 |
|
Supplemental Insurance and Mortality in Elderly Americans
Findings From a National Cohort
Mark P. Doescher, MD, MSPH;
Peter Franks, MD;
Jessica S. Banthin, PhD;
Carolyn M. Clancy, MD
Arch Fam Med. 2000;9:251-257.
ABSTRACT
| |
Context As the burden of out-of-pocket health care expenditures for Medicare beneficiaries has grown, the need to assess the relationship between uncovered costs and health outcomes has become more pressing.
Objective To assess the relationship between risk for out-of-pocket expenditures and mortality in elderly persons with private supplemental insurance.
Design Retrospective cohort study using proportional hazards survival analyses to assess mortality as a function of health insurance, adjusting for sociodemographic, access, and case mix–health status measures.
Setting The 1987 National Medical Expenditure Survey, a representative cohort of the US civilian population, linked to the National Death Index.
Participants A total of 3751 persons aged 65 years and older.
Main Outcomes Measures Five-year mortality rate.
Results After 5 years, 18.5% of persons at low risk for out-of-pocket expenditures, 22.5% of those at intermediate risk, and 22.6% of those at high risk had died. After multivariate adjustment, a significant linear trend (P = .02) toward increasing mortality with increasing risk category was observed. Compared with the low-risk group, persons in the intermediate-risk group had an adjusted hazard ratio of 1.2 (95% confidence interval, 0.9-1.6), whereas those in the high-risk group had an adjusted hazard ratio of 1.4 (95% confidence interval, 1.0-1.9).
Conclusions Increasing risk for out-of-pocket costs is associated with higher subsequent mortality among elderly Americans with supplemental private coverage. Although research is needed to identify which specific components of out-of-pocket expenditures are adversely associated with health outcomes, findings support policies to decrease out-of-pocket health care expenditures to reduce the risk for premature mortality in elderly Americans.
INTRODUCTION
EFFORTS TO control growth in costs have led to an increasing financial burden placed on elderly Medicare beneficiaries. Various cost-sharing requirements and benefit limitations in the Medicare program leave beneficiaries, especially those without supplemental coverage, vulnerable to high levels of out-of-pocket costs. Consequently, to fill some of these gaps in Medicare coverage, many beneficiaries are covered by private supplemental insurance, through employer-provided or individually purchased ("Medigap") plans.1-3 However, because premiums have risen rapidly in recent years, even beneficiaries with private supplemental plans increasingly face substantial out-of-pocket costs.3 Little is known about the impact of these shifts in costs on health outcomes.
From 1977 through 1987, out-of-pocket health care expenditures for Americans older than 65 years rose by more than 3 times the consumer price index.4 By 1996, excluding nursing home costs, Medicare beneficiaries spent an average of $2600 out-of-pocket on health care costs above what Medicare and private insurers paid.5-6 This represented 21% of household income for elderly individuals, up from 15% in 1987.5 Recent withdrawals of Medicare health maintenance organizations (HMOs) from Medicare contracts7 are likely to exacerbate the problem of out-of-pocket costs for the elderly.
Landerman et al8 recently reported that the odds of development of disability were higher among Medicare enrollees without private supplemental coverage compared with those with such coverage after adjustment for education, income, and other risk factors. Although additional data examining uncovered expenditures and health outcomes are lacking in Medicare populations, studies in persons younger than age 65 years have documented adverse associations between insurance status and health outcomes. Results from the Rand Health Insurance Experiment showed that financial barriers in the form of greater copayments resulted in clinically important adverse health outcomes for certain groups, including those with poor vision, high blood pressure, and low incomes.9 Observational studies also have documented adverse associations between lacking health insurance coverage and health outcomes. Persons losing health insurance benefits suffer measurable declines in their health,10-12 and persons lacking health insurance describe relatively lower self-reported health status13-14 and functional status.15 Lacking health insurance in persons younger than 65 years has been shown to be adversely associated with mortality.16-17
Because of the growth of the problem of out-of-pocket costs in the elderly and the limited information about possible implications of these changes, we examined the relationship between the risk for out-of-pocket expenditures for persons older than 65 years and subsequent mortality. Our study is based on linked 1987 National Medical Expenditure Survey (NMES)–National Death Index data, a representative cohort of the US population with 5 years of mortality follow-up. We examined a relatively homogeneous subset of the elderly in that sample, those with supplemental health insurance. Based on previous theoretical research,18-20 the adequacy of health insurance coverage is viewed as an enabling factor facilitating access to care. We hypothesized that, after adjusting for potential confounders, as the liability for out-of-pocket expenditures increases, the risk for mortality also increases.
