Decomposition of disparities in life expectancy with applications to administrative health claims and registry data.

Research shows that geographic disparities in life expectancy between leading and lagging states are increasing over time while racial disparities between Black and White Americans have been going down. In the 65+ age strata morbidity is the most common cause of death, making differences in morbidity and associated adverse health-related outcomes between advantaged and disadvantaged groups an important aspect of disparities in life expectancy at age 65 (LE65). In this study, we used Pollard's decomposition to evaluate the disease-related contributions to disparities in LE65 for two types of data with distinctly differing structures: population/registry and administrative claims. To do so, we analyzed Pollard's integral, which is exact by construction, and developed exact analytic solutions for both types of data without the need for numerical integration. The solutions are broadly applicable and easily implemented. Applying these solutions, we found that the largest relative contributions to geographic disparities in LE65 were chronic lower respiratory diseases, circulatory diseases, and lung cancer; and, to racial disparities: arterial hypertension, diabetes mellitus, and cerebrovascular diseases. Overall, the increase in LE65 observed over 1998-2005 and 2010-2017 was primarily due to a reduction in the contributions of acute and chronic ischemic diseases; this was partially offset by increased contributions of diseases of the nervous system including dementia and Alzheimer's disease.

A two-disease model of female breast cancer: mortality in 1969 among white females in the United States.

A mathematical model of the age distribution of breast cancer mortality was developed on the basis of the two-disease theory of breast cancer incidence. The model included representations of the time from tumor initiation to death, the competing risk effects of other disease, and differential susceptibility to each of the disease components. This model successfully predicted the single year of age frequency of breast canceomponents of this model was consistent with several epidemiologic findings. Most significantly, the age distribution of breast cancer deaths from premenopausal disease was consistent with incidence patterns in non-Western countries, where the incidence of the postmenopausal disease component was hypothesized to be lower because of nutritional differences.

Statistically adjusted estimates of geographic mortality profiles.

The spatial variation of site-specific cancer mortality rates at the county or state economic area level can provide a) insights into possible etiologic factors and b) the basis for more detailed epidemiologic studies. One difficulty with such studies, especially for rare cancer types, is that unstable local area rate estimates, resulting from small population sizes, can obscure the underlying spatial pattern of disease risk. This paper presents a methodology for producing more stable rate estimates by statistically weighting the local area rate estimate toward the experience at the national level. The methodology is illustrated by the analysis of the spatial variation of two cancer types, bladder and lung, for U.S. white males over the three decades 1950-79.

U.S. cancer mortality 1950-1978: a strategy for analyzing spatial and temporal patterns.

There are a number of technical and statistical problems in monitoring the temporal and spatial variation of local area death rates in the United States for evidence of systematically elevated risks. An analytic strategy is proposed to reduce one of the major statistical concerns, i.e., that of identifying areas with truly elevated mortality risks from a large number of local area comparisons. This analytic strategy involves two stages. The first is a procedure for examining the entire distribution of local area death rates instead of simply selecting high risk "outliers." The second is the development of an analytic procedure to relate the temporal changes in the cross-sectional distribution of local area death rates to models of the disease process operating within the populations in those areas. The procedures are applied to data on cancer mortality for the 3050 counties (or county equivalents) of the United States over the period 1950 to 1978. A number of striking mortality patterns, both within the entire United States and within various regions and states, are identified. For example, perhaps the most persistent finding was that the risk increases in the death rates for respiratory cancer mortality were due to a "catching up" of nonmetropolitan county mortality rates with metropolitan area mortality rates.

Assessment of spatial variation of risks in small populations.

Often environmental hazards are assessed by examining the spatial variation of disease-specific mortality or morbidity rates. These rates, when estimated for small local populations, can have a high degree of random variation or uncertainty associated with them. If those rate estimates are used to prioritize environmental clean-up actions or to allocate resources, then those decisions may be influenced by this high degree of uncertainty. Unfortunately, the effect of this uncertainty is not to add "random noise" into the decision-making process, but to systematically bias action toward the smallest populations where uncertainty is greatest and where extreme high and low rate deviations are most likely to be manifest by chance. We present a statistical procedure for adjusting rate estimates for differences in variability due to differentials in local area population sizes. Such adjustments produce rate estimates for areas that have better properties than the unadjusted rates for use in making statistically based decisions about the entire set of areas. Examples are provided for county variation in bladder, stomach, and lung cancer mortality rates for U.S. white males for the period 1970 to 1979.

