Progression of functional disability in patients with rheumatoid arthritis. Associations with rheumatology subspecialty care.

BACKGROUND: To determine whether patients with rheumatoid arthritis and their physicians make appropriate decisions regarding referral to rheumatologists and the need for continuing rheumatology care, we examined the relationship between the progression of functional disability in these patients and their use of rheumatology subspecialty care over time. METHODS: A cohort of 282 patients with rheumatoid arthritis was followed prospectively for up to 10 years. Participants were categorized into three subgroups based on the pattern of care received from rheumatologists over the study period: patients who were never treated by a rheumatologist; patients treated by a rheumatologist only intermittently; and patients treated by a rheumatologist at least once during each 6-month study period. The outcome was the rate of progression of functional disability, measured using the Health Assessment Questionnaire Disability Index. RESULTS: Among the 52 patients who had not been referred to a rheumatologist, 30 (58%) had rates of progression of functional disability that were stable or improving over time (rate < 0.01 Disability Index units per year), while 22 (42%) had rates that were worsening (rate > or = 0.01 Disability Index units per year). Among patients treated by rheumatologists, the average rate of progression was substantially lower among the 69 patients who were treated regularly by a rheumatologist than among 161 patients treated by rheumatologists intermittently (0.008 Disability Index units per year vs 0.020 Disability Index units per year). This difference was associated with more intensive use of second-line antirheumatic medications, and more frequent joint surgeries, among patients treated by rheumatologists on a regular basis. CONCLUSIONS: Most patients with rheumatoid arthritis in this community cohort were treated by a rheumatologist, but 42% of those not referred had progressively increasing functional disability. Among patients treated by rheumatologists, those who had continuing care from rheumatologists experienced lower rates of progression of functional disability than those who had only intermittent care. These results suggest that use of rheumatology subspecialty care is associated with better health outcomes in rheumatoid arthritis.

Costs of hepatitis C.

OBJECTIVE: To estimate the direct and indirect costs of the hepatitis C virus (HCV) in the United States in 1997. DESIGN: Aggregation and analysis of national data sets collected by the National Center for Health Statistics, the Health Care Financing Administration, and other government bureaus and private firms. To estimate costs, we used the human capital method, which decomposes costs into direct categories, such as medical expenses, and indirect categories, such as lost earnings and lost home production. We consider HCV that results in chronic liver disease separate from HCV that results in primary liver cancer. RESULTS: We estimate $5.46 billion as the cost of HCV in 1997. Costs are split as follows: 33% for direct and 67% for indirect costs. Hepatitis C virus that results in chronic liver disease contributes roughly 92% of the costs, and HCV that results in primary liver cancer contributes the remaining 8%. The total estimate of $5.46 billion is conservative, because we ignore costs associated with pain and suffering and the value of care rendered by family members. CONCLUSIONS: To our knowledge, only one estimate of the annual costs of HCV in the 1990s has appeared in the literature, $0.6 billion. However, that estimate was not supported by an explanation of the methods. Our estimate, which relies on detailed methods, is nearly 10 times the original estimate. Our estimate of $5.46 billion is on a par with the cost of asthma ($5.8 billion [1994]).

Occupational injury and illness in the United States. Estimates of costs, morbidity, and mortality.