MATERIALS AND METHODS
SAMPLE
We used data from the Household Survey component of the NMES.21 The Household Survey consisted of a 1-year, cross-sectional survey of nearly 38 000 individuals from approximately 15 000 households representing the 1987 US civilian, noninstitutionalized population. The survey used a stratified, multistage area probability design with oversampling of minorities, poor persons, disabled persons, and the elderly. Four rounds of interviews were completed during 1987, and a wide range of information regarding medical care, health expenditures, and health insurance coverage was collected. Respondents were asked to complete a self-administered questionnaire that included items on subjective health status, a checklist of medical conditions, health care access, and attitudes toward the value of health care and health insurance. The overall survey response rate for the Household Survey component was 80%.22
A unique aspect of the 1987 NMES was the detailed information on health insurance benefits. The Health Insurance Plans Survey (HIPS)23 collected health insurance policy booklets for households with private insurance in a follow-up survey of employers, insurers, and unions. Detailed benefit provisions were abstracted from the booklets, including covered services, deductibles, copayments and coinsurance rates, out-of-pocket stop-loss limits, and annual and lifetime maximum payments. Benefit provisions were abstracted separately for dozens of different services, including prescription drugs. Medigap policies were coded with respect to the gaps in Medicare coverage. The HIPS response rate was 62%. This analysis uses imputed premium and benefit data assigned to the 38% of the HIPS sample that did not respond. Respondents aged 65 years and older with private coverage supplemental to Medicare were included in the analyses. These included persons with employment-sponsored retirement coverage as well as persons with individually purchased Medigap policies. The final sample included 3751 subjects.
MEASURES
Estimated Liability for Out-of-Pocket Expenditures in the Catastrophic Illness Scenario
In our sample of Medicare beneficiaries with private supplemental health insurance coverage, policies are characterized by hundreds of specific provisions that make it difficult to measure the breadth and depth of coverage. Medigap plans designed specifically to supplement Medicare benefits cover about 53% of our sample.2 These plans vary by whether they cover the part A deductible and copayments or the part B deductible and coinsurance, and whether they provide other benefits not covered by Medicare, such as prescription drugs. In 1987, there was more variation in Medigap plans than today, since federal legislation specified 10 standardized Medigap plans.24 The rest of our sample are covered by employer-sponsored insurance, which is not usually designed to supplement Medicare. These plans are governed by more provisions than most Medigap plans. Patient cost-sharing provisions reduce coverage at the front end in the form of deductibles and copayments, which also may vary by type of service. At the back end, plans vary in the degree of protection they offer against catastrophic events. Some plans set annual or lifetime maximums that limit insurance payments after a certain amount. Other plans offer more protection to patients with serious illness by setting stop-loss limits on out-of-pocket expenses.
To simplify this complexity into one summary measure of coverage generosity, we calculated the share of total expenditures that is covered by Medicare plus private insurance and the share paid out-of-pocket by the patient for a stylized illness scenario. This approach to evaluating insurance benefits has been used elsewhere and has the advantage of summarizing the multidimensional concept of insurance plan generosity by a single continuous variable.25-26 To compute these out-of-pocket expenses, we used detailed benefit data collected from the health policy booklets of every person in our sample. Coverage information was collected on more than 16 types of health care services, including hospital facility care, inpatient and outpatient physician services, emergency department care, dental care, prescription drugs, durable medical equipment, supplies, home health care, and more. We used these data in conjunction with the claims processing program BENSIM, a component of the MEDSIM microsimulation model developed by researchers at the Agency for Health Care Policy and Research.27 The BENSIM program replicates the standard calculations of insurance payments and out-of-pocket responsibility that Medicare and private insurance plans would make for each person in our sample. Coordination of benefits between Medicare and private supplemental policies was determined by information collected from policy booklets.