A cohort analysis of U.S. stomach cancer mortality 1950-1977.

Models of human carcinogenesis, such as the multi-stage model of Armitage and Doll, are designed to explain the age increase in the incidence of cancers in individuals. As a consequence, analyses of population level age-specific death rates via such models are appropriately applied to cohort data where such data are available. In this study a multi-stage model is applied to cohort data for stomach cancer death rates in the U.S. population for nine distinct cohorts observed over a recent 28-year period (1950-1977). The multi-stage model parameters obtained from the analysis of the cohort data show significant differences from the parameters obtained from analyses of cross-sectional mortality data under the assumption of no cohort differences in age-specific stomach cancer death rates.

Population impact of mortality reduction: the effects of elimination of major causes of death on the 'saved' population.

In this paper we examine the effects on life expectancy of elimination of 4 major causes of death. Methodologically, we compare the results of cause elimination under assumptions of pattern of failure elimination and assumptions of underlying cause elimination in a modified multiple-decrement life table framework for the segment of the population impacted. The 4 diseases selected for analysis are cancer, ischaemic heart disease, stroke, and diabetes, major killers among the elderly population. The degree to which life expectancy changes occur within the population from elimination of a given cause if a function of 3 factors: 1) distribution of age at death by cause for persons who die of that cause, 2) the gain in person years lived for those 'saved' from dying from that cause which has been eliminated, and 3) the proportion of all deaths which are due to the specific cause which is eliminated. Mortality data from the 1969 U.S. multiple cause mortality tapes from NCHS are analysed to determine the impact of life expectancy for males and females of both races when one of these 4 specific causes of death is eliminated.

Longitudinal models for chronic disease risk: an evaluation of logistic multiple regression and alternatives.

The logistic multiple regression model is often used in the analysis of the relation between chronic disease risk and selected risk factors in longitudinal data. Unfortunately, the logistic function has certain properties that make it inappropriate as a mode of risk analysis for longitudinal studies. The consequences of applying the logistic function to longitudinal data is that the numerical values of logistic regression coefficients cannot be meaningfully compared between studies of different durations. Sample calculations are presented to illustrate the magnitude of the problem for a range of relative study lengths and levels of risk. Two solutions are offered for the problem. First, a series of approximations are derived which permit such comparisons if the studies are not greatly dissimilar in length. Second, if comparisons of the risk coefficients are to be made across studies of greatly dissimilar duration, it is necessary to model risk via an appropriate statistical model. Criteria for assessing the appropriateness of risk functions for the analysis of longitudinal data are proposed and alternatives evaluated.

A dynamic analysis of chronic disease development: a study of sex specific changes in coronary heart disease incidence and risk factors in Framingham.

An analysis of sex differentials in the dynamics of coronary heart disease (CHD) in the Framingham study is conducted using a methodology designed process which aids in the interpretation of the longitudinal results. This methodology permits the analysis of risk variable changes over time to be conducted independently of the analysis of the relation of risk variable values to CHD risk. The independent analyses afford a complete utilisation of information from longitudinal studies of chronic disease risk and produce insights into the dynamics of chronic disease development not available by other analytic strategies. The results of this analysis are compared with results obtained from a similarly structured multiple logistic analysis--a comparison which illustrates some of the technical and conceptual deficiencies of the often employed multiple logistic analysis when applied to longitudinal data.

A stochastic compartment model of stomach cancer with correlated waiting time distributions.

The incidence and growth rate of stomach cancer in the US population is modelled, for each sex, as a partially observed, discrete state stochastic process. Explicit evaluation of the transition rates between the states of the model is made possible by identifying them as specific functions of the time spent within each state. The functions used in the model were selected from the medical and epidemiological literature. With the model it was found possible to obtain fits to the age distribution of deaths due to stomach cancer for white males in 1975 and for selected age ranges for white females. These results suggested that the natural history of stomach cancer is different for females above and below age 65.

Risk factor dynamics, mortality and life expectancy differences between eastern and western Finland: the Finnish Cohorts of the Seven Countries Study.