OBJECTIVE: To estimate the annual incidence, the mortality and the direct and indirect costs associated with occupational injuries and illnesses in the United States in 1992. DESIGN: Aggregation and analysis of national and large regional data sets collected by the Bureau of Labor Statistics, the National Council on Compensation Insurance, the National Center for Health Statistics, the Health Care Financing Administration, and other governmental bureaus and private firms. METHODS: To assess incidence of and mortality from occupational injuries and illnesses, we reviewed data from national surveys and applied an attributable risk proportion method. To assess costs, we used the human capital method that decomposes costs into direct categories such as medical and insurance administration expenses as well as indirect categories such as lost earnings, lost home production, and lost fringe benefits. Some cost estimates were drawn from the literature while others were generated within this study. Total costs were calculated by multiplying average costs by the number of injuries and illnesses in each diagnostic category. RESULTS: Approximately 6500 job-related deaths from injury, 13.2 million nonfatal injuries, 60,300 deaths from disease, and 862,200 illnesses are estimated to occur annually in the civilian American workforce. The total direct ($65 billion) plus indirect ($106 billion) costs were estimated to be $171 billion. Injuries cost $145 billion and illnesses $26 billion. These estimates are likely to be low, because they ignore costs associated with pain and suffering as well as those of within-home care provided by family members, and because the numbers of occupational injuries and illnesses are likely to be undercounted. CONCLUSIONS: The costs of occupational injuries and illnesses are high, in sharp contrast to the limited public attention and societal resources devoted to their prevention and amelioration. Occupational injuries and illnesses are an insufficiently appreciated contributor to the total burden of health care costs in the United States.

Randomized controlled study of a retiree health promotion program. The Bank of American Study.

The initial results of a 12-month controlled trial of a health promotion program in 5686 Bank of America retirees, randomized into full program, questionnaire only, and insurance claims only groups, were analyzed to determine whether the health promotion program was effective. Comparisons were between program and questionnaire only groups for self-reported health habit changes, health risk scores, medical care utilization, and days confined to home, and between all groups for insurance claims data. The intervention, or full program, included health habit questionnaires administered every 6 months, individualized time-oriented health risk appraisals, personal recommendation letters, self-management materials, and a health promotion book. Twelve-month changes in health habits, health status, and economic variables favored the full program group in 31 of 32 comparisons and were statistically significant at the .05 level in two-tailed tests in 19 comparisons and at the .01 level in two-tailed tests in 13 comparisons. Over 12 months, overall computed health risk scores decreased by 4.3% in the full program experimental group and increased by 7.2% in the questionnaire only control group. Total direct and indirect costs decreased by 11% in the experimental group and increased by 6.3% in the questionnaire only control group. Analysis of claims data confirmed these trends. A low-cost health promotion program for retirees was effective in changing health behaviors and has potential to decrease health care utilization.

Hours at work and employment status among HIV-infected patients.

OBJECTIVE: To study differences in employment and work hours among three groups of HIV-infected and non-infected individuals. METHODS: Data on 1263 patients seen in five different sites in California were drawn from the AIDS Time-Oriented Health Outcome Study. Three groups of patients were examined: AIDS patients, HIV-positives without diagnosed AIDS, and HIV-negatives. The HIV-negative patients were used as a comparison group in comparing hours worked by all patients, whether they worked or not; the probability of working, regardless of the number of hours; and work hours only for those patients who worked. RESULTS: Adjustment for covariates in a 2-equation econometric model reduced the difference in employment rates between the AIDS patients and the other two groups, suggesting that characteristics other than AIDS status account, in part, for their low employment rates. After adjustment, we did not find any statistically significant differences in employment probabilities or work hours between the HIV-positive patients without diagnosed AIDS and the comparison group. However, AIDS patients reported approximately 14 work hours fewer (P < 0.0001) and lower probabilities of employment (P < 0.0001) than the HIV-negative comparison group among all patients with and without jobs. Moreover, among those with jobs, patients with AIDS reported approximately 3 work hours fewer per week (P = 0.0385). No statistically significant differences in work hours were found between HIV-positives without diagnosed AIDS and comparison patients. CONCLUSION: AIDS patients were less likely to be employed than either of the other groups, but crude, unadjusted unemployment rates exaggerate the effect of AIDS. For those employed, AIDS patients work only 3 h less per week than either of the other groups.

Assessing the importance of an independent variable in multiple regression: is stepwise unwise?

Two procedures for evaluating the importance of an independent variable are reviewed: the allocation of R2 method and the significance method. Multicollinearity severely hampers attempts to evaluate an independent variable using the allocation of R2 method. The significance method relies on assessing the stochastic, "magnitudinal", (or clinical) and practical significance of an independent variable and does not suffer as severely from the ravages of multicollinearity. An example of the limitations of allocations of R2 method in stepwise regression is discussed and another interpretation of the results using the significance approach is suggested.