We defined a catastrophic illness scenario corresponding to the 99th percentile of the distribution of total expenditures for all persons aged 65 years and older in the NMES sample. Thus, each person in our study sample was estimated to have a 1% chance of experiencing such an event. The total expenditure for the catastrophic illness scenario, including payments from all sources, was $76 729 in 1987 dollars. The scenario included inpatient hospital expenditures of $59 235. The rest of the expenditures were for physician services, emergency department care, physical therapy, prescription drugs, home health care, and other types of care. Simulated out-of-pocket payments were calculated based on this illness scenario and expressed as a percentage of the total illness expenditure. Simulated out-of-pocket expenditures in the event of a catastrophic illness was categorized as low (<2% of total expenditure), intermediate (2%-10% of total expenditure), and high risk (>10% of total expenditure), and also examined as a continuous variable.
Sociodemographic Factors
Several sociodemographic variables were examined to adjust for potential confounding and included age in years; level of education (less than high school, high school, and greater than high school); household income as percentage of poverty level for 1987 (poor, <100%; near poor, 100% to <125%; low income, 125% to <200%; middle income, 200%-400%; and high income, >400%); wealth, defined as asset income (annuity, dividend, estate, cash contribution, interest, non–home gain [or loss], rental, and royalty income); race or ethnicity (white and nonwhite); and marital status (married or unmarried living with a partner and single).
Regular Source of Care
A national study showed that 79% of Americans could identify a regular personal physician by name,28 so presence or absence of a self-reported personal physician was assessed.
Health Status Factors
Subjective health status was measured using items that constitute subscales of the Medical Outcomes Study general short-form health survey, a measure of the health effects of chronic disease with demonstrated reliability and validity.29-30 The subscales included 5 questions each on health perceptions, including self-reported health status (Cronbach  = .90), mental health ( = .88), physical functioning ( = .85), and 2 questions on role functioning ( = .79), and exhibited excellent internal reliability in the NMES as demonstrated by the high Cronbach scores.
Self-Reported Morbidity
Respondents were asked whether a physician had ever told them they had any of the following chronic medical diagnoses: gallbladder disease, stroke, cancer, emphysema, diabetes, arthritis, high blood pressure, heart attack, hardening of the arteries, or heart disease. From these items, a continuous scale was derived as well as a dichotomous self-reported medical diagnosis variable (any diagnoses and no diagnoses).
Body Mass Index
Body mass index (BMI; calculated as weight in kilograms divided by the square of height in meters) was calculated for each subject from self-reported weight and height and categorized to reflect extremes that have been associated with excess mortality ( 19 or  30).31-32
Medical Skepticism Scale
Ten health attitude questions, derived from a 1970 Center for Health Administration Studies–National Opinion Research Center study,33 were included in the NMES. For our report, 5 questions that contributed to reliability (Cronbach = .62) were selected to form a unidimensional scale to measure the adult respondent's attitude toward health insurance and health care. Increasing medical skepticism has been shown to be adversely associated with mortality in the NMES.34
Tobacco Use
Self-reported smoking status was categorized for each subject as current, former, or never.
Expenditure and Health Care System Utilization Measures
The NMES includes detailed corroborated data on 1987 health care expenditures and utilization. Total annual medical care expenditures were examined as a continuous variable. The number of hospitalizations was assessed as a continuous variable and also as a dichotomous variable (none or 1).