Prior studies have not accounted for male mortality being higher in east than west Finland. Efforts to identify the mechanisms producing higher mortality in the east, due primarily to cardiovascular diseases (CVD), initially focused on a search for new risk factors. An alternate approach is to examine the assumptions of the analysis. This was investigated using a model which described (a) changes in risk factors over time, (b) dependency of risk factor effects on age, and (c) interactions and nonlinear effects of risk factors on mortality. The model was applied to 25-year follow-up data from cohorts of eastern (N = 823) and western (N = 888) Finnish men using pulse pressure, diastolic blood pressure, body mass index, total cholesterol, vital capacity index, cigarette smoking, and heart rate as risk factors. At age 40, men in the west had a life expectancy 2.4 years higher. Of the difference 29% (0.7 years) was associated with area differences in risk factor means, variances, and their change with age. The remainder, 1.7 years, was associated with age differences in the relation of risk factor interactions to CVD mortality. Possible reasons for these differences, such as joint elevation of several risk factors inducing rapid progression of atherogenesis, are discussed. No significant area differences were observed for mortality from either cancer or other causes.

Predicting 10-year care requirements for older people with suspected Alzheimer's disease.

OBJECTIVE: To describe the types and costs of care received for 10 years after the identification of an older person with suspected Alzheimer's disease (AD) by using data from 3254 patients with suspected AD who participated in the National Long Term Care Survey (NLTCS). METHODS: By using a Markov model derived using grade of membership techniques, the following were determined: survival probabilities at 10 years; years of survival during the 10 years; years in institutions; years with two or more impairments in basic activities of daily living; hours of paid and informal care while the older person lived in the community; and costs of paid community, institutional, and medical care. RESULTS: Greater degrees of cognitive impairment present when AD was identified were associated with reduced predicted probability of surviving 10 years, increased predicted number of years spent in institutions, increased hours of care required while affected individuals remained in the community, and increased costs of paid community, institutional, and medical care. Substantial differences between men and women were seen: severity-adjusted 10-year costs were almost two times higher for women with AD than for men ($75,000 compared with $44,000); according to sensitivity analysis, average 10-year costs might be as high as $109,000 for women and $67,000 for men. CONCLUSIONS: AD imposes a substantial burden on older persons. Interventions that slow the progression of the disease may therefore affect community survival as well as healthcare costs.

Product liability forecasting for asbestos-related personal injury claims: a multidisciplinary approach.

This paper focuses on three aspects of forecasting models for asbestos-related disease/injuries relating to the Manville asbestos case: (1) The structure of forecasting models for asbestos-related personal injuries. (2) The epidemiologic evidence supporting the selected model structure and the constraints on the modeling assumptions imposed by that evidence. (3) The range of uncertainty associated with projections based on these forecasting models and issues relating to decision making under uncertainty.

Longevity in the united states: age and sex-specific evidence on life span limits from mortality patterns 1960-1990.

Determining the biological limits to human longevity is more difficult than for most other species because humans are long-lived. Consequently, mortality data, such as from the U.S. vital statistics system, which have been available for a long time (relative to most epidemiological studies) and have large numbers of cases, including deaths reported to advanced ages, are important in studying human longevity-though care must be exercised in dealing with error in age reporting. Furthermore, it is unlikely that free-living humans can realize as much of their biological endowment for longevity as animals living in a highly controlled experimental environment. We examined changes, 1960 to 1990, in U.S. White male and female extinct cohort life tables and age at death distributions to (a) examine evidence for the effects of a biological life span limit in current U.S. mortality patterns and (b) produce lower bound estimates of that limit.

The dynamics of dimensions of age-related disability 1982 to 1994 in the U.S. elderly population.

Declines in chronic disability were observed in the National Long Term Care Survey (NLTCS) 1982 to 1994. We analyzed the 1982, 1984, 1989, and 1994 NLTCS to identify the dimensions of chronic disability from multivariate analyses of a rich battery of measures of the ability (or inability) to perform specific activities. Changes over time in the prevalence of individual disability dimensions can be tracked to evaluate the rate of age-related losses of specific functions, 1982-1994. Seven dimensions described changes in the age dependence of 27 activities of daily living, instrumental activities of daily living, and physical performance measures in community and institutional resident elderly individuals over the 12 year period. Adjusted for age, the healthiest dimension with the best physical function experienced the largest increase in prevalence (3.3%) implying a decline in age-related disability. Disability declines were correlated with reductions in select health conditions (e.g., dementia and circulatory disease) over the study period.

Use of Medicare services before and after introduction of the prospective payment system.