Pooled time series regression analysis in longitudinal studies.

Many longitudinal studies attempt to examine changes in outcome measures over time in groups of patients. Applying conventional analytic techniques, such as a single classical linear regression model, to these data will often not result in minimum variance estimates, and may affect the results of tests of significance. Pooled time series regression analyses comprise a set of techniques that may be used in these instances to model changes in outcome measures over time. Pooling of time series data from many individuals may be done using two types of models: fixed effect models, which specify differences among individuals in separate intercept terms, and random effects models, which allow for differences among individuals by including an additional error component in the model. The choice between these alternative model specifications is guided by both theoretical and statistical considerations. This paper describes the use of pooled time series analysis, contrasts these methods with two classical linear regression approaches, and demonstrates these differences using two examples: a hypothetical study of serum glucose measurements in patients with diabetic ketoacidosis, and a longitudinal study of the development of functional disability in a cohort of patients with rheumatoid arthritis. These methods may be applicable to the study of outcomes in many chronic illnesses.

Effects of smoking and being overweight on current earnings.

Medical researchers have demonstrated the potent influences of smoking and being overweight on the deterioration of health. Economists have documented the sizable impact of poor health on earnings. We looked for statistical associations between smoking and being overweight on the one hand and current earnings on the other using a national probability sample of American workers employed 20 or more hours a week. Using multiple regressions that included a variety of control variables, for both male and female workers no strong statistical associations could be found. These results should not be interpreted as suggesting that smoking and being overweight do not have profound influences on lifetime earnings through their effects on health. The results suggest only that among workers currently employed 20 hours or more a week, no residual short-run direct association exists between smoking or being overweight and current annual earnings.

Liver cirrhosis deaths within occupations and industries in the California occupational mortality study.

Mortality rates were drawn from the California Occupational Mortality Study (COMS) to analyze liver cirrhosis deaths within occupations and industries from 1979 to 1981. Age-adjusted Standardized Mortality Rates (SMRs) were made available by the State of California for separate analyses of women, men, blacks and whites. Rankings of occupations with narrow confidence intervals were strikingly similar for blacks and whites. Within occupations, the highest female SMRs were for waitresses, telephone operators, cosmetologists, dress makers, hospital orderlies, textile workers, and laborers. The lowest female SMRs were for skilled crafts workers and teachers. High male occupations included water transportation workers, bartenders, loggers, laborers, roofers, construction workers, farm workers, iron workers, and painters. Low male occupations included teachers, physicians and dentists, managers, factory supervisors, business sales workers, heavy equipment operators, and other professionals. High female industries included eating and drinking places, laundry/dry cleaning, nursing and personal care facilities, aerospace, beauty shops, and entertainment. Low female industries included wholesale trades and education. High male industries included water transportation, military, guard services, eating and drinking places, iron and steel mills, and railroads. Low male industries included research/engineering labs, education, and computer manufacturing. This study was descriptive. It remains unknown whether certain jobs cause excessive drinking and cirrhosis, or whether people who are prone to develop cirrhosis select certain jobs.

Gender and race differences in the correlation between body mass and education in the 1971-1975 NHANES I.

