explained only by the increased use of mammography; likewise, there has been no equivalent dramatic increase in survival rates that would implicate therapy alone. Thus, indications are that both are involved in the recent rapid decline in breast cancer mortality rates in the United States.

Factors associated with surgical and radiation therapy for early stage breast cancer in older women.

BACKGROUND: In addition to demographic and health care-related characteristics, the age and physiologic status of women at the time of breast cancer diagnosis have been reported to influence receipt of standard treatments. Previous studies of the influence of age and comorbidity have not examined whether other patient-, region-, or health care-related characteristics altered the association of age and comorbidity with type of treatment received. PURPOSE: This study examined factors associated with receipt of breast-conserving surgery and radiation therapy, both of which are recommended treatments for breast cancer, among a cohort of 18,704 women aged 65 years or more who had breast cancer diagnosed during the period from 1985 through 1989. METHODS: A data file linking Medicare claims records to data from the Surveillance, Epidemiology, and End Results (SEER) Program of the U.S. National Cancer Institute was utilized. Logistic regression analysis was used to examine associations between patient, region, and hospital characteristics and the receipt of specific treatments. The likelihood test was used to assess the significance of observed associations (expressed as odds ratios [ORs]). Because of multiple comparisons, only those ORs with two-sided P values <.01 were considered statistically significant. RESULTS: The frequency of breast-conserving surgery was highest (54%) among women aged 80 years or more, who had two or more comorbid conditions and stage I disease. However, in general, the receipt of radiation therapy among women undergoing breast-conserving surgery declined markedly with age, irrespective of comorbidity status and disease stage. Between the ages of 65-69 years and 80 years or older, radiation therapy declined from 77% to 24% among women with no comorbid conditions and from 50% to 12% among women with two or more comorbid conditions. In regression models that included hospital, region, and patient characteristics as variables, age and comorbidity remained independently associated with the receipt of radiation therapy (OR = 0.12 and 95% confidence interval [CI] = 0.10-0.14 for women aged 80 years or more compared with women 65-69 years of age and OR of 0.33 [95% CI = 0.24-0.46] for women with two or more comorbid conditions versus no comorbid conditions). CONCLUSIONS: After adjustment for multiple clinical and nonclinical factors influencing treatment, chronologic age remains an important independent factor associated with the receipt of radiation therapy after breast-conserving surgery among women aged 65 years or more who were diagnosed with early stage breast cancer. IMPLICATIONS: Future studies should determine whether these differences in treatment patterns among older women result in increased morbidity (e.g., from recurrence), shortened disease-free or overall survival, or decreased quality of life.

Use of gene tests to detect hereditary predisposition to cancer: what do we know about cost effectiveness?

Mutations in 4 genes associated with DNA repair have been shown to be associated with hereditary non-polyposis colon cancer (HNPCC) in families which display unusually high risk for colorectal cancer. Laboratory tests for mutations in these genes will soon be commercially available, raising the possibility that population-wide gene testing to identify individuals with an inherited susceptibility to colorectal cancer could be conducted. The purpose of our report is to explore the economic implications of conducting a program of population-wide screening for HNPCC compared with alternative programs which would be restricted to families already known to be at high risk for HNPCC. Rather than provide a definitive answer to these questions, our purpose is to indicate priority areas of research which need to be conducted before such a definitive analysis can be done. An exploratory analysis has been conducted to determine which factors are most important in determining the cost-effectiveness of a genetic testing program for HNPCC. The base case analysis focuses on current uncertainty about the population prevalence of the HNPCC genotype and phenotype, factors which are central to the cost-effectiveness of population-wide screening. Uncertainty in parameters related to the cost and effectiveness of screening and preventive interventions for HNPCC were explored using additional sensitivity analyses. Favorable levels of cost-effectiveness for population-wide screening are achieved only when the most favorable assumptions about HNPCC prevalence and the cost and effectiveness of screening and preventive interventions are made.

Breast cancer control among the underserved--an overview.

This paper explores barriers to the use of standard screening and breast cancer treatment that result in systematic differences in health outcomes. We review available data on individual, socioeconomic, and health system determinants of access to standard breast cancer care, including screening, diagnostic, and treatment services. Based on this review, we discuss the combination of factors which result in underservice. We argue that a broad framework which considers health system and social class as well as individual factors is useful for analyzing how structures of health care delivery tend to provide less than standard care to women who are older, have less income, or are less educated, black, or Hispanic. Data collection efforts which do not include structural and socioeconomic variables may result in an incomplete or misleading understanding of the determinants of underservice. These factors also need to be considered in the design and evaluation of public health policies and interventions meant to ameliorate the effects of underservice.

Medical device vigilance at FDA.