Mortality
We used the 5-year mortality experience of respondents in NMES ascertained through the National Death Index, which is available on a public-use tape.35 The National Death Index is a computer file of all deaths in the United States since 1979 and is maintained by the National Center for Health Statistics.36 The National Death Index has been shown to be an effective and accurate means of ascertaining deaths for large-cohort studies.37-41 Deaths were ascertained if a match occurred between identifying data for persons in the Household Survey and identifying data for persons whose deaths were recorded in the National Death Index from January 1, 1987, through December 31, 1992. Identifying data, including first and last name; month, day, and year of birth; sex; and race, were available for more than 99% of the NMES sample persons. Social security number was available for 70% of the NMES sample persons.35
STATISTICAL ANALYSES
Weights provided on the public-use tapes were used to yield population estimates adjusted for survey oversampling and nonresponse. Because of the complex survey design of NMES, analyses were conducted with the statistical package SUDAAN,42 which uses the method of Taylor series linearization to produce appropriate SEs and 95% confidence intervals (CIs) in surveys involving cluster sampling. To examine the relationship between liability for out-of-pocket expenditures in the event of a catastrophic illness and total mortality, proportional hazards survival analyses were performed to obtain hazard ratios (HRs) adjusted for potential confounders. Interaction between insurance category and education, income, wealth, age, sex, race or ethnicity, and health status was assessed.
RESULTS
Baseline characteristics of the sample stratified by insurance risk category are presented in Table 1, which also provides national population estimates for each covariate. The 3 risk categories in Table 1 measure different levels of estimated liability for out-of-pocket expenditures as a percentage of total health care expenditures in the catastrophic illness scenario (low risk, <2%; intermediate risk, 2%-10%; high risk, >10%). Overall, the 3 risk groups had similar characteristics, with significant differences across groups present for only BMI and medical skepticism score. Low body mass (BMI, <20) was less likely to be reported by persons in the high-risk group (3.6%) than persons in the low (8.3%) or intermediate-risk groups (8.1%), whereas the likelihood of large body mass (BMI, >30) was similar across groups. The mean medical skepticism score was greatest in persons in the intermediate-risk group (mean score, 2.2), but was lower in persons in the high (mean score, 2.0) and low-risk (mean score, 2.1) groups. Age, sex, education and income levels, wealth, marital status, and race or ethnicity did not vary significantly across groups. More than 90% of persons in each group identified a personal physician. Smoking status did not vary substantially across groups. Self-reported health status, measured by mean Medical Outcomes Study scale scores and the mean number of medical diagnoses, was similar across groups. Mean medical expenditures that were covered by insurance did not vary significantly by risk group, nor did mean uncovered (out-of-pocket) expenditures.
|
|
|
|
Table 1. Relationships Between Liability for Out-of-Pocket Expenditures in the Catastrophic Illness Scenario and Mortality and Selected Characteristics*
|
|
|
UNADJUSTED MORTALITY
After 5 years of mortality follow-up, 18.5% of persons classified as low risk for out-of-pocket expenditures in the event of catastrophic illness, 22.5% of those classified as intermediate risk, and 22.6% of those classified as high risk had died (Table 1).
ADJUSTED MORTALITY
Examination of the measure of liability for out-of-pocket expenditures in the event of a catastrophic illness as a continuous variable revealed a significant association with mortality. An adjusted HR of 1.03 (95% CI, 1.00-1.05; P = .04) for each percentage increase in risk was found (data not tabled). Categorization of this measure into the low-, intermediate-, and high-risk groups (Table 2) showed a significant linear trend (P = .02). Persons in the high-risk group (ie, underinsured subjects) had a significantly increased risk for mortality compared with those in the low-risk group (HR, 1.4; 95% CI, 1.0-1.9), whereas persons in the intermediate-risk group had a modestly elevated, although not statistically significant, risk for mortality (HR, 1.2; 95% CI, 0.9-1.6). Assessments of interaction between liability for out-of-pocket expenditures in catastrophic illness and other covariates made no statistically significant (P<.05) contributions to the proportional hazards regression models.
|
|
|
|
Table 2. Proportional Hazards Model for Survival Time Adjusted for Baseline Characteristics*
|
|
|
COMMENT
Our observations suggest that strategies aimed at increasing out-of-pocket health care expenditures to contain Medicare costs may result in adverse health outcomes. Most Medicare beneficiaries have obtained private supplemental coverage for protection, to varying degrees, from the copayments, deductibles, and other uncovered costs of Medicare. However, even for persons with private supplemental coverage, out-of-pocket health care costs can be substantial and are rising.3, 5-6 Because our observations reveal an adverse association between liability for out-of-pocket expenditures in the event of catastrophic illness and mortality after adjustment for covariates, support is provided for policies that would reduce, rather than increase, financial barriers to care for elderly Americans.