STUDY OBJECTIVE: Differences in the correlation between body mass index and education across four gender and race groups were investigated while simultaneously accounting for occupation, income, marital status, and age. DESIGN: The study used analysis of covariance techniques to calculate average body mass and confidence intervals within education categories while simultaneously adjusting for the covariates: age, square of age, family income, marital status, and occupation. SETTING: Data were drawn from the US National Health and Nutrition Examination Survey (NHANES I), 1971-1975. NHANES I is a national probability sample designed to gather information on the non-institutionalised US civilians, ages 1-74 years. SUBJECTS: Samples of 8211 white women, 1673 black women, 6188 white men, and 1023 black men were drawn from the NHANES I, 1971-1975. MAIN RESULTS: Data in the female samples indicate a strictly inverse relation between body mass and years of schooling among white women and an inverted "U" association among black women, achieving a maximum around 8 to 11 years of schooling. In the male samples data indicate inverted "U" relations among both black and white men, reaching maxima between 12 and 15 years of schooling. The sides of the "U" curve are much steeper for black than for white men. CONCLUSIONS: The four gender/race categories display four different body mass index and education associations. These four associations are only slightly altered by simultaneously adjusting for two additional measures of socioeconomic status: occupation and income.

Gender, firm size, industry, and estimates of the value-of-life.

Two national probability samples are used to investigate four empirical questions associated with estimating the value-of-life using coefficients from wage regressions. The four questions pertain to whether wage regression coefficients on fatality rate variables are sensitive to (1) including or excluding women from the sample, (2) definitions of the fatality rate which include estimates of either male only or female only deaths, (3) removing the influence of firm size from the fatality rate or the wage rate, (4) respondents' inaccuracies in reporting their three-digit industry. Problems associated with (3) and (4) are found to be minimal. Problems associated with (1) and (2) also appear to be small if the goal is to estimate the value-of-life for men.

Medical costs in workers' compensation insurance: comment.

Professors Baker and Krueger ignore some costs associated with workers' compensation. Because of these costs, the contention that physicians willfully exploit the workers' compensation system for their own gain is questioned.

Direct and indirect effects of education on health.

In virtually every study by economists considering the determinants of an individual's health, years of schooling has stood out as an important regressor. Yet there is little agreement among economists concerning the mechanisms through which schooling affects health. This study is a first attempt to test for the direct and indirect effects of education on an individual's health. Education is assumed to enhance health directly by, for example, allowing wise use of medical care and indirectly through encouraging healthy habits and caution in the choice of occupation. Evidence from two national surveys indicates that the indirect dominate the direct effects.

Employee and job attributes as predictors of absenteeism in a national sample of workers: the importance of health and dangerous working conditions.

This study reports on research which looks for employee and job characteristics which correlate with absenteeism. A large cross-sectional national probability sample of workers employed for at least 20 hr per week is analyzed (n = 1308). The dependent variable is the number of self-reported absences during the past 14 days. Thirty-seven independent variables are considered. Ordinary Least Squares (multiple regressions), two-limit Tobits, and two-part models are used to assess the statistical and practical significance of possible covariates. Statistically significant predictors included health variables such as being overweight, complaining of insomnia, and hazardous working conditions; job characteristics such as inflexible house; and personal variables such as being a mother with small children. Variables reflecting dangerous working conditions appear to be the strongest correlates of absenteeism. Notable variables which do not predict absenteeism include age, race, wages, and job satisfaction. Future research should direct attention toward workers' health and working conditions as covariates of absenteeism, since they are strongly significant in this study and have been neglected by most absenteeism investigators.

Occupational hazards and heart attacks.

There has been limited research on occupation health hazards and heart disease. With the exception of the effects of noise, even less research has been conducted on occupational safety hazards, however. This study takes an initial look at interrelationships between safety hazards and heart disease. Fatal injury rates within industries are taken as proxies for safety hazards; fatal heart attack rates are taken as proxies for heart disease. Holding age, race, and gender constant evidence is found which is consistent with the hypothesis that safety hazards, through their effects on stress, either cause or exascerbate heart disease.

Who chooses risky jobs?

Empirical studies indicate that a disproportionate number of blacks are found in dangerous jobs. This study replicates this finding with new fatality rates within 353 occupations. A model implicating family background and race as determinants of risk and time preferences is developed to explain the result. Evidence is also produced suggesting that family background influences risk and time preferences and risky job choice. Policy implications for expenditures on Occupational Safety and Health are drawn.

Correlations between education and arthritis in the 1971-1975 NHANES I.