This overview of medical device vigilance at the U.S. Food and Drug Administration (FDA) discusses the basic reporting regulation and program at FDA. The FDA program has grown to the point where we now receive over 100,000 reports per year. How FDA examines the data patterns in overall reporting, to identify signals in the database, and the potential actions FDA takes to respond to these problems are also presented. New initiatives include, for example, new methods to triage reports and the move developing an internationally harmonised nomenclature.

Screening mammography in community practice: positive predictive value of abnormal findings and yield of follow-up diagnostic procedures.

OBJECTIVE: The purpose of this study was to gather from 50 community mammography practices that were included in the National Survey of Mammography Facilities data concerning abnormal findings on screening mammograms to determine the frequency of various recommendations made for patients who had abnormal findings and to compare these recommendations with the frequency with which the procedures were actually performed. We also determined the positive predictive value of screening mammograms (the number of cancers detected per 100 abnormal findings) and the yield (the number of cancers detected per 100 procedures done) of various diagnostic procedures done because of abnormal findings. MATERIALS AND METHODS: We identified 1717 screening mammograms done in the last half of 1991 that had abnormal findings. Radiologic recommendations and follow-up procedures, including repeat standard (screening) mammography, additional mammographic views, sonography, clinical breast examination, needle aspiration, needle biopsy, and open biopsy, were identified for all of the cases from the radiologic records, and follow-up data were obtained from referring physicians. The positive predictive value and yield in the National Survey of Mammography Facilities were compared with data from the mammography screening practice of the University of California at San Francisco (UCSF), a facility noted for its clinical efficiency. RESULTS: We estimate that 11% of all screening mammograms resulted in a recommendation for further diagnostic procedures. These 1717 mammograms with abnormal findings led to the following recommendations and procedures: repeat standard (screening) mammography, 610 (recommended)/635 (performed); additional mammographic views, 785/707; sonography, 400/345; biopsy, 189/229; and needle aspiration, 21/51. More procedures were done than were recommended in some cases because the results of certain procedures often led to the performance of other, additional procedures. The positive predictive value for screening examinations with abnormal findings was 3.5%, and the yield for open biopsy was 21%. In the UCSF data base, the positive predictive value for examinations with abnormal findings was 10%, and the yield for open biopsy was 34%. CONCLUSION: The positive predictive value for examinations with abnormal findings and the yield for diagnostic procedures performed as a result of abnormal findings in 50 community radiologic facilities were higher than those reported in some earlier studies, a fact that raised concern about the induced cost of screening mammography. However, these values were low compared with those in the UCSF data base. This fact was particularly true of repeat standard (screening) mammography.

The use of gene tests to detect hereditary predisposition to cancer: economic considerations.

Several mutations recently have been shown to be associated with hereditary nonpolyposis colon cancer (HNPCC) in families displaying unusually strong predisposition to colorectal cancer. Laboratory tests to detect such gene mutations soon will be commercially available, raising the possibility for population-wide screening. The purpose of this study was to explore the economic implications of conducting a population-wide screening for HNPCC compared with restricted screening among members of the families at high risk. An exploratory analysis was performed to determine which factors are most important in determining the cost-effectiveness of such a testing program. The base-case analysis focuses on current uncertainty about the population prevalence of the HNPCC genotype and phenotype. The uncertainty with regard to parameters related to the cost-effectiveness of screening and preventive interventions for HNPCC were explored, using additional sensitivity analyses. Cost-effectiveness is achieved for population-wide screening only when assumptions most favorable to that outcome are made regarding the HNPCC prevalence, the cost and effectiveness of screening, and preventive intervention. Information on the population prevalence of HNPCC genotypes and the penetrance of these genotypes is essential for making policy decisions. Additional information is needed regarding the determinants of the cost of genetic testing and counseling, the efficacy of preventive regimens for individuals at high risk, and the accuracy and cost of alternative methods of identifying families at risk.

Medicare payments from diagnosis to death for elderly cancer patients by stage at diagnosis.

Although extensive resources go to cancer care, national population-based data on the costs of such care at the patient level have been unavailable. Medicare payments subsequent to diagnosis of cancer for elderly enrollees with five common cancers were estimated using tumor registry data from the Surveillance, Epidemiology, and End Results Program linked to Medicare claims from 1984 to 1990. The time between diagnosis and death was divided into four phases corresponding to the clinical course of solid tumors, average payments for each phase were estimated (including payments for services not related to cancer), then phase-specific payment data were aggregated. Average payments by phase varied among cancer sites, especially in the initial care phase, where payments were highest for lung and colorectal cancers ($17,500 in 1990 dollars) and lowest for female breast cancer ($8,913). Total Medicare payments from diagnosis to death were highest for persons with bladder cancer ($57,629) and lowest for those with lung cancer ($29,184). Low payments for persons with lung cancer corresponded to brief survival times. Persons diagnosed at earlier stages incurred higher total payments between diagnosis and death than those diagnosed at later stages, reflecting their longer survival. This implies that early detection may increase total Medicare expenditures by extending beneficiaries' lives. However, Medicare payments per year of survival were lower for earlier stages. Data on Medicare payments subsequent to diagnosis of cancer are useful for identifying the cost implications of differences in treatment patterns by demographic characteristics, geography, and delivery systems; comparing the financial impact of alternative therapies; evaluating the long-term cost impacts of screening and prevention programs; and risk-adjusting payments to health plans.