The observed relationship may be the result of confounding by incompletely measured or unmeasured factors. Persons with less comprehensive supplemental coverage may value health less than those with more generous coverage. Consequently, persons placing less value on health may fail to secure comprehensive supplemental coverage and may also experience higher mortality because of individual lifestyle factors. However, we controlled for a far wider range of measures than usually are found in secondary data sets, including a medical skepticism scale measuring attitudes toward health insurance and health care, sociodemographic factors, self-reported health status, and lifestyle measures of smoking status and other unhealthy behaviors.
The inferences that can be drawn from these results are strengthened by the study design and the use of a comprehensive measure of insurance protection. The study population was limited to the relatively homogeneous group of persons covered by private insurance supplemental to Medicare. These persons are likely to be similar along a variety of unobserved factors.
The measure of the generosity of insurance protection used is less subject to endogeneity concerns than other potential measures. For example, measuring the generosity of insurance coverage by selecting a few specific provisions, such as coverage for the Medicare hospital deductible or coverage for Medicare coinsurance on physician office visits, could be directly correlated with a person's health status. Individuals in ill health in 1987 may not have been able to purchase a Medigap policy including those specific provisions, and may be more likely to experience subsequent mortality. Similarly, measuring insurance generosity by using actual out-of-pocket expenditures in 1987 (as reported in the NMES) would clearly be correlated with individuals' health status and likelihood of subsequent mortality. In contrast, the measure of insurance generosity used is less likely to be correlated with subjects' health status in 1987 because it estimated liability for out-of-pocket expenditures in the event of a simulated catastrophic illness.
The reported hazard ratio may underestimate the relationship between liability for out-of-pocket expenditures in the event of catastrophic illness and mortality. First, insurance status was measured only once, at baseline, and no account of changes into, and out of, private health plans could be performed. Migration of persons in the high-risk group to lower-risk groups and vice versa likely would bias the observed association toward the null hypothesis. Second, the variables included in the survival analysis may have resulted in overadjustment of the true relationship between insurance status and mortality. If inadequate insurance in the elderly adversely affects health, as suggested in studies examining lack of insurance in younger persons,13-14 then adjustment for baseline health status will tend to reduce the observed relationship between health insurance and mortality.
Observational studies cannot completely exclude the possibility of unmeasured confounding, nor prove causality. It is unlikely, however, that a randomized, controlled trial to assess different levels of liability for out-of-pocket expenditures in catastrophic illness and adverse health outcomes will be performed. Similarly, it is unlikely that randomized, controlled trials to assess specific aspects of uncovered expenditures, such as varying levels of prescription medication coverage, will be conducted. In the absence of prospective trials, more detailed analyses from diverse data sources are warranted.
Although the findings reported herein provide evidence that cost-sharing in the elderly is adversely associated with mortality, alternative strategies to reduce the growth of Medicare could also result in undesired outcomes. Relying on reducing payments to providers such as physicians, hospitals, and HMOs ultimately would reach limits beyond which additional cutbacks would become impossible. Medical savings accounts would be more likely to be used by relatively affluent persons, whereas the sickest persons would likely remain in the traditional fee-for-service programs, driving up fee-for-service costs.43 Although Medicare HMO enrollment offers relatively greater protection against out-of-pocket costs, cost savings appear to occur through biased selection of relatively healthy enrollees.44 If Medicare HMOs eventually were to begin enrolling elderly Americans nonselectively, an increased reliance on cost-sharing would almost certainly result.