OBJECTIVE: The case mix-adjusted pattern of use of health care services, especially posthospital care, is compared before and after the introduction of Medicare's Prospective Payment System (PPS). DATA SOURCES: The 1982 and 1984 National Long Term Care Surveys (NLTCS) linked to Medicare administrative records 1982-1986 provide health and health service use data for 12-month periods before and after the introduction of PPS. STUDY DESIGN: Case-mix differences between pre- and post-periods are controlled by using the Grade of Membership model to identify health groups from the NLTCS data. Differences in timing (e.g., hospital length of stay) were controlled using life table models estimated for each health group, that is, service use patterns pre- and post-PPS are compared within groups. PRINCIPAL FINDINGS: Hospital LOS and admission rates declined post-PPS. Changes in the timing and location of death occurred but, overall, mortality did not increase. Changes in post-acute care service use by elderly, chronically disabled Medicare beneficiaries were observed: home health service use increased overall and among the unmarried disabled population. CONCLUSIONS: PPS did not adversely affect quality of care as reflected in mortality or in hospital readmissions. Moreover, the differential use of post-acute care, and changes in hospital LOS by health group, indicate that the system responded, specific to marital status and age, to the severity of needs of chronically disabled persons.

Bioactuarial models of national mortality time series data.

The incidence and prevalence of chronic degenerative disease in America's elderly population are important determinants of the need for long-term care health services. Though a wide range of data on disease incidence and prevalence is available from a variety of different health studies, a Congressional Budget Office study (1977) concluded that data limitations are a major factor in the lack of precise national long-term care cost estimates. In this paper, we present a modeling strategy to make better use of existing data by using biomedically motivated actuarial models to integrate multiple data sources into a comprehensive model of population health dynamics. The development of a specific model for application to a disease of interest involves three distinct phases. First, biomedical evidence and data are used to specify a cohort model of chronic disease morbidity and mortality. Second, the model is fitted to cohort mortality data with estimates of its parameters being derived by maximum likelihood procedures. Third, the morbidity distribution in the national population is generated from the parameter estimates. The model is used to examine lung cancer morbidity and mortality patterns for U. S. white and non-white males in 1977. A review of these patterns suggests that, based on current concepts of lung cancer incidence and natural history, over 2 percent of white males in the United States have lung cancer at some stage of development, though most of this prevalence is pre-clinical.

Analysis of underwriting factors for AAPCC (adjusted average per capita cost).

The adjusted average per capita cost (AAPCC) formula is used to determine payment to health maintenance organizations (HMOs) by Medicare. The four original underwriting factors (i.e., age, sex, institutional status, and welfare status) for the AAPCC were calibrated from the Current Medicare Surveys for 1974-76. Those factors have been updated by various actuarial adjustments. Revised calculations of the AAPCC underwriting factors are presented using survey data from the 1984 National Long-Term Care Survey and expenditure data from the Medicare Part A and Part B bill files. Also examined is the effect on the underwriting factors of chronic functional disability, defined as having one or more chronic limitations in activities of daily living. Comparison of alternative underwriting factors is conducted by simulating the dollar impact on payment to HMOs for select enrollee populations.

Home health and skilled nursing facility use: 1982-90.

In this article, analyses are made of home health and skilled nursing facility (SNF) use for the period 1982-90 using Medicare records linked to data on community and institutional residents from the National Long-Term Care Surveys (NLTCSs) of 1982, 1984, and 1989. The combined survey and administrative data analyses are performed to ascertain how the chronic health and functional characteristics of community and institutional residents using Medicare-reimbursed services changed during the period. During this period, changes had been made in the Medicare system that affected the use of services for persons with specific health and functional problems.

Sex differences in human mortality and aging at late ages: the effect of mortality selection and state dynamics.

Models of gender differences in human mortality and aging depend on assumptions about temporal rates of physiological change. Simple models like the Gompertz fail to describe the mortality of either males or females at late ages. This suggests a need for biologically more detailed models to represent the age dependency of human mortality as well as gender differences in that age dependence. By modeling the sex-specific interaction of time-varying covariates with multiple dimensions of mortality selection, one can more accurately describe the age dependence of mortality and more complex physiological aging patterns. The multivariate model of aging changes is used to describe gender differences using data from (a) a longitudinal study of physiological changes and mortality and (b) a nationally representative longitudinal survey of changes in function and mortality.