"Eat for Health": a supermarket intervention for nutrition and cancer risk reduction.

OBJECTIVES: The results of an evaluation of "Eat for Health," a supermarket nutrition intervention, are presented. The study tested whether such a program could be successfully carried out and whether it could effect changes in knowledge, attitudes, and food purchasing behavior in line with nutrition and cancer risk reduction guidelines. METHODS: The evaluation consisted of an in-store monitoring element, an in-store and telephone consumer survey, and an analysis of sales data on selected foods. A matched-pair design, using a total of 40 stores in the intervention and comparison groups, was used. RESULTS: The intervention was successfully implemented and had limited success in changing some food purchasing behaviors. There appeared to be no effect on knowledge and attitudes except for increased awareness of a link between diet and cancer and of the program itself. CONCLUSIONS: Despite the intervention's success, limitations of the consumer survey and sales data analyses and the continuing diffusion of nutrition messages throughout society make it difficult to specify the impact of this program on consumer nutrition knowledge and behaviors.

Potential for cancer related health services research using a linked Medicare-tumor registry database.

The National Cancer Institute and the Health Care Financing Administration share a strong research interest in cancer costs, access to cancer prevention and treatment services, and cancer patient outcomes. To develop a database for such research, the two agencies have undertaken a collaborative effort to link Medicare Program data with the Surveillance, Epidemiology, and End Results (SEER) Program database. The SEER Program is a system of 9 population-based tumor registries that collect standardized clinical information on cases diagnosed in separate, geographically defined areas covering approximately 10% of the US population. Using a deterministic matching algorithm, the records of 94% of SEER registry cases diagnosed at age 65 or older between 1973 to 1989, or more than 610,000 persons, were successfully linked with Medicare claims files. The resulting database, combining clinical characteristics with information on utilization and costs, will permit the investigation of the contribution of various patient and health care setting factors to treatment patterns, costs, and medical outcomes.

A model for making project funding decisions at the National Cancer Institute.

This paper describes the development of a model for making project funding decisions at The National Cancer Institute (NCI). The American Stop Smoking Intervention Study (ASSIST) is a multiple-year, multiple-site demonstration project, aimed at reducing smoking prevalence. The initial request for ASSIST proposals was answered by about twice as many states as could be funded. Scientific peer review of the proposals was the primary criterion used for funding decisions. However, a modified Delphi process made explicit several criteria of secondary importance. A structured questionnaire identified the relative importance of these secondary criteria, some of which we incorporated into a composite preference function. We modeled the proposal funding decision as a zero-one program, and adjusted the preference function and available budget parametrically to generate many suitable outcomes. The actual funding decision, identified by our model, offers significant advantages over manually generated solutions found by experts at NCI.

Representativeness of the surveillance, epidemiology, and end results program data: recent trends in cancer mortality rates.

BACKGROUND: Mortality, incidence, and survival rates are the primary measures used by the National Cancer Institute (NCI) to monitor cancer in the United States. The Surveillance, Epidemiology, and End Results (SEER) data system collects data on all cancers diagnosed among residents in geographically defined populations, which comprise about 10% of the U.S. population. This data system is the major component of the NCI system for tracking these rates. Thus, it is important to assess the degree to which SEER data are representative of the entire U.S. population. PURPOSE: National data on mortality, but not on incidence or survival, are available from the National Center for Health Statistics. These data provide a census against which mortality data from the subset of the SEER regions may be compared. METHODS: Multivariate regression analyses of age-adjusted mortality rates from 1975 to 1988, computed for the SEER areas and for the entire United States, were performed for race- and sex-specific data from 15 cancer sites. Representativeness was evaluated by testing for differences in trends and levels between the data from the U.S. population and those from the SEER Program. RESULTS: Data from the SEER regions reflected the correct direction of trend for all sites, although some race-, sex-, and site-specific differences existed for the magnitude of the trends and levels of mortality when compared with data from the U.S. population. CONCLUSIONS: The demonstration that data from the SEER population do occasionally yield mortality rates that differ from those for the entire U.S. population suggests that data from the SEER coverage population are, in some cases, not representative of the greater U.S. population. IMPLICATIONS: This issue is of particular relevance to the interpretation of incidence measures, computed from the SEER data, for which there is no national database. Future efforts should be directed at a better understanding of how the SEER population differs from the U.S. population so that SEER rates can be adjusted to be more nationally representative.