Most health plans in the United States require beneficiaries to meet cost-sharing requirements that are unrelated to income.45 This reliance on cost-sharing is not only inequitable, but it ultimately may not result in significant slowing in Medicare spending.45-46 Although the critical issue of cost containment will not be resolved soon, the health consequences of the current economic imperative must be considered: Might our current efforts to contain costs through market-driven solutions ultimately result in more harm than good for a vulnerable segment of society?
Our study indicates that increasing risk for out-of-pocket expenditures in catastrophic illness is associated with increased subsequent mortality among elderly Americans with supplemental private coverage. The impending bankruptcy of the Medicare part A trust fund and Congressional overtures to limit Medicare eligibility to those older than 67 years suggest that Medicare coverage will be eroded further and that cost-sharing will increase. This trend may have an adverse impact on the health of elderly persons in our society, especially low-income individuals who bear the greatest burden of out-of-pocket expenditures relative to income. Research is needed to identify which specific components of uncovered expenditures in the elderly are most likely to result in reduced access to potentially useful health care interventions. Similarly, research is needed to identify which specific components of uncovered expenditures are associated with adverse health outcomes. Moreover, research is needed to examine the health impact of cost containment strategies that do not directly increase out-of-pocket expenditures, such as managed Medicare and medical savings accounts. Despite these research needs, our findings support policies to decrease out-of-pocket health care expenditures to reduce the risk for premature mortality in elderly Americans.
AUTHOR INFORMATION
Accepted for publication September 30, 1999.
This study was supported by grant R03 HS09522-01 from the Agency for Health Care Policy and Research, Rockville, Md.
Reprints: Mark P. Doescher, MD, MSPH, Department of Family Medicine, University of Washington School of Medicine, 4225 Roosevelt Way NE, Suite 308, Seattle, WA 98105-6099 (e-mail: mdoesche{at}u.washington.edu).
From the Department of Family Medicine, University of Washington School of Medicine, Seattle (Dr Doescher); the Department of Family Medicine, Primary Care Institute, University of Rochester, Rochester, NY (Dr Franks); Center for Cost and Financing Studies, (Dr Banthin) and Center for Outcomes and Effectiveness Research (Dr Clancy), Agency for Healthcare Research and Quality, Rockville, Md.
REFERENCES
| |
1. Ettner SL. Adverse selection and the purchase of Medigap insurance by the elderly. J Health Econ. 1997;16:543-562.
FULL TEXT
|
ISI
| PUBMED
2. Vistnes J, Banthin JS. The demand for Medicare supplemental insurance benefits: the role of attitudes toward medical care and risk. Inquiry. Winter 1997-1998;34:311-324.
3. Alecxih LMB, Lutzky S, Sevak P, Claxton G. Key Issues Affecting Accessibility to Medigap Insurance. Washington, DC: The Lewin Group; August 1997.
4. Taylor A, Banthin J. Changes in Out-of-Pocket Expenditures for Personal Health Services: 1977 and 1987: National Medical Expenditure Survey. Rockville, Md: Public Health Service; 1994. Agency for Health Care Policy and Research publication 94-0065; Research Finding 21.
5. Moon M, Kuntz C, Pounder L. Protecting Low-Income Medicare Beneficiaries. New York, NY: Commonwealth Fund; 1996.
6. American Association of Retired Persons. Out-of-Pocket Spending by Medicare Beneficiaries Age 65 and Older: 1997 Projections. Washington, DC: The Lewin Group; 1997.
7. Iglehart JK. The American health care system: Medicare. N Engl J Med. 1999;340:327-332.
FREE FULL TEXT
8. Landerman LR, Fillenbaum GG, Pieper CF, Maddox GL, Gold DT, Guralnik JM. Private health insurance coverage and disability among older Americans. J Gerontol B Psychol Sci Soc Sci. 1998;53(suppl):S258-S266.
9. Brook RH, Ware JE Jr, Rogers WH, et al. Does free care improve adults' health? results from a randomized controlled trial. N Engl J Med. 1983;309:1426-1434.