Data from the National Health and Nutrition Examination Survey I, 1971-1975 (NHANES I) were used to analyze associations among highest education level and arthritis. The dependent variables indicated whether the respondent had ever been diagnosed with any form of arthritis by a physician (10,678 women and 7243 men) or whether physician X-ray readings suggested arthritis of the knee (3491 women and 3119 men). These variables did not distinguish between osteo- and rheumatoid arthritis. It is likely that the great majority of the sample reporting or diagnosed with arthritis had osteoarthritis. There were strong univariate correlations between answers to the general arthritis question and the knee question on the one hand and gender, age, body mass, schooling, income and employment on the other. Respondents' education level was found to be strongly and negatively associated with self-reported arthritis in the larger samples both before and after controls were entered for employment, income and potential biological risk factors. The association between self-reported arthritis or arthritis of the knees and education was weaker for men, but not for women after employment and income were accounted for. When body mass was accounted for, the association between self-reported arthritis or arthritis of the knees and education was weaker among women but not men. Long-run preventive strategies to combat osteoarthritis ought to consider investments in education.

Tobit, fixed effects, and cohort analyses of the relationship between severity and duration of rheumatoid arthritis.

Three methodological problems are commonly faced by researchers investigating relationships between severity and duration of illness among patients with rheumatoid arthritis (RA). (1) Linear regression techniques yield biased estimates when measures of severity are continuous but range between and include limiting values such as 0 and 3. (2) Data from the same patient over time are typically pooled together with data from different patients at the same time and over time. Models are then used that do not account for the statistical problems that can result from pooling. (3) Persons with varying years of duration of disease are typically combined and analyzed without any special attention to cohort effects. Changes in severity over time for cohorts of patients with fewer than 10 years of duration may be different from changes in severity of patients with more than 20 years of duration from the onset of the disease. In this study, severity is measured by the 0-3 disability scale in the Stanford Health Assessment Questionnaire (HAQ). Duration is measured by self-report of the onset of symptoms by subjects. Popular techniques are borrowed from econometrics--Tobit, Fixed Effects, and dummy variables for Cohort Models--that were developed to address three analogous problems in economic data. The three economic techniques are applied separately and together using data collected by Arthritis, Rheumatism, and Aging Medical Information System (ARAMIS) on 330 RA patients in 1981 who were followed until 1989. Although the Tobit technique does not appear to be especially useful with these data, Fixed Effects and Cohort Models do appear to be useful.(ABSTRACT TRUNCATED AT 250 WORDS)

Public and private workers' compensation insurance.

Utilizing unique data from the federal government's Social Security office, we analyze the costs of providing workers' compensation by public vs private insurers. Our analysis has the advantage that public and private recipients are likely to be more similar than public (Medicare, Medicaid) and private recipients of health insurance in the US or Canadian economies. We find that public firms appear to provide workers' compensation insurance more efficiently than private firms.

Ranking occupations based upon the costs of job-related injuries and diseases.

In this article, we construct a ranking of occupations based upon the costs of job-related injuries and illnesses. Data are drawn from large nationally representative Bureau of Labor (BLS) data sets. Information is obtained on occupation and workers' compensation (WC) category of the injury or illness, which are then matched to information on costs. Six broad occupations and 413 specific (3-digit) occupations are ranked by total costs. Six broad and 223 specific occupations are ranked by costs per worker (average cost). Operators and laborers is the broad occupation category that contributes both the highest total and average cost. Specific occupations that contributed the most to total costs include heavy truck drivers, non-construction laborers, machine operators (not specified), occupations not classified, janitors, nursing orderlies, construction laborers, assemblers, retail sales workers (not elsewhere specified), miscellaneous machine operators, and carpenters. Occupations high on the average cost list include not-specified mechanics, general and construction laborers, press apprentices, welders, stone cutters, and warehouse workers. Although the BLS data are limited, they can be used to provide a preliminary look at which occupations are contributing the most and the least to the overall economic costs of occupational injuries and illnesses.