ABSTRACT
10. Lurie N, Ward NB, Shapiro MF, Brook RH. Termination from Medi-Cal: does it affect health? N Engl J Med. 1984;311:480-484.
ISI
| PUBMED
11. Lurie N, Ward NB, Shapiro MF, Gallego C, Vaghaiwalla R, Brook RH. Termination of Medi-Cal benefits: a follow-up study one year later. N Engl J Med. 1986;314:1266-1268.
ISI
| PUBMED
12. Fihn SD, Wicher JB. Withdrawing routine outpatient medical services: effects on access and health. J Gen Intern Med. 1988;3:356-362.
ISI
| PUBMED
13. Patrick DL, Madden CW, Diehr P, Martin DP, Cheadle A, Skillman SM. Health status and use of health services among families with and without health insurance. Med Care. 1992;30:941-949.
FULL TEXT
|
ISI
| PUBMED
14. Franks P, Clancy CM, Gold MR, Nutting PA. Health insurance and subjective health status: data from the 1987 National Medical Expenditure Survey. Am J Public Health. 1993;83:1295-1299.
FREE FULL TEXT
15. Hubbell FA, Waitzkin H, Rodriguez FI. Functional status and financial barriers to medical care among the poor. South Med J. 1990;83:548-550.
ISI
| PUBMED
16. Franks P, Clancy CM, Gold MR. Health insurance and mortality: evidence from a national cohort. JAMA. 1993;270:737-741.
FREE FULL TEXT
17. Sorlie PD, Johnson NJ, Backlund E, Bradham DD. Mortality in the uninsured compared with that in persons with public and private health insurance. Arch Intern Med. 1994;154:2409-2416.
FREE FULL TEXT
18. Andersen R. A Behavioral Model of Families' Use of Health Services. Chicago, Ill: University of Chicago, Center for Health Administration Studies; 1968. Research Series 25.
19. Aday LA, Andersen R. Development of Indices of Access to Medical Care. Ann Arbor, Mich: Health Administration Press; 1975.
20. Millman ML, ed. Access to Health Care in America: Committee on Monitoring Access to Personal Health Care Services. Washington, DC: National Academy Press; 1993.
21. Edwards WS, Berlin M. Questionnaires and Data Collection Methods for the Household Survey and the Survey of American Indians and Alaska Natives. Rockville, Md: US Dept of Health and Human Services; 1989. Department of Health and Human Services publication (PHS)89-3450.
22. National Center for Health Services Research and Health Care Technology Assessment. National Medical Expenditure Survey Methods 2. Rockville, Md: Public Health Service; 1989.
23. Emmons C, Hill C. Questionnaires and Data Collection Methods for the Health Insurance Plans Survey: National Medical Expenditure Survey Methods 5. Rockville, Md: Public Health Service, Agency for Health Care Policy and Research; February 1994. Agency for Health Care Policy and Research publication 94-0016.
24. Omnibus Reconciliation Act of 1990, Pub L No. 101-508, 105 Stat 1388 (1990).
25. Short PF, Banthin JS. New estimates of the uninsured younger than 65 years. JAMA. 1995;274:1302-1306.
FREE FULL TEXT
26. Zuvekas SH, Banthin JS, Selden TM. Mental health parity: what are the gaps in coverage? J Ment Health Policy Econ. 1998;1:135-146.
FULL TEXT
| PUBMED
27. Moeller JF, Banthin JS, Beauregard KM, Cohen SB, Selden TM, Zabinski D. Medical Expenditure MicroSIMulation Model (MEDSIM): A Brief Overview of the Model and Database. Rockville, Md: Center for Cost and Financing Studies, Agency for Health Care Policy and Research; 1996.
28. Spiegel JS, Rubenstein LV, Scott B, Brook RH. Who is the primary physician? N Engl J Med. 1983;308:1208-1212.
ABSTRACT
29. Stewart AL, Hays RD, Ware JE Jr. The MOS Short-Form General Health Survey: reliability and validity in a patient population. Med Care. 1988;26:724-735.
ISI
| PUBMED
30. Stewart AL, Greenfield S, Hays RD, et al. Functional status and well-being of patients with chronic conditions: results from the Medical Outcomes Study. JAMA. 1989;262:907-913.
FREE FULL TEXT
31. Harris T, Cook EF, Garrison R, Higgins M, Kannel W, Goldman L. Body mass index and mortality among nonsmoking older persons: the Framingham Study. JAMA. 1988;259:1520-1524.
FREE FULL TEXT
32. Cornoni-Huntley JC, Harris TB, Everett DF, et al. An overview of body weight of older persons, including the impact on mortality: the National Health and Nutrition Examination Survey I–Epidemiologic Follow-Up Study. J Clin Epidemiol. 1991;44:743-753.
FULL TEXT
|
ISI
| PUBMED
33. Andersen R, Lion J, Anderson OW. Two Decades of Health Services Research: Social Survey Trends in Use and Expenditures. Cambridge, Mass: Ballinger Publishing Co; 1976.
34. Fiscella K, Franks P, Clancy CM, Doescher MP, Bathin JS. Does skepticism towards medical care predict mortality? Med Care. 1999;37:409-414.
FULL TEXT
|
ISI
| PUBMED
35. Agency for Health Care Policy and Research. National Medical Expenditure Survey, 1987: Data From the Household Survey, Health Insurance Plans Survey, Survey of American Indians and Alaska Natives, and Institutional Population Component [Research File 40R]. Rockville, Md: Agency for Health Care Policy and Research; 1997.
36. National Center for Health Statistics. National Death Index Users Manual. Hyattsville, Md: US Dept of Health and Human Services; 1990. Department of Health and Human Services publication (PHS) 90-1148.
37. Wentworth DN, Neaton JD, Rasmussen WL. An evaluation of the Social Security Administration master beneficiary record file and the National Death Index in the ascertainment of vital status. Am J Public Health. 1983;73:1270-1274.
FREE FULL TEXT
38. Stampfer MJ, Willett WC, Speizer FE, et al. Test of the National Death Index. Am J Epidemiol. 1984;119:837-839.
FREE FULL TEXT
39. Williams BC, Demitrack LB, Fries BE. The accuracy of the National Death Index when personal identifiers other than Social Security number are used. Am J Public Health. 1992;82:1145-1147.
FREE FULL TEXT
40. Calle EE, Terrell DD. Utility of the National Death Index for ascertainment of mortality among cancer prevention study II participants. Am J Epidemiol. 1993;137:235-241.
FREE FULL TEXT
41. Liao Y, Cooper RX, Cao G, et al. Mortality patterns among adult Hispanics: findings from the NHIS, 1986 to 1990. Am J Public Health. 1998;88:227-232.
FREE FULL TEXT
42. Research Triangle Institute. SUDAAN: Professional Software for Survey Data Analysis, Version 6.34. Research Triangle Park, NC: Research Triangle Institute; 1993.
43. Angell M. Fixing Medicare. N Engl J Med. 1997;337:192-195.
FREE FULL TEXT
44. Morgan RO, Virnig BA, DeVito CA, Persily NA. The Medicare-HMO revolving door: the healthy go in and the sick go out. N Engl J Med. 1997;337:169-175.
FREE FULL TEXT
45. Rice T, Thorpe KE. Income-related cost sharing in health insurance. Health Aff (Millwood). Spring 1993;12:21-39.
ABSTRACT
46. Chulis GS, Eppig FJ, Hogan MO, Waldo DR, Arnett III RH. Why more cost-sharing won't slow Medicare spending. J Am Health Policy. 1993;3:15-20.
RELATED ARTICLE
THE Archives of Familly Medicine Continuing Medical Education Program
Arch Fam Med. 2000;9(3):267-269.
FULL TEXT
THIS ARTICLE HAS BEEN CITED BY OTHER ARTICLES
Book Review: The Impact of Cost-Sharing on Appropriate Utilization and Health Status: A Review of the Literature on Seniors
Rice and Matsuoka
Med Care Res Rev 2004;61:415-452.
ABSTRACT
|
